Show More
@@ -1,762 +1,762 b'' | |||||
1 | /* |
|
1 | /* | |
2 | parsers.c - efficient content parsing |
|
2 | parsers.c - efficient content parsing | |
3 |
|
3 | |||
4 | Copyright 2008 Matt Mackall <mpm@selenic.com> and others |
|
4 | Copyright 2008 Matt Mackall <mpm@selenic.com> and others | |
5 |
|
5 | |||
6 | This software may be used and distributed according to the terms of |
|
6 | This software may be used and distributed according to the terms of | |
7 | the GNU General Public License, incorporated herein by reference. |
|
7 | the GNU General Public License, incorporated herein by reference. | |
8 | */ |
|
8 | */ | |
9 |
|
9 | |||
10 | #define PY_SSIZE_T_CLEAN |
|
10 | #define PY_SSIZE_T_CLEAN | |
11 | #include <Python.h> |
|
11 | #include <Python.h> | |
12 | #include <ctype.h> |
|
12 | #include <ctype.h> | |
13 | #include <stddef.h> |
|
13 | #include <stddef.h> | |
14 | #include <string.h> |
|
14 | #include <string.h> | |
15 |
|
15 | |||
16 | #include "bitmanipulation.h" |
|
16 | #include "bitmanipulation.h" | |
17 | #include "charencode.h" |
|
17 | #include "charencode.h" | |
18 | #include "util.h" |
|
18 | #include "util.h" | |
19 |
|
19 | |||
20 | #ifdef IS_PY3K |
|
20 | #ifdef IS_PY3K | |
21 | /* The mapping of Python types is meant to be temporary to get Python |
|
21 | /* The mapping of Python types is meant to be temporary to get Python | |
22 | * 3 to compile. We should remove this once Python 3 support is fully |
|
22 | * 3 to compile. We should remove this once Python 3 support is fully | |
23 | * supported and proper types are used in the extensions themselves. */ |
|
23 | * supported and proper types are used in the extensions themselves. */ | |
24 | #define PyInt_Check PyLong_Check |
|
24 | #define PyInt_Check PyLong_Check | |
25 | #define PyInt_FromLong PyLong_FromLong |
|
25 | #define PyInt_FromLong PyLong_FromLong | |
26 | #define PyInt_FromSsize_t PyLong_FromSsize_t |
|
26 | #define PyInt_FromSsize_t PyLong_FromSsize_t | |
27 | #define PyInt_AsLong PyLong_AsLong |
|
27 | #define PyInt_AsLong PyLong_AsLong | |
28 | #endif |
|
28 | #endif | |
29 |
|
29 | |||
30 | static const char *const versionerrortext = "Python minor version mismatch"; |
|
30 | static const char *const versionerrortext = "Python minor version mismatch"; | |
31 |
|
31 | |||
32 | static PyObject *dict_new_presized(PyObject *self, PyObject *args) |
|
32 | static PyObject *dict_new_presized(PyObject *self, PyObject *args) | |
33 | { |
|
33 | { | |
34 | Py_ssize_t expected_size; |
|
34 | Py_ssize_t expected_size; | |
35 |
|
35 | |||
36 | if (!PyArg_ParseTuple(args, "n:make_presized_dict", &expected_size)) { |
|
36 | if (!PyArg_ParseTuple(args, "n:make_presized_dict", &expected_size)) { | |
37 | return NULL; |
|
37 | return NULL; | |
38 | } |
|
38 | } | |
39 |
|
39 | |||
40 | return _dict_new_presized(expected_size); |
|
40 | return _dict_new_presized(expected_size); | |
41 | } |
|
41 | } | |
42 |
|
42 | |||
43 | static inline dirstateTupleObject *make_dirstate_tuple(char state, int mode, |
|
43 | static inline dirstateTupleObject *make_dirstate_tuple(char state, int mode, | |
44 | int size, int mtime) |
|
44 | int size, int mtime) | |
45 | { |
|
45 | { | |
46 | dirstateTupleObject *t = |
|
46 | dirstateTupleObject *t = | |
47 | PyObject_New(dirstateTupleObject, &dirstateTupleType); |
|
47 | PyObject_New(dirstateTupleObject, &dirstateTupleType); | |
48 | if (!t) { |
|
48 | if (!t) { | |
49 | return NULL; |
|
49 | return NULL; | |
50 | } |
|
50 | } | |
51 | t->state = state; |
|
51 | t->state = state; | |
52 | t->mode = mode; |
|
52 | t->mode = mode; | |
53 | t->size = size; |
|
53 | t->size = size; | |
54 | t->mtime = mtime; |
|
54 | t->mtime = mtime; | |
55 | return t; |
|
55 | return t; | |
56 | } |
|
56 | } | |
57 |
|
57 | |||
58 | static PyObject *dirstate_tuple_new(PyTypeObject *subtype, PyObject *args, |
|
58 | static PyObject *dirstate_tuple_new(PyTypeObject *subtype, PyObject *args, | |
59 | PyObject *kwds) |
|
59 | PyObject *kwds) | |
60 | { |
|
60 | { | |
61 | /* We do all the initialization here and not a tp_init function because |
|
61 | /* We do all the initialization here and not a tp_init function because | |
62 | * dirstate_tuple is immutable. */ |
|
62 | * dirstate_tuple is immutable. */ | |
63 | dirstateTupleObject *t; |
|
63 | dirstateTupleObject *t; | |
64 | char state; |
|
64 | char state; | |
65 | int size, mode, mtime; |
|
65 | int size, mode, mtime; | |
66 | if (!PyArg_ParseTuple(args, "ciii", &state, &mode, &size, &mtime)) { |
|
66 | if (!PyArg_ParseTuple(args, "ciii", &state, &mode, &size, &mtime)) { | |
67 | return NULL; |
|
67 | return NULL; | |
68 | } |
|
68 | } | |
69 |
|
69 | |||
70 | t = (dirstateTupleObject *)subtype->tp_alloc(subtype, 1); |
|
70 | t = (dirstateTupleObject *)subtype->tp_alloc(subtype, 1); | |
71 | if (!t) { |
|
71 | if (!t) { | |
72 | return NULL; |
|
72 | return NULL; | |
73 | } |
|
73 | } | |
74 | t->state = state; |
|
74 | t->state = state; | |
75 | t->mode = mode; |
|
75 | t->mode = mode; | |
76 | t->size = size; |
|
76 | t->size = size; | |
77 | t->mtime = mtime; |
|
77 | t->mtime = mtime; | |
78 |
|
78 | |||
79 | return (PyObject *)t; |
|
79 | return (PyObject *)t; | |
80 | } |
|
80 | } | |
81 |
|
81 | |||
82 | static void dirstate_tuple_dealloc(PyObject *o) |
|
82 | static void dirstate_tuple_dealloc(PyObject *o) | |
83 | { |
|
83 | { | |
84 | PyObject_Del(o); |
|
84 | PyObject_Del(o); | |
85 | } |
|
85 | } | |
86 |
|
86 | |||
87 | static Py_ssize_t dirstate_tuple_length(PyObject *o) |
|
87 | static Py_ssize_t dirstate_tuple_length(PyObject *o) | |
88 | { |
|
88 | { | |
89 | return 4; |
|
89 | return 4; | |
90 | } |
|
90 | } | |
91 |
|
91 | |||
92 | static PyObject *dirstate_tuple_item(PyObject *o, Py_ssize_t i) |
|
92 | static PyObject *dirstate_tuple_item(PyObject *o, Py_ssize_t i) | |
93 | { |
|
93 | { | |
94 | dirstateTupleObject *t = (dirstateTupleObject *)o; |
|
94 | dirstateTupleObject *t = (dirstateTupleObject *)o; | |
95 | switch (i) { |
|
95 | switch (i) { | |
96 | case 0: |
|
96 | case 0: | |
97 | return PyBytes_FromStringAndSize(&t->state, 1); |
|
97 | return PyBytes_FromStringAndSize(&t->state, 1); | |
98 | case 1: |
|
98 | case 1: | |
99 | return PyInt_FromLong(t->mode); |
|
99 | return PyInt_FromLong(t->mode); | |
100 | case 2: |
|
100 | case 2: | |
101 | return PyInt_FromLong(t->size); |
|
101 | return PyInt_FromLong(t->size); | |
102 | case 3: |
|
102 | case 3: | |
103 | return PyInt_FromLong(t->mtime); |
|
103 | return PyInt_FromLong(t->mtime); | |
104 | default: |
|
104 | default: | |
105 | PyErr_SetString(PyExc_IndexError, "index out of range"); |
|
105 | PyErr_SetString(PyExc_IndexError, "index out of range"); | |
106 | return NULL; |
|
106 | return NULL; | |
107 | } |
|
107 | } | |
108 | } |
|
108 | } | |
109 |
|
109 | |||
110 | static PySequenceMethods dirstate_tuple_sq = { |
|
110 | static PySequenceMethods dirstate_tuple_sq = { | |
111 | dirstate_tuple_length, /* sq_length */ |
|
111 | dirstate_tuple_length, /* sq_length */ | |
112 | 0, /* sq_concat */ |
|
112 | 0, /* sq_concat */ | |
113 | 0, /* sq_repeat */ |
|
113 | 0, /* sq_repeat */ | |
114 | dirstate_tuple_item, /* sq_item */ |
|
114 | dirstate_tuple_item, /* sq_item */ | |
115 | 0, /* sq_ass_item */ |
|
115 | 0, /* sq_ass_item */ | |
116 | 0, /* sq_contains */ |
|
116 | 0, /* sq_contains */ | |
117 | 0, /* sq_inplace_concat */ |
|
117 | 0, /* sq_inplace_concat */ | |
118 | 0 /* sq_inplace_repeat */ |
|
118 | 0 /* sq_inplace_repeat */ | |
119 | }; |
|
119 | }; | |
120 |
|
120 | |||
121 | PyTypeObject dirstateTupleType = { |
|
121 | PyTypeObject dirstateTupleType = { | |
122 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ |
|
122 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ | |
123 | "dirstate_tuple", /* tp_name */ |
|
123 | "dirstate_tuple", /* tp_name */ | |
124 | sizeof(dirstateTupleObject), /* tp_basicsize */ |
|
124 | sizeof(dirstateTupleObject), /* tp_basicsize */ | |
125 | 0, /* tp_itemsize */ |
|
125 | 0, /* tp_itemsize */ | |
126 | (destructor)dirstate_tuple_dealloc, /* tp_dealloc */ |
|
126 | (destructor)dirstate_tuple_dealloc, /* tp_dealloc */ | |
127 | 0, /* tp_print */ |
|
127 | 0, /* tp_print */ | |
128 | 0, /* tp_getattr */ |
|
128 | 0, /* tp_getattr */ | |
129 | 0, /* tp_setattr */ |
|
129 | 0, /* tp_setattr */ | |
130 | 0, /* tp_compare */ |
|
130 | 0, /* tp_compare */ | |
131 | 0, /* tp_repr */ |
|
131 | 0, /* tp_repr */ | |
132 | 0, /* tp_as_number */ |
|
132 | 0, /* tp_as_number */ | |
133 | &dirstate_tuple_sq, /* tp_as_sequence */ |
|
133 | &dirstate_tuple_sq, /* tp_as_sequence */ | |
134 | 0, /* tp_as_mapping */ |
|
134 | 0, /* tp_as_mapping */ | |
135 | 0, /* tp_hash */ |
|
135 | 0, /* tp_hash */ | |
136 | 0, /* tp_call */ |
|
136 | 0, /* tp_call */ | |
137 | 0, /* tp_str */ |
|
137 | 0, /* tp_str */ | |
138 | 0, /* tp_getattro */ |
|
138 | 0, /* tp_getattro */ | |
139 | 0, /* tp_setattro */ |
|
139 | 0, /* tp_setattro */ | |
140 | 0, /* tp_as_buffer */ |
|
140 | 0, /* tp_as_buffer */ | |
141 | Py_TPFLAGS_DEFAULT, /* tp_flags */ |
|
141 | Py_TPFLAGS_DEFAULT, /* tp_flags */ | |
142 | "dirstate tuple", /* tp_doc */ |
|
142 | "dirstate tuple", /* tp_doc */ | |
143 | 0, /* tp_traverse */ |
|
143 | 0, /* tp_traverse */ | |
144 | 0, /* tp_clear */ |
|
144 | 0, /* tp_clear */ | |
145 | 0, /* tp_richcompare */ |
|
145 | 0, /* tp_richcompare */ | |
146 | 0, /* tp_weaklistoffset */ |
|
146 | 0, /* tp_weaklistoffset */ | |
147 | 0, /* tp_iter */ |
|
147 | 0, /* tp_iter */ | |
148 | 0, /* tp_iternext */ |
|
148 | 0, /* tp_iternext */ | |
149 | 0, /* tp_methods */ |
|
149 | 0, /* tp_methods */ | |
150 | 0, /* tp_members */ |
|
150 | 0, /* tp_members */ | |
151 | 0, /* tp_getset */ |
|
151 | 0, /* tp_getset */ | |
152 | 0, /* tp_base */ |
|
152 | 0, /* tp_base */ | |
153 | 0, /* tp_dict */ |
|
153 | 0, /* tp_dict */ | |
154 | 0, /* tp_descr_get */ |
|
154 | 0, /* tp_descr_get */ | |
155 | 0, /* tp_descr_set */ |
|
155 | 0, /* tp_descr_set */ | |
156 | 0, /* tp_dictoffset */ |
|
156 | 0, /* tp_dictoffset */ | |
157 | 0, /* tp_init */ |
|
157 | 0, /* tp_init */ | |
158 | 0, /* tp_alloc */ |
|
158 | 0, /* tp_alloc */ | |
159 | dirstate_tuple_new, /* tp_new */ |
|
159 | dirstate_tuple_new, /* tp_new */ | |
160 | }; |
|
160 | }; | |
161 |
|
161 | |||
162 | static PyObject *parse_dirstate(PyObject *self, PyObject *args) |
|
162 | static PyObject *parse_dirstate(PyObject *self, PyObject *args) | |
163 | { |
|
163 | { | |
164 | PyObject *dmap, *cmap, *parents = NULL, *ret = NULL; |
|
164 | PyObject *dmap, *cmap, *parents = NULL, *ret = NULL; | |
165 | PyObject *fname = NULL, *cname = NULL, *entry = NULL; |
|
165 | PyObject *fname = NULL, *cname = NULL, *entry = NULL; | |
166 | char state, *cur, *str, *cpos; |
|
166 | char state, *cur, *str, *cpos; | |
167 | int mode, size, mtime; |
|
167 | int mode, size, mtime; | |
168 | unsigned int flen, pos = 40; |
|
168 | unsigned int flen, pos = 40; | |
169 | Py_ssize_t len = 40; |
|
169 | Py_ssize_t len = 40; | |
170 | Py_ssize_t readlen; |
|
170 | Py_ssize_t readlen; | |
171 |
|
171 | |||
172 | if (!PyArg_ParseTuple( |
|
172 | if (!PyArg_ParseTuple( | |
173 | args, PY23("O!O!s#:parse_dirstate", "O!O!y#:parse_dirstate"), |
|
173 | args, PY23("O!O!s#:parse_dirstate", "O!O!y#:parse_dirstate"), | |
174 | &PyDict_Type, &dmap, &PyDict_Type, &cmap, &str, &readlen)) { |
|
174 | &PyDict_Type, &dmap, &PyDict_Type, &cmap, &str, &readlen)) { | |
175 | goto quit; |
|
175 | goto quit; | |
176 | } |
|
176 | } | |
177 |
|
177 | |||
178 | len = readlen; |
|
178 | len = readlen; | |
179 |
|
179 | |||
180 | /* read parents */ |
|
180 | /* read parents */ | |
181 | if (len < 40) { |
|
181 | if (len < 40) { | |
182 | PyErr_SetString(PyExc_ValueError, |
|
182 | PyErr_SetString(PyExc_ValueError, | |
183 | "too little data for parents"); |
|
183 | "too little data for parents"); | |
184 | goto quit; |
|
184 | goto quit; | |
185 | } |
|
185 | } | |
186 |
|
186 | |||
187 | parents = Py_BuildValue(PY23("s#s#", "y#y#"), str, (Py_ssize_t)20, |
|
187 | parents = Py_BuildValue(PY23("s#s#", "y#y#"), str, (Py_ssize_t)20, | |
188 | str + 20, (Py_ssize_t)20); |
|
188 | str + 20, (Py_ssize_t)20); | |
189 | if (!parents) { |
|
189 | if (!parents) { | |
190 | goto quit; |
|
190 | goto quit; | |
191 | } |
|
191 | } | |
192 |
|
192 | |||
193 | /* read filenames */ |
|
193 | /* read filenames */ | |
194 | while (pos >= 40 && pos < len) { |
|
194 | while (pos >= 40 && pos < len) { | |
195 | if (pos + 17 > len) { |
|
195 | if (pos + 17 > len) { | |
196 | PyErr_SetString(PyExc_ValueError, |
|
196 | PyErr_SetString(PyExc_ValueError, | |
197 | "overflow in dirstate"); |
|
197 | "overflow in dirstate"); | |
198 | goto quit; |
|
198 | goto quit; | |
199 | } |
|
199 | } | |
200 | cur = str + pos; |
|
200 | cur = str + pos; | |
201 | /* unpack header */ |
|
201 | /* unpack header */ | |
202 | state = *cur; |
|
202 | state = *cur; | |
203 | mode = getbe32(cur + 1); |
|
203 | mode = getbe32(cur + 1); | |
204 | size = getbe32(cur + 5); |
|
204 | size = getbe32(cur + 5); | |
205 | mtime = getbe32(cur + 9); |
|
205 | mtime = getbe32(cur + 9); | |
206 | flen = getbe32(cur + 13); |
|
206 | flen = getbe32(cur + 13); | |
207 | pos += 17; |
|
207 | pos += 17; | |
208 | cur += 17; |
|
208 | cur += 17; | |
209 | if (flen > len - pos) { |
|
209 | if (flen > len - pos) { | |
210 | PyErr_SetString(PyExc_ValueError, |
|
210 | PyErr_SetString(PyExc_ValueError, | |
211 | "overflow in dirstate"); |
|
211 | "overflow in dirstate"); | |
212 | goto quit; |
|
212 | goto quit; | |
213 | } |
|
213 | } | |
214 |
|
214 | |||
215 | entry = |
|
215 | entry = | |
216 | (PyObject *)make_dirstate_tuple(state, mode, size, mtime); |
|
216 | (PyObject *)make_dirstate_tuple(state, mode, size, mtime); | |
217 | cpos = memchr(cur, 0, flen); |
|
217 | cpos = memchr(cur, 0, flen); | |
218 | if (cpos) { |
|
218 | if (cpos) { | |
219 | fname = PyBytes_FromStringAndSize(cur, cpos - cur); |
|
219 | fname = PyBytes_FromStringAndSize(cur, cpos - cur); | |
220 | cname = PyBytes_FromStringAndSize( |
|
220 | cname = PyBytes_FromStringAndSize( | |
221 | cpos + 1, flen - (cpos - cur) - 1); |
|
221 | cpos + 1, flen - (cpos - cur) - 1); | |
222 | if (!fname || !cname || |
|
222 | if (!fname || !cname || | |
223 | PyDict_SetItem(cmap, fname, cname) == -1 || |
|
223 | PyDict_SetItem(cmap, fname, cname) == -1 || | |
224 | PyDict_SetItem(dmap, fname, entry) == -1) { |
|
224 | PyDict_SetItem(dmap, fname, entry) == -1) { | |
225 | goto quit; |
|
225 | goto quit; | |
226 | } |
|
226 | } | |
227 | Py_DECREF(cname); |
|
227 | Py_DECREF(cname); | |
228 | } else { |
|
228 | } else { | |
229 | fname = PyBytes_FromStringAndSize(cur, flen); |
|
229 | fname = PyBytes_FromStringAndSize(cur, flen); | |
230 | if (!fname || |
|
230 | if (!fname || | |
231 | PyDict_SetItem(dmap, fname, entry) == -1) { |
|
231 | PyDict_SetItem(dmap, fname, entry) == -1) { | |
232 | goto quit; |
|
232 | goto quit; | |
233 | } |
|
233 | } | |
234 | } |
|
234 | } | |
235 | Py_DECREF(fname); |
|
235 | Py_DECREF(fname); | |
236 | Py_DECREF(entry); |
|
236 | Py_DECREF(entry); | |
237 | fname = cname = entry = NULL; |
|
237 | fname = cname = entry = NULL; | |
238 | pos += flen; |
|
238 | pos += flen; | |
239 | } |
|
239 | } | |
240 |
|
240 | |||
241 | ret = parents; |
|
241 | ret = parents; | |
242 | Py_INCREF(ret); |
|
242 | Py_INCREF(ret); | |
243 | quit: |
|
243 | quit: | |
244 | Py_XDECREF(fname); |
|
244 | Py_XDECREF(fname); | |
245 | Py_XDECREF(cname); |
|
245 | Py_XDECREF(cname); | |
246 | Py_XDECREF(entry); |
|
246 | Py_XDECREF(entry); | |
247 | Py_XDECREF(parents); |
|
247 | Py_XDECREF(parents); | |
248 | return ret; |
|
248 | return ret; | |
249 | } |
|
249 | } | |
250 |
|
250 | |||
251 | /* |
|
251 | /* | |
252 | * Build a set of non-normal and other parent entries from the dirstate dmap |
|
252 | * Build a set of non-normal and other parent entries from the dirstate dmap | |
253 | */ |
|
253 | */ | |
254 | static PyObject *nonnormalotherparententries(PyObject *self, PyObject *args) |
|
254 | static PyObject *nonnormalotherparententries(PyObject *self, PyObject *args) | |
255 | { |
|
255 | { | |
256 | PyObject *dmap, *fname, *v; |
|
256 | PyObject *dmap, *fname, *v; | |
257 | PyObject *nonnset = NULL, *otherpset = NULL, *result = NULL; |
|
257 | PyObject *nonnset = NULL, *otherpset = NULL, *result = NULL; | |
258 | Py_ssize_t pos; |
|
258 | Py_ssize_t pos; | |
259 |
|
259 | |||
260 | if (!PyArg_ParseTuple(args, "O!:nonnormalentries", &PyDict_Type, |
|
260 | if (!PyArg_ParseTuple(args, "O!:nonnormalentries", &PyDict_Type, | |
261 | &dmap)) { |
|
261 | &dmap)) { | |
262 | goto bail; |
|
262 | goto bail; | |
263 | } |
|
263 | } | |
264 |
|
264 | |||
265 | nonnset = PySet_New(NULL); |
|
265 | nonnset = PySet_New(NULL); | |
266 | if (nonnset == NULL) { |
|
266 | if (nonnset == NULL) { | |
267 | goto bail; |
|
267 | goto bail; | |
268 | } |
|
268 | } | |
269 |
|
269 | |||
270 | otherpset = PySet_New(NULL); |
|
270 | otherpset = PySet_New(NULL); | |
271 | if (otherpset == NULL) { |
|
271 | if (otherpset == NULL) { | |
272 | goto bail; |
|
272 | goto bail; | |
273 | } |
|
273 | } | |
274 |
|
274 | |||
275 | pos = 0; |
|
275 | pos = 0; | |
276 | while (PyDict_Next(dmap, &pos, &fname, &v)) { |
|
276 | while (PyDict_Next(dmap, &pos, &fname, &v)) { | |
277 | dirstateTupleObject *t; |
|
277 | dirstateTupleObject *t; | |
278 | if (!dirstate_tuple_check(v)) { |
|
278 | if (!dirstate_tuple_check(v)) { | |
279 | PyErr_SetString(PyExc_TypeError, |
|
279 | PyErr_SetString(PyExc_TypeError, | |
280 | "expected a dirstate tuple"); |
|
280 | "expected a dirstate tuple"); | |
281 | goto bail; |
|
281 | goto bail; | |
282 | } |
|
282 | } | |
283 | t = (dirstateTupleObject *)v; |
|
283 | t = (dirstateTupleObject *)v; | |
284 |
|
284 | |||
285 | if (t->state == 'n' && t->size == -2) { |
|
285 | if (t->state == 'n' && t->size == -2) { | |
286 | if (PySet_Add(otherpset, fname) == -1) { |
|
286 | if (PySet_Add(otherpset, fname) == -1) { | |
287 | goto bail; |
|
287 | goto bail; | |
288 | } |
|
288 | } | |
289 | } |
|
289 | } | |
290 |
|
290 | |||
291 | if (t->state == 'n' && t->mtime != -1) { |
|
291 | if (t->state == 'n' && t->mtime != -1) { | |
292 | continue; |
|
292 | continue; | |
293 | } |
|
293 | } | |
294 | if (PySet_Add(nonnset, fname) == -1) { |
|
294 | if (PySet_Add(nonnset, fname) == -1) { | |
295 | goto bail; |
|
295 | goto bail; | |
296 | } |
|
296 | } | |
297 | } |
|
297 | } | |
298 |
|
298 | |||
299 | result = Py_BuildValue("(OO)", nonnset, otherpset); |
|
299 | result = Py_BuildValue("(OO)", nonnset, otherpset); | |
300 | if (result == NULL) { |
|
300 | if (result == NULL) { | |
301 | goto bail; |
|
301 | goto bail; | |
302 | } |
|
302 | } | |
303 | Py_DECREF(nonnset); |
|
303 | Py_DECREF(nonnset); | |
304 | Py_DECREF(otherpset); |
|
304 | Py_DECREF(otherpset); | |
305 | return result; |
|
305 | return result; | |
306 | bail: |
|
306 | bail: | |
307 | Py_XDECREF(nonnset); |
|
307 | Py_XDECREF(nonnset); | |
308 | Py_XDECREF(otherpset); |
|
308 | Py_XDECREF(otherpset); | |
309 | Py_XDECREF(result); |
|
309 | Py_XDECREF(result); | |
310 | return NULL; |
|
310 | return NULL; | |
311 | } |
|
311 | } | |
312 |
|
312 | |||
313 | /* |
|
313 | /* | |
314 | * Efficiently pack a dirstate object into its on-disk format. |
|
314 | * Efficiently pack a dirstate object into its on-disk format. | |
315 | */ |
|
315 | */ | |
316 | static PyObject *pack_dirstate(PyObject *self, PyObject *args) |
|
316 | static PyObject *pack_dirstate(PyObject *self, PyObject *args) | |
317 | { |
|
317 | { | |
318 | PyObject *packobj = NULL; |
|
318 | PyObject *packobj = NULL; | |
319 | PyObject *map, *copymap, *pl, *mtime_unset = NULL; |
|
319 | PyObject *map, *copymap, *pl, *mtime_unset = NULL; | |
320 | Py_ssize_t nbytes, pos, l; |
|
320 | Py_ssize_t nbytes, pos, l; | |
321 | PyObject *k, *v = NULL, *pn; |
|
321 | PyObject *k, *v = NULL, *pn; | |
322 | char *p, *s; |
|
322 | char *p, *s; | |
323 | int now; |
|
323 | int now; | |
324 |
|
324 | |||
325 | if (!PyArg_ParseTuple(args, "O!O!O!i:pack_dirstate", &PyDict_Type, &map, |
|
325 | if (!PyArg_ParseTuple(args, "O!O!O!i:pack_dirstate", &PyDict_Type, &map, | |
326 | &PyDict_Type, ©map, &PyTuple_Type, &pl, |
|
326 | &PyDict_Type, ©map, &PyTuple_Type, &pl, | |
327 | &now)) { |
|
327 | &now)) { | |
328 | return NULL; |
|
328 | return NULL; | |
329 | } |
|
329 | } | |
330 |
|
330 | |||
331 | if (PyTuple_Size(pl) != 2) { |
|
331 | if (PyTuple_Size(pl) != 2) { | |
332 | PyErr_SetString(PyExc_TypeError, "expected 2-element tuple"); |
|
332 | PyErr_SetString(PyExc_TypeError, "expected 2-element tuple"); | |
333 | return NULL; |
|
333 | return NULL; | |
334 | } |
|
334 | } | |
335 |
|
335 | |||
336 | /* Figure out how much we need to allocate. */ |
|
336 | /* Figure out how much we need to allocate. */ | |
337 | for (nbytes = 40, pos = 0; PyDict_Next(map, &pos, &k, &v);) { |
|
337 | for (nbytes = 40, pos = 0; PyDict_Next(map, &pos, &k, &v);) { | |
338 | PyObject *c; |
|
338 | PyObject *c; | |
339 | if (!PyBytes_Check(k)) { |
|
339 | if (!PyBytes_Check(k)) { | |
340 | PyErr_SetString(PyExc_TypeError, "expected string key"); |
|
340 | PyErr_SetString(PyExc_TypeError, "expected string key"); | |
341 | goto bail; |
|
341 | goto bail; | |
342 | } |
|
342 | } | |
343 | nbytes += PyBytes_GET_SIZE(k) + 17; |
|
343 | nbytes += PyBytes_GET_SIZE(k) + 17; | |
344 | c = PyDict_GetItem(copymap, k); |
|
344 | c = PyDict_GetItem(copymap, k); | |
345 | if (c) { |
|
345 | if (c) { | |
346 | if (!PyBytes_Check(c)) { |
|
346 | if (!PyBytes_Check(c)) { | |
347 | PyErr_SetString(PyExc_TypeError, |
|
347 | PyErr_SetString(PyExc_TypeError, | |
348 | "expected string key"); |
|
348 | "expected string key"); | |
349 | goto bail; |
|
349 | goto bail; | |
350 | } |
|
350 | } | |
351 | nbytes += PyBytes_GET_SIZE(c) + 1; |
|
351 | nbytes += PyBytes_GET_SIZE(c) + 1; | |
352 | } |
|
352 | } | |
353 | } |
|
353 | } | |
354 |
|
354 | |||
355 | packobj = PyBytes_FromStringAndSize(NULL, nbytes); |
|
355 | packobj = PyBytes_FromStringAndSize(NULL, nbytes); | |
356 | if (packobj == NULL) { |
|
356 | if (packobj == NULL) { | |
357 | goto bail; |
|
357 | goto bail; | |
358 | } |
|
358 | } | |
359 |
|
359 | |||
360 | p = PyBytes_AS_STRING(packobj); |
|
360 | p = PyBytes_AS_STRING(packobj); | |
361 |
|
361 | |||
362 | pn = PyTuple_GET_ITEM(pl, 0); |
|
362 | pn = PyTuple_GET_ITEM(pl, 0); | |
363 | if (PyBytes_AsStringAndSize(pn, &s, &l) == -1 || l != 20) { |
|
363 | if (PyBytes_AsStringAndSize(pn, &s, &l) == -1 || l != 20) { | |
364 | PyErr_SetString(PyExc_TypeError, "expected a 20-byte hash"); |
|
364 | PyErr_SetString(PyExc_TypeError, "expected a 20-byte hash"); | |
365 | goto bail; |
|
365 | goto bail; | |
366 | } |
|
366 | } | |
367 | memcpy(p, s, l); |
|
367 | memcpy(p, s, l); | |
368 | p += 20; |
|
368 | p += 20; | |
369 | pn = PyTuple_GET_ITEM(pl, 1); |
|
369 | pn = PyTuple_GET_ITEM(pl, 1); | |
370 | if (PyBytes_AsStringAndSize(pn, &s, &l) == -1 || l != 20) { |
|
370 | if (PyBytes_AsStringAndSize(pn, &s, &l) == -1 || l != 20) { | |
371 | PyErr_SetString(PyExc_TypeError, "expected a 20-byte hash"); |
|
371 | PyErr_SetString(PyExc_TypeError, "expected a 20-byte hash"); | |
372 | goto bail; |
|
372 | goto bail; | |
373 | } |
|
373 | } | |
374 | memcpy(p, s, l); |
|
374 | memcpy(p, s, l); | |
375 | p += 20; |
|
375 | p += 20; | |
376 |
|
376 | |||
377 | for (pos = 0; PyDict_Next(map, &pos, &k, &v);) { |
|
377 | for (pos = 0; PyDict_Next(map, &pos, &k, &v);) { | |
378 | dirstateTupleObject *tuple; |
|
378 | dirstateTupleObject *tuple; | |
379 | char state; |
|
379 | char state; | |
380 | int mode, size, mtime; |
|
380 | int mode, size, mtime; | |
381 | Py_ssize_t len, l; |
|
381 | Py_ssize_t len, l; | |
382 | PyObject *o; |
|
382 | PyObject *o; | |
383 | char *t; |
|
383 | char *t; | |
384 |
|
384 | |||
385 | if (!dirstate_tuple_check(v)) { |
|
385 | if (!dirstate_tuple_check(v)) { | |
386 | PyErr_SetString(PyExc_TypeError, |
|
386 | PyErr_SetString(PyExc_TypeError, | |
387 | "expected a dirstate tuple"); |
|
387 | "expected a dirstate tuple"); | |
388 | goto bail; |
|
388 | goto bail; | |
389 | } |
|
389 | } | |
390 | tuple = (dirstateTupleObject *)v; |
|
390 | tuple = (dirstateTupleObject *)v; | |
391 |
|
391 | |||
392 | state = tuple->state; |
|
392 | state = tuple->state; | |
393 | mode = tuple->mode; |
|
393 | mode = tuple->mode; | |
394 | size = tuple->size; |
|
394 | size = tuple->size; | |
395 | mtime = tuple->mtime; |
|
395 | mtime = tuple->mtime; | |
396 | if (state == 'n' && mtime == now) { |
|
396 | if (state == 'n' && mtime == now) { | |
397 | /* See pure/parsers.py:pack_dirstate for why we do |
|
397 | /* See pure/parsers.py:pack_dirstate for why we do | |
398 | * this. */ |
|
398 | * this. */ | |
399 | mtime = -1; |
|
399 | mtime = -1; | |
400 | mtime_unset = (PyObject *)make_dirstate_tuple( |
|
400 | mtime_unset = (PyObject *)make_dirstate_tuple( | |
401 | state, mode, size, mtime); |
|
401 | state, mode, size, mtime); | |
402 | if (!mtime_unset) { |
|
402 | if (!mtime_unset) { | |
403 | goto bail; |
|
403 | goto bail; | |
404 | } |
|
404 | } | |
405 | if (PyDict_SetItem(map, k, mtime_unset) == -1) { |
|
405 | if (PyDict_SetItem(map, k, mtime_unset) == -1) { | |
406 | goto bail; |
|
406 | goto bail; | |
407 | } |
|
407 | } | |
408 | Py_DECREF(mtime_unset); |
|
408 | Py_DECREF(mtime_unset); | |
409 | mtime_unset = NULL; |
|
409 | mtime_unset = NULL; | |
410 | } |
|
410 | } | |
411 | *p++ = state; |
|
411 | *p++ = state; | |
412 | putbe32((uint32_t)mode, p); |
|
412 | putbe32((uint32_t)mode, p); | |
413 | putbe32((uint32_t)size, p + 4); |
|
413 | putbe32((uint32_t)size, p + 4); | |
414 | putbe32((uint32_t)mtime, p + 8); |
|
414 | putbe32((uint32_t)mtime, p + 8); | |
415 | t = p + 12; |
|
415 | t = p + 12; | |
416 | p += 16; |
|
416 | p += 16; | |
417 | len = PyBytes_GET_SIZE(k); |
|
417 | len = PyBytes_GET_SIZE(k); | |
418 | memcpy(p, PyBytes_AS_STRING(k), len); |
|
418 | memcpy(p, PyBytes_AS_STRING(k), len); | |
419 | p += len; |
|
419 | p += len; | |
420 | o = PyDict_GetItem(copymap, k); |
|
420 | o = PyDict_GetItem(copymap, k); | |
421 | if (o) { |
|
421 | if (o) { | |
422 | *p++ = '\0'; |
|
422 | *p++ = '\0'; | |
423 | l = PyBytes_GET_SIZE(o); |
|
423 | l = PyBytes_GET_SIZE(o); | |
424 | memcpy(p, PyBytes_AS_STRING(o), l); |
|
424 | memcpy(p, PyBytes_AS_STRING(o), l); | |
425 | p += l; |
|
425 | p += l; | |
426 | len += l + 1; |
|
426 | len += l + 1; | |
427 | } |
|
427 | } | |
428 | putbe32((uint32_t)len, t); |
|
428 | putbe32((uint32_t)len, t); | |
429 | } |
|
429 | } | |
430 |
|
430 | |||
431 | pos = p - PyBytes_AS_STRING(packobj); |
|
431 | pos = p - PyBytes_AS_STRING(packobj); | |
432 | if (pos != nbytes) { |
|
432 | if (pos != nbytes) { | |
433 | PyErr_Format(PyExc_SystemError, "bad dirstate size: %ld != %ld", |
|
433 | PyErr_Format(PyExc_SystemError, "bad dirstate size: %ld != %ld", | |
434 | (long)pos, (long)nbytes); |
|
434 | (long)pos, (long)nbytes); | |
435 | goto bail; |
|
435 | goto bail; | |
436 | } |
|
436 | } | |
437 |
|
437 | |||
438 | return packobj; |
|
438 | return packobj; | |
439 | bail: |
|
439 | bail: | |
440 | Py_XDECREF(mtime_unset); |
|
440 | Py_XDECREF(mtime_unset); | |
441 | Py_XDECREF(packobj); |
|
441 | Py_XDECREF(packobj); | |
442 | Py_XDECREF(v); |
|
442 | Py_XDECREF(v); | |
443 | return NULL; |
|
443 | return NULL; | |
444 | } |
|
444 | } | |
445 |
|
445 | |||
446 | #define BUMPED_FIX 1 |
|
446 | #define BUMPED_FIX 1 | |
447 | #define USING_SHA_256 2 |
|
447 | #define USING_SHA_256 2 | |
448 | #define FM1_HEADER_SIZE (4 + 8 + 2 + 2 + 1 + 1 + 1) |
|
448 | #define FM1_HEADER_SIZE (4 + 8 + 2 + 2 + 1 + 1 + 1) | |
449 |
|
449 | |||
450 | static PyObject *readshas(const char *source, unsigned char num, |
|
450 | static PyObject *readshas(const char *source, unsigned char num, | |
451 | Py_ssize_t hashwidth) |
|
451 | Py_ssize_t hashwidth) | |
452 | { |
|
452 | { | |
453 | int i; |
|
453 | int i; | |
454 | PyObject *list = PyTuple_New(num); |
|
454 | PyObject *list = PyTuple_New(num); | |
455 | if (list == NULL) { |
|
455 | if (list == NULL) { | |
456 | return NULL; |
|
456 | return NULL; | |
457 | } |
|
457 | } | |
458 | for (i = 0; i < num; i++) { |
|
458 | for (i = 0; i < num; i++) { | |
459 | PyObject *hash = PyBytes_FromStringAndSize(source, hashwidth); |
|
459 | PyObject *hash = PyBytes_FromStringAndSize(source, hashwidth); | |
460 | if (hash == NULL) { |
|
460 | if (hash == NULL) { | |
461 | Py_DECREF(list); |
|
461 | Py_DECREF(list); | |
462 | return NULL; |
|
462 | return NULL; | |
463 | } |
|
463 | } | |
464 | PyTuple_SET_ITEM(list, i, hash); |
|
464 | PyTuple_SET_ITEM(list, i, hash); | |
465 | source += hashwidth; |
|
465 | source += hashwidth; | |
466 | } |
|
466 | } | |
467 | return list; |
|
467 | return list; | |
468 | } |
|
468 | } | |
469 |
|
469 | |||
470 | static PyObject *fm1readmarker(const char *databegin, const char *dataend, |
|
470 | static PyObject *fm1readmarker(const char *databegin, const char *dataend, | |
471 | uint32_t *msize) |
|
471 | uint32_t *msize) | |
472 | { |
|
472 | { | |
473 | const char *data = databegin; |
|
473 | const char *data = databegin; | |
474 | const char *meta; |
|
474 | const char *meta; | |
475 |
|
475 | |||
476 | double mtime; |
|
476 | double mtime; | |
477 | int16_t tz; |
|
477 | int16_t tz; | |
478 | uint16_t flags; |
|
478 | uint16_t flags; | |
479 | unsigned char nsuccs, nparents, nmetadata; |
|
479 | unsigned char nsuccs, nparents, nmetadata; | |
480 | Py_ssize_t hashwidth = 20; |
|
480 | Py_ssize_t hashwidth = 20; | |
481 |
|
481 | |||
482 | PyObject *prec = NULL, *parents = NULL, *succs = NULL; |
|
482 | PyObject *prec = NULL, *parents = NULL, *succs = NULL; | |
483 | PyObject *metadata = NULL, *ret = NULL; |
|
483 | PyObject *metadata = NULL, *ret = NULL; | |
484 | int i; |
|
484 | int i; | |
485 |
|
485 | |||
486 | if (data + FM1_HEADER_SIZE > dataend) { |
|
486 | if (data + FM1_HEADER_SIZE > dataend) { | |
487 | goto overflow; |
|
487 | goto overflow; | |
488 | } |
|
488 | } | |
489 |
|
489 | |||
490 | *msize = getbe32(data); |
|
490 | *msize = getbe32(data); | |
491 | data += 4; |
|
491 | data += 4; | |
492 | mtime = getbefloat64(data); |
|
492 | mtime = getbefloat64(data); | |
493 | data += 8; |
|
493 | data += 8; | |
494 | tz = getbeint16(data); |
|
494 | tz = getbeint16(data); | |
495 | data += 2; |
|
495 | data += 2; | |
496 | flags = getbeuint16(data); |
|
496 | flags = getbeuint16(data); | |
497 | data += 2; |
|
497 | data += 2; | |
498 |
|
498 | |||
499 | if (flags & USING_SHA_256) { |
|
499 | if (flags & USING_SHA_256) { | |
500 | hashwidth = 32; |
|
500 | hashwidth = 32; | |
501 | } |
|
501 | } | |
502 |
|
502 | |||
503 | nsuccs = (unsigned char)(*data++); |
|
503 | nsuccs = (unsigned char)(*data++); | |
504 | nparents = (unsigned char)(*data++); |
|
504 | nparents = (unsigned char)(*data++); | |
505 | nmetadata = (unsigned char)(*data++); |
|
505 | nmetadata = (unsigned char)(*data++); | |
506 |
|
506 | |||
507 | if (databegin + *msize > dataend) { |
|
507 | if (databegin + *msize > dataend) { | |
508 | goto overflow; |
|
508 | goto overflow; | |
509 | } |
|
509 | } | |
510 | dataend = databegin + *msize; /* narrow down to marker size */ |
|
510 | dataend = databegin + *msize; /* narrow down to marker size */ | |
511 |
|
511 | |||
512 | if (data + hashwidth > dataend) { |
|
512 | if (data + hashwidth > dataend) { | |
513 | goto overflow; |
|
513 | goto overflow; | |
514 | } |
|
514 | } | |
515 | prec = PyBytes_FromStringAndSize(data, hashwidth); |
|
515 | prec = PyBytes_FromStringAndSize(data, hashwidth); | |
516 | data += hashwidth; |
|
516 | data += hashwidth; | |
517 | if (prec == NULL) { |
|
517 | if (prec == NULL) { | |
518 | goto bail; |
|
518 | goto bail; | |
519 | } |
|
519 | } | |
520 |
|
520 | |||
521 | if (data + nsuccs * hashwidth > dataend) { |
|
521 | if (data + nsuccs * hashwidth > dataend) { | |
522 | goto overflow; |
|
522 | goto overflow; | |
523 | } |
|
523 | } | |
524 | succs = readshas(data, nsuccs, hashwidth); |
|
524 | succs = readshas(data, nsuccs, hashwidth); | |
525 | if (succs == NULL) { |
|
525 | if (succs == NULL) { | |
526 | goto bail; |
|
526 | goto bail; | |
527 | } |
|
527 | } | |
528 | data += nsuccs * hashwidth; |
|
528 | data += nsuccs * hashwidth; | |
529 |
|
529 | |||
530 | if (nparents == 1 || nparents == 2) { |
|
530 | if (nparents == 1 || nparents == 2) { | |
531 | if (data + nparents * hashwidth > dataend) { |
|
531 | if (data + nparents * hashwidth > dataend) { | |
532 | goto overflow; |
|
532 | goto overflow; | |
533 | } |
|
533 | } | |
534 | parents = readshas(data, nparents, hashwidth); |
|
534 | parents = readshas(data, nparents, hashwidth); | |
535 | if (parents == NULL) { |
|
535 | if (parents == NULL) { | |
536 | goto bail; |
|
536 | goto bail; | |
537 | } |
|
537 | } | |
538 | data += nparents * hashwidth; |
|
538 | data += nparents * hashwidth; | |
539 | } else { |
|
539 | } else { | |
540 | parents = Py_None; |
|
540 | parents = Py_None; | |
541 | Py_INCREF(parents); |
|
541 | Py_INCREF(parents); | |
542 | } |
|
542 | } | |
543 |
|
543 | |||
544 | if (data + 2 * nmetadata > dataend) { |
|
544 | if (data + 2 * nmetadata > dataend) { | |
545 | goto overflow; |
|
545 | goto overflow; | |
546 | } |
|
546 | } | |
547 | meta = data + (2 * nmetadata); |
|
547 | meta = data + (2 * nmetadata); | |
548 | metadata = PyTuple_New(nmetadata); |
|
548 | metadata = PyTuple_New(nmetadata); | |
549 | if (metadata == NULL) { |
|
549 | if (metadata == NULL) { | |
550 | goto bail; |
|
550 | goto bail; | |
551 | } |
|
551 | } | |
552 | for (i = 0; i < nmetadata; i++) { |
|
552 | for (i = 0; i < nmetadata; i++) { | |
553 | PyObject *tmp, *left = NULL, *right = NULL; |
|
553 | PyObject *tmp, *left = NULL, *right = NULL; | |
554 | Py_ssize_t leftsize = (unsigned char)(*data++); |
|
554 | Py_ssize_t leftsize = (unsigned char)(*data++); | |
555 | Py_ssize_t rightsize = (unsigned char)(*data++); |
|
555 | Py_ssize_t rightsize = (unsigned char)(*data++); | |
556 | if (meta + leftsize + rightsize > dataend) { |
|
556 | if (meta + leftsize + rightsize > dataend) { | |
557 | goto overflow; |
|
557 | goto overflow; | |
558 | } |
|
558 | } | |
559 | left = PyBytes_FromStringAndSize(meta, leftsize); |
|
559 | left = PyBytes_FromStringAndSize(meta, leftsize); | |
560 | meta += leftsize; |
|
560 | meta += leftsize; | |
561 | right = PyBytes_FromStringAndSize(meta, rightsize); |
|
561 | right = PyBytes_FromStringAndSize(meta, rightsize); | |
562 | meta += rightsize; |
|
562 | meta += rightsize; | |
563 | tmp = PyTuple_New(2); |
|
563 | tmp = PyTuple_New(2); | |
564 | if (!left || !right || !tmp) { |
|
564 | if (!left || !right || !tmp) { | |
565 | Py_XDECREF(left); |
|
565 | Py_XDECREF(left); | |
566 | Py_XDECREF(right); |
|
566 | Py_XDECREF(right); | |
567 | Py_XDECREF(tmp); |
|
567 | Py_XDECREF(tmp); | |
568 | goto bail; |
|
568 | goto bail; | |
569 | } |
|
569 | } | |
570 | PyTuple_SET_ITEM(tmp, 0, left); |
|
570 | PyTuple_SET_ITEM(tmp, 0, left); | |
571 | PyTuple_SET_ITEM(tmp, 1, right); |
|
571 | PyTuple_SET_ITEM(tmp, 1, right); | |
572 | PyTuple_SET_ITEM(metadata, i, tmp); |
|
572 | PyTuple_SET_ITEM(metadata, i, tmp); | |
573 | } |
|
573 | } | |
574 | ret = Py_BuildValue("(OOHO(di)O)", prec, succs, flags, metadata, mtime, |
|
574 | ret = Py_BuildValue("(OOHO(di)O)", prec, succs, flags, metadata, mtime, | |
575 | (int)tz * 60, parents); |
|
575 | (int)tz * 60, parents); | |
576 | goto bail; /* return successfully */ |
|
576 | goto bail; /* return successfully */ | |
577 |
|
577 | |||
578 | overflow: |
|
578 | overflow: | |
579 | PyErr_SetString(PyExc_ValueError, "overflow in obsstore"); |
|
579 | PyErr_SetString(PyExc_ValueError, "overflow in obsstore"); | |
580 | bail: |
|
580 | bail: | |
581 | Py_XDECREF(prec); |
|
581 | Py_XDECREF(prec); | |
582 | Py_XDECREF(succs); |
|
582 | Py_XDECREF(succs); | |
583 | Py_XDECREF(metadata); |
|
583 | Py_XDECREF(metadata); | |
584 | Py_XDECREF(parents); |
|
584 | Py_XDECREF(parents); | |
585 | return ret; |
|
585 | return ret; | |
586 | } |
|
586 | } | |
587 |
|
587 | |||
588 | static PyObject *fm1readmarkers(PyObject *self, PyObject *args) |
|
588 | static PyObject *fm1readmarkers(PyObject *self, PyObject *args) | |
589 | { |
|
589 | { | |
590 | const char *data, *dataend; |
|
590 | const char *data, *dataend; | |
591 | Py_ssize_t datalen, offset, stop; |
|
591 | Py_ssize_t datalen, offset, stop; | |
592 | PyObject *markers = NULL; |
|
592 | PyObject *markers = NULL; | |
593 |
|
593 | |||
594 | if (!PyArg_ParseTuple(args, PY23("s#nn", "y#nn"), &data, &datalen, |
|
594 | if (!PyArg_ParseTuple(args, PY23("s#nn", "y#nn"), &data, &datalen, | |
595 | &offset, &stop)) { |
|
595 | &offset, &stop)) { | |
596 | return NULL; |
|
596 | return NULL; | |
597 | } |
|
597 | } | |
598 | if (offset < 0) { |
|
598 | if (offset < 0) { | |
599 | PyErr_SetString(PyExc_ValueError, |
|
599 | PyErr_SetString(PyExc_ValueError, | |
600 | "invalid negative offset in fm1readmarkers"); |
|
600 | "invalid negative offset in fm1readmarkers"); | |
601 | return NULL; |
|
601 | return NULL; | |
602 | } |
|
602 | } | |
603 | if (stop > datalen) { |
|
603 | if (stop > datalen) { | |
604 | PyErr_SetString( |
|
604 | PyErr_SetString( | |
605 | PyExc_ValueError, |
|
605 | PyExc_ValueError, | |
606 | "stop longer than data length in fm1readmarkers"); |
|
606 | "stop longer than data length in fm1readmarkers"); | |
607 | return NULL; |
|
607 | return NULL; | |
608 | } |
|
608 | } | |
609 | dataend = data + datalen; |
|
609 | dataend = data + datalen; | |
610 | data += offset; |
|
610 | data += offset; | |
611 | markers = PyList_New(0); |
|
611 | markers = PyList_New(0); | |
612 | if (!markers) { |
|
612 | if (!markers) { | |
613 | return NULL; |
|
613 | return NULL; | |
614 | } |
|
614 | } | |
615 | while (offset < stop) { |
|
615 | while (offset < stop) { | |
616 | uint32_t msize; |
|
616 | uint32_t msize; | |
617 | int error; |
|
617 | int error; | |
618 | PyObject *record = fm1readmarker(data, dataend, &msize); |
|
618 | PyObject *record = fm1readmarker(data, dataend, &msize); | |
619 | if (!record) { |
|
619 | if (!record) { | |
620 | goto bail; |
|
620 | goto bail; | |
621 | } |
|
621 | } | |
622 | error = PyList_Append(markers, record); |
|
622 | error = PyList_Append(markers, record); | |
623 | Py_DECREF(record); |
|
623 | Py_DECREF(record); | |
624 | if (error) { |
|
624 | if (error) { | |
625 | goto bail; |
|
625 | goto bail; | |
626 | } |
|
626 | } | |
627 | data += msize; |
|
627 | data += msize; | |
628 | offset += msize; |
|
628 | offset += msize; | |
629 | } |
|
629 | } | |
630 | return markers; |
|
630 | return markers; | |
631 | bail: |
|
631 | bail: | |
632 | Py_DECREF(markers); |
|
632 | Py_DECREF(markers); | |
633 | return NULL; |
|
633 | return NULL; | |
634 | } |
|
634 | } | |
635 |
|
635 | |||
636 | static char parsers_doc[] = "Efficient content parsing."; |
|
636 | static char parsers_doc[] = "Efficient content parsing."; | |
637 |
|
637 | |||
638 | PyObject *encodedir(PyObject *self, PyObject *args); |
|
638 | PyObject *encodedir(PyObject *self, PyObject *args); | |
639 | PyObject *pathencode(PyObject *self, PyObject *args); |
|
639 | PyObject *pathencode(PyObject *self, PyObject *args); | |
640 | PyObject *lowerencode(PyObject *self, PyObject *args); |
|
640 | PyObject *lowerencode(PyObject *self, PyObject *args); | |
641 | PyObject *parse_index2(PyObject *self, PyObject *args); |
|
641 | PyObject *parse_index2(PyObject *self, PyObject *args); | |
642 |
|
642 | |||
643 | static PyMethodDef methods[] = { |
|
643 | static PyMethodDef methods[] = { | |
644 | {"pack_dirstate", pack_dirstate, METH_VARARGS, "pack a dirstate\n"}, |
|
644 | {"pack_dirstate", pack_dirstate, METH_VARARGS, "pack a dirstate\n"}, | |
645 | {"nonnormalotherparententries", nonnormalotherparententries, METH_VARARGS, |
|
645 | {"nonnormalotherparententries", nonnormalotherparententries, METH_VARARGS, | |
646 | "create a set containing non-normal and other parent entries of given " |
|
646 | "create a set containing non-normal and other parent entries of given " | |
647 | "dirstate\n"}, |
|
647 | "dirstate\n"}, | |
648 | {"parse_dirstate", parse_dirstate, METH_VARARGS, "parse a dirstate\n"}, |
|
648 | {"parse_dirstate", parse_dirstate, METH_VARARGS, "parse a dirstate\n"}, | |
649 | {"parse_index2", parse_index2, METH_VARARGS, "parse a revlog index\n"}, |
|
649 | {"parse_index2", parse_index2, METH_VARARGS, "parse a revlog index\n"}, | |
650 | {"isasciistr", isasciistr, METH_VARARGS, "check if an ASCII string\n"}, |
|
650 | {"isasciistr", isasciistr, METH_VARARGS, "check if an ASCII string\n"}, | |
651 | {"asciilower", asciilower, METH_VARARGS, "lowercase an ASCII string\n"}, |
|
651 | {"asciilower", asciilower, METH_VARARGS, "lowercase an ASCII string\n"}, | |
652 | {"asciiupper", asciiupper, METH_VARARGS, "uppercase an ASCII string\n"}, |
|
652 | {"asciiupper", asciiupper, METH_VARARGS, "uppercase an ASCII string\n"}, | |
653 | {"dict_new_presized", dict_new_presized, METH_VARARGS, |
|
653 | {"dict_new_presized", dict_new_presized, METH_VARARGS, | |
654 | "construct a dict with an expected size\n"}, |
|
654 | "construct a dict with an expected size\n"}, | |
655 | {"make_file_foldmap", make_file_foldmap, METH_VARARGS, |
|
655 | {"make_file_foldmap", make_file_foldmap, METH_VARARGS, | |
656 | "make file foldmap\n"}, |
|
656 | "make file foldmap\n"}, | |
657 | {"jsonescapeu8fast", jsonescapeu8fast, METH_VARARGS, |
|
657 | {"jsonescapeu8fast", jsonescapeu8fast, METH_VARARGS, | |
658 | "escape a UTF-8 byte string to JSON (fast path)\n"}, |
|
658 | "escape a UTF-8 byte string to JSON (fast path)\n"}, | |
659 | {"encodedir", encodedir, METH_VARARGS, "encodedir a path\n"}, |
|
659 | {"encodedir", encodedir, METH_VARARGS, "encodedir a path\n"}, | |
660 | {"pathencode", pathencode, METH_VARARGS, "fncache-encode a path\n"}, |
|
660 | {"pathencode", pathencode, METH_VARARGS, "fncache-encode a path\n"}, | |
661 | {"lowerencode", lowerencode, METH_VARARGS, "lower-encode a path\n"}, |
|
661 | {"lowerencode", lowerencode, METH_VARARGS, "lower-encode a path\n"}, | |
662 | {"fm1readmarkers", fm1readmarkers, METH_VARARGS, |
|
662 | {"fm1readmarkers", fm1readmarkers, METH_VARARGS, | |
663 | "parse v1 obsolete markers\n"}, |
|
663 | "parse v1 obsolete markers\n"}, | |
664 | {NULL, NULL}}; |
|
664 | {NULL, NULL}}; | |
665 |
|
665 | |||
666 | void dirs_module_init(PyObject *mod); |
|
666 | void dirs_module_init(PyObject *mod); | |
667 | void manifest_module_init(PyObject *mod); |
|
667 | void manifest_module_init(PyObject *mod); | |
668 | void revlog_module_init(PyObject *mod); |
|
668 | void revlog_module_init(PyObject *mod); | |
669 |
|
669 | |||
670 |
static const int version = 1 |
|
670 | static const int version = 14; | |
671 |
|
671 | |||
672 | static void module_init(PyObject *mod) |
|
672 | static void module_init(PyObject *mod) | |
673 | { |
|
673 | { | |
674 | PyObject *capsule = NULL; |
|
674 | PyObject *capsule = NULL; | |
675 | PyModule_AddIntConstant(mod, "version", version); |
|
675 | PyModule_AddIntConstant(mod, "version", version); | |
676 |
|
676 | |||
677 | /* This module constant has two purposes. First, it lets us unit test |
|
677 | /* This module constant has two purposes. First, it lets us unit test | |
678 | * the ImportError raised without hard-coding any error text. This |
|
678 | * the ImportError raised without hard-coding any error text. This | |
679 | * means we can change the text in the future without breaking tests, |
|
679 | * means we can change the text in the future without breaking tests, | |
680 | * even across changesets without a recompile. Second, its presence |
|
680 | * even across changesets without a recompile. Second, its presence | |
681 | * can be used to determine whether the version-checking logic is |
|
681 | * can be used to determine whether the version-checking logic is | |
682 | * present, which also helps in testing across changesets without a |
|
682 | * present, which also helps in testing across changesets without a | |
683 | * recompile. Note that this means the pure-Python version of parsers |
|
683 | * recompile. Note that this means the pure-Python version of parsers | |
684 | * should not have this module constant. */ |
|
684 | * should not have this module constant. */ | |
685 | PyModule_AddStringConstant(mod, "versionerrortext", versionerrortext); |
|
685 | PyModule_AddStringConstant(mod, "versionerrortext", versionerrortext); | |
686 |
|
686 | |||
687 | dirs_module_init(mod); |
|
687 | dirs_module_init(mod); | |
688 | manifest_module_init(mod); |
|
688 | manifest_module_init(mod); | |
689 | revlog_module_init(mod); |
|
689 | revlog_module_init(mod); | |
690 |
|
690 | |||
691 | capsule = PyCapsule_New( |
|
691 | capsule = PyCapsule_New( | |
692 | make_dirstate_tuple, |
|
692 | make_dirstate_tuple, | |
693 | "mercurial.cext.parsers.make_dirstate_tuple_CAPI", NULL); |
|
693 | "mercurial.cext.parsers.make_dirstate_tuple_CAPI", NULL); | |
694 | if (capsule != NULL) |
|
694 | if (capsule != NULL) | |
695 | PyModule_AddObject(mod, "make_dirstate_tuple_CAPI", capsule); |
|
695 | PyModule_AddObject(mod, "make_dirstate_tuple_CAPI", capsule); | |
696 |
|
696 | |||
697 | if (PyType_Ready(&dirstateTupleType) < 0) { |
|
697 | if (PyType_Ready(&dirstateTupleType) < 0) { | |
698 | return; |
|
698 | return; | |
699 | } |
|
699 | } | |
700 | Py_INCREF(&dirstateTupleType); |
|
700 | Py_INCREF(&dirstateTupleType); | |
701 | PyModule_AddObject(mod, "dirstatetuple", |
|
701 | PyModule_AddObject(mod, "dirstatetuple", | |
702 | (PyObject *)&dirstateTupleType); |
|
702 | (PyObject *)&dirstateTupleType); | |
703 | } |
|
703 | } | |
704 |
|
704 | |||
705 | static int check_python_version(void) |
|
705 | static int check_python_version(void) | |
706 | { |
|
706 | { | |
707 | PyObject *sys = PyImport_ImportModule("sys"), *ver; |
|
707 | PyObject *sys = PyImport_ImportModule("sys"), *ver; | |
708 | long hexversion; |
|
708 | long hexversion; | |
709 | if (!sys) { |
|
709 | if (!sys) { | |
710 | return -1; |
|
710 | return -1; | |
711 | } |
|
711 | } | |
712 | ver = PyObject_GetAttrString(sys, "hexversion"); |
|
712 | ver = PyObject_GetAttrString(sys, "hexversion"); | |
713 | Py_DECREF(sys); |
|
713 | Py_DECREF(sys); | |
714 | if (!ver) { |
|
714 | if (!ver) { | |
715 | return -1; |
|
715 | return -1; | |
716 | } |
|
716 | } | |
717 | hexversion = PyInt_AsLong(ver); |
|
717 | hexversion = PyInt_AsLong(ver); | |
718 | Py_DECREF(ver); |
|
718 | Py_DECREF(ver); | |
719 | /* sys.hexversion is a 32-bit number by default, so the -1 case |
|
719 | /* sys.hexversion is a 32-bit number by default, so the -1 case | |
720 | * should only occur in unusual circumstances (e.g. if sys.hexversion |
|
720 | * should only occur in unusual circumstances (e.g. if sys.hexversion | |
721 | * is manually set to an invalid value). */ |
|
721 | * is manually set to an invalid value). */ | |
722 | if ((hexversion == -1) || (hexversion >> 16 != PY_VERSION_HEX >> 16)) { |
|
722 | if ((hexversion == -1) || (hexversion >> 16 != PY_VERSION_HEX >> 16)) { | |
723 | PyErr_Format(PyExc_ImportError, |
|
723 | PyErr_Format(PyExc_ImportError, | |
724 | "%s: The Mercurial extension " |
|
724 | "%s: The Mercurial extension " | |
725 | "modules were compiled with Python " PY_VERSION |
|
725 | "modules were compiled with Python " PY_VERSION | |
726 | ", but " |
|
726 | ", but " | |
727 | "Mercurial is currently using Python with " |
|
727 | "Mercurial is currently using Python with " | |
728 | "sys.hexversion=%ld: " |
|
728 | "sys.hexversion=%ld: " | |
729 | "Python %s\n at: %s", |
|
729 | "Python %s\n at: %s", | |
730 | versionerrortext, hexversion, Py_GetVersion(), |
|
730 | versionerrortext, hexversion, Py_GetVersion(), | |
731 | Py_GetProgramFullPath()); |
|
731 | Py_GetProgramFullPath()); | |
732 | return -1; |
|
732 | return -1; | |
733 | } |
|
733 | } | |
734 | return 0; |
|
734 | return 0; | |
735 | } |
|
735 | } | |
736 |
|
736 | |||
737 | #ifdef IS_PY3K |
|
737 | #ifdef IS_PY3K | |
738 | static struct PyModuleDef parsers_module = {PyModuleDef_HEAD_INIT, "parsers", |
|
738 | static struct PyModuleDef parsers_module = {PyModuleDef_HEAD_INIT, "parsers", | |
739 | parsers_doc, -1, methods}; |
|
739 | parsers_doc, -1, methods}; | |
740 |
|
740 | |||
741 | PyMODINIT_FUNC PyInit_parsers(void) |
|
741 | PyMODINIT_FUNC PyInit_parsers(void) | |
742 | { |
|
742 | { | |
743 | PyObject *mod; |
|
743 | PyObject *mod; | |
744 |
|
744 | |||
745 | if (check_python_version() == -1) |
|
745 | if (check_python_version() == -1) | |
746 | return NULL; |
|
746 | return NULL; | |
747 | mod = PyModule_Create(&parsers_module); |
|
747 | mod = PyModule_Create(&parsers_module); | |
748 | module_init(mod); |
|
748 | module_init(mod); | |
749 | return mod; |
|
749 | return mod; | |
750 | } |
|
750 | } | |
751 | #else |
|
751 | #else | |
752 | PyMODINIT_FUNC initparsers(void) |
|
752 | PyMODINIT_FUNC initparsers(void) | |
753 | { |
|
753 | { | |
754 | PyObject *mod; |
|
754 | PyObject *mod; | |
755 |
|
755 | |||
756 | if (check_python_version() == -1) { |
|
756 | if (check_python_version() == -1) { | |
757 | return; |
|
757 | return; | |
758 | } |
|
758 | } | |
759 | mod = Py_InitModule3("parsers", methods, parsers_doc); |
|
759 | mod = Py_InitModule3("parsers", methods, parsers_doc); | |
760 | module_init(mod); |
|
760 | module_init(mod); | |
761 | } |
|
761 | } | |
762 | #endif |
|
762 | #endif |
@@ -1,3041 +1,3051 b'' | |||||
1 | /* |
|
1 | /* | |
2 | parsers.c - efficient content parsing |
|
2 | parsers.c - efficient content parsing | |
3 |
|
3 | |||
4 | Copyright 2008 Matt Mackall <mpm@selenic.com> and others |
|
4 | Copyright 2008 Matt Mackall <mpm@selenic.com> and others | |
5 |
|
5 | |||
6 | This software may be used and distributed according to the terms of |
|
6 | This software may be used and distributed according to the terms of | |
7 | the GNU General Public License, incorporated herein by reference. |
|
7 | the GNU General Public License, incorporated herein by reference. | |
8 | */ |
|
8 | */ | |
9 |
|
9 | |||
10 | #define PY_SSIZE_T_CLEAN |
|
10 | #define PY_SSIZE_T_CLEAN | |
11 | #include <Python.h> |
|
11 | #include <Python.h> | |
12 | #include <assert.h> |
|
12 | #include <assert.h> | |
13 | #include <ctype.h> |
|
13 | #include <ctype.h> | |
14 | #include <limits.h> |
|
14 | #include <limits.h> | |
15 | #include <stddef.h> |
|
15 | #include <stddef.h> | |
16 | #include <stdlib.h> |
|
16 | #include <stdlib.h> | |
17 | #include <string.h> |
|
17 | #include <string.h> | |
18 |
|
18 | |||
19 | #include "bitmanipulation.h" |
|
19 | #include "bitmanipulation.h" | |
20 | #include "charencode.h" |
|
20 | #include "charencode.h" | |
21 | #include "revlog.h" |
|
21 | #include "revlog.h" | |
22 | #include "util.h" |
|
22 | #include "util.h" | |
23 |
|
23 | |||
24 | #ifdef IS_PY3K |
|
24 | #ifdef IS_PY3K | |
25 | /* The mapping of Python types is meant to be temporary to get Python |
|
25 | /* The mapping of Python types is meant to be temporary to get Python | |
26 | * 3 to compile. We should remove this once Python 3 support is fully |
|
26 | * 3 to compile. We should remove this once Python 3 support is fully | |
27 | * supported and proper types are used in the extensions themselves. */ |
|
27 | * supported and proper types are used in the extensions themselves. */ | |
28 | #define PyInt_Check PyLong_Check |
|
28 | #define PyInt_Check PyLong_Check | |
29 | #define PyInt_FromLong PyLong_FromLong |
|
29 | #define PyInt_FromLong PyLong_FromLong | |
30 | #define PyInt_FromSsize_t PyLong_FromSsize_t |
|
30 | #define PyInt_FromSsize_t PyLong_FromSsize_t | |
31 | #define PyInt_AsLong PyLong_AsLong |
|
31 | #define PyInt_AsLong PyLong_AsLong | |
32 | #endif |
|
32 | #endif | |
33 |
|
33 | |||
34 | typedef struct indexObjectStruct indexObject; |
|
34 | typedef struct indexObjectStruct indexObject; | |
35 |
|
35 | |||
36 | typedef struct { |
|
36 | typedef struct { | |
37 | int children[16]; |
|
37 | int children[16]; | |
38 | } nodetreenode; |
|
38 | } nodetreenode; | |
39 |
|
39 | |||
40 | /* |
|
40 | /* | |
41 | * A base-16 trie for fast node->rev mapping. |
|
41 | * A base-16 trie for fast node->rev mapping. | |
42 | * |
|
42 | * | |
43 | * Positive value is index of the next node in the trie |
|
43 | * Positive value is index of the next node in the trie | |
44 | * Negative value is a leaf: -(rev + 2) |
|
44 | * Negative value is a leaf: -(rev + 2) | |
45 | * Zero is empty |
|
45 | * Zero is empty | |
46 | */ |
|
46 | */ | |
47 | typedef struct { |
|
47 | typedef struct { | |
48 | indexObject *index; |
|
48 | indexObject *index; | |
49 | nodetreenode *nodes; |
|
49 | nodetreenode *nodes; | |
50 | unsigned length; /* # nodes in use */ |
|
50 | unsigned length; /* # nodes in use */ | |
51 | unsigned capacity; /* # nodes allocated */ |
|
51 | unsigned capacity; /* # nodes allocated */ | |
52 | int depth; /* maximum depth of tree */ |
|
52 | int depth; /* maximum depth of tree */ | |
53 | int splits; /* # splits performed */ |
|
53 | int splits; /* # splits performed */ | |
54 | } nodetree; |
|
54 | } nodetree; | |
55 |
|
55 | |||
56 | typedef struct { |
|
56 | typedef struct { | |
57 | PyObject_HEAD /* ; */ |
|
57 | PyObject_HEAD /* ; */ | |
58 | nodetree nt; |
|
58 | nodetree nt; | |
59 | } nodetreeObject; |
|
59 | } nodetreeObject; | |
60 |
|
60 | |||
61 | /* |
|
61 | /* | |
62 | * This class has two behaviors. |
|
62 | * This class has two behaviors. | |
63 | * |
|
63 | * | |
64 | * When used in a list-like way (with integer keys), we decode an |
|
64 | * When used in a list-like way (with integer keys), we decode an | |
65 | * entry in a RevlogNG index file on demand. Our last entry is a |
|
65 | * entry in a RevlogNG index file on demand. Our last entry is a | |
66 | * sentinel, always a nullid. We have limited support for |
|
66 | * sentinel, always a nullid. We have limited support for | |
67 | * integer-keyed insert and delete, only at elements right before the |
|
67 | * integer-keyed insert and delete, only at elements right before the | |
68 | * sentinel. |
|
68 | * sentinel. | |
69 | * |
|
69 | * | |
70 | * With string keys, we lazily perform a reverse mapping from node to |
|
70 | * With string keys, we lazily perform a reverse mapping from node to | |
71 | * rev, using a base-16 trie. |
|
71 | * rev, using a base-16 trie. | |
72 | */ |
|
72 | */ | |
73 | struct indexObjectStruct { |
|
73 | struct indexObjectStruct { | |
74 | PyObject_HEAD |
|
74 | PyObject_HEAD | |
75 | /* Type-specific fields go here. */ |
|
75 | /* Type-specific fields go here. */ | |
76 | PyObject *data; /* raw bytes of index */ |
|
76 | PyObject *data; /* raw bytes of index */ | |
77 | Py_buffer buf; /* buffer of data */ |
|
77 | Py_buffer buf; /* buffer of data */ | |
78 | PyObject **cache; /* cached tuples */ |
|
78 | PyObject **cache; /* cached tuples */ | |
79 | const char **offsets; /* populated on demand */ |
|
79 | const char **offsets; /* populated on demand */ | |
80 | Py_ssize_t raw_length; /* original number of elements */ |
|
80 | Py_ssize_t raw_length; /* original number of elements */ | |
81 | Py_ssize_t length; /* current number of elements */ |
|
81 | Py_ssize_t length; /* current number of elements */ | |
82 | PyObject *added; /* populated on demand */ |
|
82 | PyObject *added; /* populated on demand */ | |
83 | PyObject *headrevs; /* cache, invalidated on changes */ |
|
83 | PyObject *headrevs; /* cache, invalidated on changes */ | |
84 | PyObject *filteredrevs; /* filtered revs set */ |
|
84 | PyObject *filteredrevs; /* filtered revs set */ | |
85 | nodetree nt; /* base-16 trie */ |
|
85 | nodetree nt; /* base-16 trie */ | |
86 | int ntinitialized; /* 0 or 1 */ |
|
86 | int ntinitialized; /* 0 or 1 */ | |
87 | int ntrev; /* last rev scanned */ |
|
87 | int ntrev; /* last rev scanned */ | |
88 | int ntlookups; /* # lookups */ |
|
88 | int ntlookups; /* # lookups */ | |
89 | int ntmisses; /* # lookups that miss the cache */ |
|
89 | int ntmisses; /* # lookups that miss the cache */ | |
90 | int inlined; |
|
90 | int inlined; | |
91 | }; |
|
91 | }; | |
92 |
|
92 | |||
93 | static Py_ssize_t index_length(const indexObject *self) |
|
93 | static Py_ssize_t index_length(const indexObject *self) | |
94 | { |
|
94 | { | |
95 | if (self->added == NULL) |
|
95 | if (self->added == NULL) | |
96 | return self->length; |
|
96 | return self->length; | |
97 | return self->length + PyList_GET_SIZE(self->added); |
|
97 | return self->length + PyList_GET_SIZE(self->added); | |
98 | } |
|
98 | } | |
99 |
|
99 | |||
100 | static PyObject *nullentry = NULL; |
|
100 | static PyObject *nullentry = NULL; | |
101 | static const char nullid[20] = {0}; |
|
101 | static const char nullid[20] = {0}; | |
102 | static const Py_ssize_t nullrev = -1; |
|
102 | static const Py_ssize_t nullrev = -1; | |
103 |
|
103 | |||
104 | static Py_ssize_t inline_scan(indexObject *self, const char **offsets); |
|
104 | static Py_ssize_t inline_scan(indexObject *self, const char **offsets); | |
105 |
|
105 | |||
106 | #if LONG_MAX == 0x7fffffffL |
|
106 | #if LONG_MAX == 0x7fffffffL | |
107 | static const char *const tuple_format = PY23("Kiiiiiis#", "Kiiiiiiy#"); |
|
107 | static const char *const tuple_format = PY23("Kiiiiiis#", "Kiiiiiiy#"); | |
108 | #else |
|
108 | #else | |
109 | static const char *const tuple_format = PY23("kiiiiiis#", "kiiiiiiy#"); |
|
109 | static const char *const tuple_format = PY23("kiiiiiis#", "kiiiiiiy#"); | |
110 | #endif |
|
110 | #endif | |
111 |
|
111 | |||
112 | /* A RevlogNG v1 index entry is 64 bytes long. */ |
|
112 | /* A RevlogNG v1 index entry is 64 bytes long. */ | |
113 | static const long v1_hdrsize = 64; |
|
113 | static const long v1_hdrsize = 64; | |
114 |
|
114 | |||
115 | static void raise_revlog_error(void) |
|
115 | static void raise_revlog_error(void) | |
116 | { |
|
116 | { | |
117 | PyObject *mod = NULL, *dict = NULL, *errclass = NULL; |
|
117 | PyObject *mod = NULL, *dict = NULL, *errclass = NULL; | |
118 |
|
118 | |||
119 | mod = PyImport_ImportModule("mercurial.error"); |
|
119 | mod = PyImport_ImportModule("mercurial.error"); | |
120 | if (mod == NULL) { |
|
120 | if (mod == NULL) { | |
121 | goto cleanup; |
|
121 | goto cleanup; | |
122 | } |
|
122 | } | |
123 |
|
123 | |||
124 | dict = PyModule_GetDict(mod); |
|
124 | dict = PyModule_GetDict(mod); | |
125 | if (dict == NULL) { |
|
125 | if (dict == NULL) { | |
126 | goto cleanup; |
|
126 | goto cleanup; | |
127 | } |
|
127 | } | |
128 | Py_INCREF(dict); |
|
128 | Py_INCREF(dict); | |
129 |
|
129 | |||
130 | errclass = PyDict_GetItemString(dict, "RevlogError"); |
|
130 | errclass = PyDict_GetItemString(dict, "RevlogError"); | |
131 | if (errclass == NULL) { |
|
131 | if (errclass == NULL) { | |
132 | PyErr_SetString(PyExc_SystemError, |
|
132 | PyErr_SetString(PyExc_SystemError, | |
133 | "could not find RevlogError"); |
|
133 | "could not find RevlogError"); | |
134 | goto cleanup; |
|
134 | goto cleanup; | |
135 | } |
|
135 | } | |
136 |
|
136 | |||
137 | /* value of exception is ignored by callers */ |
|
137 | /* value of exception is ignored by callers */ | |
138 | PyErr_SetString(errclass, "RevlogError"); |
|
138 | PyErr_SetString(errclass, "RevlogError"); | |
139 |
|
139 | |||
140 | cleanup: |
|
140 | cleanup: | |
141 | Py_XDECREF(dict); |
|
141 | Py_XDECREF(dict); | |
142 | Py_XDECREF(mod); |
|
142 | Py_XDECREF(mod); | |
143 | } |
|
143 | } | |
144 |
|
144 | |||
145 | /* |
|
145 | /* | |
146 | * Return a pointer to the beginning of a RevlogNG record. |
|
146 | * Return a pointer to the beginning of a RevlogNG record. | |
147 | */ |
|
147 | */ | |
148 | static const char *index_deref(indexObject *self, Py_ssize_t pos) |
|
148 | static const char *index_deref(indexObject *self, Py_ssize_t pos) | |
149 | { |
|
149 | { | |
150 | if (self->inlined && pos > 0) { |
|
150 | if (self->inlined && pos > 0) { | |
151 | if (self->offsets == NULL) { |
|
151 | if (self->offsets == NULL) { | |
152 | self->offsets = PyMem_Malloc(self->raw_length * |
|
152 | self->offsets = PyMem_Malloc(self->raw_length * | |
153 | sizeof(*self->offsets)); |
|
153 | sizeof(*self->offsets)); | |
154 | if (self->offsets == NULL) |
|
154 | if (self->offsets == NULL) | |
155 | return (const char *)PyErr_NoMemory(); |
|
155 | return (const char *)PyErr_NoMemory(); | |
156 | inline_scan(self, self->offsets); |
|
156 | inline_scan(self, self->offsets); | |
157 | } |
|
157 | } | |
158 | return self->offsets[pos]; |
|
158 | return self->offsets[pos]; | |
159 | } |
|
159 | } | |
160 |
|
160 | |||
161 | return (const char *)(self->buf.buf) + pos * v1_hdrsize; |
|
161 | return (const char *)(self->buf.buf) + pos * v1_hdrsize; | |
162 | } |
|
162 | } | |
163 |
|
163 | |||
164 | /* |
|
164 | /* | |
165 | * Get parents of the given rev. |
|
165 | * Get parents of the given rev. | |
166 | * |
|
166 | * | |
167 | * The specified rev must be valid and must not be nullrev. A returned |
|
167 | * The specified rev must be valid and must not be nullrev. A returned | |
168 | * parent revision may be nullrev, but is guaranteed to be in valid range. |
|
168 | * parent revision may be nullrev, but is guaranteed to be in valid range. | |
169 | */ |
|
169 | */ | |
170 | static inline int index_get_parents(indexObject *self, Py_ssize_t rev, int *ps, |
|
170 | static inline int index_get_parents(indexObject *self, Py_ssize_t rev, int *ps, | |
171 | int maxrev) |
|
171 | int maxrev) | |
172 | { |
|
172 | { | |
173 | if (rev >= self->length) { |
|
173 | if (rev >= self->length) { | |
174 | long tmp; |
|
174 | long tmp; | |
175 | PyObject *tuple = |
|
175 | PyObject *tuple = | |
176 | PyList_GET_ITEM(self->added, rev - self->length); |
|
176 | PyList_GET_ITEM(self->added, rev - self->length); | |
177 | if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 5), &tmp)) { |
|
177 | if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 5), &tmp)) { | |
178 | return -1; |
|
178 | return -1; | |
179 | } |
|
179 | } | |
180 | ps[0] = (int)tmp; |
|
180 | ps[0] = (int)tmp; | |
181 | if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 6), &tmp)) { |
|
181 | if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 6), &tmp)) { | |
182 | return -1; |
|
182 | return -1; | |
183 | } |
|
183 | } | |
184 | ps[1] = (int)tmp; |
|
184 | ps[1] = (int)tmp; | |
185 | } else { |
|
185 | } else { | |
186 | const char *data = index_deref(self, rev); |
|
186 | const char *data = index_deref(self, rev); | |
187 | ps[0] = getbe32(data + 24); |
|
187 | ps[0] = getbe32(data + 24); | |
188 | ps[1] = getbe32(data + 28); |
|
188 | ps[1] = getbe32(data + 28); | |
189 | } |
|
189 | } | |
190 | /* If index file is corrupted, ps[] may point to invalid revisions. So |
|
190 | /* If index file is corrupted, ps[] may point to invalid revisions. So | |
191 | * there is a risk of buffer overflow to trust them unconditionally. */ |
|
191 | * there is a risk of buffer overflow to trust them unconditionally. */ | |
192 | if (ps[0] < -1 || ps[0] > maxrev || ps[1] < -1 || ps[1] > maxrev) { |
|
192 | if (ps[0] < -1 || ps[0] > maxrev || ps[1] < -1 || ps[1] > maxrev) { | |
193 | PyErr_SetString(PyExc_ValueError, "parent out of range"); |
|
193 | PyErr_SetString(PyExc_ValueError, "parent out of range"); | |
194 | return -1; |
|
194 | return -1; | |
195 | } |
|
195 | } | |
196 | return 0; |
|
196 | return 0; | |
197 | } |
|
197 | } | |
198 |
|
198 | |||
199 | /* |
|
199 | /* | |
200 | * Get parents of the given rev. |
|
200 | * Get parents of the given rev. | |
201 | * |
|
201 | * | |
202 | * If the specified rev is out of range, IndexError will be raised. If the |
|
202 | * If the specified rev is out of range, IndexError will be raised. If the | |
203 | * revlog entry is corrupted, ValueError may be raised. |
|
203 | * revlog entry is corrupted, ValueError may be raised. | |
204 | * |
|
204 | * | |
205 | * Returns 0 on success or -1 on failure. |
|
205 | * Returns 0 on success or -1 on failure. | |
206 | */ |
|
206 | */ | |
207 | int HgRevlogIndex_GetParents(PyObject *op, int rev, int *ps) |
|
207 | int HgRevlogIndex_GetParents(PyObject *op, int rev, int *ps) | |
208 | { |
|
208 | { | |
209 | int tiprev; |
|
209 | int tiprev; | |
210 | if (!op || !HgRevlogIndex_Check(op) || !ps) { |
|
210 | if (!op || !HgRevlogIndex_Check(op) || !ps) { | |
211 | PyErr_BadInternalCall(); |
|
211 | PyErr_BadInternalCall(); | |
212 | return -1; |
|
212 | return -1; | |
213 | } |
|
213 | } | |
214 | tiprev = (int)index_length((indexObject *)op) - 1; |
|
214 | tiprev = (int)index_length((indexObject *)op) - 1; | |
215 | if (rev < -1 || rev > tiprev) { |
|
215 | if (rev < -1 || rev > tiprev) { | |
216 | PyErr_Format(PyExc_IndexError, "rev out of range: %d", rev); |
|
216 | PyErr_Format(PyExc_IndexError, "rev out of range: %d", rev); | |
217 | return -1; |
|
217 | return -1; | |
218 | } else if (rev == -1) { |
|
218 | } else if (rev == -1) { | |
219 | ps[0] = ps[1] = -1; |
|
219 | ps[0] = ps[1] = -1; | |
220 | return 0; |
|
220 | return 0; | |
221 | } else { |
|
221 | } else { | |
222 | return index_get_parents((indexObject *)op, rev, ps, tiprev); |
|
222 | return index_get_parents((indexObject *)op, rev, ps, tiprev); | |
223 | } |
|
223 | } | |
224 | } |
|
224 | } | |
225 |
|
225 | |||
226 | static inline int64_t index_get_start(indexObject *self, Py_ssize_t rev) |
|
226 | static inline int64_t index_get_start(indexObject *self, Py_ssize_t rev) | |
227 | { |
|
227 | { | |
228 | uint64_t offset; |
|
228 | uint64_t offset; | |
229 | if (rev == nullrev) { |
|
229 | if (rev == nullrev) { | |
230 | return 0; |
|
230 | return 0; | |
231 | } |
|
231 | } | |
232 | if (rev >= self->length) { |
|
232 | if (rev >= self->length) { | |
233 | PyObject *tuple; |
|
233 | PyObject *tuple; | |
234 | PyObject *pylong; |
|
234 | PyObject *pylong; | |
235 | PY_LONG_LONG tmp; |
|
235 | PY_LONG_LONG tmp; | |
236 | tuple = PyList_GET_ITEM(self->added, rev - self->length); |
|
236 | tuple = PyList_GET_ITEM(self->added, rev - self->length); | |
237 | pylong = PyTuple_GET_ITEM(tuple, 0); |
|
237 | pylong = PyTuple_GET_ITEM(tuple, 0); | |
238 | tmp = PyLong_AsLongLong(pylong); |
|
238 | tmp = PyLong_AsLongLong(pylong); | |
239 | if (tmp == -1 && PyErr_Occurred()) { |
|
239 | if (tmp == -1 && PyErr_Occurred()) { | |
240 | return -1; |
|
240 | return -1; | |
241 | } |
|
241 | } | |
242 | if (tmp < 0) { |
|
242 | if (tmp < 0) { | |
243 | PyErr_Format(PyExc_OverflowError, |
|
243 | PyErr_Format(PyExc_OverflowError, | |
244 | "revlog entry size out of bound (%lld)", |
|
244 | "revlog entry size out of bound (%lld)", | |
245 | (long long)tmp); |
|
245 | (long long)tmp); | |
246 | return -1; |
|
246 | return -1; | |
247 | } |
|
247 | } | |
248 | offset = (uint64_t)tmp; |
|
248 | offset = (uint64_t)tmp; | |
249 | } else { |
|
249 | } else { | |
250 | const char *data = index_deref(self, rev); |
|
250 | const char *data = index_deref(self, rev); | |
251 | offset = getbe32(data + 4); |
|
251 | offset = getbe32(data + 4); | |
252 | if (rev == 0) { |
|
252 | if (rev == 0) { | |
253 | /* mask out version number for the first entry */ |
|
253 | /* mask out version number for the first entry */ | |
254 | offset &= 0xFFFF; |
|
254 | offset &= 0xFFFF; | |
255 | } else { |
|
255 | } else { | |
256 | uint32_t offset_high = getbe32(data); |
|
256 | uint32_t offset_high = getbe32(data); | |
257 | offset |= ((uint64_t)offset_high) << 32; |
|
257 | offset |= ((uint64_t)offset_high) << 32; | |
258 | } |
|
258 | } | |
259 | } |
|
259 | } | |
260 | return (int64_t)(offset >> 16); |
|
260 | return (int64_t)(offset >> 16); | |
261 | } |
|
261 | } | |
262 |
|
262 | |||
263 | static inline int index_get_length(indexObject *self, Py_ssize_t rev) |
|
263 | static inline int index_get_length(indexObject *self, Py_ssize_t rev) | |
264 | { |
|
264 | { | |
265 | if (rev == nullrev) { |
|
265 | if (rev == nullrev) { | |
266 | return 0; |
|
266 | return 0; | |
267 | } |
|
267 | } | |
268 | if (rev >= self->length) { |
|
268 | if (rev >= self->length) { | |
269 | PyObject *tuple; |
|
269 | PyObject *tuple; | |
270 | PyObject *pylong; |
|
270 | PyObject *pylong; | |
271 | long ret; |
|
271 | long ret; | |
272 | tuple = PyList_GET_ITEM(self->added, rev - self->length); |
|
272 | tuple = PyList_GET_ITEM(self->added, rev - self->length); | |
273 | pylong = PyTuple_GET_ITEM(tuple, 1); |
|
273 | pylong = PyTuple_GET_ITEM(tuple, 1); | |
274 | ret = PyInt_AsLong(pylong); |
|
274 | ret = PyInt_AsLong(pylong); | |
275 | if (ret == -1 && PyErr_Occurred()) { |
|
275 | if (ret == -1 && PyErr_Occurred()) { | |
276 | return -1; |
|
276 | return -1; | |
277 | } |
|
277 | } | |
278 | if (ret < 0 || ret > (long)INT_MAX) { |
|
278 | if (ret < 0 || ret > (long)INT_MAX) { | |
279 | PyErr_Format(PyExc_OverflowError, |
|
279 | PyErr_Format(PyExc_OverflowError, | |
280 | "revlog entry size out of bound (%ld)", |
|
280 | "revlog entry size out of bound (%ld)", | |
281 | ret); |
|
281 | ret); | |
282 | return -1; |
|
282 | return -1; | |
283 | } |
|
283 | } | |
284 | return (int)ret; |
|
284 | return (int)ret; | |
285 | } else { |
|
285 | } else { | |
286 | const char *data = index_deref(self, rev); |
|
286 | const char *data = index_deref(self, rev); | |
287 | int tmp = (int)getbe32(data + 8); |
|
287 | int tmp = (int)getbe32(data + 8); | |
288 | if (tmp < 0) { |
|
288 | if (tmp < 0) { | |
289 | PyErr_Format(PyExc_OverflowError, |
|
289 | PyErr_Format(PyExc_OverflowError, | |
290 | "revlog entry size out of bound (%d)", |
|
290 | "revlog entry size out of bound (%d)", | |
291 | tmp); |
|
291 | tmp); | |
292 | return -1; |
|
292 | return -1; | |
293 | } |
|
293 | } | |
294 | return tmp; |
|
294 | return tmp; | |
295 | } |
|
295 | } | |
296 | } |
|
296 | } | |
297 |
|
297 | |||
298 | /* |
|
298 | /* | |
299 | * RevlogNG format (all in big endian, data may be inlined): |
|
299 | * RevlogNG format (all in big endian, data may be inlined): | |
300 | * 6 bytes: offset |
|
300 | * 6 bytes: offset | |
301 | * 2 bytes: flags |
|
301 | * 2 bytes: flags | |
302 | * 4 bytes: compressed length |
|
302 | * 4 bytes: compressed length | |
303 | * 4 bytes: uncompressed length |
|
303 | * 4 bytes: uncompressed length | |
304 | * 4 bytes: base revision |
|
304 | * 4 bytes: base revision | |
305 | * 4 bytes: link revision |
|
305 | * 4 bytes: link revision | |
306 | * 4 bytes: parent 1 revision |
|
306 | * 4 bytes: parent 1 revision | |
307 | * 4 bytes: parent 2 revision |
|
307 | * 4 bytes: parent 2 revision | |
308 | * 32 bytes: nodeid (only 20 bytes used) |
|
308 | * 32 bytes: nodeid (only 20 bytes used) | |
309 | */ |
|
309 | */ | |
310 | static PyObject *index_get(indexObject *self, Py_ssize_t pos) |
|
310 | static PyObject *index_get(indexObject *self, Py_ssize_t pos) | |
311 | { |
|
311 | { | |
312 | uint64_t offset_flags; |
|
312 | uint64_t offset_flags; | |
313 | int comp_len, uncomp_len, base_rev, link_rev, parent_1, parent_2; |
|
313 | int comp_len, uncomp_len, base_rev, link_rev, parent_1, parent_2; | |
314 | const char *c_node_id; |
|
314 | const char *c_node_id; | |
315 | const char *data; |
|
315 | const char *data; | |
316 | Py_ssize_t length = index_length(self); |
|
316 | Py_ssize_t length = index_length(self); | |
317 | PyObject *entry; |
|
317 | PyObject *entry; | |
318 |
|
318 | |||
319 | if (pos == nullrev) { |
|
319 | if (pos == nullrev) { | |
320 | Py_INCREF(nullentry); |
|
320 | Py_INCREF(nullentry); | |
321 | return nullentry; |
|
321 | return nullentry; | |
322 | } |
|
322 | } | |
323 |
|
323 | |||
324 | if (pos < 0 || pos >= length) { |
|
324 | if (pos < 0 || pos >= length) { | |
325 | PyErr_SetString(PyExc_IndexError, "revlog index out of range"); |
|
325 | PyErr_SetString(PyExc_IndexError, "revlog index out of range"); | |
326 | return NULL; |
|
326 | return NULL; | |
327 | } |
|
327 | } | |
328 |
|
328 | |||
329 | if (pos >= self->length) { |
|
329 | if (pos >= self->length) { | |
330 | PyObject *obj; |
|
330 | PyObject *obj; | |
331 | obj = PyList_GET_ITEM(self->added, pos - self->length); |
|
331 | obj = PyList_GET_ITEM(self->added, pos - self->length); | |
332 | Py_INCREF(obj); |
|
332 | Py_INCREF(obj); | |
333 | return obj; |
|
333 | return obj; | |
334 | } |
|
334 | } | |
335 |
|
335 | |||
336 | if (self->cache) { |
|
336 | if (self->cache) { | |
337 | if (self->cache[pos]) { |
|
337 | if (self->cache[pos]) { | |
338 | Py_INCREF(self->cache[pos]); |
|
338 | Py_INCREF(self->cache[pos]); | |
339 | return self->cache[pos]; |
|
339 | return self->cache[pos]; | |
340 | } |
|
340 | } | |
341 | } else { |
|
341 | } else { | |
342 | self->cache = calloc(self->raw_length, sizeof(PyObject *)); |
|
342 | self->cache = calloc(self->raw_length, sizeof(PyObject *)); | |
343 | if (self->cache == NULL) |
|
343 | if (self->cache == NULL) | |
344 | return PyErr_NoMemory(); |
|
344 | return PyErr_NoMemory(); | |
345 | } |
|
345 | } | |
346 |
|
346 | |||
347 | data = index_deref(self, pos); |
|
347 | data = index_deref(self, pos); | |
348 | if (data == NULL) |
|
348 | if (data == NULL) | |
349 | return NULL; |
|
349 | return NULL; | |
350 |
|
350 | |||
351 | offset_flags = getbe32(data + 4); |
|
351 | offset_flags = getbe32(data + 4); | |
352 | if (pos == 0) /* mask out version number for the first entry */ |
|
352 | if (pos == 0) /* mask out version number for the first entry */ | |
353 | offset_flags &= 0xFFFF; |
|
353 | offset_flags &= 0xFFFF; | |
354 | else { |
|
354 | else { | |
355 | uint32_t offset_high = getbe32(data); |
|
355 | uint32_t offset_high = getbe32(data); | |
356 | offset_flags |= ((uint64_t)offset_high) << 32; |
|
356 | offset_flags |= ((uint64_t)offset_high) << 32; | |
357 | } |
|
357 | } | |
358 |
|
358 | |||
359 | comp_len = getbe32(data + 8); |
|
359 | comp_len = getbe32(data + 8); | |
360 | uncomp_len = getbe32(data + 12); |
|
360 | uncomp_len = getbe32(data + 12); | |
361 | base_rev = getbe32(data + 16); |
|
361 | base_rev = getbe32(data + 16); | |
362 | link_rev = getbe32(data + 20); |
|
362 | link_rev = getbe32(data + 20); | |
363 | parent_1 = getbe32(data + 24); |
|
363 | parent_1 = getbe32(data + 24); | |
364 | parent_2 = getbe32(data + 28); |
|
364 | parent_2 = getbe32(data + 28); | |
365 | c_node_id = data + 32; |
|
365 | c_node_id = data + 32; | |
366 |
|
366 | |||
367 | entry = Py_BuildValue(tuple_format, offset_flags, comp_len, uncomp_len, |
|
367 | entry = Py_BuildValue(tuple_format, offset_flags, comp_len, uncomp_len, | |
368 | base_rev, link_rev, parent_1, parent_2, c_node_id, |
|
368 | base_rev, link_rev, parent_1, parent_2, c_node_id, | |
369 | (Py_ssize_t)20); |
|
369 | (Py_ssize_t)20); | |
370 |
|
370 | |||
371 | if (entry) { |
|
371 | if (entry) { | |
372 | PyObject_GC_UnTrack(entry); |
|
372 | PyObject_GC_UnTrack(entry); | |
373 | Py_INCREF(entry); |
|
373 | Py_INCREF(entry); | |
374 | } |
|
374 | } | |
375 |
|
375 | |||
376 | self->cache[pos] = entry; |
|
376 | self->cache[pos] = entry; | |
377 |
|
377 | |||
378 | return entry; |
|
378 | return entry; | |
379 | } |
|
379 | } | |
380 |
|
380 | |||
381 | /* |
|
381 | /* | |
382 | * Return the 20-byte SHA of the node corresponding to the given rev. |
|
382 | * Return the 20-byte SHA of the node corresponding to the given rev. | |
383 | */ |
|
383 | */ | |
384 | static const char *index_node(indexObject *self, Py_ssize_t pos) |
|
384 | static const char *index_node(indexObject *self, Py_ssize_t pos) | |
385 | { |
|
385 | { | |
386 | Py_ssize_t length = index_length(self); |
|
386 | Py_ssize_t length = index_length(self); | |
387 | const char *data; |
|
387 | const char *data; | |
388 |
|
388 | |||
389 | if (pos == nullrev) |
|
389 | if (pos == nullrev) | |
390 | return nullid; |
|
390 | return nullid; | |
391 |
|
391 | |||
392 | if (pos >= length) |
|
392 | if (pos >= length) | |
393 | return NULL; |
|
393 | return NULL; | |
394 |
|
394 | |||
395 | if (pos >= self->length) { |
|
395 | if (pos >= self->length) { | |
396 | PyObject *tuple, *str; |
|
396 | PyObject *tuple, *str; | |
397 | tuple = PyList_GET_ITEM(self->added, pos - self->length); |
|
397 | tuple = PyList_GET_ITEM(self->added, pos - self->length); | |
398 | str = PyTuple_GetItem(tuple, 7); |
|
398 | str = PyTuple_GetItem(tuple, 7); | |
399 | return str ? PyBytes_AS_STRING(str) : NULL; |
|
399 | return str ? PyBytes_AS_STRING(str) : NULL; | |
400 | } |
|
400 | } | |
401 |
|
401 | |||
402 | data = index_deref(self, pos); |
|
402 | data = index_deref(self, pos); | |
403 | return data ? data + 32 : NULL; |
|
403 | return data ? data + 32 : NULL; | |
404 | } |
|
404 | } | |
405 |
|
405 | |||
406 | /* |
|
406 | /* | |
407 | * Return the 20-byte SHA of the node corresponding to the given rev. The |
|
407 | * Return the 20-byte SHA of the node corresponding to the given rev. The | |
408 | * rev is assumed to be existing. If not, an exception is set. |
|
408 | * rev is assumed to be existing. If not, an exception is set. | |
409 | */ |
|
409 | */ | |
410 | static const char *index_node_existing(indexObject *self, Py_ssize_t pos) |
|
410 | static const char *index_node_existing(indexObject *self, Py_ssize_t pos) | |
411 | { |
|
411 | { | |
412 | const char *node = index_node(self, pos); |
|
412 | const char *node = index_node(self, pos); | |
413 | if (node == NULL) { |
|
413 | if (node == NULL) { | |
414 | PyErr_Format(PyExc_IndexError, "could not access rev %d", |
|
414 | PyErr_Format(PyExc_IndexError, "could not access rev %d", | |
415 | (int)pos); |
|
415 | (int)pos); | |
416 | } |
|
416 | } | |
417 | return node; |
|
417 | return node; | |
418 | } |
|
418 | } | |
419 |
|
419 | |||
420 | static int nt_insert(nodetree *self, const char *node, int rev); |
|
420 | static int nt_insert(nodetree *self, const char *node, int rev); | |
421 |
|
421 | |||
422 | static int node_check(PyObject *obj, char **node) |
|
422 | static int node_check(PyObject *obj, char **node) | |
423 | { |
|
423 | { | |
424 | Py_ssize_t nodelen; |
|
424 | Py_ssize_t nodelen; | |
425 | if (PyBytes_AsStringAndSize(obj, node, &nodelen) == -1) |
|
425 | if (PyBytes_AsStringAndSize(obj, node, &nodelen) == -1) | |
426 | return -1; |
|
426 | return -1; | |
427 | if (nodelen == 20) |
|
427 | if (nodelen == 20) | |
428 | return 0; |
|
428 | return 0; | |
429 | PyErr_SetString(PyExc_ValueError, "20-byte hash required"); |
|
429 | PyErr_SetString(PyExc_ValueError, "20-byte hash required"); | |
430 | return -1; |
|
430 | return -1; | |
431 | } |
|
431 | } | |
432 |
|
432 | |||
433 | static PyObject *index_append(indexObject *self, PyObject *obj) |
|
433 | static PyObject *index_append(indexObject *self, PyObject *obj) | |
434 | { |
|
434 | { | |
435 | char *node; |
|
435 | char *node; | |
436 | Py_ssize_t len; |
|
436 | Py_ssize_t len; | |
437 |
|
437 | |||
438 | if (!PyTuple_Check(obj) || PyTuple_GET_SIZE(obj) != 8) { |
|
438 | if (!PyTuple_Check(obj) || PyTuple_GET_SIZE(obj) != 8) { | |
439 | PyErr_SetString(PyExc_TypeError, "8-tuple required"); |
|
439 | PyErr_SetString(PyExc_TypeError, "8-tuple required"); | |
440 | return NULL; |
|
440 | return NULL; | |
441 | } |
|
441 | } | |
442 |
|
442 | |||
443 | if (node_check(PyTuple_GET_ITEM(obj, 7), &node) == -1) |
|
443 | if (node_check(PyTuple_GET_ITEM(obj, 7), &node) == -1) | |
444 | return NULL; |
|
444 | return NULL; | |
445 |
|
445 | |||
446 | len = index_length(self); |
|
446 | len = index_length(self); | |
447 |
|
447 | |||
448 | if (self->added == NULL) { |
|
448 | if (self->added == NULL) { | |
449 | self->added = PyList_New(0); |
|
449 | self->added = PyList_New(0); | |
450 | if (self->added == NULL) |
|
450 | if (self->added == NULL) | |
451 | return NULL; |
|
451 | return NULL; | |
452 | } |
|
452 | } | |
453 |
|
453 | |||
454 | if (PyList_Append(self->added, obj) == -1) |
|
454 | if (PyList_Append(self->added, obj) == -1) | |
455 | return NULL; |
|
455 | return NULL; | |
456 |
|
456 | |||
457 | if (self->ntinitialized) |
|
457 | if (self->ntinitialized) | |
458 | nt_insert(&self->nt, node, (int)len); |
|
458 | nt_insert(&self->nt, node, (int)len); | |
459 |
|
459 | |||
460 | Py_CLEAR(self->headrevs); |
|
460 | Py_CLEAR(self->headrevs); | |
461 | Py_RETURN_NONE; |
|
461 | Py_RETURN_NONE; | |
462 | } |
|
462 | } | |
463 |
|
463 | |||
464 | static PyObject *index_stats(indexObject *self) |
|
464 | static PyObject *index_stats(indexObject *self) | |
465 | { |
|
465 | { | |
466 | PyObject *obj = PyDict_New(); |
|
466 | PyObject *obj = PyDict_New(); | |
467 | PyObject *s = NULL; |
|
467 | PyObject *s = NULL; | |
468 | PyObject *t = NULL; |
|
468 | PyObject *t = NULL; | |
469 |
|
469 | |||
470 | if (obj == NULL) |
|
470 | if (obj == NULL) | |
471 | return NULL; |
|
471 | return NULL; | |
472 |
|
472 | |||
473 | #define istat(__n, __d) \ |
|
473 | #define istat(__n, __d) \ | |
474 | do { \ |
|
474 | do { \ | |
475 | s = PyBytes_FromString(__d); \ |
|
475 | s = PyBytes_FromString(__d); \ | |
476 | t = PyInt_FromSsize_t(self->__n); \ |
|
476 | t = PyInt_FromSsize_t(self->__n); \ | |
477 | if (!s || !t) \ |
|
477 | if (!s || !t) \ | |
478 | goto bail; \ |
|
478 | goto bail; \ | |
479 | if (PyDict_SetItem(obj, s, t) == -1) \ |
|
479 | if (PyDict_SetItem(obj, s, t) == -1) \ | |
480 | goto bail; \ |
|
480 | goto bail; \ | |
481 | Py_CLEAR(s); \ |
|
481 | Py_CLEAR(s); \ | |
482 | Py_CLEAR(t); \ |
|
482 | Py_CLEAR(t); \ | |
483 | } while (0) |
|
483 | } while (0) | |
484 |
|
484 | |||
485 | if (self->added) { |
|
485 | if (self->added) { | |
486 | Py_ssize_t len = PyList_GET_SIZE(self->added); |
|
486 | Py_ssize_t len = PyList_GET_SIZE(self->added); | |
487 | s = PyBytes_FromString("index entries added"); |
|
487 | s = PyBytes_FromString("index entries added"); | |
488 | t = PyInt_FromSsize_t(len); |
|
488 | t = PyInt_FromSsize_t(len); | |
489 | if (!s || !t) |
|
489 | if (!s || !t) | |
490 | goto bail; |
|
490 | goto bail; | |
491 | if (PyDict_SetItem(obj, s, t) == -1) |
|
491 | if (PyDict_SetItem(obj, s, t) == -1) | |
492 | goto bail; |
|
492 | goto bail; | |
493 | Py_CLEAR(s); |
|
493 | Py_CLEAR(s); | |
494 | Py_CLEAR(t); |
|
494 | Py_CLEAR(t); | |
495 | } |
|
495 | } | |
496 |
|
496 | |||
497 | if (self->raw_length != self->length) |
|
497 | if (self->raw_length != self->length) | |
498 | istat(raw_length, "revs on disk"); |
|
498 | istat(raw_length, "revs on disk"); | |
499 | istat(length, "revs in memory"); |
|
499 | istat(length, "revs in memory"); | |
500 | istat(ntlookups, "node trie lookups"); |
|
500 | istat(ntlookups, "node trie lookups"); | |
501 | istat(ntmisses, "node trie misses"); |
|
501 | istat(ntmisses, "node trie misses"); | |
502 | istat(ntrev, "node trie last rev scanned"); |
|
502 | istat(ntrev, "node trie last rev scanned"); | |
503 | if (self->ntinitialized) { |
|
503 | if (self->ntinitialized) { | |
504 | istat(nt.capacity, "node trie capacity"); |
|
504 | istat(nt.capacity, "node trie capacity"); | |
505 | istat(nt.depth, "node trie depth"); |
|
505 | istat(nt.depth, "node trie depth"); | |
506 | istat(nt.length, "node trie count"); |
|
506 | istat(nt.length, "node trie count"); | |
507 | istat(nt.splits, "node trie splits"); |
|
507 | istat(nt.splits, "node trie splits"); | |
508 | } |
|
508 | } | |
509 |
|
509 | |||
510 | #undef istat |
|
510 | #undef istat | |
511 |
|
511 | |||
512 | return obj; |
|
512 | return obj; | |
513 |
|
513 | |||
514 | bail: |
|
514 | bail: | |
515 | Py_XDECREF(obj); |
|
515 | Py_XDECREF(obj); | |
516 | Py_XDECREF(s); |
|
516 | Py_XDECREF(s); | |
517 | Py_XDECREF(t); |
|
517 | Py_XDECREF(t); | |
518 | return NULL; |
|
518 | return NULL; | |
519 | } |
|
519 | } | |
520 |
|
520 | |||
521 | /* |
|
521 | /* | |
522 | * When we cache a list, we want to be sure the caller can't mutate |
|
522 | * When we cache a list, we want to be sure the caller can't mutate | |
523 | * the cached copy. |
|
523 | * the cached copy. | |
524 | */ |
|
524 | */ | |
525 | static PyObject *list_copy(PyObject *list) |
|
525 | static PyObject *list_copy(PyObject *list) | |
526 | { |
|
526 | { | |
527 | Py_ssize_t len = PyList_GET_SIZE(list); |
|
527 | Py_ssize_t len = PyList_GET_SIZE(list); | |
528 | PyObject *newlist = PyList_New(len); |
|
528 | PyObject *newlist = PyList_New(len); | |
529 | Py_ssize_t i; |
|
529 | Py_ssize_t i; | |
530 |
|
530 | |||
531 | if (newlist == NULL) |
|
531 | if (newlist == NULL) | |
532 | return NULL; |
|
532 | return NULL; | |
533 |
|
533 | |||
534 | for (i = 0; i < len; i++) { |
|
534 | for (i = 0; i < len; i++) { | |
535 | PyObject *obj = PyList_GET_ITEM(list, i); |
|
535 | PyObject *obj = PyList_GET_ITEM(list, i); | |
536 | Py_INCREF(obj); |
|
536 | Py_INCREF(obj); | |
537 | PyList_SET_ITEM(newlist, i, obj); |
|
537 | PyList_SET_ITEM(newlist, i, obj); | |
538 | } |
|
538 | } | |
539 |
|
539 | |||
540 | return newlist; |
|
540 | return newlist; | |
541 | } |
|
541 | } | |
542 |
|
542 | |||
543 | static int check_filter(PyObject *filter, Py_ssize_t arg) |
|
543 | static int check_filter(PyObject *filter, Py_ssize_t arg) | |
544 | { |
|
544 | { | |
545 | if (filter) { |
|
545 | if (filter) { | |
546 | PyObject *arglist, *result; |
|
546 | PyObject *arglist, *result; | |
547 | int isfiltered; |
|
547 | int isfiltered; | |
548 |
|
548 | |||
549 | arglist = Py_BuildValue("(n)", arg); |
|
549 | arglist = Py_BuildValue("(n)", arg); | |
550 | if (!arglist) { |
|
550 | if (!arglist) { | |
551 | return -1; |
|
551 | return -1; | |
552 | } |
|
552 | } | |
553 |
|
553 | |||
554 | result = PyEval_CallObject(filter, arglist); |
|
554 | result = PyEval_CallObject(filter, arglist); | |
555 | Py_DECREF(arglist); |
|
555 | Py_DECREF(arglist); | |
556 | if (!result) { |
|
556 | if (!result) { | |
557 | return -1; |
|
557 | return -1; | |
558 | } |
|
558 | } | |
559 |
|
559 | |||
560 | /* PyObject_IsTrue returns 1 if true, 0 if false, -1 if error, |
|
560 | /* PyObject_IsTrue returns 1 if true, 0 if false, -1 if error, | |
561 | * same as this function, so we can just return it directly.*/ |
|
561 | * same as this function, so we can just return it directly.*/ | |
562 | isfiltered = PyObject_IsTrue(result); |
|
562 | isfiltered = PyObject_IsTrue(result); | |
563 | Py_DECREF(result); |
|
563 | Py_DECREF(result); | |
564 | return isfiltered; |
|
564 | return isfiltered; | |
565 | } else { |
|
565 | } else { | |
566 | return 0; |
|
566 | return 0; | |
567 | } |
|
567 | } | |
568 | } |
|
568 | } | |
569 |
|
569 | |||
570 | static Py_ssize_t add_roots_get_min(indexObject *self, PyObject *list, |
|
570 | static Py_ssize_t add_roots_get_min(indexObject *self, PyObject *list, | |
571 | Py_ssize_t marker, char *phases) |
|
571 | Py_ssize_t marker, char *phases) | |
572 | { |
|
572 | { | |
573 | PyObject *iter = NULL; |
|
573 | PyObject *iter = NULL; | |
574 | PyObject *iter_item = NULL; |
|
574 | PyObject *iter_item = NULL; | |
575 | Py_ssize_t min_idx = index_length(self) + 2; |
|
575 | Py_ssize_t min_idx = index_length(self) + 2; | |
576 | long iter_item_long; |
|
576 | long iter_item_long; | |
577 |
|
577 | |||
578 | if (PyList_GET_SIZE(list) != 0) { |
|
578 | if (PyList_GET_SIZE(list) != 0) { | |
579 | iter = PyObject_GetIter(list); |
|
579 | iter = PyObject_GetIter(list); | |
580 | if (iter == NULL) |
|
580 | if (iter == NULL) | |
581 | return -2; |
|
581 | return -2; | |
582 | while ((iter_item = PyIter_Next(iter))) { |
|
582 | while ((iter_item = PyIter_Next(iter))) { | |
583 | if (!pylong_to_long(iter_item, &iter_item_long)) { |
|
583 | if (!pylong_to_long(iter_item, &iter_item_long)) { | |
584 | Py_DECREF(iter_item); |
|
584 | Py_DECREF(iter_item); | |
585 | return -2; |
|
585 | return -2; | |
586 | } |
|
586 | } | |
587 | Py_DECREF(iter_item); |
|
587 | Py_DECREF(iter_item); | |
588 | if (iter_item_long < min_idx) |
|
588 | if (iter_item_long < min_idx) | |
589 | min_idx = iter_item_long; |
|
589 | min_idx = iter_item_long; | |
590 | phases[iter_item_long] = (char)marker; |
|
590 | phases[iter_item_long] = (char)marker; | |
591 | } |
|
591 | } | |
592 | Py_DECREF(iter); |
|
592 | Py_DECREF(iter); | |
593 | } |
|
593 | } | |
594 |
|
594 | |||
595 | return min_idx; |
|
595 | return min_idx; | |
596 | } |
|
596 | } | |
597 |
|
597 | |||
598 | static inline void set_phase_from_parents(char *phases, int parent_1, |
|
598 | static inline void set_phase_from_parents(char *phases, int parent_1, | |
599 | int parent_2, Py_ssize_t i) |
|
599 | int parent_2, Py_ssize_t i) | |
600 | { |
|
600 | { | |
601 | if (parent_1 >= 0 && phases[parent_1] > phases[i]) |
|
601 | if (parent_1 >= 0 && phases[parent_1] > phases[i]) | |
602 | phases[i] = phases[parent_1]; |
|
602 | phases[i] = phases[parent_1]; | |
603 | if (parent_2 >= 0 && phases[parent_2] > phases[i]) |
|
603 | if (parent_2 >= 0 && phases[parent_2] > phases[i]) | |
604 | phases[i] = phases[parent_2]; |
|
604 | phases[i] = phases[parent_2]; | |
605 | } |
|
605 | } | |
606 |
|
606 | |||
607 | static PyObject *reachableroots2(indexObject *self, PyObject *args) |
|
607 | static PyObject *reachableroots2(indexObject *self, PyObject *args) | |
608 | { |
|
608 | { | |
609 |
|
609 | |||
610 | /* Input */ |
|
610 | /* Input */ | |
611 | long minroot; |
|
611 | long minroot; | |
612 | PyObject *includepatharg = NULL; |
|
612 | PyObject *includepatharg = NULL; | |
613 | int includepath = 0; |
|
613 | int includepath = 0; | |
614 | /* heads and roots are lists */ |
|
614 | /* heads and roots are lists */ | |
615 | PyObject *heads = NULL; |
|
615 | PyObject *heads = NULL; | |
616 | PyObject *roots = NULL; |
|
616 | PyObject *roots = NULL; | |
617 | PyObject *reachable = NULL; |
|
617 | PyObject *reachable = NULL; | |
618 |
|
618 | |||
619 | PyObject *val; |
|
619 | PyObject *val; | |
620 | Py_ssize_t len = index_length(self); |
|
620 | Py_ssize_t len = index_length(self); | |
621 | long revnum; |
|
621 | long revnum; | |
622 | Py_ssize_t k; |
|
622 | Py_ssize_t k; | |
623 | Py_ssize_t i; |
|
623 | Py_ssize_t i; | |
624 | Py_ssize_t l; |
|
624 | Py_ssize_t l; | |
625 | int r; |
|
625 | int r; | |
626 | int parents[2]; |
|
626 | int parents[2]; | |
627 |
|
627 | |||
628 | /* Internal data structure: |
|
628 | /* Internal data structure: | |
629 | * tovisit: array of length len+1 (all revs + nullrev), filled upto |
|
629 | * tovisit: array of length len+1 (all revs + nullrev), filled upto | |
630 | * lentovisit |
|
630 | * lentovisit | |
631 | * |
|
631 | * | |
632 | * revstates: array of length len+1 (all revs + nullrev) */ |
|
632 | * revstates: array of length len+1 (all revs + nullrev) */ | |
633 | int *tovisit = NULL; |
|
633 | int *tovisit = NULL; | |
634 | long lentovisit = 0; |
|
634 | long lentovisit = 0; | |
635 | enum { RS_SEEN = 1, RS_ROOT = 2, RS_REACHABLE = 4 }; |
|
635 | enum { RS_SEEN = 1, RS_ROOT = 2, RS_REACHABLE = 4 }; | |
636 | char *revstates = NULL; |
|
636 | char *revstates = NULL; | |
637 |
|
637 | |||
638 | /* Get arguments */ |
|
638 | /* Get arguments */ | |
639 | if (!PyArg_ParseTuple(args, "lO!O!O!", &minroot, &PyList_Type, &heads, |
|
639 | if (!PyArg_ParseTuple(args, "lO!O!O!", &minroot, &PyList_Type, &heads, | |
640 | &PyList_Type, &roots, &PyBool_Type, |
|
640 | &PyList_Type, &roots, &PyBool_Type, | |
641 | &includepatharg)) |
|
641 | &includepatharg)) | |
642 | goto bail; |
|
642 | goto bail; | |
643 |
|
643 | |||
644 | if (includepatharg == Py_True) |
|
644 | if (includepatharg == Py_True) | |
645 | includepath = 1; |
|
645 | includepath = 1; | |
646 |
|
646 | |||
647 | /* Initialize return set */ |
|
647 | /* Initialize return set */ | |
648 | reachable = PyList_New(0); |
|
648 | reachable = PyList_New(0); | |
649 | if (reachable == NULL) |
|
649 | if (reachable == NULL) | |
650 | goto bail; |
|
650 | goto bail; | |
651 |
|
651 | |||
652 | /* Initialize internal datastructures */ |
|
652 | /* Initialize internal datastructures */ | |
653 | tovisit = (int *)malloc((len + 1) * sizeof(int)); |
|
653 | tovisit = (int *)malloc((len + 1) * sizeof(int)); | |
654 | if (tovisit == NULL) { |
|
654 | if (tovisit == NULL) { | |
655 | PyErr_NoMemory(); |
|
655 | PyErr_NoMemory(); | |
656 | goto bail; |
|
656 | goto bail; | |
657 | } |
|
657 | } | |
658 |
|
658 | |||
659 | revstates = (char *)calloc(len + 1, 1); |
|
659 | revstates = (char *)calloc(len + 1, 1); | |
660 | if (revstates == NULL) { |
|
660 | if (revstates == NULL) { | |
661 | PyErr_NoMemory(); |
|
661 | PyErr_NoMemory(); | |
662 | goto bail; |
|
662 | goto bail; | |
663 | } |
|
663 | } | |
664 |
|
664 | |||
665 | l = PyList_GET_SIZE(roots); |
|
665 | l = PyList_GET_SIZE(roots); | |
666 | for (i = 0; i < l; i++) { |
|
666 | for (i = 0; i < l; i++) { | |
667 | revnum = PyInt_AsLong(PyList_GET_ITEM(roots, i)); |
|
667 | revnum = PyInt_AsLong(PyList_GET_ITEM(roots, i)); | |
668 | if (revnum == -1 && PyErr_Occurred()) |
|
668 | if (revnum == -1 && PyErr_Occurred()) | |
669 | goto bail; |
|
669 | goto bail; | |
670 | /* If root is out of range, e.g. wdir(), it must be unreachable |
|
670 | /* If root is out of range, e.g. wdir(), it must be unreachable | |
671 | * from heads. So we can just ignore it. */ |
|
671 | * from heads. So we can just ignore it. */ | |
672 | if (revnum + 1 < 0 || revnum + 1 >= len + 1) |
|
672 | if (revnum + 1 < 0 || revnum + 1 >= len + 1) | |
673 | continue; |
|
673 | continue; | |
674 | revstates[revnum + 1] |= RS_ROOT; |
|
674 | revstates[revnum + 1] |= RS_ROOT; | |
675 | } |
|
675 | } | |
676 |
|
676 | |||
677 | /* Populate tovisit with all the heads */ |
|
677 | /* Populate tovisit with all the heads */ | |
678 | l = PyList_GET_SIZE(heads); |
|
678 | l = PyList_GET_SIZE(heads); | |
679 | for (i = 0; i < l; i++) { |
|
679 | for (i = 0; i < l; i++) { | |
680 | revnum = PyInt_AsLong(PyList_GET_ITEM(heads, i)); |
|
680 | revnum = PyInt_AsLong(PyList_GET_ITEM(heads, i)); | |
681 | if (revnum == -1 && PyErr_Occurred()) |
|
681 | if (revnum == -1 && PyErr_Occurred()) | |
682 | goto bail; |
|
682 | goto bail; | |
683 | if (revnum + 1 < 0 || revnum + 1 >= len + 1) { |
|
683 | if (revnum + 1 < 0 || revnum + 1 >= len + 1) { | |
684 | PyErr_SetString(PyExc_IndexError, "head out of range"); |
|
684 | PyErr_SetString(PyExc_IndexError, "head out of range"); | |
685 | goto bail; |
|
685 | goto bail; | |
686 | } |
|
686 | } | |
687 | if (!(revstates[revnum + 1] & RS_SEEN)) { |
|
687 | if (!(revstates[revnum + 1] & RS_SEEN)) { | |
688 | tovisit[lentovisit++] = (int)revnum; |
|
688 | tovisit[lentovisit++] = (int)revnum; | |
689 | revstates[revnum + 1] |= RS_SEEN; |
|
689 | revstates[revnum + 1] |= RS_SEEN; | |
690 | } |
|
690 | } | |
691 | } |
|
691 | } | |
692 |
|
692 | |||
693 | /* Visit the tovisit list and find the reachable roots */ |
|
693 | /* Visit the tovisit list and find the reachable roots */ | |
694 | k = 0; |
|
694 | k = 0; | |
695 | while (k < lentovisit) { |
|
695 | while (k < lentovisit) { | |
696 | /* Add the node to reachable if it is a root*/ |
|
696 | /* Add the node to reachable if it is a root*/ | |
697 | revnum = tovisit[k++]; |
|
697 | revnum = tovisit[k++]; | |
698 | if (revstates[revnum + 1] & RS_ROOT) { |
|
698 | if (revstates[revnum + 1] & RS_ROOT) { | |
699 | revstates[revnum + 1] |= RS_REACHABLE; |
|
699 | revstates[revnum + 1] |= RS_REACHABLE; | |
700 | val = PyInt_FromLong(revnum); |
|
700 | val = PyInt_FromLong(revnum); | |
701 | if (val == NULL) |
|
701 | if (val == NULL) | |
702 | goto bail; |
|
702 | goto bail; | |
703 | r = PyList_Append(reachable, val); |
|
703 | r = PyList_Append(reachable, val); | |
704 | Py_DECREF(val); |
|
704 | Py_DECREF(val); | |
705 | if (r < 0) |
|
705 | if (r < 0) | |
706 | goto bail; |
|
706 | goto bail; | |
707 | if (includepath == 0) |
|
707 | if (includepath == 0) | |
708 | continue; |
|
708 | continue; | |
709 | } |
|
709 | } | |
710 |
|
710 | |||
711 | /* Add its parents to the list of nodes to visit */ |
|
711 | /* Add its parents to the list of nodes to visit */ | |
712 | if (revnum == nullrev) |
|
712 | if (revnum == nullrev) | |
713 | continue; |
|
713 | continue; | |
714 | r = index_get_parents(self, revnum, parents, (int)len - 1); |
|
714 | r = index_get_parents(self, revnum, parents, (int)len - 1); | |
715 | if (r < 0) |
|
715 | if (r < 0) | |
716 | goto bail; |
|
716 | goto bail; | |
717 | for (i = 0; i < 2; i++) { |
|
717 | for (i = 0; i < 2; i++) { | |
718 | if (!(revstates[parents[i] + 1] & RS_SEEN) && |
|
718 | if (!(revstates[parents[i] + 1] & RS_SEEN) && | |
719 | parents[i] >= minroot) { |
|
719 | parents[i] >= minroot) { | |
720 | tovisit[lentovisit++] = parents[i]; |
|
720 | tovisit[lentovisit++] = parents[i]; | |
721 | revstates[parents[i] + 1] |= RS_SEEN; |
|
721 | revstates[parents[i] + 1] |= RS_SEEN; | |
722 | } |
|
722 | } | |
723 | } |
|
723 | } | |
724 | } |
|
724 | } | |
725 |
|
725 | |||
726 | /* Find all the nodes in between the roots we found and the heads |
|
726 | /* Find all the nodes in between the roots we found and the heads | |
727 | * and add them to the reachable set */ |
|
727 | * and add them to the reachable set */ | |
728 | if (includepath == 1) { |
|
728 | if (includepath == 1) { | |
729 | long minidx = minroot; |
|
729 | long minidx = minroot; | |
730 | if (minidx < 0) |
|
730 | if (minidx < 0) | |
731 | minidx = 0; |
|
731 | minidx = 0; | |
732 | for (i = minidx; i < len; i++) { |
|
732 | for (i = minidx; i < len; i++) { | |
733 | if (!(revstates[i + 1] & RS_SEEN)) |
|
733 | if (!(revstates[i + 1] & RS_SEEN)) | |
734 | continue; |
|
734 | continue; | |
735 | r = index_get_parents(self, i, parents, (int)len - 1); |
|
735 | r = index_get_parents(self, i, parents, (int)len - 1); | |
736 | /* Corrupted index file, error is set from |
|
736 | /* Corrupted index file, error is set from | |
737 | * index_get_parents */ |
|
737 | * index_get_parents */ | |
738 | if (r < 0) |
|
738 | if (r < 0) | |
739 | goto bail; |
|
739 | goto bail; | |
740 | if (((revstates[parents[0] + 1] | |
|
740 | if (((revstates[parents[0] + 1] | | |
741 | revstates[parents[1] + 1]) & |
|
741 | revstates[parents[1] + 1]) & | |
742 | RS_REACHABLE) && |
|
742 | RS_REACHABLE) && | |
743 | !(revstates[i + 1] & RS_REACHABLE)) { |
|
743 | !(revstates[i + 1] & RS_REACHABLE)) { | |
744 | revstates[i + 1] |= RS_REACHABLE; |
|
744 | revstates[i + 1] |= RS_REACHABLE; | |
745 | val = PyInt_FromSsize_t(i); |
|
745 | val = PyInt_FromSsize_t(i); | |
746 | if (val == NULL) |
|
746 | if (val == NULL) | |
747 | goto bail; |
|
747 | goto bail; | |
748 | r = PyList_Append(reachable, val); |
|
748 | r = PyList_Append(reachable, val); | |
749 | Py_DECREF(val); |
|
749 | Py_DECREF(val); | |
750 | if (r < 0) |
|
750 | if (r < 0) | |
751 | goto bail; |
|
751 | goto bail; | |
752 | } |
|
752 | } | |
753 | } |
|
753 | } | |
754 | } |
|
754 | } | |
755 |
|
755 | |||
756 | free(revstates); |
|
756 | free(revstates); | |
757 | free(tovisit); |
|
757 | free(tovisit); | |
758 | return reachable; |
|
758 | return reachable; | |
759 | bail: |
|
759 | bail: | |
760 | Py_XDECREF(reachable); |
|
760 | Py_XDECREF(reachable); | |
761 | free(revstates); |
|
761 | free(revstates); | |
762 | free(tovisit); |
|
762 | free(tovisit); | |
763 | return NULL; |
|
763 | return NULL; | |
764 | } |
|
764 | } | |
765 |
|
765 | |||
766 | static PyObject *compute_phases_map_sets(indexObject *self, PyObject *args) |
|
766 | static PyObject *compute_phases_map_sets(indexObject *self, PyObject *args) | |
767 | { |
|
767 | { | |
768 | PyObject *roots = Py_None; |
|
768 | PyObject *roots = Py_None; | |
769 | PyObject *ret = NULL; |
|
769 | PyObject *ret = NULL; | |
770 | PyObject *phasessize = NULL; |
|
770 | PyObject *phasessize = NULL; | |
771 | PyObject *phaseroots = NULL; |
|
771 | PyObject *phaseroots = NULL; | |
772 | PyObject *phaseset = NULL; |
|
772 | PyObject *phaseset = NULL; | |
773 | PyObject *phasessetlist = NULL; |
|
773 | PyObject *phasessetlist = NULL; | |
774 | PyObject *rev = NULL; |
|
774 | PyObject *rev = NULL; | |
775 | Py_ssize_t len = index_length(self); |
|
775 | Py_ssize_t len = index_length(self); | |
776 | Py_ssize_t numphase = 0; |
|
776 | Py_ssize_t numphase = 0; | |
777 | Py_ssize_t minrevallphases = 0; |
|
777 | Py_ssize_t minrevallphases = 0; | |
778 | Py_ssize_t minrevphase = 0; |
|
778 | Py_ssize_t minrevphase = 0; | |
779 | Py_ssize_t i = 0; |
|
779 | Py_ssize_t i = 0; | |
780 | char *phases = NULL; |
|
780 | char *phases = NULL; | |
781 | long phase; |
|
781 | long phase; | |
782 |
|
782 | |||
783 | if (!PyArg_ParseTuple(args, "O", &roots)) |
|
783 | if (!PyArg_ParseTuple(args, "O", &roots)) | |
784 | goto done; |
|
784 | goto done; | |
785 | if (roots == NULL || !PyList_Check(roots)) { |
|
785 | if (roots == NULL || !PyList_Check(roots)) { | |
786 | PyErr_SetString(PyExc_TypeError, "roots must be a list"); |
|
786 | PyErr_SetString(PyExc_TypeError, "roots must be a list"); | |
787 | goto done; |
|
787 | goto done; | |
788 | } |
|
788 | } | |
789 |
|
789 | |||
790 | phases = calloc( |
|
790 | phases = calloc( | |
791 | len, 1); /* phase per rev: {0: public, 1: draft, 2: secret} */ |
|
791 | len, 1); /* phase per rev: {0: public, 1: draft, 2: secret} */ | |
792 | if (phases == NULL) { |
|
792 | if (phases == NULL) { | |
793 | PyErr_NoMemory(); |
|
793 | PyErr_NoMemory(); | |
794 | goto done; |
|
794 | goto done; | |
795 | } |
|
795 | } | |
796 | /* Put the phase information of all the roots in phases */ |
|
796 | /* Put the phase information of all the roots in phases */ | |
797 | numphase = PyList_GET_SIZE(roots) + 1; |
|
797 | numphase = PyList_GET_SIZE(roots) + 1; | |
798 | minrevallphases = len + 1; |
|
798 | minrevallphases = len + 1; | |
799 | phasessetlist = PyList_New(numphase); |
|
799 | phasessetlist = PyList_New(numphase); | |
800 | if (phasessetlist == NULL) |
|
800 | if (phasessetlist == NULL) | |
801 | goto done; |
|
801 | goto done; | |
802 |
|
802 | |||
803 | PyList_SET_ITEM(phasessetlist, 0, Py_None); |
|
803 | PyList_SET_ITEM(phasessetlist, 0, Py_None); | |
804 | Py_INCREF(Py_None); |
|
804 | Py_INCREF(Py_None); | |
805 |
|
805 | |||
806 | for (i = 0; i < numphase - 1; i++) { |
|
806 | for (i = 0; i < numphase - 1; i++) { | |
807 | phaseroots = PyList_GET_ITEM(roots, i); |
|
807 | phaseroots = PyList_GET_ITEM(roots, i); | |
808 | phaseset = PySet_New(NULL); |
|
808 | phaseset = PySet_New(NULL); | |
809 | if (phaseset == NULL) |
|
809 | if (phaseset == NULL) | |
810 | goto release; |
|
810 | goto release; | |
811 | PyList_SET_ITEM(phasessetlist, i + 1, phaseset); |
|
811 | PyList_SET_ITEM(phasessetlist, i + 1, phaseset); | |
812 | if (!PyList_Check(phaseroots)) { |
|
812 | if (!PyList_Check(phaseroots)) { | |
813 | PyErr_SetString(PyExc_TypeError, |
|
813 | PyErr_SetString(PyExc_TypeError, | |
814 | "roots item must be a list"); |
|
814 | "roots item must be a list"); | |
815 | goto release; |
|
815 | goto release; | |
816 | } |
|
816 | } | |
817 | minrevphase = |
|
817 | minrevphase = | |
818 | add_roots_get_min(self, phaseroots, i + 1, phases); |
|
818 | add_roots_get_min(self, phaseroots, i + 1, phases); | |
819 | if (minrevphase == -2) /* Error from add_roots_get_min */ |
|
819 | if (minrevphase == -2) /* Error from add_roots_get_min */ | |
820 | goto release; |
|
820 | goto release; | |
821 | minrevallphases = MIN(minrevallphases, minrevphase); |
|
821 | minrevallphases = MIN(minrevallphases, minrevphase); | |
822 | } |
|
822 | } | |
823 | /* Propagate the phase information from the roots to the revs */ |
|
823 | /* Propagate the phase information from the roots to the revs */ | |
824 | if (minrevallphases != -1) { |
|
824 | if (minrevallphases != -1) { | |
825 | int parents[2]; |
|
825 | int parents[2]; | |
826 | for (i = minrevallphases; i < len; i++) { |
|
826 | for (i = minrevallphases; i < len; i++) { | |
827 | if (index_get_parents(self, i, parents, (int)len - 1) < |
|
827 | if (index_get_parents(self, i, parents, (int)len - 1) < | |
828 | 0) |
|
828 | 0) | |
829 | goto release; |
|
829 | goto release; | |
830 | set_phase_from_parents(phases, parents[0], parents[1], |
|
830 | set_phase_from_parents(phases, parents[0], parents[1], | |
831 | i); |
|
831 | i); | |
832 | } |
|
832 | } | |
833 | } |
|
833 | } | |
834 | /* Transform phase list to a python list */ |
|
834 | /* Transform phase list to a python list */ | |
835 | phasessize = PyInt_FromSsize_t(len); |
|
835 | phasessize = PyInt_FromSsize_t(len); | |
836 | if (phasessize == NULL) |
|
836 | if (phasessize == NULL) | |
837 | goto release; |
|
837 | goto release; | |
838 | for (i = 0; i < len; i++) { |
|
838 | for (i = 0; i < len; i++) { | |
839 | phase = phases[i]; |
|
839 | phase = phases[i]; | |
840 | /* We only store the sets of phase for non public phase, the |
|
840 | /* We only store the sets of phase for non public phase, the | |
841 | * public phase is computed as a difference */ |
|
841 | * public phase is computed as a difference */ | |
842 | if (phase != 0) { |
|
842 | if (phase != 0) { | |
843 | phaseset = PyList_GET_ITEM(phasessetlist, phase); |
|
843 | phaseset = PyList_GET_ITEM(phasessetlist, phase); | |
844 | rev = PyInt_FromSsize_t(i); |
|
844 | rev = PyInt_FromSsize_t(i); | |
845 | if (rev == NULL) |
|
845 | if (rev == NULL) | |
846 | goto release; |
|
846 | goto release; | |
847 | PySet_Add(phaseset, rev); |
|
847 | PySet_Add(phaseset, rev); | |
848 | Py_XDECREF(rev); |
|
848 | Py_XDECREF(rev); | |
849 | } |
|
849 | } | |
850 | } |
|
850 | } | |
851 | ret = PyTuple_Pack(2, phasessize, phasessetlist); |
|
851 | ret = PyTuple_Pack(2, phasessize, phasessetlist); | |
852 |
|
852 | |||
853 | release: |
|
853 | release: | |
854 | Py_XDECREF(phasessize); |
|
854 | Py_XDECREF(phasessize); | |
855 | Py_XDECREF(phasessetlist); |
|
855 | Py_XDECREF(phasessetlist); | |
856 | done: |
|
856 | done: | |
857 | free(phases); |
|
857 | free(phases); | |
858 | return ret; |
|
858 | return ret; | |
859 | } |
|
859 | } | |
860 |
|
860 | |||
861 | static PyObject *index_headrevs(indexObject *self, PyObject *args) |
|
861 | static PyObject *index_headrevs(indexObject *self, PyObject *args) | |
862 | { |
|
862 | { | |
863 | Py_ssize_t i, j, len; |
|
863 | Py_ssize_t i, j, len; | |
864 | char *nothead = NULL; |
|
864 | char *nothead = NULL; | |
865 | PyObject *heads = NULL; |
|
865 | PyObject *heads = NULL; | |
866 | PyObject *filter = NULL; |
|
866 | PyObject *filter = NULL; | |
867 | PyObject *filteredrevs = Py_None; |
|
867 | PyObject *filteredrevs = Py_None; | |
868 |
|
868 | |||
869 | if (!PyArg_ParseTuple(args, "|O", &filteredrevs)) { |
|
869 | if (!PyArg_ParseTuple(args, "|O", &filteredrevs)) { | |
870 | return NULL; |
|
870 | return NULL; | |
871 | } |
|
871 | } | |
872 |
|
872 | |||
873 | if (self->headrevs && filteredrevs == self->filteredrevs) |
|
873 | if (self->headrevs && filteredrevs == self->filteredrevs) | |
874 | return list_copy(self->headrevs); |
|
874 | return list_copy(self->headrevs); | |
875 |
|
875 | |||
876 | Py_DECREF(self->filteredrevs); |
|
876 | Py_DECREF(self->filteredrevs); | |
877 | self->filteredrevs = filteredrevs; |
|
877 | self->filteredrevs = filteredrevs; | |
878 | Py_INCREF(filteredrevs); |
|
878 | Py_INCREF(filteredrevs); | |
879 |
|
879 | |||
880 | if (filteredrevs != Py_None) { |
|
880 | if (filteredrevs != Py_None) { | |
881 | filter = PyObject_GetAttrString(filteredrevs, "__contains__"); |
|
881 | filter = PyObject_GetAttrString(filteredrevs, "__contains__"); | |
882 | if (!filter) { |
|
882 | if (!filter) { | |
883 | PyErr_SetString( |
|
883 | PyErr_SetString( | |
884 | PyExc_TypeError, |
|
884 | PyExc_TypeError, | |
885 | "filteredrevs has no attribute __contains__"); |
|
885 | "filteredrevs has no attribute __contains__"); | |
886 | goto bail; |
|
886 | goto bail; | |
887 | } |
|
887 | } | |
888 | } |
|
888 | } | |
889 |
|
889 | |||
890 | len = index_length(self); |
|
890 | len = index_length(self); | |
891 | heads = PyList_New(0); |
|
891 | heads = PyList_New(0); | |
892 | if (heads == NULL) |
|
892 | if (heads == NULL) | |
893 | goto bail; |
|
893 | goto bail; | |
894 | if (len == 0) { |
|
894 | if (len == 0) { | |
895 | PyObject *nullid = PyInt_FromLong(-1); |
|
895 | PyObject *nullid = PyInt_FromLong(-1); | |
896 | if (nullid == NULL || PyList_Append(heads, nullid) == -1) { |
|
896 | if (nullid == NULL || PyList_Append(heads, nullid) == -1) { | |
897 | Py_XDECREF(nullid); |
|
897 | Py_XDECREF(nullid); | |
898 | goto bail; |
|
898 | goto bail; | |
899 | } |
|
899 | } | |
900 | goto done; |
|
900 | goto done; | |
901 | } |
|
901 | } | |
902 |
|
902 | |||
903 | nothead = calloc(len, 1); |
|
903 | nothead = calloc(len, 1); | |
904 | if (nothead == NULL) { |
|
904 | if (nothead == NULL) { | |
905 | PyErr_NoMemory(); |
|
905 | PyErr_NoMemory(); | |
906 | goto bail; |
|
906 | goto bail; | |
907 | } |
|
907 | } | |
908 |
|
908 | |||
909 | for (i = len - 1; i >= 0; i--) { |
|
909 | for (i = len - 1; i >= 0; i--) { | |
910 | int isfiltered; |
|
910 | int isfiltered; | |
911 | int parents[2]; |
|
911 | int parents[2]; | |
912 |
|
912 | |||
913 | /* If nothead[i] == 1, it means we've seen an unfiltered child |
|
913 | /* If nothead[i] == 1, it means we've seen an unfiltered child | |
914 | * of this node already, and therefore this node is not |
|
914 | * of this node already, and therefore this node is not | |
915 | * filtered. So we can skip the expensive check_filter step. |
|
915 | * filtered. So we can skip the expensive check_filter step. | |
916 | */ |
|
916 | */ | |
917 | if (nothead[i] != 1) { |
|
917 | if (nothead[i] != 1) { | |
918 | isfiltered = check_filter(filter, i); |
|
918 | isfiltered = check_filter(filter, i); | |
919 | if (isfiltered == -1) { |
|
919 | if (isfiltered == -1) { | |
920 | PyErr_SetString(PyExc_TypeError, |
|
920 | PyErr_SetString(PyExc_TypeError, | |
921 | "unable to check filter"); |
|
921 | "unable to check filter"); | |
922 | goto bail; |
|
922 | goto bail; | |
923 | } |
|
923 | } | |
924 |
|
924 | |||
925 | if (isfiltered) { |
|
925 | if (isfiltered) { | |
926 | nothead[i] = 1; |
|
926 | nothead[i] = 1; | |
927 | continue; |
|
927 | continue; | |
928 | } |
|
928 | } | |
929 | } |
|
929 | } | |
930 |
|
930 | |||
931 | if (index_get_parents(self, i, parents, (int)len - 1) < 0) |
|
931 | if (index_get_parents(self, i, parents, (int)len - 1) < 0) | |
932 | goto bail; |
|
932 | goto bail; | |
933 | for (j = 0; j < 2; j++) { |
|
933 | for (j = 0; j < 2; j++) { | |
934 | if (parents[j] >= 0) |
|
934 | if (parents[j] >= 0) | |
935 | nothead[parents[j]] = 1; |
|
935 | nothead[parents[j]] = 1; | |
936 | } |
|
936 | } | |
937 | } |
|
937 | } | |
938 |
|
938 | |||
939 | for (i = 0; i < len; i++) { |
|
939 | for (i = 0; i < len; i++) { | |
940 | PyObject *head; |
|
940 | PyObject *head; | |
941 |
|
941 | |||
942 | if (nothead[i]) |
|
942 | if (nothead[i]) | |
943 | continue; |
|
943 | continue; | |
944 | head = PyInt_FromSsize_t(i); |
|
944 | head = PyInt_FromSsize_t(i); | |
945 | if (head == NULL || PyList_Append(heads, head) == -1) { |
|
945 | if (head == NULL || PyList_Append(heads, head) == -1) { | |
946 | Py_XDECREF(head); |
|
946 | Py_XDECREF(head); | |
947 | goto bail; |
|
947 | goto bail; | |
948 | } |
|
948 | } | |
949 | } |
|
949 | } | |
950 |
|
950 | |||
951 | done: |
|
951 | done: | |
952 | self->headrevs = heads; |
|
952 | self->headrevs = heads; | |
953 | Py_XDECREF(filter); |
|
953 | Py_XDECREF(filter); | |
954 | free(nothead); |
|
954 | free(nothead); | |
955 | return list_copy(self->headrevs); |
|
955 | return list_copy(self->headrevs); | |
956 | bail: |
|
956 | bail: | |
957 | Py_XDECREF(filter); |
|
957 | Py_XDECREF(filter); | |
958 | Py_XDECREF(heads); |
|
958 | Py_XDECREF(heads); | |
959 | free(nothead); |
|
959 | free(nothead); | |
960 | return NULL; |
|
960 | return NULL; | |
961 | } |
|
961 | } | |
962 |
|
962 | |||
963 | /** |
|
963 | /** | |
964 | * Obtain the base revision index entry. |
|
964 | * Obtain the base revision index entry. | |
965 | * |
|
965 | * | |
966 | * Callers must ensure that rev >= 0 or illegal memory access may occur. |
|
966 | * Callers must ensure that rev >= 0 or illegal memory access may occur. | |
967 | */ |
|
967 | */ | |
968 | static inline int index_baserev(indexObject *self, int rev) |
|
968 | static inline int index_baserev(indexObject *self, int rev) | |
969 | { |
|
969 | { | |
970 | const char *data; |
|
970 | const char *data; | |
971 | int result; |
|
971 | int result; | |
972 |
|
972 | |||
973 | if (rev >= self->length) { |
|
973 | if (rev >= self->length) { | |
974 | PyObject *tuple = |
|
974 | PyObject *tuple = | |
975 | PyList_GET_ITEM(self->added, rev - self->length); |
|
975 | PyList_GET_ITEM(self->added, rev - self->length); | |
976 | long ret; |
|
976 | long ret; | |
977 | if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 3), &ret)) { |
|
977 | if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 3), &ret)) { | |
978 | return -2; |
|
978 | return -2; | |
979 | } |
|
979 | } | |
980 | result = (int)ret; |
|
980 | result = (int)ret; | |
981 | } else { |
|
981 | } else { | |
982 | data = index_deref(self, rev); |
|
982 | data = index_deref(self, rev); | |
983 | if (data == NULL) { |
|
983 | if (data == NULL) { | |
984 | return -2; |
|
984 | return -2; | |
985 | } |
|
985 | } | |
986 |
|
986 | |||
987 | result = getbe32(data + 16); |
|
987 | result = getbe32(data + 16); | |
988 | } |
|
988 | } | |
989 | if (result > rev) { |
|
989 | if (result > rev) { | |
990 | PyErr_Format( |
|
990 | PyErr_Format( | |
991 | PyExc_ValueError, |
|
991 | PyExc_ValueError, | |
992 | "corrupted revlog, revision base above revision: %d, %d", |
|
992 | "corrupted revlog, revision base above revision: %d, %d", | |
993 | rev, result); |
|
993 | rev, result); | |
994 | return -2; |
|
994 | return -2; | |
995 | } |
|
995 | } | |
996 | if (result < -1) { |
|
996 | if (result < -1) { | |
997 | PyErr_Format( |
|
997 | PyErr_Format( | |
998 | PyExc_ValueError, |
|
998 | PyExc_ValueError, | |
999 | "corrupted revlog, revision base out of range: %d, %d", rev, |
|
999 | "corrupted revlog, revision base out of range: %d, %d", rev, | |
1000 | result); |
|
1000 | result); | |
1001 | return -2; |
|
1001 | return -2; | |
1002 | } |
|
1002 | } | |
1003 | return result; |
|
1003 | return result; | |
1004 | } |
|
1004 | } | |
1005 |
|
1005 | |||
1006 | /** |
|
1006 | /** | |
1007 | * Find if a revision is a snapshot or not |
|
1007 | * Find if a revision is a snapshot or not | |
1008 | * |
|
1008 | * | |
1009 | * Only relevant for sparse-revlog case. |
|
1009 | * Only relevant for sparse-revlog case. | |
1010 | * Callers must ensure that rev is in a valid range. |
|
1010 | * Callers must ensure that rev is in a valid range. | |
1011 | */ |
|
1011 | */ | |
1012 | static int index_issnapshotrev(indexObject *self, Py_ssize_t rev) |
|
1012 | static int index_issnapshotrev(indexObject *self, Py_ssize_t rev) | |
1013 | { |
|
1013 | { | |
1014 | int ps[2]; |
|
1014 | int ps[2]; | |
1015 | Py_ssize_t base; |
|
1015 | Py_ssize_t base; | |
1016 | while (rev >= 0) { |
|
1016 | while (rev >= 0) { | |
1017 | base = (Py_ssize_t)index_baserev(self, rev); |
|
1017 | base = (Py_ssize_t)index_baserev(self, rev); | |
1018 | if (base == rev) { |
|
1018 | if (base == rev) { | |
1019 | base = -1; |
|
1019 | base = -1; | |
1020 | } |
|
1020 | } | |
1021 | if (base == -2) { |
|
1021 | if (base == -2) { | |
1022 | assert(PyErr_Occurred()); |
|
1022 | assert(PyErr_Occurred()); | |
1023 | return -1; |
|
1023 | return -1; | |
1024 | } |
|
1024 | } | |
1025 | if (base == -1) { |
|
1025 | if (base == -1) { | |
1026 | return 1; |
|
1026 | return 1; | |
1027 | } |
|
1027 | } | |
1028 | if (index_get_parents(self, rev, ps, (int)rev) < 0) { |
|
1028 | if (index_get_parents(self, rev, ps, (int)rev) < 0) { | |
1029 | assert(PyErr_Occurred()); |
|
1029 | assert(PyErr_Occurred()); | |
1030 | return -1; |
|
1030 | return -1; | |
1031 | }; |
|
1031 | }; | |
1032 | if (base == ps[0] || base == ps[1]) { |
|
1032 | if (base == ps[0] || base == ps[1]) { | |
1033 | return 0; |
|
1033 | return 0; | |
1034 | } |
|
1034 | } | |
1035 | rev = base; |
|
1035 | rev = base; | |
1036 | } |
|
1036 | } | |
1037 | return rev == -1; |
|
1037 | return rev == -1; | |
1038 | } |
|
1038 | } | |
1039 |
|
1039 | |||
1040 | static PyObject *index_issnapshot(indexObject *self, PyObject *value) |
|
1040 | static PyObject *index_issnapshot(indexObject *self, PyObject *value) | |
1041 | { |
|
1041 | { | |
1042 | long rev; |
|
1042 | long rev; | |
1043 | int issnap; |
|
1043 | int issnap; | |
1044 | Py_ssize_t length = index_length(self); |
|
1044 | Py_ssize_t length = index_length(self); | |
1045 |
|
1045 | |||
1046 | if (!pylong_to_long(value, &rev)) { |
|
1046 | if (!pylong_to_long(value, &rev)) { | |
1047 | return NULL; |
|
1047 | return NULL; | |
1048 | } |
|
1048 | } | |
1049 | if (rev < -1 || rev >= length) { |
|
1049 | if (rev < -1 || rev >= length) { | |
1050 | PyErr_Format(PyExc_ValueError, "revlog index out of range: %ld", |
|
1050 | PyErr_Format(PyExc_ValueError, "revlog index out of range: %ld", | |
1051 | rev); |
|
1051 | rev); | |
1052 | return NULL; |
|
1052 | return NULL; | |
1053 | }; |
|
1053 | }; | |
1054 | issnap = index_issnapshotrev(self, (Py_ssize_t)rev); |
|
1054 | issnap = index_issnapshotrev(self, (Py_ssize_t)rev); | |
1055 | if (issnap < 0) { |
|
1055 | if (issnap < 0) { | |
1056 | return NULL; |
|
1056 | return NULL; | |
1057 | }; |
|
1057 | }; | |
1058 | return PyBool_FromLong((long)issnap); |
|
1058 | return PyBool_FromLong((long)issnap); | |
1059 | } |
|
1059 | } | |
1060 |
|
1060 | |||
1061 | static PyObject *index_findsnapshots(indexObject *self, PyObject *args) |
|
1061 | static PyObject *index_findsnapshots(indexObject *self, PyObject *args) | |
1062 | { |
|
1062 | { | |
1063 | Py_ssize_t start_rev; |
|
1063 | Py_ssize_t start_rev; | |
1064 | PyObject *cache; |
|
1064 | PyObject *cache; | |
1065 | Py_ssize_t base; |
|
1065 | Py_ssize_t base; | |
1066 | Py_ssize_t rev; |
|
1066 | Py_ssize_t rev; | |
1067 | PyObject *key = NULL; |
|
1067 | PyObject *key = NULL; | |
1068 | PyObject *value = NULL; |
|
1068 | PyObject *value = NULL; | |
1069 | const Py_ssize_t length = index_length(self); |
|
1069 | const Py_ssize_t length = index_length(self); | |
1070 | if (!PyArg_ParseTuple(args, "O!n", &PyDict_Type, &cache, &start_rev)) { |
|
1070 | if (!PyArg_ParseTuple(args, "O!n", &PyDict_Type, &cache, &start_rev)) { | |
1071 | return NULL; |
|
1071 | return NULL; | |
1072 | } |
|
1072 | } | |
1073 | for (rev = start_rev; rev < length; rev++) { |
|
1073 | for (rev = start_rev; rev < length; rev++) { | |
1074 | int issnap; |
|
1074 | int issnap; | |
1075 | PyObject *allvalues = NULL; |
|
1075 | PyObject *allvalues = NULL; | |
1076 | issnap = index_issnapshotrev(self, rev); |
|
1076 | issnap = index_issnapshotrev(self, rev); | |
1077 | if (issnap < 0) { |
|
1077 | if (issnap < 0) { | |
1078 | goto bail; |
|
1078 | goto bail; | |
1079 | } |
|
1079 | } | |
1080 | if (issnap == 0) { |
|
1080 | if (issnap == 0) { | |
1081 | continue; |
|
1081 | continue; | |
1082 | } |
|
1082 | } | |
1083 | base = (Py_ssize_t)index_baserev(self, rev); |
|
1083 | base = (Py_ssize_t)index_baserev(self, rev); | |
1084 | if (base == rev) { |
|
1084 | if (base == rev) { | |
1085 | base = -1; |
|
1085 | base = -1; | |
1086 | } |
|
1086 | } | |
1087 | if (base == -2) { |
|
1087 | if (base == -2) { | |
1088 | assert(PyErr_Occurred()); |
|
1088 | assert(PyErr_Occurred()); | |
1089 | goto bail; |
|
1089 | goto bail; | |
1090 | } |
|
1090 | } | |
1091 | key = PyInt_FromSsize_t(base); |
|
1091 | key = PyInt_FromSsize_t(base); | |
1092 | allvalues = PyDict_GetItem(cache, key); |
|
1092 | allvalues = PyDict_GetItem(cache, key); | |
1093 | if (allvalues == NULL && PyErr_Occurred()) { |
|
1093 | if (allvalues == NULL && PyErr_Occurred()) { | |
1094 | goto bail; |
|
1094 | goto bail; | |
1095 | } |
|
1095 | } | |
1096 | if (allvalues == NULL) { |
|
1096 | if (allvalues == NULL) { | |
1097 | int r; |
|
1097 | int r; | |
1098 | allvalues = PyList_New(0); |
|
1098 | allvalues = PyList_New(0); | |
1099 | if (!allvalues) { |
|
1099 | if (!allvalues) { | |
1100 | goto bail; |
|
1100 | goto bail; | |
1101 | } |
|
1101 | } | |
1102 | r = PyDict_SetItem(cache, key, allvalues); |
|
1102 | r = PyDict_SetItem(cache, key, allvalues); | |
1103 | Py_DECREF(allvalues); |
|
1103 | Py_DECREF(allvalues); | |
1104 | if (r < 0) { |
|
1104 | if (r < 0) { | |
1105 | goto bail; |
|
1105 | goto bail; | |
1106 | } |
|
1106 | } | |
1107 | } |
|
1107 | } | |
1108 | value = PyInt_FromSsize_t(rev); |
|
1108 | value = PyInt_FromSsize_t(rev); | |
1109 | if (PyList_Append(allvalues, value)) { |
|
1109 | if (PyList_Append(allvalues, value)) { | |
1110 | goto bail; |
|
1110 | goto bail; | |
1111 | } |
|
1111 | } | |
1112 | Py_CLEAR(key); |
|
1112 | Py_CLEAR(key); | |
1113 | Py_CLEAR(value); |
|
1113 | Py_CLEAR(value); | |
1114 | } |
|
1114 | } | |
1115 | Py_RETURN_NONE; |
|
1115 | Py_RETURN_NONE; | |
1116 | bail: |
|
1116 | bail: | |
1117 | Py_XDECREF(key); |
|
1117 | Py_XDECREF(key); | |
1118 | Py_XDECREF(value); |
|
1118 | Py_XDECREF(value); | |
1119 | return NULL; |
|
1119 | return NULL; | |
1120 | } |
|
1120 | } | |
1121 |
|
1121 | |||
1122 | static PyObject *index_deltachain(indexObject *self, PyObject *args) |
|
1122 | static PyObject *index_deltachain(indexObject *self, PyObject *args) | |
1123 | { |
|
1123 | { | |
1124 | int rev, generaldelta; |
|
1124 | int rev, generaldelta; | |
1125 | PyObject *stoparg; |
|
1125 | PyObject *stoparg; | |
1126 | int stoprev, iterrev, baserev = -1; |
|
1126 | int stoprev, iterrev, baserev = -1; | |
1127 | int stopped; |
|
1127 | int stopped; | |
1128 | PyObject *chain = NULL, *result = NULL; |
|
1128 | PyObject *chain = NULL, *result = NULL; | |
1129 | const Py_ssize_t length = index_length(self); |
|
1129 | const Py_ssize_t length = index_length(self); | |
1130 |
|
1130 | |||
1131 | if (!PyArg_ParseTuple(args, "iOi", &rev, &stoparg, &generaldelta)) { |
|
1131 | if (!PyArg_ParseTuple(args, "iOi", &rev, &stoparg, &generaldelta)) { | |
1132 | return NULL; |
|
1132 | return NULL; | |
1133 | } |
|
1133 | } | |
1134 |
|
1134 | |||
1135 | if (PyInt_Check(stoparg)) { |
|
1135 | if (PyInt_Check(stoparg)) { | |
1136 | stoprev = (int)PyInt_AsLong(stoparg); |
|
1136 | stoprev = (int)PyInt_AsLong(stoparg); | |
1137 | if (stoprev == -1 && PyErr_Occurred()) { |
|
1137 | if (stoprev == -1 && PyErr_Occurred()) { | |
1138 | return NULL; |
|
1138 | return NULL; | |
1139 | } |
|
1139 | } | |
1140 | } else if (stoparg == Py_None) { |
|
1140 | } else if (stoparg == Py_None) { | |
1141 | stoprev = -2; |
|
1141 | stoprev = -2; | |
1142 | } else { |
|
1142 | } else { | |
1143 | PyErr_SetString(PyExc_ValueError, |
|
1143 | PyErr_SetString(PyExc_ValueError, | |
1144 | "stoprev must be integer or None"); |
|
1144 | "stoprev must be integer or None"); | |
1145 | return NULL; |
|
1145 | return NULL; | |
1146 | } |
|
1146 | } | |
1147 |
|
1147 | |||
1148 | if (rev < 0 || rev >= length) { |
|
1148 | if (rev < 0 || rev >= length) { | |
1149 | PyErr_SetString(PyExc_ValueError, "revlog index out of range"); |
|
1149 | PyErr_SetString(PyExc_ValueError, "revlog index out of range"); | |
1150 | return NULL; |
|
1150 | return NULL; | |
1151 | } |
|
1151 | } | |
1152 |
|
1152 | |||
1153 | chain = PyList_New(0); |
|
1153 | chain = PyList_New(0); | |
1154 | if (chain == NULL) { |
|
1154 | if (chain == NULL) { | |
1155 | return NULL; |
|
1155 | return NULL; | |
1156 | } |
|
1156 | } | |
1157 |
|
1157 | |||
1158 | baserev = index_baserev(self, rev); |
|
1158 | baserev = index_baserev(self, rev); | |
1159 |
|
1159 | |||
1160 | /* This should never happen. */ |
|
1160 | /* This should never happen. */ | |
1161 | if (baserev <= -2) { |
|
1161 | if (baserev <= -2) { | |
1162 | /* Error should be set by index_deref() */ |
|
1162 | /* Error should be set by index_deref() */ | |
1163 | assert(PyErr_Occurred()); |
|
1163 | assert(PyErr_Occurred()); | |
1164 | goto bail; |
|
1164 | goto bail; | |
1165 | } |
|
1165 | } | |
1166 |
|
1166 | |||
1167 | iterrev = rev; |
|
1167 | iterrev = rev; | |
1168 |
|
1168 | |||
1169 | while (iterrev != baserev && iterrev != stoprev) { |
|
1169 | while (iterrev != baserev && iterrev != stoprev) { | |
1170 | PyObject *value = PyInt_FromLong(iterrev); |
|
1170 | PyObject *value = PyInt_FromLong(iterrev); | |
1171 | if (value == NULL) { |
|
1171 | if (value == NULL) { | |
1172 | goto bail; |
|
1172 | goto bail; | |
1173 | } |
|
1173 | } | |
1174 | if (PyList_Append(chain, value)) { |
|
1174 | if (PyList_Append(chain, value)) { | |
1175 | Py_DECREF(value); |
|
1175 | Py_DECREF(value); | |
1176 | goto bail; |
|
1176 | goto bail; | |
1177 | } |
|
1177 | } | |
1178 | Py_DECREF(value); |
|
1178 | Py_DECREF(value); | |
1179 |
|
1179 | |||
1180 | if (generaldelta) { |
|
1180 | if (generaldelta) { | |
1181 | iterrev = baserev; |
|
1181 | iterrev = baserev; | |
1182 | } else { |
|
1182 | } else { | |
1183 | iterrev--; |
|
1183 | iterrev--; | |
1184 | } |
|
1184 | } | |
1185 |
|
1185 | |||
1186 | if (iterrev < 0) { |
|
1186 | if (iterrev < 0) { | |
1187 | break; |
|
1187 | break; | |
1188 | } |
|
1188 | } | |
1189 |
|
1189 | |||
1190 | if (iterrev >= length) { |
|
1190 | if (iterrev >= length) { | |
1191 | PyErr_SetString(PyExc_IndexError, |
|
1191 | PyErr_SetString(PyExc_IndexError, | |
1192 | "revision outside index"); |
|
1192 | "revision outside index"); | |
1193 | return NULL; |
|
1193 | return NULL; | |
1194 | } |
|
1194 | } | |
1195 |
|
1195 | |||
1196 | baserev = index_baserev(self, iterrev); |
|
1196 | baserev = index_baserev(self, iterrev); | |
1197 |
|
1197 | |||
1198 | /* This should never happen. */ |
|
1198 | /* This should never happen. */ | |
1199 | if (baserev <= -2) { |
|
1199 | if (baserev <= -2) { | |
1200 | /* Error should be set by index_deref() */ |
|
1200 | /* Error should be set by index_deref() */ | |
1201 | assert(PyErr_Occurred()); |
|
1201 | assert(PyErr_Occurred()); | |
1202 | goto bail; |
|
1202 | goto bail; | |
1203 | } |
|
1203 | } | |
1204 | } |
|
1204 | } | |
1205 |
|
1205 | |||
1206 | if (iterrev == stoprev) { |
|
1206 | if (iterrev == stoprev) { | |
1207 | stopped = 1; |
|
1207 | stopped = 1; | |
1208 | } else { |
|
1208 | } else { | |
1209 | PyObject *value = PyInt_FromLong(iterrev); |
|
1209 | PyObject *value = PyInt_FromLong(iterrev); | |
1210 | if (value == NULL) { |
|
1210 | if (value == NULL) { | |
1211 | goto bail; |
|
1211 | goto bail; | |
1212 | } |
|
1212 | } | |
1213 | if (PyList_Append(chain, value)) { |
|
1213 | if (PyList_Append(chain, value)) { | |
1214 | Py_DECREF(value); |
|
1214 | Py_DECREF(value); | |
1215 | goto bail; |
|
1215 | goto bail; | |
1216 | } |
|
1216 | } | |
1217 | Py_DECREF(value); |
|
1217 | Py_DECREF(value); | |
1218 |
|
1218 | |||
1219 | stopped = 0; |
|
1219 | stopped = 0; | |
1220 | } |
|
1220 | } | |
1221 |
|
1221 | |||
1222 | if (PyList_Reverse(chain)) { |
|
1222 | if (PyList_Reverse(chain)) { | |
1223 | goto bail; |
|
1223 | goto bail; | |
1224 | } |
|
1224 | } | |
1225 |
|
1225 | |||
1226 | result = Py_BuildValue("OO", chain, stopped ? Py_True : Py_False); |
|
1226 | result = Py_BuildValue("OO", chain, stopped ? Py_True : Py_False); | |
1227 | Py_DECREF(chain); |
|
1227 | Py_DECREF(chain); | |
1228 | return result; |
|
1228 | return result; | |
1229 |
|
1229 | |||
1230 | bail: |
|
1230 | bail: | |
1231 | Py_DECREF(chain); |
|
1231 | Py_DECREF(chain); | |
1232 | return NULL; |
|
1232 | return NULL; | |
1233 | } |
|
1233 | } | |
1234 |
|
1234 | |||
1235 | static inline int64_t |
|
1235 | static inline int64_t | |
1236 | index_segment_span(indexObject *self, Py_ssize_t start_rev, Py_ssize_t end_rev) |
|
1236 | index_segment_span(indexObject *self, Py_ssize_t start_rev, Py_ssize_t end_rev) | |
1237 | { |
|
1237 | { | |
1238 | int64_t start_offset; |
|
1238 | int64_t start_offset; | |
1239 | int64_t end_offset; |
|
1239 | int64_t end_offset; | |
1240 | int end_size; |
|
1240 | int end_size; | |
1241 | start_offset = index_get_start(self, start_rev); |
|
1241 | start_offset = index_get_start(self, start_rev); | |
1242 | if (start_offset < 0) { |
|
1242 | if (start_offset < 0) { | |
1243 | return -1; |
|
1243 | return -1; | |
1244 | } |
|
1244 | } | |
1245 | end_offset = index_get_start(self, end_rev); |
|
1245 | end_offset = index_get_start(self, end_rev); | |
1246 | if (end_offset < 0) { |
|
1246 | if (end_offset < 0) { | |
1247 | return -1; |
|
1247 | return -1; | |
1248 | } |
|
1248 | } | |
1249 | end_size = index_get_length(self, end_rev); |
|
1249 | end_size = index_get_length(self, end_rev); | |
1250 | if (end_size < 0) { |
|
1250 | if (end_size < 0) { | |
1251 | return -1; |
|
1251 | return -1; | |
1252 | } |
|
1252 | } | |
1253 | if (end_offset < start_offset) { |
|
1253 | if (end_offset < start_offset) { | |
1254 | PyErr_Format(PyExc_ValueError, |
|
1254 | PyErr_Format(PyExc_ValueError, | |
1255 | "corrupted revlog index: inconsistent offset " |
|
1255 | "corrupted revlog index: inconsistent offset " | |
1256 | "between revisions (%zd) and (%zd)", |
|
1256 | "between revisions (%zd) and (%zd)", | |
1257 | start_rev, end_rev); |
|
1257 | start_rev, end_rev); | |
1258 | return -1; |
|
1258 | return -1; | |
1259 | } |
|
1259 | } | |
1260 | return (end_offset - start_offset) + (int64_t)end_size; |
|
1260 | return (end_offset - start_offset) + (int64_t)end_size; | |
1261 | } |
|
1261 | } | |
1262 |
|
1262 | |||
1263 | /* returns endidx so that revs[startidx:endidx] has no empty trailing revs */ |
|
1263 | /* returns endidx so that revs[startidx:endidx] has no empty trailing revs */ | |
1264 | static Py_ssize_t trim_endidx(indexObject *self, const Py_ssize_t *revs, |
|
1264 | static Py_ssize_t trim_endidx(indexObject *self, const Py_ssize_t *revs, | |
1265 | Py_ssize_t startidx, Py_ssize_t endidx) |
|
1265 | Py_ssize_t startidx, Py_ssize_t endidx) | |
1266 | { |
|
1266 | { | |
1267 | int length; |
|
1267 | int length; | |
1268 | while (endidx > 1 && endidx > startidx) { |
|
1268 | while (endidx > 1 && endidx > startidx) { | |
1269 | length = index_get_length(self, revs[endidx - 1]); |
|
1269 | length = index_get_length(self, revs[endidx - 1]); | |
1270 | if (length < 0) { |
|
1270 | if (length < 0) { | |
1271 | return -1; |
|
1271 | return -1; | |
1272 | } |
|
1272 | } | |
1273 | if (length != 0) { |
|
1273 | if (length != 0) { | |
1274 | break; |
|
1274 | break; | |
1275 | } |
|
1275 | } | |
1276 | endidx -= 1; |
|
1276 | endidx -= 1; | |
1277 | } |
|
1277 | } | |
1278 | return endidx; |
|
1278 | return endidx; | |
1279 | } |
|
1279 | } | |
1280 |
|
1280 | |||
1281 | struct Gap { |
|
1281 | struct Gap { | |
1282 | int64_t size; |
|
1282 | int64_t size; | |
1283 | Py_ssize_t idx; |
|
1283 | Py_ssize_t idx; | |
1284 | }; |
|
1284 | }; | |
1285 |
|
1285 | |||
1286 | static int gap_compare(const void *left, const void *right) |
|
1286 | static int gap_compare(const void *left, const void *right) | |
1287 | { |
|
1287 | { | |
1288 | const struct Gap *l_left = ((const struct Gap *)left); |
|
1288 | const struct Gap *l_left = ((const struct Gap *)left); | |
1289 | const struct Gap *l_right = ((const struct Gap *)right); |
|
1289 | const struct Gap *l_right = ((const struct Gap *)right); | |
1290 | if (l_left->size < l_right->size) { |
|
1290 | if (l_left->size < l_right->size) { | |
1291 | return -1; |
|
1291 | return -1; | |
1292 | } else if (l_left->size > l_right->size) { |
|
1292 | } else if (l_left->size > l_right->size) { | |
1293 | return 1; |
|
1293 | return 1; | |
1294 | } |
|
1294 | } | |
1295 | return 0; |
|
1295 | return 0; | |
1296 | } |
|
1296 | } | |
1297 | static int Py_ssize_t_compare(const void *left, const void *right) |
|
1297 | static int Py_ssize_t_compare(const void *left, const void *right) | |
1298 | { |
|
1298 | { | |
1299 | const Py_ssize_t l_left = *(const Py_ssize_t *)left; |
|
1299 | const Py_ssize_t l_left = *(const Py_ssize_t *)left; | |
1300 | const Py_ssize_t l_right = *(const Py_ssize_t *)right; |
|
1300 | const Py_ssize_t l_right = *(const Py_ssize_t *)right; | |
1301 | if (l_left < l_right) { |
|
1301 | if (l_left < l_right) { | |
1302 | return -1; |
|
1302 | return -1; | |
1303 | } else if (l_left > l_right) { |
|
1303 | } else if (l_left > l_right) { | |
1304 | return 1; |
|
1304 | return 1; | |
1305 | } |
|
1305 | } | |
1306 | return 0; |
|
1306 | return 0; | |
1307 | } |
|
1307 | } | |
1308 |
|
1308 | |||
1309 | static PyObject *index_slicechunktodensity(indexObject *self, PyObject *args) |
|
1309 | static PyObject *index_slicechunktodensity(indexObject *self, PyObject *args) | |
1310 | { |
|
1310 | { | |
1311 | /* method arguments */ |
|
1311 | /* method arguments */ | |
1312 | PyObject *list_revs = NULL; /* revisions in the chain */ |
|
1312 | PyObject *list_revs = NULL; /* revisions in the chain */ | |
1313 | double targetdensity = 0; /* min density to achieve */ |
|
1313 | double targetdensity = 0; /* min density to achieve */ | |
1314 | Py_ssize_t mingapsize = 0; /* threshold to ignore gaps */ |
|
1314 | Py_ssize_t mingapsize = 0; /* threshold to ignore gaps */ | |
1315 |
|
1315 | |||
1316 | /* other core variables */ |
|
1316 | /* other core variables */ | |
1317 | Py_ssize_t idxlen = index_length(self); |
|
1317 | Py_ssize_t idxlen = index_length(self); | |
1318 | Py_ssize_t i; /* used for various iteration */ |
|
1318 | Py_ssize_t i; /* used for various iteration */ | |
1319 | PyObject *result = NULL; /* the final return of the function */ |
|
1319 | PyObject *result = NULL; /* the final return of the function */ | |
1320 |
|
1320 | |||
1321 | /* generic information about the delta chain being slice */ |
|
1321 | /* generic information about the delta chain being slice */ | |
1322 | Py_ssize_t num_revs = 0; /* size of the full delta chain */ |
|
1322 | Py_ssize_t num_revs = 0; /* size of the full delta chain */ | |
1323 | Py_ssize_t *revs = NULL; /* native array of revision in the chain */ |
|
1323 | Py_ssize_t *revs = NULL; /* native array of revision in the chain */ | |
1324 | int64_t chainpayload = 0; /* sum of all delta in the chain */ |
|
1324 | int64_t chainpayload = 0; /* sum of all delta in the chain */ | |
1325 | int64_t deltachainspan = 0; /* distance from first byte to last byte */ |
|
1325 | int64_t deltachainspan = 0; /* distance from first byte to last byte */ | |
1326 |
|
1326 | |||
1327 | /* variable used for slicing the delta chain */ |
|
1327 | /* variable used for slicing the delta chain */ | |
1328 | int64_t readdata = 0; /* amount of data currently planned to be read */ |
|
1328 | int64_t readdata = 0; /* amount of data currently planned to be read */ | |
1329 | double density = 0; /* ration of payload data compared to read ones */ |
|
1329 | double density = 0; /* ration of payload data compared to read ones */ | |
1330 | int64_t previous_end; |
|
1330 | int64_t previous_end; | |
1331 | struct Gap *gaps = NULL; /* array of notable gap in the chain */ |
|
1331 | struct Gap *gaps = NULL; /* array of notable gap in the chain */ | |
1332 | Py_ssize_t num_gaps = |
|
1332 | Py_ssize_t num_gaps = | |
1333 | 0; /* total number of notable gap recorded so far */ |
|
1333 | 0; /* total number of notable gap recorded so far */ | |
1334 | Py_ssize_t *selected_indices = NULL; /* indices of gap skipped over */ |
|
1334 | Py_ssize_t *selected_indices = NULL; /* indices of gap skipped over */ | |
1335 | Py_ssize_t num_selected = 0; /* number of gaps skipped */ |
|
1335 | Py_ssize_t num_selected = 0; /* number of gaps skipped */ | |
1336 | PyObject *chunk = NULL; /* individual slice */ |
|
1336 | PyObject *chunk = NULL; /* individual slice */ | |
1337 | PyObject *allchunks = NULL; /* all slices */ |
|
1337 | PyObject *allchunks = NULL; /* all slices */ | |
1338 | Py_ssize_t previdx; |
|
1338 | Py_ssize_t previdx; | |
1339 |
|
1339 | |||
1340 | /* parsing argument */ |
|
1340 | /* parsing argument */ | |
1341 | if (!PyArg_ParseTuple(args, "O!dn", &PyList_Type, &list_revs, |
|
1341 | if (!PyArg_ParseTuple(args, "O!dn", &PyList_Type, &list_revs, | |
1342 | &targetdensity, &mingapsize)) { |
|
1342 | &targetdensity, &mingapsize)) { | |
1343 | goto bail; |
|
1343 | goto bail; | |
1344 | } |
|
1344 | } | |
1345 |
|
1345 | |||
1346 | /* If the delta chain contains a single element, we do not need slicing |
|
1346 | /* If the delta chain contains a single element, we do not need slicing | |
1347 | */ |
|
1347 | */ | |
1348 | num_revs = PyList_GET_SIZE(list_revs); |
|
1348 | num_revs = PyList_GET_SIZE(list_revs); | |
1349 | if (num_revs <= 1) { |
|
1349 | if (num_revs <= 1) { | |
1350 | result = PyTuple_Pack(1, list_revs); |
|
1350 | result = PyTuple_Pack(1, list_revs); | |
1351 | goto done; |
|
1351 | goto done; | |
1352 | } |
|
1352 | } | |
1353 |
|
1353 | |||
1354 | /* Turn the python list into a native integer array (for efficiency) */ |
|
1354 | /* Turn the python list into a native integer array (for efficiency) */ | |
1355 | revs = (Py_ssize_t *)calloc(num_revs, sizeof(Py_ssize_t)); |
|
1355 | revs = (Py_ssize_t *)calloc(num_revs, sizeof(Py_ssize_t)); | |
1356 | if (revs == NULL) { |
|
1356 | if (revs == NULL) { | |
1357 | PyErr_NoMemory(); |
|
1357 | PyErr_NoMemory(); | |
1358 | goto bail; |
|
1358 | goto bail; | |
1359 | } |
|
1359 | } | |
1360 | for (i = 0; i < num_revs; i++) { |
|
1360 | for (i = 0; i < num_revs; i++) { | |
1361 | Py_ssize_t revnum = PyInt_AsLong(PyList_GET_ITEM(list_revs, i)); |
|
1361 | Py_ssize_t revnum = PyInt_AsLong(PyList_GET_ITEM(list_revs, i)); | |
1362 | if (revnum == -1 && PyErr_Occurred()) { |
|
1362 | if (revnum == -1 && PyErr_Occurred()) { | |
1363 | goto bail; |
|
1363 | goto bail; | |
1364 | } |
|
1364 | } | |
1365 | if (revnum < nullrev || revnum >= idxlen) { |
|
1365 | if (revnum < nullrev || revnum >= idxlen) { | |
1366 | PyErr_Format(PyExc_IndexError, |
|
1366 | PyErr_Format(PyExc_IndexError, | |
1367 | "index out of range: %zd", revnum); |
|
1367 | "index out of range: %zd", revnum); | |
1368 | goto bail; |
|
1368 | goto bail; | |
1369 | } |
|
1369 | } | |
1370 | revs[i] = revnum; |
|
1370 | revs[i] = revnum; | |
1371 | } |
|
1371 | } | |
1372 |
|
1372 | |||
1373 | /* Compute and check various property of the unsliced delta chain */ |
|
1373 | /* Compute and check various property of the unsliced delta chain */ | |
1374 | deltachainspan = index_segment_span(self, revs[0], revs[num_revs - 1]); |
|
1374 | deltachainspan = index_segment_span(self, revs[0], revs[num_revs - 1]); | |
1375 | if (deltachainspan < 0) { |
|
1375 | if (deltachainspan < 0) { | |
1376 | goto bail; |
|
1376 | goto bail; | |
1377 | } |
|
1377 | } | |
1378 |
|
1378 | |||
1379 | if (deltachainspan <= mingapsize) { |
|
1379 | if (deltachainspan <= mingapsize) { | |
1380 | result = PyTuple_Pack(1, list_revs); |
|
1380 | result = PyTuple_Pack(1, list_revs); | |
1381 | goto done; |
|
1381 | goto done; | |
1382 | } |
|
1382 | } | |
1383 | chainpayload = 0; |
|
1383 | chainpayload = 0; | |
1384 | for (i = 0; i < num_revs; i++) { |
|
1384 | for (i = 0; i < num_revs; i++) { | |
1385 | int tmp = index_get_length(self, revs[i]); |
|
1385 | int tmp = index_get_length(self, revs[i]); | |
1386 | if (tmp < 0) { |
|
1386 | if (tmp < 0) { | |
1387 | goto bail; |
|
1387 | goto bail; | |
1388 | } |
|
1388 | } | |
1389 | chainpayload += tmp; |
|
1389 | chainpayload += tmp; | |
1390 | } |
|
1390 | } | |
1391 |
|
1391 | |||
1392 | readdata = deltachainspan; |
|
1392 | readdata = deltachainspan; | |
1393 | density = 1.0; |
|
1393 | density = 1.0; | |
1394 |
|
1394 | |||
1395 | if (0 < deltachainspan) { |
|
1395 | if (0 < deltachainspan) { | |
1396 | density = (double)chainpayload / (double)deltachainspan; |
|
1396 | density = (double)chainpayload / (double)deltachainspan; | |
1397 | } |
|
1397 | } | |
1398 |
|
1398 | |||
1399 | if (density >= targetdensity) { |
|
1399 | if (density >= targetdensity) { | |
1400 | result = PyTuple_Pack(1, list_revs); |
|
1400 | result = PyTuple_Pack(1, list_revs); | |
1401 | goto done; |
|
1401 | goto done; | |
1402 | } |
|
1402 | } | |
1403 |
|
1403 | |||
1404 | /* if chain is too sparse, look for relevant gaps */ |
|
1404 | /* if chain is too sparse, look for relevant gaps */ | |
1405 | gaps = (struct Gap *)calloc(num_revs, sizeof(struct Gap)); |
|
1405 | gaps = (struct Gap *)calloc(num_revs, sizeof(struct Gap)); | |
1406 | if (gaps == NULL) { |
|
1406 | if (gaps == NULL) { | |
1407 | PyErr_NoMemory(); |
|
1407 | PyErr_NoMemory(); | |
1408 | goto bail; |
|
1408 | goto bail; | |
1409 | } |
|
1409 | } | |
1410 |
|
1410 | |||
1411 | previous_end = -1; |
|
1411 | previous_end = -1; | |
1412 | for (i = 0; i < num_revs; i++) { |
|
1412 | for (i = 0; i < num_revs; i++) { | |
1413 | int64_t revstart; |
|
1413 | int64_t revstart; | |
1414 | int revsize; |
|
1414 | int revsize; | |
1415 | revstart = index_get_start(self, revs[i]); |
|
1415 | revstart = index_get_start(self, revs[i]); | |
1416 | if (revstart < 0) { |
|
1416 | if (revstart < 0) { | |
1417 | goto bail; |
|
1417 | goto bail; | |
1418 | }; |
|
1418 | }; | |
1419 | revsize = index_get_length(self, revs[i]); |
|
1419 | revsize = index_get_length(self, revs[i]); | |
1420 | if (revsize < 0) { |
|
1420 | if (revsize < 0) { | |
1421 | goto bail; |
|
1421 | goto bail; | |
1422 | }; |
|
1422 | }; | |
1423 | if (revsize == 0) { |
|
1423 | if (revsize == 0) { | |
1424 | continue; |
|
1424 | continue; | |
1425 | } |
|
1425 | } | |
1426 | if (previous_end >= 0) { |
|
1426 | if (previous_end >= 0) { | |
1427 | int64_t gapsize = revstart - previous_end; |
|
1427 | int64_t gapsize = revstart - previous_end; | |
1428 | if (gapsize > mingapsize) { |
|
1428 | if (gapsize > mingapsize) { | |
1429 | gaps[num_gaps].size = gapsize; |
|
1429 | gaps[num_gaps].size = gapsize; | |
1430 | gaps[num_gaps].idx = i; |
|
1430 | gaps[num_gaps].idx = i; | |
1431 | num_gaps += 1; |
|
1431 | num_gaps += 1; | |
1432 | } |
|
1432 | } | |
1433 | } |
|
1433 | } | |
1434 | previous_end = revstart + revsize; |
|
1434 | previous_end = revstart + revsize; | |
1435 | } |
|
1435 | } | |
1436 | if (num_gaps == 0) { |
|
1436 | if (num_gaps == 0) { | |
1437 | result = PyTuple_Pack(1, list_revs); |
|
1437 | result = PyTuple_Pack(1, list_revs); | |
1438 | goto done; |
|
1438 | goto done; | |
1439 | } |
|
1439 | } | |
1440 | qsort(gaps, num_gaps, sizeof(struct Gap), &gap_compare); |
|
1440 | qsort(gaps, num_gaps, sizeof(struct Gap), &gap_compare); | |
1441 |
|
1441 | |||
1442 | /* Slice the largest gap first, they improve the density the most */ |
|
1442 | /* Slice the largest gap first, they improve the density the most */ | |
1443 | selected_indices = |
|
1443 | selected_indices = | |
1444 | (Py_ssize_t *)malloc((num_gaps + 1) * sizeof(Py_ssize_t)); |
|
1444 | (Py_ssize_t *)malloc((num_gaps + 1) * sizeof(Py_ssize_t)); | |
1445 | if (selected_indices == NULL) { |
|
1445 | if (selected_indices == NULL) { | |
1446 | PyErr_NoMemory(); |
|
1446 | PyErr_NoMemory(); | |
1447 | goto bail; |
|
1447 | goto bail; | |
1448 | } |
|
1448 | } | |
1449 |
|
1449 | |||
1450 | for (i = num_gaps - 1; i >= 0; i--) { |
|
1450 | for (i = num_gaps - 1; i >= 0; i--) { | |
1451 | selected_indices[num_selected] = gaps[i].idx; |
|
1451 | selected_indices[num_selected] = gaps[i].idx; | |
1452 | readdata -= gaps[i].size; |
|
1452 | readdata -= gaps[i].size; | |
1453 | num_selected += 1; |
|
1453 | num_selected += 1; | |
1454 | if (readdata <= 0) { |
|
1454 | if (readdata <= 0) { | |
1455 | density = 1.0; |
|
1455 | density = 1.0; | |
1456 | } else { |
|
1456 | } else { | |
1457 | density = (double)chainpayload / (double)readdata; |
|
1457 | density = (double)chainpayload / (double)readdata; | |
1458 | } |
|
1458 | } | |
1459 | if (density >= targetdensity) { |
|
1459 | if (density >= targetdensity) { | |
1460 | break; |
|
1460 | break; | |
1461 | } |
|
1461 | } | |
1462 | } |
|
1462 | } | |
1463 | qsort(selected_indices, num_selected, sizeof(Py_ssize_t), |
|
1463 | qsort(selected_indices, num_selected, sizeof(Py_ssize_t), | |
1464 | &Py_ssize_t_compare); |
|
1464 | &Py_ssize_t_compare); | |
1465 |
|
1465 | |||
1466 | /* create the resulting slice */ |
|
1466 | /* create the resulting slice */ | |
1467 | allchunks = PyList_New(0); |
|
1467 | allchunks = PyList_New(0); | |
1468 | if (allchunks == NULL) { |
|
1468 | if (allchunks == NULL) { | |
1469 | goto bail; |
|
1469 | goto bail; | |
1470 | } |
|
1470 | } | |
1471 | previdx = 0; |
|
1471 | previdx = 0; | |
1472 | selected_indices[num_selected] = num_revs; |
|
1472 | selected_indices[num_selected] = num_revs; | |
1473 | for (i = 0; i <= num_selected; i++) { |
|
1473 | for (i = 0; i <= num_selected; i++) { | |
1474 | Py_ssize_t idx = selected_indices[i]; |
|
1474 | Py_ssize_t idx = selected_indices[i]; | |
1475 | Py_ssize_t endidx = trim_endidx(self, revs, previdx, idx); |
|
1475 | Py_ssize_t endidx = trim_endidx(self, revs, previdx, idx); | |
1476 | if (endidx < 0) { |
|
1476 | if (endidx < 0) { | |
1477 | goto bail; |
|
1477 | goto bail; | |
1478 | } |
|
1478 | } | |
1479 | if (previdx < endidx) { |
|
1479 | if (previdx < endidx) { | |
1480 | chunk = PyList_GetSlice(list_revs, previdx, endidx); |
|
1480 | chunk = PyList_GetSlice(list_revs, previdx, endidx); | |
1481 | if (chunk == NULL) { |
|
1481 | if (chunk == NULL) { | |
1482 | goto bail; |
|
1482 | goto bail; | |
1483 | } |
|
1483 | } | |
1484 | if (PyList_Append(allchunks, chunk) == -1) { |
|
1484 | if (PyList_Append(allchunks, chunk) == -1) { | |
1485 | goto bail; |
|
1485 | goto bail; | |
1486 | } |
|
1486 | } | |
1487 | Py_DECREF(chunk); |
|
1487 | Py_DECREF(chunk); | |
1488 | chunk = NULL; |
|
1488 | chunk = NULL; | |
1489 | } |
|
1489 | } | |
1490 | previdx = idx; |
|
1490 | previdx = idx; | |
1491 | } |
|
1491 | } | |
1492 | result = allchunks; |
|
1492 | result = allchunks; | |
1493 | goto done; |
|
1493 | goto done; | |
1494 |
|
1494 | |||
1495 | bail: |
|
1495 | bail: | |
1496 | Py_XDECREF(allchunks); |
|
1496 | Py_XDECREF(allchunks); | |
1497 | Py_XDECREF(chunk); |
|
1497 | Py_XDECREF(chunk); | |
1498 | done: |
|
1498 | done: | |
1499 | free(revs); |
|
1499 | free(revs); | |
1500 | free(gaps); |
|
1500 | free(gaps); | |
1501 | free(selected_indices); |
|
1501 | free(selected_indices); | |
1502 | return result; |
|
1502 | return result; | |
1503 | } |
|
1503 | } | |
1504 |
|
1504 | |||
1505 | static inline int nt_level(const char *node, Py_ssize_t level) |
|
1505 | static inline int nt_level(const char *node, Py_ssize_t level) | |
1506 | { |
|
1506 | { | |
1507 | int v = node[level >> 1]; |
|
1507 | int v = node[level >> 1]; | |
1508 | if (!(level & 1)) |
|
1508 | if (!(level & 1)) | |
1509 | v >>= 4; |
|
1509 | v >>= 4; | |
1510 | return v & 0xf; |
|
1510 | return v & 0xf; | |
1511 | } |
|
1511 | } | |
1512 |
|
1512 | |||
1513 | /* |
|
1513 | /* | |
1514 | * Return values: |
|
1514 | * Return values: | |
1515 | * |
|
1515 | * | |
1516 | * -4: match is ambiguous (multiple candidates) |
|
1516 | * -4: match is ambiguous (multiple candidates) | |
1517 | * -2: not found |
|
1517 | * -2: not found | |
1518 | * rest: valid rev |
|
1518 | * rest: valid rev | |
1519 | */ |
|
1519 | */ | |
1520 | static int nt_find(nodetree *self, const char *node, Py_ssize_t nodelen, |
|
1520 | static int nt_find(nodetree *self, const char *node, Py_ssize_t nodelen, | |
1521 | int hex) |
|
1521 | int hex) | |
1522 | { |
|
1522 | { | |
1523 | int (*getnybble)(const char *, Py_ssize_t) = hex ? hexdigit : nt_level; |
|
1523 | int (*getnybble)(const char *, Py_ssize_t) = hex ? hexdigit : nt_level; | |
1524 | int level, maxlevel, off; |
|
1524 | int level, maxlevel, off; | |
1525 |
|
1525 | |||
1526 | if (nodelen == 20 && node[0] == '\0' && memcmp(node, nullid, 20) == 0) |
|
1526 | if (nodelen == 20 && node[0] == '\0' && memcmp(node, nullid, 20) == 0) | |
1527 | return -1; |
|
1527 | return -1; | |
1528 |
|
1528 | |||
1529 | if (hex) |
|
1529 | if (hex) | |
1530 | maxlevel = nodelen > 40 ? 40 : (int)nodelen; |
|
1530 | maxlevel = nodelen > 40 ? 40 : (int)nodelen; | |
1531 | else |
|
1531 | else | |
1532 | maxlevel = nodelen > 20 ? 40 : ((int)nodelen * 2); |
|
1532 | maxlevel = nodelen > 20 ? 40 : ((int)nodelen * 2); | |
1533 |
|
1533 | |||
1534 | for (level = off = 0; level < maxlevel; level++) { |
|
1534 | for (level = off = 0; level < maxlevel; level++) { | |
1535 | int k = getnybble(node, level); |
|
1535 | int k = getnybble(node, level); | |
1536 | nodetreenode *n = &self->nodes[off]; |
|
1536 | nodetreenode *n = &self->nodes[off]; | |
1537 | int v = n->children[k]; |
|
1537 | int v = n->children[k]; | |
1538 |
|
1538 | |||
1539 | if (v < 0) { |
|
1539 | if (v < 0) { | |
1540 | const char *n; |
|
1540 | const char *n; | |
1541 | Py_ssize_t i; |
|
1541 | Py_ssize_t i; | |
1542 |
|
1542 | |||
1543 | v = -(v + 2); |
|
1543 | v = -(v + 2); | |
1544 | n = index_node(self->index, v); |
|
1544 | n = index_node(self->index, v); | |
1545 | if (n == NULL) |
|
1545 | if (n == NULL) | |
1546 | return -2; |
|
1546 | return -2; | |
1547 | for (i = level; i < maxlevel; i++) |
|
1547 | for (i = level; i < maxlevel; i++) | |
1548 | if (getnybble(node, i) != nt_level(n, i)) |
|
1548 | if (getnybble(node, i) != nt_level(n, i)) | |
1549 | return -2; |
|
1549 | return -2; | |
1550 | return v; |
|
1550 | return v; | |
1551 | } |
|
1551 | } | |
1552 | if (v == 0) |
|
1552 | if (v == 0) | |
1553 | return -2; |
|
1553 | return -2; | |
1554 | off = v; |
|
1554 | off = v; | |
1555 | } |
|
1555 | } | |
1556 | /* multiple matches against an ambiguous prefix */ |
|
1556 | /* multiple matches against an ambiguous prefix */ | |
1557 | return -4; |
|
1557 | return -4; | |
1558 | } |
|
1558 | } | |
1559 |
|
1559 | |||
1560 | static int nt_new(nodetree *self) |
|
1560 | static int nt_new(nodetree *self) | |
1561 | { |
|
1561 | { | |
1562 | if (self->length == self->capacity) { |
|
1562 | if (self->length == self->capacity) { | |
1563 | unsigned newcapacity; |
|
1563 | unsigned newcapacity; | |
1564 | nodetreenode *newnodes; |
|
1564 | nodetreenode *newnodes; | |
1565 | newcapacity = self->capacity * 2; |
|
1565 | newcapacity = self->capacity * 2; | |
1566 | if (newcapacity >= INT_MAX / sizeof(nodetreenode)) { |
|
1566 | if (newcapacity >= INT_MAX / sizeof(nodetreenode)) { | |
1567 | PyErr_SetString(PyExc_MemoryError, |
|
1567 | PyErr_SetString(PyExc_MemoryError, | |
1568 | "overflow in nt_new"); |
|
1568 | "overflow in nt_new"); | |
1569 | return -1; |
|
1569 | return -1; | |
1570 | } |
|
1570 | } | |
1571 | newnodes = |
|
1571 | newnodes = | |
1572 | realloc(self->nodes, newcapacity * sizeof(nodetreenode)); |
|
1572 | realloc(self->nodes, newcapacity * sizeof(nodetreenode)); | |
1573 | if (newnodes == NULL) { |
|
1573 | if (newnodes == NULL) { | |
1574 | PyErr_SetString(PyExc_MemoryError, "out of memory"); |
|
1574 | PyErr_SetString(PyExc_MemoryError, "out of memory"); | |
1575 | return -1; |
|
1575 | return -1; | |
1576 | } |
|
1576 | } | |
1577 | self->capacity = newcapacity; |
|
1577 | self->capacity = newcapacity; | |
1578 | self->nodes = newnodes; |
|
1578 | self->nodes = newnodes; | |
1579 | memset(&self->nodes[self->length], 0, |
|
1579 | memset(&self->nodes[self->length], 0, | |
1580 | sizeof(nodetreenode) * (self->capacity - self->length)); |
|
1580 | sizeof(nodetreenode) * (self->capacity - self->length)); | |
1581 | } |
|
1581 | } | |
1582 | return self->length++; |
|
1582 | return self->length++; | |
1583 | } |
|
1583 | } | |
1584 |
|
1584 | |||
1585 | static int nt_insert(nodetree *self, const char *node, int rev) |
|
1585 | static int nt_insert(nodetree *self, const char *node, int rev) | |
1586 | { |
|
1586 | { | |
1587 | int level = 0; |
|
1587 | int level = 0; | |
1588 | int off = 0; |
|
1588 | int off = 0; | |
1589 |
|
1589 | |||
1590 | while (level < 40) { |
|
1590 | while (level < 40) { | |
1591 | int k = nt_level(node, level); |
|
1591 | int k = nt_level(node, level); | |
1592 | nodetreenode *n; |
|
1592 | nodetreenode *n; | |
1593 | int v; |
|
1593 | int v; | |
1594 |
|
1594 | |||
1595 | n = &self->nodes[off]; |
|
1595 | n = &self->nodes[off]; | |
1596 | v = n->children[k]; |
|
1596 | v = n->children[k]; | |
1597 |
|
1597 | |||
1598 | if (v == 0) { |
|
1598 | if (v == 0) { | |
1599 | n->children[k] = -rev - 2; |
|
1599 | n->children[k] = -rev - 2; | |
1600 | return 0; |
|
1600 | return 0; | |
1601 | } |
|
1601 | } | |
1602 | if (v < 0) { |
|
1602 | if (v < 0) { | |
1603 | const char *oldnode = |
|
1603 | const char *oldnode = | |
1604 | index_node_existing(self->index, -(v + 2)); |
|
1604 | index_node_existing(self->index, -(v + 2)); | |
1605 | int noff; |
|
1605 | int noff; | |
1606 |
|
1606 | |||
1607 | if (oldnode == NULL) |
|
1607 | if (oldnode == NULL) | |
1608 | return -1; |
|
1608 | return -1; | |
1609 | if (!memcmp(oldnode, node, 20)) { |
|
1609 | if (!memcmp(oldnode, node, 20)) { | |
1610 | n->children[k] = -rev - 2; |
|
1610 | n->children[k] = -rev - 2; | |
1611 | return 0; |
|
1611 | return 0; | |
1612 | } |
|
1612 | } | |
1613 | noff = nt_new(self); |
|
1613 | noff = nt_new(self); | |
1614 | if (noff == -1) |
|
1614 | if (noff == -1) | |
1615 | return -1; |
|
1615 | return -1; | |
1616 | /* self->nodes may have been changed by realloc */ |
|
1616 | /* self->nodes may have been changed by realloc */ | |
1617 | self->nodes[off].children[k] = noff; |
|
1617 | self->nodes[off].children[k] = noff; | |
1618 | off = noff; |
|
1618 | off = noff; | |
1619 | n = &self->nodes[off]; |
|
1619 | n = &self->nodes[off]; | |
1620 | n->children[nt_level(oldnode, ++level)] = v; |
|
1620 | n->children[nt_level(oldnode, ++level)] = v; | |
1621 | if (level > self->depth) |
|
1621 | if (level > self->depth) | |
1622 | self->depth = level; |
|
1622 | self->depth = level; | |
1623 | self->splits += 1; |
|
1623 | self->splits += 1; | |
1624 | } else { |
|
1624 | } else { | |
1625 | level += 1; |
|
1625 | level += 1; | |
1626 | off = v; |
|
1626 | off = v; | |
1627 | } |
|
1627 | } | |
1628 | } |
|
1628 | } | |
1629 |
|
1629 | |||
1630 | return -1; |
|
1630 | return -1; | |
1631 | } |
|
1631 | } | |
1632 |
|
1632 | |||
1633 | static PyObject *ntobj_insert(nodetreeObject *self, PyObject *args) |
|
1633 | static PyObject *ntobj_insert(nodetreeObject *self, PyObject *args) | |
1634 | { |
|
1634 | { | |
1635 | Py_ssize_t rev; |
|
1635 | Py_ssize_t rev; | |
1636 | const char *node; |
|
1636 | const char *node; | |
1637 | Py_ssize_t length; |
|
1637 | Py_ssize_t length; | |
1638 | if (!PyArg_ParseTuple(args, "n", &rev)) |
|
1638 | if (!PyArg_ParseTuple(args, "n", &rev)) | |
1639 | return NULL; |
|
1639 | return NULL; | |
1640 | length = index_length(self->nt.index); |
|
1640 | length = index_length(self->nt.index); | |
1641 | if (rev < 0 || rev >= length) { |
|
1641 | if (rev < 0 || rev >= length) { | |
1642 | PyErr_SetString(PyExc_ValueError, "revlog index out of range"); |
|
1642 | PyErr_SetString(PyExc_ValueError, "revlog index out of range"); | |
1643 | return NULL; |
|
1643 | return NULL; | |
1644 | } |
|
1644 | } | |
1645 | node = index_node_existing(self->nt.index, rev); |
|
1645 | node = index_node_existing(self->nt.index, rev); | |
1646 | if (nt_insert(&self->nt, node, (int)rev) == -1) |
|
1646 | if (nt_insert(&self->nt, node, (int)rev) == -1) | |
1647 | return NULL; |
|
1647 | return NULL; | |
1648 | Py_RETURN_NONE; |
|
1648 | Py_RETURN_NONE; | |
1649 | } |
|
1649 | } | |
1650 |
|
1650 | |||
1651 | static int nt_delete_node(nodetree *self, const char *node) |
|
1651 | static int nt_delete_node(nodetree *self, const char *node) | |
1652 | { |
|
1652 | { | |
1653 | /* rev==-2 happens to get encoded as 0, which is interpreted as not set |
|
1653 | /* rev==-2 happens to get encoded as 0, which is interpreted as not set | |
1654 | */ |
|
1654 | */ | |
1655 | return nt_insert(self, node, -2); |
|
1655 | return nt_insert(self, node, -2); | |
1656 | } |
|
1656 | } | |
1657 |
|
1657 | |||
1658 | static int nt_init(nodetree *self, indexObject *index, unsigned capacity) |
|
1658 | static int nt_init(nodetree *self, indexObject *index, unsigned capacity) | |
1659 | { |
|
1659 | { | |
1660 | /* Initialize before overflow-checking to avoid nt_dealloc() crash. */ |
|
1660 | /* Initialize before overflow-checking to avoid nt_dealloc() crash. */ | |
1661 | self->nodes = NULL; |
|
1661 | self->nodes = NULL; | |
1662 |
|
1662 | |||
1663 | self->index = index; |
|
1663 | self->index = index; | |
1664 | /* The input capacity is in terms of revisions, while the field is in |
|
1664 | /* The input capacity is in terms of revisions, while the field is in | |
1665 | * terms of nodetree nodes. */ |
|
1665 | * terms of nodetree nodes. */ | |
1666 | self->capacity = (capacity < 4 ? 4 : capacity / 2); |
|
1666 | self->capacity = (capacity < 4 ? 4 : capacity / 2); | |
1667 | self->depth = 0; |
|
1667 | self->depth = 0; | |
1668 | self->splits = 0; |
|
1668 | self->splits = 0; | |
1669 | if ((size_t)self->capacity > INT_MAX / sizeof(nodetreenode)) { |
|
1669 | if ((size_t)self->capacity > INT_MAX / sizeof(nodetreenode)) { | |
1670 | PyErr_SetString(PyExc_ValueError, "overflow in init_nt"); |
|
1670 | PyErr_SetString(PyExc_ValueError, "overflow in init_nt"); | |
1671 | return -1; |
|
1671 | return -1; | |
1672 | } |
|
1672 | } | |
1673 | self->nodes = calloc(self->capacity, sizeof(nodetreenode)); |
|
1673 | self->nodes = calloc(self->capacity, sizeof(nodetreenode)); | |
1674 | if (self->nodes == NULL) { |
|
1674 | if (self->nodes == NULL) { | |
1675 | PyErr_NoMemory(); |
|
1675 | PyErr_NoMemory(); | |
1676 | return -1; |
|
1676 | return -1; | |
1677 | } |
|
1677 | } | |
1678 | self->length = 1; |
|
1678 | self->length = 1; | |
1679 | return 0; |
|
1679 | return 0; | |
1680 | } |
|
1680 | } | |
1681 |
|
1681 | |||
1682 | static int ntobj_init(nodetreeObject *self, PyObject *args) |
|
1682 | static int ntobj_init(nodetreeObject *self, PyObject *args) | |
1683 | { |
|
1683 | { | |
1684 | PyObject *index; |
|
1684 | PyObject *index; | |
1685 | unsigned capacity; |
|
1685 | unsigned capacity; | |
1686 | if (!PyArg_ParseTuple(args, "O!I", &HgRevlogIndex_Type, &index, |
|
1686 | if (!PyArg_ParseTuple(args, "O!I", &HgRevlogIndex_Type, &index, | |
1687 | &capacity)) |
|
1687 | &capacity)) | |
1688 | return -1; |
|
1688 | return -1; | |
1689 | Py_INCREF(index); |
|
1689 | Py_INCREF(index); | |
1690 | return nt_init(&self->nt, (indexObject *)index, capacity); |
|
1690 | return nt_init(&self->nt, (indexObject *)index, capacity); | |
1691 | } |
|
1691 | } | |
1692 |
|
1692 | |||
1693 | static int nt_partialmatch(nodetree *self, const char *node, Py_ssize_t nodelen) |
|
1693 | static int nt_partialmatch(nodetree *self, const char *node, Py_ssize_t nodelen) | |
1694 | { |
|
1694 | { | |
1695 | return nt_find(self, node, nodelen, 1); |
|
1695 | return nt_find(self, node, nodelen, 1); | |
1696 | } |
|
1696 | } | |
1697 |
|
1697 | |||
1698 | /* |
|
1698 | /* | |
1699 | * Find the length of the shortest unique prefix of node. |
|
1699 | * Find the length of the shortest unique prefix of node. | |
1700 | * |
|
1700 | * | |
1701 | * Return values: |
|
1701 | * Return values: | |
1702 | * |
|
1702 | * | |
1703 | * -3: error (exception set) |
|
1703 | * -3: error (exception set) | |
1704 | * -2: not found (no exception set) |
|
1704 | * -2: not found (no exception set) | |
1705 | * rest: length of shortest prefix |
|
1705 | * rest: length of shortest prefix | |
1706 | */ |
|
1706 | */ | |
1707 | static int nt_shortest(nodetree *self, const char *node) |
|
1707 | static int nt_shortest(nodetree *self, const char *node) | |
1708 | { |
|
1708 | { | |
1709 | int level, off; |
|
1709 | int level, off; | |
1710 |
|
1710 | |||
1711 | for (level = off = 0; level < 40; level++) { |
|
1711 | for (level = off = 0; level < 40; level++) { | |
1712 | int k, v; |
|
1712 | int k, v; | |
1713 | nodetreenode *n = &self->nodes[off]; |
|
1713 | nodetreenode *n = &self->nodes[off]; | |
1714 | k = nt_level(node, level); |
|
1714 | k = nt_level(node, level); | |
1715 | v = n->children[k]; |
|
1715 | v = n->children[k]; | |
1716 | if (v < 0) { |
|
1716 | if (v < 0) { | |
1717 | const char *n; |
|
1717 | const char *n; | |
1718 | v = -(v + 2); |
|
1718 | v = -(v + 2); | |
1719 | n = index_node_existing(self->index, v); |
|
1719 | n = index_node_existing(self->index, v); | |
1720 | if (n == NULL) |
|
1720 | if (n == NULL) | |
1721 | return -3; |
|
1721 | return -3; | |
1722 | if (memcmp(node, n, 20) != 0) |
|
1722 | if (memcmp(node, n, 20) != 0) | |
1723 | /* |
|
1723 | /* | |
1724 | * Found a unique prefix, but it wasn't for the |
|
1724 | * Found a unique prefix, but it wasn't for the | |
1725 | * requested node (i.e the requested node does |
|
1725 | * requested node (i.e the requested node does | |
1726 | * not exist). |
|
1726 | * not exist). | |
1727 | */ |
|
1727 | */ | |
1728 | return -2; |
|
1728 | return -2; | |
1729 | return level + 1; |
|
1729 | return level + 1; | |
1730 | } |
|
1730 | } | |
1731 | if (v == 0) |
|
1731 | if (v == 0) | |
1732 | return -2; |
|
1732 | return -2; | |
1733 | off = v; |
|
1733 | off = v; | |
1734 | } |
|
1734 | } | |
1735 | /* |
|
1735 | /* | |
1736 | * The node was still not unique after 40 hex digits, so this won't |
|
1736 | * The node was still not unique after 40 hex digits, so this won't | |
1737 | * happen. Also, if we get here, then there's a programming error in |
|
1737 | * happen. Also, if we get here, then there's a programming error in | |
1738 | * this file that made us insert a node longer than 40 hex digits. |
|
1738 | * this file that made us insert a node longer than 40 hex digits. | |
1739 | */ |
|
1739 | */ | |
1740 | PyErr_SetString(PyExc_Exception, "broken node tree"); |
|
1740 | PyErr_SetString(PyExc_Exception, "broken node tree"); | |
1741 | return -3; |
|
1741 | return -3; | |
1742 | } |
|
1742 | } | |
1743 |
|
1743 | |||
1744 | static PyObject *ntobj_shortest(nodetreeObject *self, PyObject *args) |
|
1744 | static PyObject *ntobj_shortest(nodetreeObject *self, PyObject *args) | |
1745 | { |
|
1745 | { | |
1746 | PyObject *val; |
|
1746 | PyObject *val; | |
1747 | char *node; |
|
1747 | char *node; | |
1748 | int length; |
|
1748 | int length; | |
1749 |
|
1749 | |||
1750 | if (!PyArg_ParseTuple(args, "O", &val)) |
|
1750 | if (!PyArg_ParseTuple(args, "O", &val)) | |
1751 | return NULL; |
|
1751 | return NULL; | |
1752 | if (node_check(val, &node) == -1) |
|
1752 | if (node_check(val, &node) == -1) | |
1753 | return NULL; |
|
1753 | return NULL; | |
1754 |
|
1754 | |||
1755 | length = nt_shortest(&self->nt, node); |
|
1755 | length = nt_shortest(&self->nt, node); | |
1756 | if (length == -3) |
|
1756 | if (length == -3) | |
1757 | return NULL; |
|
1757 | return NULL; | |
1758 | if (length == -2) { |
|
1758 | if (length == -2) { | |
1759 | raise_revlog_error(); |
|
1759 | raise_revlog_error(); | |
1760 | return NULL; |
|
1760 | return NULL; | |
1761 | } |
|
1761 | } | |
1762 | return PyInt_FromLong(length); |
|
1762 | return PyInt_FromLong(length); | |
1763 | } |
|
1763 | } | |
1764 |
|
1764 | |||
1765 | static void nt_dealloc(nodetree *self) |
|
1765 | static void nt_dealloc(nodetree *self) | |
1766 | { |
|
1766 | { | |
1767 | free(self->nodes); |
|
1767 | free(self->nodes); | |
1768 | self->nodes = NULL; |
|
1768 | self->nodes = NULL; | |
1769 | } |
|
1769 | } | |
1770 |
|
1770 | |||
1771 | static void ntobj_dealloc(nodetreeObject *self) |
|
1771 | static void ntobj_dealloc(nodetreeObject *self) | |
1772 | { |
|
1772 | { | |
1773 | Py_XDECREF(self->nt.index); |
|
1773 | Py_XDECREF(self->nt.index); | |
1774 | nt_dealloc(&self->nt); |
|
1774 | nt_dealloc(&self->nt); | |
1775 | PyObject_Del(self); |
|
1775 | PyObject_Del(self); | |
1776 | } |
|
1776 | } | |
1777 |
|
1777 | |||
1778 | static PyMethodDef ntobj_methods[] = { |
|
1778 | static PyMethodDef ntobj_methods[] = { | |
1779 | {"insert", (PyCFunction)ntobj_insert, METH_VARARGS, |
|
1779 | {"insert", (PyCFunction)ntobj_insert, METH_VARARGS, | |
1780 | "insert an index entry"}, |
|
1780 | "insert an index entry"}, | |
1781 | {"shortest", (PyCFunction)ntobj_shortest, METH_VARARGS, |
|
1781 | {"shortest", (PyCFunction)ntobj_shortest, METH_VARARGS, | |
1782 | "find length of shortest hex nodeid of a binary ID"}, |
|
1782 | "find length of shortest hex nodeid of a binary ID"}, | |
1783 | {NULL} /* Sentinel */ |
|
1783 | {NULL} /* Sentinel */ | |
1784 | }; |
|
1784 | }; | |
1785 |
|
1785 | |||
1786 | static PyTypeObject nodetreeType = { |
|
1786 | static PyTypeObject nodetreeType = { | |
1787 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ |
|
1787 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ | |
1788 | "parsers.nodetree", /* tp_name */ |
|
1788 | "parsers.nodetree", /* tp_name */ | |
1789 | sizeof(nodetreeObject), /* tp_basicsize */ |
|
1789 | sizeof(nodetreeObject), /* tp_basicsize */ | |
1790 | 0, /* tp_itemsize */ |
|
1790 | 0, /* tp_itemsize */ | |
1791 | (destructor)ntobj_dealloc, /* tp_dealloc */ |
|
1791 | (destructor)ntobj_dealloc, /* tp_dealloc */ | |
1792 | 0, /* tp_print */ |
|
1792 | 0, /* tp_print */ | |
1793 | 0, /* tp_getattr */ |
|
1793 | 0, /* tp_getattr */ | |
1794 | 0, /* tp_setattr */ |
|
1794 | 0, /* tp_setattr */ | |
1795 | 0, /* tp_compare */ |
|
1795 | 0, /* tp_compare */ | |
1796 | 0, /* tp_repr */ |
|
1796 | 0, /* tp_repr */ | |
1797 | 0, /* tp_as_number */ |
|
1797 | 0, /* tp_as_number */ | |
1798 | 0, /* tp_as_sequence */ |
|
1798 | 0, /* tp_as_sequence */ | |
1799 | 0, /* tp_as_mapping */ |
|
1799 | 0, /* tp_as_mapping */ | |
1800 | 0, /* tp_hash */ |
|
1800 | 0, /* tp_hash */ | |
1801 | 0, /* tp_call */ |
|
1801 | 0, /* tp_call */ | |
1802 | 0, /* tp_str */ |
|
1802 | 0, /* tp_str */ | |
1803 | 0, /* tp_getattro */ |
|
1803 | 0, /* tp_getattro */ | |
1804 | 0, /* tp_setattro */ |
|
1804 | 0, /* tp_setattro */ | |
1805 | 0, /* tp_as_buffer */ |
|
1805 | 0, /* tp_as_buffer */ | |
1806 | Py_TPFLAGS_DEFAULT, /* tp_flags */ |
|
1806 | Py_TPFLAGS_DEFAULT, /* tp_flags */ | |
1807 | "nodetree", /* tp_doc */ |
|
1807 | "nodetree", /* tp_doc */ | |
1808 | 0, /* tp_traverse */ |
|
1808 | 0, /* tp_traverse */ | |
1809 | 0, /* tp_clear */ |
|
1809 | 0, /* tp_clear */ | |
1810 | 0, /* tp_richcompare */ |
|
1810 | 0, /* tp_richcompare */ | |
1811 | 0, /* tp_weaklistoffset */ |
|
1811 | 0, /* tp_weaklistoffset */ | |
1812 | 0, /* tp_iter */ |
|
1812 | 0, /* tp_iter */ | |
1813 | 0, /* tp_iternext */ |
|
1813 | 0, /* tp_iternext */ | |
1814 | ntobj_methods, /* tp_methods */ |
|
1814 | ntobj_methods, /* tp_methods */ | |
1815 | 0, /* tp_members */ |
|
1815 | 0, /* tp_members */ | |
1816 | 0, /* tp_getset */ |
|
1816 | 0, /* tp_getset */ | |
1817 | 0, /* tp_base */ |
|
1817 | 0, /* tp_base */ | |
1818 | 0, /* tp_dict */ |
|
1818 | 0, /* tp_dict */ | |
1819 | 0, /* tp_descr_get */ |
|
1819 | 0, /* tp_descr_get */ | |
1820 | 0, /* tp_descr_set */ |
|
1820 | 0, /* tp_descr_set */ | |
1821 | 0, /* tp_dictoffset */ |
|
1821 | 0, /* tp_dictoffset */ | |
1822 | (initproc)ntobj_init, /* tp_init */ |
|
1822 | (initproc)ntobj_init, /* tp_init */ | |
1823 | 0, /* tp_alloc */ |
|
1823 | 0, /* tp_alloc */ | |
1824 | }; |
|
1824 | }; | |
1825 |
|
1825 | |||
1826 | static int index_init_nt(indexObject *self) |
|
1826 | static int index_init_nt(indexObject *self) | |
1827 | { |
|
1827 | { | |
1828 | if (!self->ntinitialized) { |
|
1828 | if (!self->ntinitialized) { | |
1829 | if (nt_init(&self->nt, self, (int)self->raw_length) == -1) { |
|
1829 | if (nt_init(&self->nt, self, (int)self->raw_length) == -1) { | |
1830 | nt_dealloc(&self->nt); |
|
1830 | nt_dealloc(&self->nt); | |
1831 | return -1; |
|
1831 | return -1; | |
1832 | } |
|
1832 | } | |
1833 | if (nt_insert(&self->nt, nullid, -1) == -1) { |
|
1833 | if (nt_insert(&self->nt, nullid, -1) == -1) { | |
1834 | nt_dealloc(&self->nt); |
|
1834 | nt_dealloc(&self->nt); | |
1835 | return -1; |
|
1835 | return -1; | |
1836 | } |
|
1836 | } | |
1837 | self->ntinitialized = 1; |
|
1837 | self->ntinitialized = 1; | |
1838 | self->ntrev = (int)index_length(self); |
|
1838 | self->ntrev = (int)index_length(self); | |
1839 | self->ntlookups = 1; |
|
1839 | self->ntlookups = 1; | |
1840 | self->ntmisses = 0; |
|
1840 | self->ntmisses = 0; | |
1841 | } |
|
1841 | } | |
1842 | return 0; |
|
1842 | return 0; | |
1843 | } |
|
1843 | } | |
1844 |
|
1844 | |||
1845 | /* |
|
1845 | /* | |
1846 | * Return values: |
|
1846 | * Return values: | |
1847 | * |
|
1847 | * | |
1848 | * -3: error (exception set) |
|
1848 | * -3: error (exception set) | |
1849 | * -2: not found (no exception set) |
|
1849 | * -2: not found (no exception set) | |
1850 | * rest: valid rev |
|
1850 | * rest: valid rev | |
1851 | */ |
|
1851 | */ | |
1852 | static int index_find_node(indexObject *self, const char *node, |
|
1852 | static int index_find_node(indexObject *self, const char *node, | |
1853 | Py_ssize_t nodelen) |
|
1853 | Py_ssize_t nodelen) | |
1854 | { |
|
1854 | { | |
1855 | int rev; |
|
1855 | int rev; | |
1856 |
|
1856 | |||
1857 | if (index_init_nt(self) == -1) |
|
1857 | if (index_init_nt(self) == -1) | |
1858 | return -3; |
|
1858 | return -3; | |
1859 |
|
1859 | |||
1860 | self->ntlookups++; |
|
1860 | self->ntlookups++; | |
1861 | rev = nt_find(&self->nt, node, nodelen, 0); |
|
1861 | rev = nt_find(&self->nt, node, nodelen, 0); | |
1862 | if (rev >= -1) |
|
1862 | if (rev >= -1) | |
1863 | return rev; |
|
1863 | return rev; | |
1864 |
|
1864 | |||
1865 | /* |
|
1865 | /* | |
1866 | * For the first handful of lookups, we scan the entire index, |
|
1866 | * For the first handful of lookups, we scan the entire index, | |
1867 | * and cache only the matching nodes. This optimizes for cases |
|
1867 | * and cache only the matching nodes. This optimizes for cases | |
1868 | * like "hg tip", where only a few nodes are accessed. |
|
1868 | * like "hg tip", where only a few nodes are accessed. | |
1869 | * |
|
1869 | * | |
1870 | * After that, we cache every node we visit, using a single |
|
1870 | * After that, we cache every node we visit, using a single | |
1871 | * scan amortized over multiple lookups. This gives the best |
|
1871 | * scan amortized over multiple lookups. This gives the best | |
1872 | * bulk performance, e.g. for "hg log". |
|
1872 | * bulk performance, e.g. for "hg log". | |
1873 | */ |
|
1873 | */ | |
1874 | if (self->ntmisses++ < 4) { |
|
1874 | if (self->ntmisses++ < 4) { | |
1875 | for (rev = self->ntrev - 1; rev >= 0; rev--) { |
|
1875 | for (rev = self->ntrev - 1; rev >= 0; rev--) { | |
1876 | const char *n = index_node_existing(self, rev); |
|
1876 | const char *n = index_node_existing(self, rev); | |
1877 | if (n == NULL) |
|
1877 | if (n == NULL) | |
1878 | return -3; |
|
1878 | return -3; | |
1879 | if (memcmp(node, n, nodelen > 20 ? 20 : nodelen) == 0) { |
|
1879 | if (memcmp(node, n, nodelen > 20 ? 20 : nodelen) == 0) { | |
1880 | if (nt_insert(&self->nt, n, rev) == -1) |
|
1880 | if (nt_insert(&self->nt, n, rev) == -1) | |
1881 | return -3; |
|
1881 | return -3; | |
1882 | break; |
|
1882 | break; | |
1883 | } |
|
1883 | } | |
1884 | } |
|
1884 | } | |
1885 | } else { |
|
1885 | } else { | |
1886 | for (rev = self->ntrev - 1; rev >= 0; rev--) { |
|
1886 | for (rev = self->ntrev - 1; rev >= 0; rev--) { | |
1887 | const char *n = index_node_existing(self, rev); |
|
1887 | const char *n = index_node_existing(self, rev); | |
1888 | if (n == NULL) |
|
1888 | if (n == NULL) | |
1889 | return -3; |
|
1889 | return -3; | |
1890 | if (nt_insert(&self->nt, n, rev) == -1) { |
|
1890 | if (nt_insert(&self->nt, n, rev) == -1) { | |
1891 | self->ntrev = rev + 1; |
|
1891 | self->ntrev = rev + 1; | |
1892 | return -3; |
|
1892 | return -3; | |
1893 | } |
|
1893 | } | |
1894 | if (memcmp(node, n, nodelen > 20 ? 20 : nodelen) == 0) { |
|
1894 | if (memcmp(node, n, nodelen > 20 ? 20 : nodelen) == 0) { | |
1895 | break; |
|
1895 | break; | |
1896 | } |
|
1896 | } | |
1897 | } |
|
1897 | } | |
1898 | self->ntrev = rev; |
|
1898 | self->ntrev = rev; | |
1899 | } |
|
1899 | } | |
1900 |
|
1900 | |||
1901 | if (rev >= 0) |
|
1901 | if (rev >= 0) | |
1902 | return rev; |
|
1902 | return rev; | |
1903 | return -2; |
|
1903 | return -2; | |
1904 | } |
|
1904 | } | |
1905 |
|
1905 | |||
1906 | static PyObject *index_getitem(indexObject *self, PyObject *value) |
|
1906 | static PyObject *index_getitem(indexObject *self, PyObject *value) | |
1907 | { |
|
1907 | { | |
1908 | char *node; |
|
1908 | char *node; | |
1909 | int rev; |
|
1909 | int rev; | |
1910 |
|
1910 | |||
1911 | if (PyInt_Check(value)) { |
|
1911 | if (PyInt_Check(value)) { | |
1912 | long idx; |
|
1912 | long idx; | |
1913 | if (!pylong_to_long(value, &idx)) { |
|
1913 | if (!pylong_to_long(value, &idx)) { | |
1914 | return NULL; |
|
1914 | return NULL; | |
1915 | } |
|
1915 | } | |
1916 | return index_get(self, idx); |
|
1916 | return index_get(self, idx); | |
1917 | } |
|
1917 | } | |
1918 |
|
1918 | |||
1919 | if (node_check(value, &node) == -1) |
|
1919 | if (node_check(value, &node) == -1) | |
1920 | return NULL; |
|
1920 | return NULL; | |
1921 | rev = index_find_node(self, node, 20); |
|
1921 | rev = index_find_node(self, node, 20); | |
1922 | if (rev >= -1) |
|
1922 | if (rev >= -1) | |
1923 | return PyInt_FromLong(rev); |
|
1923 | return PyInt_FromLong(rev); | |
1924 | if (rev == -2) |
|
1924 | if (rev == -2) | |
1925 | raise_revlog_error(); |
|
1925 | raise_revlog_error(); | |
1926 | return NULL; |
|
1926 | return NULL; | |
1927 | } |
|
1927 | } | |
1928 |
|
1928 | |||
1929 | /* |
|
1929 | /* | |
1930 | * Fully populate the radix tree. |
|
1930 | * Fully populate the radix tree. | |
1931 | */ |
|
1931 | */ | |
1932 | static int index_populate_nt(indexObject *self) |
|
1932 | static int index_populate_nt(indexObject *self) | |
1933 | { |
|
1933 | { | |
1934 | int rev; |
|
1934 | int rev; | |
1935 | if (self->ntrev > 0) { |
|
1935 | if (self->ntrev > 0) { | |
1936 | for (rev = self->ntrev - 1; rev >= 0; rev--) { |
|
1936 | for (rev = self->ntrev - 1; rev >= 0; rev--) { | |
1937 | const char *n = index_node_existing(self, rev); |
|
1937 | const char *n = index_node_existing(self, rev); | |
1938 | if (n == NULL) |
|
1938 | if (n == NULL) | |
1939 | return -1; |
|
1939 | return -1; | |
1940 | if (nt_insert(&self->nt, n, rev) == -1) |
|
1940 | if (nt_insert(&self->nt, n, rev) == -1) | |
1941 | return -1; |
|
1941 | return -1; | |
1942 | } |
|
1942 | } | |
1943 | self->ntrev = -1; |
|
1943 | self->ntrev = -1; | |
1944 | } |
|
1944 | } | |
1945 | return 0; |
|
1945 | return 0; | |
1946 | } |
|
1946 | } | |
1947 |
|
1947 | |||
1948 | static PyObject *index_partialmatch(indexObject *self, PyObject *args) |
|
1948 | static PyObject *index_partialmatch(indexObject *self, PyObject *args) | |
1949 | { |
|
1949 | { | |
1950 | const char *fullnode; |
|
1950 | const char *fullnode; | |
1951 | Py_ssize_t nodelen; |
|
1951 | Py_ssize_t nodelen; | |
1952 | char *node; |
|
1952 | char *node; | |
1953 | int rev, i; |
|
1953 | int rev, i; | |
1954 |
|
1954 | |||
1955 | if (!PyArg_ParseTuple(args, PY23("s#", "y#"), &node, &nodelen)) |
|
1955 | if (!PyArg_ParseTuple(args, PY23("s#", "y#"), &node, &nodelen)) | |
1956 | return NULL; |
|
1956 | return NULL; | |
1957 |
|
1957 | |||
1958 | if (nodelen < 1) { |
|
1958 | if (nodelen < 1) { | |
1959 | PyErr_SetString(PyExc_ValueError, "key too short"); |
|
1959 | PyErr_SetString(PyExc_ValueError, "key too short"); | |
1960 | return NULL; |
|
1960 | return NULL; | |
1961 | } |
|
1961 | } | |
1962 |
|
1962 | |||
1963 | if (nodelen > 40) { |
|
1963 | if (nodelen > 40) { | |
1964 | PyErr_SetString(PyExc_ValueError, "key too long"); |
|
1964 | PyErr_SetString(PyExc_ValueError, "key too long"); | |
1965 | return NULL; |
|
1965 | return NULL; | |
1966 | } |
|
1966 | } | |
1967 |
|
1967 | |||
1968 | for (i = 0; i < nodelen; i++) |
|
1968 | for (i = 0; i < nodelen; i++) | |
1969 | hexdigit(node, i); |
|
1969 | hexdigit(node, i); | |
1970 | if (PyErr_Occurred()) { |
|
1970 | if (PyErr_Occurred()) { | |
1971 | /* input contains non-hex characters */ |
|
1971 | /* input contains non-hex characters */ | |
1972 | PyErr_Clear(); |
|
1972 | PyErr_Clear(); | |
1973 | Py_RETURN_NONE; |
|
1973 | Py_RETURN_NONE; | |
1974 | } |
|
1974 | } | |
1975 |
|
1975 | |||
1976 | if (index_init_nt(self) == -1) |
|
1976 | if (index_init_nt(self) == -1) | |
1977 | return NULL; |
|
1977 | return NULL; | |
1978 | if (index_populate_nt(self) == -1) |
|
1978 | if (index_populate_nt(self) == -1) | |
1979 | return NULL; |
|
1979 | return NULL; | |
1980 | rev = nt_partialmatch(&self->nt, node, nodelen); |
|
1980 | rev = nt_partialmatch(&self->nt, node, nodelen); | |
1981 |
|
1981 | |||
1982 | switch (rev) { |
|
1982 | switch (rev) { | |
1983 | case -4: |
|
1983 | case -4: | |
1984 | raise_revlog_error(); |
|
1984 | raise_revlog_error(); | |
1985 | return NULL; |
|
1985 | return NULL; | |
1986 | case -2: |
|
1986 | case -2: | |
1987 | Py_RETURN_NONE; |
|
1987 | Py_RETURN_NONE; | |
1988 | case -1: |
|
1988 | case -1: | |
1989 | return PyBytes_FromStringAndSize(nullid, 20); |
|
1989 | return PyBytes_FromStringAndSize(nullid, 20); | |
1990 | } |
|
1990 | } | |
1991 |
|
1991 | |||
1992 | fullnode = index_node_existing(self, rev); |
|
1992 | fullnode = index_node_existing(self, rev); | |
1993 | if (fullnode == NULL) { |
|
1993 | if (fullnode == NULL) { | |
1994 | return NULL; |
|
1994 | return NULL; | |
1995 | } |
|
1995 | } | |
1996 | return PyBytes_FromStringAndSize(fullnode, 20); |
|
1996 | return PyBytes_FromStringAndSize(fullnode, 20); | |
1997 | } |
|
1997 | } | |
1998 |
|
1998 | |||
1999 | static PyObject *index_shortest(indexObject *self, PyObject *args) |
|
1999 | static PyObject *index_shortest(indexObject *self, PyObject *args) | |
2000 | { |
|
2000 | { | |
2001 | PyObject *val; |
|
2001 | PyObject *val; | |
2002 | char *node; |
|
2002 | char *node; | |
2003 | int length; |
|
2003 | int length; | |
2004 |
|
2004 | |||
2005 | if (!PyArg_ParseTuple(args, "O", &val)) |
|
2005 | if (!PyArg_ParseTuple(args, "O", &val)) | |
2006 | return NULL; |
|
2006 | return NULL; | |
2007 | if (node_check(val, &node) == -1) |
|
2007 | if (node_check(val, &node) == -1) | |
2008 | return NULL; |
|
2008 | return NULL; | |
2009 |
|
2009 | |||
2010 | self->ntlookups++; |
|
2010 | self->ntlookups++; | |
2011 | if (index_init_nt(self) == -1) |
|
2011 | if (index_init_nt(self) == -1) | |
2012 | return NULL; |
|
2012 | return NULL; | |
2013 | if (index_populate_nt(self) == -1) |
|
2013 | if (index_populate_nt(self) == -1) | |
2014 | return NULL; |
|
2014 | return NULL; | |
2015 | length = nt_shortest(&self->nt, node); |
|
2015 | length = nt_shortest(&self->nt, node); | |
2016 | if (length == -3) |
|
2016 | if (length == -3) | |
2017 | return NULL; |
|
2017 | return NULL; | |
2018 | if (length == -2) { |
|
2018 | if (length == -2) { | |
2019 | raise_revlog_error(); |
|
2019 | raise_revlog_error(); | |
2020 | return NULL; |
|
2020 | return NULL; | |
2021 | } |
|
2021 | } | |
2022 | return PyInt_FromLong(length); |
|
2022 | return PyInt_FromLong(length); | |
2023 | } |
|
2023 | } | |
2024 |
|
2024 | |||
2025 | static PyObject *index_m_get(indexObject *self, PyObject *args) |
|
2025 | static PyObject *index_m_get(indexObject *self, PyObject *args) | |
2026 | { |
|
2026 | { | |
2027 | PyObject *val; |
|
2027 | PyObject *val; | |
2028 | char *node; |
|
2028 | char *node; | |
2029 | int rev; |
|
2029 | int rev; | |
2030 |
|
2030 | |||
2031 | if (!PyArg_ParseTuple(args, "O", &val)) |
|
2031 | if (!PyArg_ParseTuple(args, "O", &val)) | |
2032 | return NULL; |
|
2032 | return NULL; | |
2033 | if (node_check(val, &node) == -1) |
|
2033 | if (node_check(val, &node) == -1) | |
2034 | return NULL; |
|
2034 | return NULL; | |
2035 | rev = index_find_node(self, node, 20); |
|
2035 | rev = index_find_node(self, node, 20); | |
2036 | if (rev == -3) |
|
2036 | if (rev == -3) | |
2037 | return NULL; |
|
2037 | return NULL; | |
2038 | if (rev == -2) |
|
2038 | if (rev == -2) | |
2039 | Py_RETURN_NONE; |
|
2039 | Py_RETURN_NONE; | |
2040 | return PyInt_FromLong(rev); |
|
2040 | return PyInt_FromLong(rev); | |
2041 | } |
|
2041 | } | |
2042 |
|
2042 | |||
2043 | static int index_contains(indexObject *self, PyObject *value) |
|
2043 | static int index_contains(indexObject *self, PyObject *value) | |
2044 | { |
|
2044 | { | |
2045 | char *node; |
|
2045 | char *node; | |
2046 |
|
2046 | |||
2047 | if (PyInt_Check(value)) { |
|
2047 | if (PyInt_Check(value)) { | |
2048 | long rev; |
|
2048 | long rev; | |
2049 | if (!pylong_to_long(value, &rev)) { |
|
2049 | if (!pylong_to_long(value, &rev)) { | |
2050 | return -1; |
|
2050 | return -1; | |
2051 | } |
|
2051 | } | |
2052 | return rev >= -1 && rev < index_length(self); |
|
2052 | return rev >= -1 && rev < index_length(self); | |
2053 | } |
|
2053 | } | |
2054 |
|
2054 | |||
2055 | if (node_check(value, &node) == -1) |
|
2055 | if (node_check(value, &node) == -1) | |
2056 | return -1; |
|
2056 | return -1; | |
2057 |
|
2057 | |||
2058 | switch (index_find_node(self, node, 20)) { |
|
2058 | switch (index_find_node(self, node, 20)) { | |
2059 | case -3: |
|
2059 | case -3: | |
2060 | return -1; |
|
2060 | return -1; | |
2061 | case -2: |
|
2061 | case -2: | |
2062 | return 0; |
|
2062 | return 0; | |
2063 | default: |
|
2063 | default: | |
2064 | return 1; |
|
2064 | return 1; | |
2065 | } |
|
2065 | } | |
2066 | } |
|
2066 | } | |
2067 |
|
2067 | |||
|
2068 | static PyObject *index_m_has_node(indexObject *self, PyObject *args) | |||
|
2069 | { | |||
|
2070 | int ret = index_contains(self, args); | |||
|
2071 | if (ret < 0) | |||
|
2072 | return NULL; | |||
|
2073 | return PyBool_FromLong((long)ret); | |||
|
2074 | } | |||
|
2075 | ||||
2068 | typedef uint64_t bitmask; |
|
2076 | typedef uint64_t bitmask; | |
2069 |
|
2077 | |||
2070 | /* |
|
2078 | /* | |
2071 | * Given a disjoint set of revs, return all candidates for the |
|
2079 | * Given a disjoint set of revs, return all candidates for the | |
2072 | * greatest common ancestor. In revset notation, this is the set |
|
2080 | * greatest common ancestor. In revset notation, this is the set | |
2073 | * "heads(::a and ::b and ...)" |
|
2081 | * "heads(::a and ::b and ...)" | |
2074 | */ |
|
2082 | */ | |
2075 | static PyObject *find_gca_candidates(indexObject *self, const int *revs, |
|
2083 | static PyObject *find_gca_candidates(indexObject *self, const int *revs, | |
2076 | int revcount) |
|
2084 | int revcount) | |
2077 | { |
|
2085 | { | |
2078 | const bitmask allseen = (1ull << revcount) - 1; |
|
2086 | const bitmask allseen = (1ull << revcount) - 1; | |
2079 | const bitmask poison = 1ull << revcount; |
|
2087 | const bitmask poison = 1ull << revcount; | |
2080 | PyObject *gca = PyList_New(0); |
|
2088 | PyObject *gca = PyList_New(0); | |
2081 | int i, v, interesting; |
|
2089 | int i, v, interesting; | |
2082 | int maxrev = -1; |
|
2090 | int maxrev = -1; | |
2083 | bitmask sp; |
|
2091 | bitmask sp; | |
2084 | bitmask *seen; |
|
2092 | bitmask *seen; | |
2085 |
|
2093 | |||
2086 | if (gca == NULL) |
|
2094 | if (gca == NULL) | |
2087 | return PyErr_NoMemory(); |
|
2095 | return PyErr_NoMemory(); | |
2088 |
|
2096 | |||
2089 | for (i = 0; i < revcount; i++) { |
|
2097 | for (i = 0; i < revcount; i++) { | |
2090 | if (revs[i] > maxrev) |
|
2098 | if (revs[i] > maxrev) | |
2091 | maxrev = revs[i]; |
|
2099 | maxrev = revs[i]; | |
2092 | } |
|
2100 | } | |
2093 |
|
2101 | |||
2094 | seen = calloc(sizeof(*seen), maxrev + 1); |
|
2102 | seen = calloc(sizeof(*seen), maxrev + 1); | |
2095 | if (seen == NULL) { |
|
2103 | if (seen == NULL) { | |
2096 | Py_DECREF(gca); |
|
2104 | Py_DECREF(gca); | |
2097 | return PyErr_NoMemory(); |
|
2105 | return PyErr_NoMemory(); | |
2098 | } |
|
2106 | } | |
2099 |
|
2107 | |||
2100 | for (i = 0; i < revcount; i++) |
|
2108 | for (i = 0; i < revcount; i++) | |
2101 | seen[revs[i]] = 1ull << i; |
|
2109 | seen[revs[i]] = 1ull << i; | |
2102 |
|
2110 | |||
2103 | interesting = revcount; |
|
2111 | interesting = revcount; | |
2104 |
|
2112 | |||
2105 | for (v = maxrev; v >= 0 && interesting; v--) { |
|
2113 | for (v = maxrev; v >= 0 && interesting; v--) { | |
2106 | bitmask sv = seen[v]; |
|
2114 | bitmask sv = seen[v]; | |
2107 | int parents[2]; |
|
2115 | int parents[2]; | |
2108 |
|
2116 | |||
2109 | if (!sv) |
|
2117 | if (!sv) | |
2110 | continue; |
|
2118 | continue; | |
2111 |
|
2119 | |||
2112 | if (sv < poison) { |
|
2120 | if (sv < poison) { | |
2113 | interesting -= 1; |
|
2121 | interesting -= 1; | |
2114 | if (sv == allseen) { |
|
2122 | if (sv == allseen) { | |
2115 | PyObject *obj = PyInt_FromLong(v); |
|
2123 | PyObject *obj = PyInt_FromLong(v); | |
2116 | if (obj == NULL) |
|
2124 | if (obj == NULL) | |
2117 | goto bail; |
|
2125 | goto bail; | |
2118 | if (PyList_Append(gca, obj) == -1) { |
|
2126 | if (PyList_Append(gca, obj) == -1) { | |
2119 | Py_DECREF(obj); |
|
2127 | Py_DECREF(obj); | |
2120 | goto bail; |
|
2128 | goto bail; | |
2121 | } |
|
2129 | } | |
2122 | sv |= poison; |
|
2130 | sv |= poison; | |
2123 | for (i = 0; i < revcount; i++) { |
|
2131 | for (i = 0; i < revcount; i++) { | |
2124 | if (revs[i] == v) |
|
2132 | if (revs[i] == v) | |
2125 | goto done; |
|
2133 | goto done; | |
2126 | } |
|
2134 | } | |
2127 | } |
|
2135 | } | |
2128 | } |
|
2136 | } | |
2129 | if (index_get_parents(self, v, parents, maxrev) < 0) |
|
2137 | if (index_get_parents(self, v, parents, maxrev) < 0) | |
2130 | goto bail; |
|
2138 | goto bail; | |
2131 |
|
2139 | |||
2132 | for (i = 0; i < 2; i++) { |
|
2140 | for (i = 0; i < 2; i++) { | |
2133 | int p = parents[i]; |
|
2141 | int p = parents[i]; | |
2134 | if (p == -1) |
|
2142 | if (p == -1) | |
2135 | continue; |
|
2143 | continue; | |
2136 | sp = seen[p]; |
|
2144 | sp = seen[p]; | |
2137 | if (sv < poison) { |
|
2145 | if (sv < poison) { | |
2138 | if (sp == 0) { |
|
2146 | if (sp == 0) { | |
2139 | seen[p] = sv; |
|
2147 | seen[p] = sv; | |
2140 | interesting++; |
|
2148 | interesting++; | |
2141 | } else if (sp != sv) |
|
2149 | } else if (sp != sv) | |
2142 | seen[p] |= sv; |
|
2150 | seen[p] |= sv; | |
2143 | } else { |
|
2151 | } else { | |
2144 | if (sp && sp < poison) |
|
2152 | if (sp && sp < poison) | |
2145 | interesting--; |
|
2153 | interesting--; | |
2146 | seen[p] = sv; |
|
2154 | seen[p] = sv; | |
2147 | } |
|
2155 | } | |
2148 | } |
|
2156 | } | |
2149 | } |
|
2157 | } | |
2150 |
|
2158 | |||
2151 | done: |
|
2159 | done: | |
2152 | free(seen); |
|
2160 | free(seen); | |
2153 | return gca; |
|
2161 | return gca; | |
2154 | bail: |
|
2162 | bail: | |
2155 | free(seen); |
|
2163 | free(seen); | |
2156 | Py_XDECREF(gca); |
|
2164 | Py_XDECREF(gca); | |
2157 | return NULL; |
|
2165 | return NULL; | |
2158 | } |
|
2166 | } | |
2159 |
|
2167 | |||
2160 | /* |
|
2168 | /* | |
2161 | * Given a disjoint set of revs, return the subset with the longest |
|
2169 | * Given a disjoint set of revs, return the subset with the longest | |
2162 | * path to the root. |
|
2170 | * path to the root. | |
2163 | */ |
|
2171 | */ | |
2164 | static PyObject *find_deepest(indexObject *self, PyObject *revs) |
|
2172 | static PyObject *find_deepest(indexObject *self, PyObject *revs) | |
2165 | { |
|
2173 | { | |
2166 | const Py_ssize_t revcount = PyList_GET_SIZE(revs); |
|
2174 | const Py_ssize_t revcount = PyList_GET_SIZE(revs); | |
2167 | static const Py_ssize_t capacity = 24; |
|
2175 | static const Py_ssize_t capacity = 24; | |
2168 | int *depth, *interesting = NULL; |
|
2176 | int *depth, *interesting = NULL; | |
2169 | int i, j, v, ninteresting; |
|
2177 | int i, j, v, ninteresting; | |
2170 | PyObject *dict = NULL, *keys = NULL; |
|
2178 | PyObject *dict = NULL, *keys = NULL; | |
2171 | long *seen = NULL; |
|
2179 | long *seen = NULL; | |
2172 | int maxrev = -1; |
|
2180 | int maxrev = -1; | |
2173 | long final; |
|
2181 | long final; | |
2174 |
|
2182 | |||
2175 | if (revcount > capacity) { |
|
2183 | if (revcount > capacity) { | |
2176 | PyErr_Format(PyExc_OverflowError, |
|
2184 | PyErr_Format(PyExc_OverflowError, | |
2177 | "bitset size (%ld) > capacity (%ld)", |
|
2185 | "bitset size (%ld) > capacity (%ld)", | |
2178 | (long)revcount, (long)capacity); |
|
2186 | (long)revcount, (long)capacity); | |
2179 | return NULL; |
|
2187 | return NULL; | |
2180 | } |
|
2188 | } | |
2181 |
|
2189 | |||
2182 | for (i = 0; i < revcount; i++) { |
|
2190 | for (i = 0; i < revcount; i++) { | |
2183 | int n = (int)PyInt_AsLong(PyList_GET_ITEM(revs, i)); |
|
2191 | int n = (int)PyInt_AsLong(PyList_GET_ITEM(revs, i)); | |
2184 | if (n > maxrev) |
|
2192 | if (n > maxrev) | |
2185 | maxrev = n; |
|
2193 | maxrev = n; | |
2186 | } |
|
2194 | } | |
2187 |
|
2195 | |||
2188 | depth = calloc(sizeof(*depth), maxrev + 1); |
|
2196 | depth = calloc(sizeof(*depth), maxrev + 1); | |
2189 | if (depth == NULL) |
|
2197 | if (depth == NULL) | |
2190 | return PyErr_NoMemory(); |
|
2198 | return PyErr_NoMemory(); | |
2191 |
|
2199 | |||
2192 | seen = calloc(sizeof(*seen), maxrev + 1); |
|
2200 | seen = calloc(sizeof(*seen), maxrev + 1); | |
2193 | if (seen == NULL) { |
|
2201 | if (seen == NULL) { | |
2194 | PyErr_NoMemory(); |
|
2202 | PyErr_NoMemory(); | |
2195 | goto bail; |
|
2203 | goto bail; | |
2196 | } |
|
2204 | } | |
2197 |
|
2205 | |||
2198 | interesting = calloc(sizeof(*interesting), ((size_t)1) << revcount); |
|
2206 | interesting = calloc(sizeof(*interesting), ((size_t)1) << revcount); | |
2199 | if (interesting == NULL) { |
|
2207 | if (interesting == NULL) { | |
2200 | PyErr_NoMemory(); |
|
2208 | PyErr_NoMemory(); | |
2201 | goto bail; |
|
2209 | goto bail; | |
2202 | } |
|
2210 | } | |
2203 |
|
2211 | |||
2204 | if (PyList_Sort(revs) == -1) |
|
2212 | if (PyList_Sort(revs) == -1) | |
2205 | goto bail; |
|
2213 | goto bail; | |
2206 |
|
2214 | |||
2207 | for (i = 0; i < revcount; i++) { |
|
2215 | for (i = 0; i < revcount; i++) { | |
2208 | int n = (int)PyInt_AsLong(PyList_GET_ITEM(revs, i)); |
|
2216 | int n = (int)PyInt_AsLong(PyList_GET_ITEM(revs, i)); | |
2209 | long b = 1l << i; |
|
2217 | long b = 1l << i; | |
2210 | depth[n] = 1; |
|
2218 | depth[n] = 1; | |
2211 | seen[n] = b; |
|
2219 | seen[n] = b; | |
2212 | interesting[b] = 1; |
|
2220 | interesting[b] = 1; | |
2213 | } |
|
2221 | } | |
2214 |
|
2222 | |||
2215 | /* invariant: ninteresting is the number of non-zero entries in |
|
2223 | /* invariant: ninteresting is the number of non-zero entries in | |
2216 | * interesting. */ |
|
2224 | * interesting. */ | |
2217 | ninteresting = (int)revcount; |
|
2225 | ninteresting = (int)revcount; | |
2218 |
|
2226 | |||
2219 | for (v = maxrev; v >= 0 && ninteresting > 1; v--) { |
|
2227 | for (v = maxrev; v >= 0 && ninteresting > 1; v--) { | |
2220 | int dv = depth[v]; |
|
2228 | int dv = depth[v]; | |
2221 | int parents[2]; |
|
2229 | int parents[2]; | |
2222 | long sv; |
|
2230 | long sv; | |
2223 |
|
2231 | |||
2224 | if (dv == 0) |
|
2232 | if (dv == 0) | |
2225 | continue; |
|
2233 | continue; | |
2226 |
|
2234 | |||
2227 | sv = seen[v]; |
|
2235 | sv = seen[v]; | |
2228 | if (index_get_parents(self, v, parents, maxrev) < 0) |
|
2236 | if (index_get_parents(self, v, parents, maxrev) < 0) | |
2229 | goto bail; |
|
2237 | goto bail; | |
2230 |
|
2238 | |||
2231 | for (i = 0; i < 2; i++) { |
|
2239 | for (i = 0; i < 2; i++) { | |
2232 | int p = parents[i]; |
|
2240 | int p = parents[i]; | |
2233 | long sp; |
|
2241 | long sp; | |
2234 | int dp; |
|
2242 | int dp; | |
2235 |
|
2243 | |||
2236 | if (p == -1) |
|
2244 | if (p == -1) | |
2237 | continue; |
|
2245 | continue; | |
2238 |
|
2246 | |||
2239 | dp = depth[p]; |
|
2247 | dp = depth[p]; | |
2240 | sp = seen[p]; |
|
2248 | sp = seen[p]; | |
2241 | if (dp <= dv) { |
|
2249 | if (dp <= dv) { | |
2242 | depth[p] = dv + 1; |
|
2250 | depth[p] = dv + 1; | |
2243 | if (sp != sv) { |
|
2251 | if (sp != sv) { | |
2244 | interesting[sv] += 1; |
|
2252 | interesting[sv] += 1; | |
2245 | seen[p] = sv; |
|
2253 | seen[p] = sv; | |
2246 | if (sp) { |
|
2254 | if (sp) { | |
2247 | interesting[sp] -= 1; |
|
2255 | interesting[sp] -= 1; | |
2248 | if (interesting[sp] == 0) |
|
2256 | if (interesting[sp] == 0) | |
2249 | ninteresting -= 1; |
|
2257 | ninteresting -= 1; | |
2250 | } |
|
2258 | } | |
2251 | } |
|
2259 | } | |
2252 | } else if (dv == dp - 1) { |
|
2260 | } else if (dv == dp - 1) { | |
2253 | long nsp = sp | sv; |
|
2261 | long nsp = sp | sv; | |
2254 | if (nsp == sp) |
|
2262 | if (nsp == sp) | |
2255 | continue; |
|
2263 | continue; | |
2256 | seen[p] = nsp; |
|
2264 | seen[p] = nsp; | |
2257 | interesting[sp] -= 1; |
|
2265 | interesting[sp] -= 1; | |
2258 | if (interesting[sp] == 0) |
|
2266 | if (interesting[sp] == 0) | |
2259 | ninteresting -= 1; |
|
2267 | ninteresting -= 1; | |
2260 | if (interesting[nsp] == 0) |
|
2268 | if (interesting[nsp] == 0) | |
2261 | ninteresting += 1; |
|
2269 | ninteresting += 1; | |
2262 | interesting[nsp] += 1; |
|
2270 | interesting[nsp] += 1; | |
2263 | } |
|
2271 | } | |
2264 | } |
|
2272 | } | |
2265 | interesting[sv] -= 1; |
|
2273 | interesting[sv] -= 1; | |
2266 | if (interesting[sv] == 0) |
|
2274 | if (interesting[sv] == 0) | |
2267 | ninteresting -= 1; |
|
2275 | ninteresting -= 1; | |
2268 | } |
|
2276 | } | |
2269 |
|
2277 | |||
2270 | final = 0; |
|
2278 | final = 0; | |
2271 | j = ninteresting; |
|
2279 | j = ninteresting; | |
2272 | for (i = 0; i < (int)(2 << revcount) && j > 0; i++) { |
|
2280 | for (i = 0; i < (int)(2 << revcount) && j > 0; i++) { | |
2273 | if (interesting[i] == 0) |
|
2281 | if (interesting[i] == 0) | |
2274 | continue; |
|
2282 | continue; | |
2275 | final |= i; |
|
2283 | final |= i; | |
2276 | j -= 1; |
|
2284 | j -= 1; | |
2277 | } |
|
2285 | } | |
2278 | if (final == 0) { |
|
2286 | if (final == 0) { | |
2279 | keys = PyList_New(0); |
|
2287 | keys = PyList_New(0); | |
2280 | goto bail; |
|
2288 | goto bail; | |
2281 | } |
|
2289 | } | |
2282 |
|
2290 | |||
2283 | dict = PyDict_New(); |
|
2291 | dict = PyDict_New(); | |
2284 | if (dict == NULL) |
|
2292 | if (dict == NULL) | |
2285 | goto bail; |
|
2293 | goto bail; | |
2286 |
|
2294 | |||
2287 | for (i = 0; i < revcount; i++) { |
|
2295 | for (i = 0; i < revcount; i++) { | |
2288 | PyObject *key; |
|
2296 | PyObject *key; | |
2289 |
|
2297 | |||
2290 | if ((final & (1 << i)) == 0) |
|
2298 | if ((final & (1 << i)) == 0) | |
2291 | continue; |
|
2299 | continue; | |
2292 |
|
2300 | |||
2293 | key = PyList_GET_ITEM(revs, i); |
|
2301 | key = PyList_GET_ITEM(revs, i); | |
2294 | Py_INCREF(key); |
|
2302 | Py_INCREF(key); | |
2295 | Py_INCREF(Py_None); |
|
2303 | Py_INCREF(Py_None); | |
2296 | if (PyDict_SetItem(dict, key, Py_None) == -1) { |
|
2304 | if (PyDict_SetItem(dict, key, Py_None) == -1) { | |
2297 | Py_DECREF(key); |
|
2305 | Py_DECREF(key); | |
2298 | Py_DECREF(Py_None); |
|
2306 | Py_DECREF(Py_None); | |
2299 | goto bail; |
|
2307 | goto bail; | |
2300 | } |
|
2308 | } | |
2301 | } |
|
2309 | } | |
2302 |
|
2310 | |||
2303 | keys = PyDict_Keys(dict); |
|
2311 | keys = PyDict_Keys(dict); | |
2304 |
|
2312 | |||
2305 | bail: |
|
2313 | bail: | |
2306 | free(depth); |
|
2314 | free(depth); | |
2307 | free(seen); |
|
2315 | free(seen); | |
2308 | free(interesting); |
|
2316 | free(interesting); | |
2309 | Py_XDECREF(dict); |
|
2317 | Py_XDECREF(dict); | |
2310 |
|
2318 | |||
2311 | return keys; |
|
2319 | return keys; | |
2312 | } |
|
2320 | } | |
2313 |
|
2321 | |||
2314 | /* |
|
2322 | /* | |
2315 | * Given a (possibly overlapping) set of revs, return all the |
|
2323 | * Given a (possibly overlapping) set of revs, return all the | |
2316 | * common ancestors heads: heads(::args[0] and ::a[1] and ...) |
|
2324 | * common ancestors heads: heads(::args[0] and ::a[1] and ...) | |
2317 | */ |
|
2325 | */ | |
2318 | static PyObject *index_commonancestorsheads(indexObject *self, PyObject *args) |
|
2326 | static PyObject *index_commonancestorsheads(indexObject *self, PyObject *args) | |
2319 | { |
|
2327 | { | |
2320 | PyObject *ret = NULL; |
|
2328 | PyObject *ret = NULL; | |
2321 | Py_ssize_t argcount, i, len; |
|
2329 | Py_ssize_t argcount, i, len; | |
2322 | bitmask repeat = 0; |
|
2330 | bitmask repeat = 0; | |
2323 | int revcount = 0; |
|
2331 | int revcount = 0; | |
2324 | int *revs; |
|
2332 | int *revs; | |
2325 |
|
2333 | |||
2326 | argcount = PySequence_Length(args); |
|
2334 | argcount = PySequence_Length(args); | |
2327 | revs = PyMem_Malloc(argcount * sizeof(*revs)); |
|
2335 | revs = PyMem_Malloc(argcount * sizeof(*revs)); | |
2328 | if (argcount > 0 && revs == NULL) |
|
2336 | if (argcount > 0 && revs == NULL) | |
2329 | return PyErr_NoMemory(); |
|
2337 | return PyErr_NoMemory(); | |
2330 | len = index_length(self); |
|
2338 | len = index_length(self); | |
2331 |
|
2339 | |||
2332 | for (i = 0; i < argcount; i++) { |
|
2340 | for (i = 0; i < argcount; i++) { | |
2333 | static const int capacity = 24; |
|
2341 | static const int capacity = 24; | |
2334 | PyObject *obj = PySequence_GetItem(args, i); |
|
2342 | PyObject *obj = PySequence_GetItem(args, i); | |
2335 | bitmask x; |
|
2343 | bitmask x; | |
2336 | long val; |
|
2344 | long val; | |
2337 |
|
2345 | |||
2338 | if (!PyInt_Check(obj)) { |
|
2346 | if (!PyInt_Check(obj)) { | |
2339 | PyErr_SetString(PyExc_TypeError, |
|
2347 | PyErr_SetString(PyExc_TypeError, | |
2340 | "arguments must all be ints"); |
|
2348 | "arguments must all be ints"); | |
2341 | Py_DECREF(obj); |
|
2349 | Py_DECREF(obj); | |
2342 | goto bail; |
|
2350 | goto bail; | |
2343 | } |
|
2351 | } | |
2344 | val = PyInt_AsLong(obj); |
|
2352 | val = PyInt_AsLong(obj); | |
2345 | Py_DECREF(obj); |
|
2353 | Py_DECREF(obj); | |
2346 | if (val == -1) { |
|
2354 | if (val == -1) { | |
2347 | ret = PyList_New(0); |
|
2355 | ret = PyList_New(0); | |
2348 | goto done; |
|
2356 | goto done; | |
2349 | } |
|
2357 | } | |
2350 | if (val < 0 || val >= len) { |
|
2358 | if (val < 0 || val >= len) { | |
2351 | PyErr_SetString(PyExc_IndexError, "index out of range"); |
|
2359 | PyErr_SetString(PyExc_IndexError, "index out of range"); | |
2352 | goto bail; |
|
2360 | goto bail; | |
2353 | } |
|
2361 | } | |
2354 | /* this cheesy bloom filter lets us avoid some more |
|
2362 | /* this cheesy bloom filter lets us avoid some more | |
2355 | * expensive duplicate checks in the common set-is-disjoint |
|
2363 | * expensive duplicate checks in the common set-is-disjoint | |
2356 | * case */ |
|
2364 | * case */ | |
2357 | x = 1ull << (val & 0x3f); |
|
2365 | x = 1ull << (val & 0x3f); | |
2358 | if (repeat & x) { |
|
2366 | if (repeat & x) { | |
2359 | int k; |
|
2367 | int k; | |
2360 | for (k = 0; k < revcount; k++) { |
|
2368 | for (k = 0; k < revcount; k++) { | |
2361 | if (val == revs[k]) |
|
2369 | if (val == revs[k]) | |
2362 | goto duplicate; |
|
2370 | goto duplicate; | |
2363 | } |
|
2371 | } | |
2364 | } else |
|
2372 | } else | |
2365 | repeat |= x; |
|
2373 | repeat |= x; | |
2366 | if (revcount >= capacity) { |
|
2374 | if (revcount >= capacity) { | |
2367 | PyErr_Format(PyExc_OverflowError, |
|
2375 | PyErr_Format(PyExc_OverflowError, | |
2368 | "bitset size (%d) > capacity (%d)", |
|
2376 | "bitset size (%d) > capacity (%d)", | |
2369 | revcount, capacity); |
|
2377 | revcount, capacity); | |
2370 | goto bail; |
|
2378 | goto bail; | |
2371 | } |
|
2379 | } | |
2372 | revs[revcount++] = (int)val; |
|
2380 | revs[revcount++] = (int)val; | |
2373 | duplicate:; |
|
2381 | duplicate:; | |
2374 | } |
|
2382 | } | |
2375 |
|
2383 | |||
2376 | if (revcount == 0) { |
|
2384 | if (revcount == 0) { | |
2377 | ret = PyList_New(0); |
|
2385 | ret = PyList_New(0); | |
2378 | goto done; |
|
2386 | goto done; | |
2379 | } |
|
2387 | } | |
2380 | if (revcount == 1) { |
|
2388 | if (revcount == 1) { | |
2381 | PyObject *obj; |
|
2389 | PyObject *obj; | |
2382 | ret = PyList_New(1); |
|
2390 | ret = PyList_New(1); | |
2383 | if (ret == NULL) |
|
2391 | if (ret == NULL) | |
2384 | goto bail; |
|
2392 | goto bail; | |
2385 | obj = PyInt_FromLong(revs[0]); |
|
2393 | obj = PyInt_FromLong(revs[0]); | |
2386 | if (obj == NULL) |
|
2394 | if (obj == NULL) | |
2387 | goto bail; |
|
2395 | goto bail; | |
2388 | PyList_SET_ITEM(ret, 0, obj); |
|
2396 | PyList_SET_ITEM(ret, 0, obj); | |
2389 | goto done; |
|
2397 | goto done; | |
2390 | } |
|
2398 | } | |
2391 |
|
2399 | |||
2392 | ret = find_gca_candidates(self, revs, revcount); |
|
2400 | ret = find_gca_candidates(self, revs, revcount); | |
2393 | if (ret == NULL) |
|
2401 | if (ret == NULL) | |
2394 | goto bail; |
|
2402 | goto bail; | |
2395 |
|
2403 | |||
2396 | done: |
|
2404 | done: | |
2397 | PyMem_Free(revs); |
|
2405 | PyMem_Free(revs); | |
2398 | return ret; |
|
2406 | return ret; | |
2399 |
|
2407 | |||
2400 | bail: |
|
2408 | bail: | |
2401 | PyMem_Free(revs); |
|
2409 | PyMem_Free(revs); | |
2402 | Py_XDECREF(ret); |
|
2410 | Py_XDECREF(ret); | |
2403 | return NULL; |
|
2411 | return NULL; | |
2404 | } |
|
2412 | } | |
2405 |
|
2413 | |||
2406 | /* |
|
2414 | /* | |
2407 | * Given a (possibly overlapping) set of revs, return the greatest |
|
2415 | * Given a (possibly overlapping) set of revs, return the greatest | |
2408 | * common ancestors: those with the longest path to the root. |
|
2416 | * common ancestors: those with the longest path to the root. | |
2409 | */ |
|
2417 | */ | |
2410 | static PyObject *index_ancestors(indexObject *self, PyObject *args) |
|
2418 | static PyObject *index_ancestors(indexObject *self, PyObject *args) | |
2411 | { |
|
2419 | { | |
2412 | PyObject *ret; |
|
2420 | PyObject *ret; | |
2413 | PyObject *gca = index_commonancestorsheads(self, args); |
|
2421 | PyObject *gca = index_commonancestorsheads(self, args); | |
2414 | if (gca == NULL) |
|
2422 | if (gca == NULL) | |
2415 | return NULL; |
|
2423 | return NULL; | |
2416 |
|
2424 | |||
2417 | if (PyList_GET_SIZE(gca) <= 1) { |
|
2425 | if (PyList_GET_SIZE(gca) <= 1) { | |
2418 | return gca; |
|
2426 | return gca; | |
2419 | } |
|
2427 | } | |
2420 |
|
2428 | |||
2421 | ret = find_deepest(self, gca); |
|
2429 | ret = find_deepest(self, gca); | |
2422 | Py_DECREF(gca); |
|
2430 | Py_DECREF(gca); | |
2423 | return ret; |
|
2431 | return ret; | |
2424 | } |
|
2432 | } | |
2425 |
|
2433 | |||
2426 | /* |
|
2434 | /* | |
2427 | * Invalidate any trie entries introduced by added revs. |
|
2435 | * Invalidate any trie entries introduced by added revs. | |
2428 | */ |
|
2436 | */ | |
2429 | static void index_invalidate_added(indexObject *self, Py_ssize_t start) |
|
2437 | static void index_invalidate_added(indexObject *self, Py_ssize_t start) | |
2430 | { |
|
2438 | { | |
2431 | Py_ssize_t i, len = PyList_GET_SIZE(self->added); |
|
2439 | Py_ssize_t i, len = PyList_GET_SIZE(self->added); | |
2432 |
|
2440 | |||
2433 | for (i = start; i < len; i++) { |
|
2441 | for (i = start; i < len; i++) { | |
2434 | PyObject *tuple = PyList_GET_ITEM(self->added, i); |
|
2442 | PyObject *tuple = PyList_GET_ITEM(self->added, i); | |
2435 | PyObject *node = PyTuple_GET_ITEM(tuple, 7); |
|
2443 | PyObject *node = PyTuple_GET_ITEM(tuple, 7); | |
2436 |
|
2444 | |||
2437 | nt_delete_node(&self->nt, PyBytes_AS_STRING(node)); |
|
2445 | nt_delete_node(&self->nt, PyBytes_AS_STRING(node)); | |
2438 | } |
|
2446 | } | |
2439 |
|
2447 | |||
2440 | if (start == 0) |
|
2448 | if (start == 0) | |
2441 | Py_CLEAR(self->added); |
|
2449 | Py_CLEAR(self->added); | |
2442 | } |
|
2450 | } | |
2443 |
|
2451 | |||
2444 | /* |
|
2452 | /* | |
2445 | * Delete a numeric range of revs, which must be at the end of the |
|
2453 | * Delete a numeric range of revs, which must be at the end of the | |
2446 | * range, but exclude the sentinel nullid entry. |
|
2454 | * range, but exclude the sentinel nullid entry. | |
2447 | */ |
|
2455 | */ | |
2448 | static int index_slice_del(indexObject *self, PyObject *item) |
|
2456 | static int index_slice_del(indexObject *self, PyObject *item) | |
2449 | { |
|
2457 | { | |
2450 | Py_ssize_t start, stop, step, slicelength; |
|
2458 | Py_ssize_t start, stop, step, slicelength; | |
2451 | Py_ssize_t length = index_length(self) + 1; |
|
2459 | Py_ssize_t length = index_length(self) + 1; | |
2452 | int ret = 0; |
|
2460 | int ret = 0; | |
2453 |
|
2461 | |||
2454 | /* Argument changed from PySliceObject* to PyObject* in Python 3. */ |
|
2462 | /* Argument changed from PySliceObject* to PyObject* in Python 3. */ | |
2455 | #ifdef IS_PY3K |
|
2463 | #ifdef IS_PY3K | |
2456 | if (PySlice_GetIndicesEx(item, length, &start, &stop, &step, |
|
2464 | if (PySlice_GetIndicesEx(item, length, &start, &stop, &step, | |
2457 | &slicelength) < 0) |
|
2465 | &slicelength) < 0) | |
2458 | #else |
|
2466 | #else | |
2459 | if (PySlice_GetIndicesEx((PySliceObject *)item, length, &start, &stop, |
|
2467 | if (PySlice_GetIndicesEx((PySliceObject *)item, length, &start, &stop, | |
2460 | &step, &slicelength) < 0) |
|
2468 | &step, &slicelength) < 0) | |
2461 | #endif |
|
2469 | #endif | |
2462 | return -1; |
|
2470 | return -1; | |
2463 |
|
2471 | |||
2464 | if (slicelength <= 0) |
|
2472 | if (slicelength <= 0) | |
2465 | return 0; |
|
2473 | return 0; | |
2466 |
|
2474 | |||
2467 | if ((step < 0 && start < stop) || (step > 0 && start > stop)) |
|
2475 | if ((step < 0 && start < stop) || (step > 0 && start > stop)) | |
2468 | stop = start; |
|
2476 | stop = start; | |
2469 |
|
2477 | |||
2470 | if (step < 0) { |
|
2478 | if (step < 0) { | |
2471 | stop = start + 1; |
|
2479 | stop = start + 1; | |
2472 | start = stop + step * (slicelength - 1) - 1; |
|
2480 | start = stop + step * (slicelength - 1) - 1; | |
2473 | step = -step; |
|
2481 | step = -step; | |
2474 | } |
|
2482 | } | |
2475 |
|
2483 | |||
2476 | if (step != 1) { |
|
2484 | if (step != 1) { | |
2477 | PyErr_SetString(PyExc_ValueError, |
|
2485 | PyErr_SetString(PyExc_ValueError, | |
2478 | "revlog index delete requires step size of 1"); |
|
2486 | "revlog index delete requires step size of 1"); | |
2479 | return -1; |
|
2487 | return -1; | |
2480 | } |
|
2488 | } | |
2481 |
|
2489 | |||
2482 | if (stop != length - 1) { |
|
2490 | if (stop != length - 1) { | |
2483 | PyErr_SetString(PyExc_IndexError, |
|
2491 | PyErr_SetString(PyExc_IndexError, | |
2484 | "revlog index deletion indices are invalid"); |
|
2492 | "revlog index deletion indices are invalid"); | |
2485 | return -1; |
|
2493 | return -1; | |
2486 | } |
|
2494 | } | |
2487 |
|
2495 | |||
2488 | if (start < self->length) { |
|
2496 | if (start < self->length) { | |
2489 | if (self->ntinitialized) { |
|
2497 | if (self->ntinitialized) { | |
2490 | Py_ssize_t i; |
|
2498 | Py_ssize_t i; | |
2491 |
|
2499 | |||
2492 | for (i = start; i < self->length; i++) { |
|
2500 | for (i = start; i < self->length; i++) { | |
2493 | const char *node = index_node_existing(self, i); |
|
2501 | const char *node = index_node_existing(self, i); | |
2494 | if (node == NULL) |
|
2502 | if (node == NULL) | |
2495 | return -1; |
|
2503 | return -1; | |
2496 |
|
2504 | |||
2497 | nt_delete_node(&self->nt, node); |
|
2505 | nt_delete_node(&self->nt, node); | |
2498 | } |
|
2506 | } | |
2499 | if (self->added) |
|
2507 | if (self->added) | |
2500 | index_invalidate_added(self, 0); |
|
2508 | index_invalidate_added(self, 0); | |
2501 | if (self->ntrev > start) |
|
2509 | if (self->ntrev > start) | |
2502 | self->ntrev = (int)start; |
|
2510 | self->ntrev = (int)start; | |
2503 | } |
|
2511 | } | |
2504 | self->length = start; |
|
2512 | self->length = start; | |
2505 | if (start < self->raw_length) { |
|
2513 | if (start < self->raw_length) { | |
2506 | if (self->cache) { |
|
2514 | if (self->cache) { | |
2507 | Py_ssize_t i; |
|
2515 | Py_ssize_t i; | |
2508 | for (i = start; i < self->raw_length; i++) |
|
2516 | for (i = start; i < self->raw_length; i++) | |
2509 | Py_CLEAR(self->cache[i]); |
|
2517 | Py_CLEAR(self->cache[i]); | |
2510 | } |
|
2518 | } | |
2511 | self->raw_length = start; |
|
2519 | self->raw_length = start; | |
2512 | } |
|
2520 | } | |
2513 | goto done; |
|
2521 | goto done; | |
2514 | } |
|
2522 | } | |
2515 |
|
2523 | |||
2516 | if (self->ntinitialized) { |
|
2524 | if (self->ntinitialized) { | |
2517 | index_invalidate_added(self, start - self->length); |
|
2525 | index_invalidate_added(self, start - self->length); | |
2518 | if (self->ntrev > start) |
|
2526 | if (self->ntrev > start) | |
2519 | self->ntrev = (int)start; |
|
2527 | self->ntrev = (int)start; | |
2520 | } |
|
2528 | } | |
2521 | if (self->added) |
|
2529 | if (self->added) | |
2522 | ret = PyList_SetSlice(self->added, start - self->length, |
|
2530 | ret = PyList_SetSlice(self->added, start - self->length, | |
2523 | PyList_GET_SIZE(self->added), NULL); |
|
2531 | PyList_GET_SIZE(self->added), NULL); | |
2524 | done: |
|
2532 | done: | |
2525 | Py_CLEAR(self->headrevs); |
|
2533 | Py_CLEAR(self->headrevs); | |
2526 | return ret; |
|
2534 | return ret; | |
2527 | } |
|
2535 | } | |
2528 |
|
2536 | |||
2529 | /* |
|
2537 | /* | |
2530 | * Supported ops: |
|
2538 | * Supported ops: | |
2531 | * |
|
2539 | * | |
2532 | * slice deletion |
|
2540 | * slice deletion | |
2533 | * string assignment (extend node->rev mapping) |
|
2541 | * string assignment (extend node->rev mapping) | |
2534 | * string deletion (shrink node->rev mapping) |
|
2542 | * string deletion (shrink node->rev mapping) | |
2535 | */ |
|
2543 | */ | |
2536 | static int index_assign_subscript(indexObject *self, PyObject *item, |
|
2544 | static int index_assign_subscript(indexObject *self, PyObject *item, | |
2537 | PyObject *value) |
|
2545 | PyObject *value) | |
2538 | { |
|
2546 | { | |
2539 | char *node; |
|
2547 | char *node; | |
2540 | long rev; |
|
2548 | long rev; | |
2541 |
|
2549 | |||
2542 | if (PySlice_Check(item) && value == NULL) |
|
2550 | if (PySlice_Check(item) && value == NULL) | |
2543 | return index_slice_del(self, item); |
|
2551 | return index_slice_del(self, item); | |
2544 |
|
2552 | |||
2545 | if (node_check(item, &node) == -1) |
|
2553 | if (node_check(item, &node) == -1) | |
2546 | return -1; |
|
2554 | return -1; | |
2547 |
|
2555 | |||
2548 | if (value == NULL) |
|
2556 | if (value == NULL) | |
2549 | return self->ntinitialized ? nt_delete_node(&self->nt, node) |
|
2557 | return self->ntinitialized ? nt_delete_node(&self->nt, node) | |
2550 | : 0; |
|
2558 | : 0; | |
2551 | rev = PyInt_AsLong(value); |
|
2559 | rev = PyInt_AsLong(value); | |
2552 | if (rev > INT_MAX || rev < 0) { |
|
2560 | if (rev > INT_MAX || rev < 0) { | |
2553 | if (!PyErr_Occurred()) |
|
2561 | if (!PyErr_Occurred()) | |
2554 | PyErr_SetString(PyExc_ValueError, "rev out of range"); |
|
2562 | PyErr_SetString(PyExc_ValueError, "rev out of range"); | |
2555 | return -1; |
|
2563 | return -1; | |
2556 | } |
|
2564 | } | |
2557 |
|
2565 | |||
2558 | if (index_init_nt(self) == -1) |
|
2566 | if (index_init_nt(self) == -1) | |
2559 | return -1; |
|
2567 | return -1; | |
2560 | return nt_insert(&self->nt, node, (int)rev); |
|
2568 | return nt_insert(&self->nt, node, (int)rev); | |
2561 | } |
|
2569 | } | |
2562 |
|
2570 | |||
2563 | /* |
|
2571 | /* | |
2564 | * Find all RevlogNG entries in an index that has inline data. Update |
|
2572 | * Find all RevlogNG entries in an index that has inline data. Update | |
2565 | * the optional "offsets" table with those entries. |
|
2573 | * the optional "offsets" table with those entries. | |
2566 | */ |
|
2574 | */ | |
2567 | static Py_ssize_t inline_scan(indexObject *self, const char **offsets) |
|
2575 | static Py_ssize_t inline_scan(indexObject *self, const char **offsets) | |
2568 | { |
|
2576 | { | |
2569 | const char *data = (const char *)self->buf.buf; |
|
2577 | const char *data = (const char *)self->buf.buf; | |
2570 | Py_ssize_t pos = 0; |
|
2578 | Py_ssize_t pos = 0; | |
2571 | Py_ssize_t end = self->buf.len; |
|
2579 | Py_ssize_t end = self->buf.len; | |
2572 | long incr = v1_hdrsize; |
|
2580 | long incr = v1_hdrsize; | |
2573 | Py_ssize_t len = 0; |
|
2581 | Py_ssize_t len = 0; | |
2574 |
|
2582 | |||
2575 | while (pos + v1_hdrsize <= end && pos >= 0) { |
|
2583 | while (pos + v1_hdrsize <= end && pos >= 0) { | |
2576 | uint32_t comp_len; |
|
2584 | uint32_t comp_len; | |
2577 | /* 3rd element of header is length of compressed inline data */ |
|
2585 | /* 3rd element of header is length of compressed inline data */ | |
2578 | comp_len = getbe32(data + pos + 8); |
|
2586 | comp_len = getbe32(data + pos + 8); | |
2579 | incr = v1_hdrsize + comp_len; |
|
2587 | incr = v1_hdrsize + comp_len; | |
2580 | if (offsets) |
|
2588 | if (offsets) | |
2581 | offsets[len] = data + pos; |
|
2589 | offsets[len] = data + pos; | |
2582 | len++; |
|
2590 | len++; | |
2583 | pos += incr; |
|
2591 | pos += incr; | |
2584 | } |
|
2592 | } | |
2585 |
|
2593 | |||
2586 | if (pos != end) { |
|
2594 | if (pos != end) { | |
2587 | if (!PyErr_Occurred()) |
|
2595 | if (!PyErr_Occurred()) | |
2588 | PyErr_SetString(PyExc_ValueError, "corrupt index file"); |
|
2596 | PyErr_SetString(PyExc_ValueError, "corrupt index file"); | |
2589 | return -1; |
|
2597 | return -1; | |
2590 | } |
|
2598 | } | |
2591 |
|
2599 | |||
2592 | return len; |
|
2600 | return len; | |
2593 | } |
|
2601 | } | |
2594 |
|
2602 | |||
2595 | static int index_init(indexObject *self, PyObject *args) |
|
2603 | static int index_init(indexObject *self, PyObject *args) | |
2596 | { |
|
2604 | { | |
2597 | PyObject *data_obj, *inlined_obj; |
|
2605 | PyObject *data_obj, *inlined_obj; | |
2598 | Py_ssize_t size; |
|
2606 | Py_ssize_t size; | |
2599 |
|
2607 | |||
2600 | /* Initialize before argument-checking to avoid index_dealloc() crash. |
|
2608 | /* Initialize before argument-checking to avoid index_dealloc() crash. | |
2601 | */ |
|
2609 | */ | |
2602 | self->raw_length = 0; |
|
2610 | self->raw_length = 0; | |
2603 | self->added = NULL; |
|
2611 | self->added = NULL; | |
2604 | self->cache = NULL; |
|
2612 | self->cache = NULL; | |
2605 | self->data = NULL; |
|
2613 | self->data = NULL; | |
2606 | memset(&self->buf, 0, sizeof(self->buf)); |
|
2614 | memset(&self->buf, 0, sizeof(self->buf)); | |
2607 | self->headrevs = NULL; |
|
2615 | self->headrevs = NULL; | |
2608 | self->filteredrevs = Py_None; |
|
2616 | self->filteredrevs = Py_None; | |
2609 | Py_INCREF(Py_None); |
|
2617 | Py_INCREF(Py_None); | |
2610 | self->ntinitialized = 0; |
|
2618 | self->ntinitialized = 0; | |
2611 | self->offsets = NULL; |
|
2619 | self->offsets = NULL; | |
2612 |
|
2620 | |||
2613 | if (!PyArg_ParseTuple(args, "OO", &data_obj, &inlined_obj)) |
|
2621 | if (!PyArg_ParseTuple(args, "OO", &data_obj, &inlined_obj)) | |
2614 | return -1; |
|
2622 | return -1; | |
2615 | if (!PyObject_CheckBuffer(data_obj)) { |
|
2623 | if (!PyObject_CheckBuffer(data_obj)) { | |
2616 | PyErr_SetString(PyExc_TypeError, |
|
2624 | PyErr_SetString(PyExc_TypeError, | |
2617 | "data does not support buffer interface"); |
|
2625 | "data does not support buffer interface"); | |
2618 | return -1; |
|
2626 | return -1; | |
2619 | } |
|
2627 | } | |
2620 |
|
2628 | |||
2621 | if (PyObject_GetBuffer(data_obj, &self->buf, PyBUF_SIMPLE) == -1) |
|
2629 | if (PyObject_GetBuffer(data_obj, &self->buf, PyBUF_SIMPLE) == -1) | |
2622 | return -1; |
|
2630 | return -1; | |
2623 | size = self->buf.len; |
|
2631 | size = self->buf.len; | |
2624 |
|
2632 | |||
2625 | self->inlined = inlined_obj && PyObject_IsTrue(inlined_obj); |
|
2633 | self->inlined = inlined_obj && PyObject_IsTrue(inlined_obj); | |
2626 | self->data = data_obj; |
|
2634 | self->data = data_obj; | |
2627 |
|
2635 | |||
2628 | self->ntlookups = self->ntmisses = 0; |
|
2636 | self->ntlookups = self->ntmisses = 0; | |
2629 | self->ntrev = -1; |
|
2637 | self->ntrev = -1; | |
2630 | Py_INCREF(self->data); |
|
2638 | Py_INCREF(self->data); | |
2631 |
|
2639 | |||
2632 | if (self->inlined) { |
|
2640 | if (self->inlined) { | |
2633 | Py_ssize_t len = inline_scan(self, NULL); |
|
2641 | Py_ssize_t len = inline_scan(self, NULL); | |
2634 | if (len == -1) |
|
2642 | if (len == -1) | |
2635 | goto bail; |
|
2643 | goto bail; | |
2636 | self->raw_length = len; |
|
2644 | self->raw_length = len; | |
2637 | self->length = len; |
|
2645 | self->length = len; | |
2638 | } else { |
|
2646 | } else { | |
2639 | if (size % v1_hdrsize) { |
|
2647 | if (size % v1_hdrsize) { | |
2640 | PyErr_SetString(PyExc_ValueError, "corrupt index file"); |
|
2648 | PyErr_SetString(PyExc_ValueError, "corrupt index file"); | |
2641 | goto bail; |
|
2649 | goto bail; | |
2642 | } |
|
2650 | } | |
2643 | self->raw_length = size / v1_hdrsize; |
|
2651 | self->raw_length = size / v1_hdrsize; | |
2644 | self->length = self->raw_length; |
|
2652 | self->length = self->raw_length; | |
2645 | } |
|
2653 | } | |
2646 |
|
2654 | |||
2647 | return 0; |
|
2655 | return 0; | |
2648 | bail: |
|
2656 | bail: | |
2649 | return -1; |
|
2657 | return -1; | |
2650 | } |
|
2658 | } | |
2651 |
|
2659 | |||
2652 | static PyObject *index_nodemap(indexObject *self) |
|
2660 | static PyObject *index_nodemap(indexObject *self) | |
2653 | { |
|
2661 | { | |
2654 | Py_INCREF(self); |
|
2662 | Py_INCREF(self); | |
2655 | return (PyObject *)self; |
|
2663 | return (PyObject *)self; | |
2656 | } |
|
2664 | } | |
2657 |
|
2665 | |||
2658 | static void _index_clearcaches(indexObject *self) |
|
2666 | static void _index_clearcaches(indexObject *self) | |
2659 | { |
|
2667 | { | |
2660 | if (self->cache) { |
|
2668 | if (self->cache) { | |
2661 | Py_ssize_t i; |
|
2669 | Py_ssize_t i; | |
2662 |
|
2670 | |||
2663 | for (i = 0; i < self->raw_length; i++) |
|
2671 | for (i = 0; i < self->raw_length; i++) | |
2664 | Py_CLEAR(self->cache[i]); |
|
2672 | Py_CLEAR(self->cache[i]); | |
2665 | free(self->cache); |
|
2673 | free(self->cache); | |
2666 | self->cache = NULL; |
|
2674 | self->cache = NULL; | |
2667 | } |
|
2675 | } | |
2668 | if (self->offsets) { |
|
2676 | if (self->offsets) { | |
2669 | PyMem_Free((void *)self->offsets); |
|
2677 | PyMem_Free((void *)self->offsets); | |
2670 | self->offsets = NULL; |
|
2678 | self->offsets = NULL; | |
2671 | } |
|
2679 | } | |
2672 | if (self->ntinitialized) { |
|
2680 | if (self->ntinitialized) { | |
2673 | nt_dealloc(&self->nt); |
|
2681 | nt_dealloc(&self->nt); | |
2674 | } |
|
2682 | } | |
2675 | self->ntinitialized = 0; |
|
2683 | self->ntinitialized = 0; | |
2676 | Py_CLEAR(self->headrevs); |
|
2684 | Py_CLEAR(self->headrevs); | |
2677 | } |
|
2685 | } | |
2678 |
|
2686 | |||
2679 | static PyObject *index_clearcaches(indexObject *self) |
|
2687 | static PyObject *index_clearcaches(indexObject *self) | |
2680 | { |
|
2688 | { | |
2681 | _index_clearcaches(self); |
|
2689 | _index_clearcaches(self); | |
2682 | self->ntrev = -1; |
|
2690 | self->ntrev = -1; | |
2683 | self->ntlookups = self->ntmisses = 0; |
|
2691 | self->ntlookups = self->ntmisses = 0; | |
2684 | Py_RETURN_NONE; |
|
2692 | Py_RETURN_NONE; | |
2685 | } |
|
2693 | } | |
2686 |
|
2694 | |||
2687 | static void index_dealloc(indexObject *self) |
|
2695 | static void index_dealloc(indexObject *self) | |
2688 | { |
|
2696 | { | |
2689 | _index_clearcaches(self); |
|
2697 | _index_clearcaches(self); | |
2690 | Py_XDECREF(self->filteredrevs); |
|
2698 | Py_XDECREF(self->filteredrevs); | |
2691 | if (self->buf.buf) { |
|
2699 | if (self->buf.buf) { | |
2692 | PyBuffer_Release(&self->buf); |
|
2700 | PyBuffer_Release(&self->buf); | |
2693 | memset(&self->buf, 0, sizeof(self->buf)); |
|
2701 | memset(&self->buf, 0, sizeof(self->buf)); | |
2694 | } |
|
2702 | } | |
2695 | Py_XDECREF(self->data); |
|
2703 | Py_XDECREF(self->data); | |
2696 | Py_XDECREF(self->added); |
|
2704 | Py_XDECREF(self->added); | |
2697 | PyObject_Del(self); |
|
2705 | PyObject_Del(self); | |
2698 | } |
|
2706 | } | |
2699 |
|
2707 | |||
2700 | static PySequenceMethods index_sequence_methods = { |
|
2708 | static PySequenceMethods index_sequence_methods = { | |
2701 | (lenfunc)index_length, /* sq_length */ |
|
2709 | (lenfunc)index_length, /* sq_length */ | |
2702 | 0, /* sq_concat */ |
|
2710 | 0, /* sq_concat */ | |
2703 | 0, /* sq_repeat */ |
|
2711 | 0, /* sq_repeat */ | |
2704 | (ssizeargfunc)index_get, /* sq_item */ |
|
2712 | (ssizeargfunc)index_get, /* sq_item */ | |
2705 | 0, /* sq_slice */ |
|
2713 | 0, /* sq_slice */ | |
2706 | 0, /* sq_ass_item */ |
|
2714 | 0, /* sq_ass_item */ | |
2707 | 0, /* sq_ass_slice */ |
|
2715 | 0, /* sq_ass_slice */ | |
2708 | (objobjproc)index_contains, /* sq_contains */ |
|
2716 | (objobjproc)index_contains, /* sq_contains */ | |
2709 | }; |
|
2717 | }; | |
2710 |
|
2718 | |||
2711 | static PyMappingMethods index_mapping_methods = { |
|
2719 | static PyMappingMethods index_mapping_methods = { | |
2712 | (lenfunc)index_length, /* mp_length */ |
|
2720 | (lenfunc)index_length, /* mp_length */ | |
2713 | (binaryfunc)index_getitem, /* mp_subscript */ |
|
2721 | (binaryfunc)index_getitem, /* mp_subscript */ | |
2714 | (objobjargproc)index_assign_subscript, /* mp_ass_subscript */ |
|
2722 | (objobjargproc)index_assign_subscript, /* mp_ass_subscript */ | |
2715 | }; |
|
2723 | }; | |
2716 |
|
2724 | |||
2717 | static PyMethodDef index_methods[] = { |
|
2725 | static PyMethodDef index_methods[] = { | |
2718 | {"ancestors", (PyCFunction)index_ancestors, METH_VARARGS, |
|
2726 | {"ancestors", (PyCFunction)index_ancestors, METH_VARARGS, | |
2719 | "return the gca set of the given revs"}, |
|
2727 | "return the gca set of the given revs"}, | |
2720 | {"commonancestorsheads", (PyCFunction)index_commonancestorsheads, |
|
2728 | {"commonancestorsheads", (PyCFunction)index_commonancestorsheads, | |
2721 | METH_VARARGS, |
|
2729 | METH_VARARGS, | |
2722 | "return the heads of the common ancestors of the given revs"}, |
|
2730 | "return the heads of the common ancestors of the given revs"}, | |
2723 | {"clearcaches", (PyCFunction)index_clearcaches, METH_NOARGS, |
|
2731 | {"clearcaches", (PyCFunction)index_clearcaches, METH_NOARGS, | |
2724 | "clear the index caches"}, |
|
2732 | "clear the index caches"}, | |
2725 | {"get", (PyCFunction)index_m_get, METH_VARARGS, "get an index entry"}, |
|
2733 | {"get", (PyCFunction)index_m_get, METH_VARARGS, "get an index entry"}, | |
|
2734 | {"has_node", (PyCFunction)index_m_has_node, METH_O, | |||
|
2735 | "return True if the node exist in the index"}, | |||
2726 | {"computephasesmapsets", (PyCFunction)compute_phases_map_sets, METH_VARARGS, |
|
2736 | {"computephasesmapsets", (PyCFunction)compute_phases_map_sets, METH_VARARGS, | |
2727 | "compute phases"}, |
|
2737 | "compute phases"}, | |
2728 | {"reachableroots2", (PyCFunction)reachableroots2, METH_VARARGS, |
|
2738 | {"reachableroots2", (PyCFunction)reachableroots2, METH_VARARGS, | |
2729 | "reachableroots"}, |
|
2739 | "reachableroots"}, | |
2730 | {"headrevs", (PyCFunction)index_headrevs, METH_VARARGS, |
|
2740 | {"headrevs", (PyCFunction)index_headrevs, METH_VARARGS, | |
2731 | "get head revisions"}, /* Can do filtering since 3.2 */ |
|
2741 | "get head revisions"}, /* Can do filtering since 3.2 */ | |
2732 | {"headrevsfiltered", (PyCFunction)index_headrevs, METH_VARARGS, |
|
2742 | {"headrevsfiltered", (PyCFunction)index_headrevs, METH_VARARGS, | |
2733 | "get filtered head revisions"}, /* Can always do filtering */ |
|
2743 | "get filtered head revisions"}, /* Can always do filtering */ | |
2734 | {"issnapshot", (PyCFunction)index_issnapshot, METH_O, |
|
2744 | {"issnapshot", (PyCFunction)index_issnapshot, METH_O, | |
2735 | "True if the object is a snapshot"}, |
|
2745 | "True if the object is a snapshot"}, | |
2736 | {"findsnapshots", (PyCFunction)index_findsnapshots, METH_VARARGS, |
|
2746 | {"findsnapshots", (PyCFunction)index_findsnapshots, METH_VARARGS, | |
2737 | "Gather snapshot data in a cache dict"}, |
|
2747 | "Gather snapshot data in a cache dict"}, | |
2738 | {"deltachain", (PyCFunction)index_deltachain, METH_VARARGS, |
|
2748 | {"deltachain", (PyCFunction)index_deltachain, METH_VARARGS, | |
2739 | "determine revisions with deltas to reconstruct fulltext"}, |
|
2749 | "determine revisions with deltas to reconstruct fulltext"}, | |
2740 | {"slicechunktodensity", (PyCFunction)index_slicechunktodensity, |
|
2750 | {"slicechunktodensity", (PyCFunction)index_slicechunktodensity, | |
2741 | METH_VARARGS, "determine revisions with deltas to reconstruct fulltext"}, |
|
2751 | METH_VARARGS, "determine revisions with deltas to reconstruct fulltext"}, | |
2742 | {"append", (PyCFunction)index_append, METH_O, "append an index entry"}, |
|
2752 | {"append", (PyCFunction)index_append, METH_O, "append an index entry"}, | |
2743 | {"partialmatch", (PyCFunction)index_partialmatch, METH_VARARGS, |
|
2753 | {"partialmatch", (PyCFunction)index_partialmatch, METH_VARARGS, | |
2744 | "match a potentially ambiguous node ID"}, |
|
2754 | "match a potentially ambiguous node ID"}, | |
2745 | {"shortest", (PyCFunction)index_shortest, METH_VARARGS, |
|
2755 | {"shortest", (PyCFunction)index_shortest, METH_VARARGS, | |
2746 | "find length of shortest hex nodeid of a binary ID"}, |
|
2756 | "find length of shortest hex nodeid of a binary ID"}, | |
2747 | {"stats", (PyCFunction)index_stats, METH_NOARGS, "stats for the index"}, |
|
2757 | {"stats", (PyCFunction)index_stats, METH_NOARGS, "stats for the index"}, | |
2748 | {NULL} /* Sentinel */ |
|
2758 | {NULL} /* Sentinel */ | |
2749 | }; |
|
2759 | }; | |
2750 |
|
2760 | |||
2751 | static PyGetSetDef index_getset[] = { |
|
2761 | static PyGetSetDef index_getset[] = { | |
2752 | {"nodemap", (getter)index_nodemap, NULL, "nodemap", NULL}, |
|
2762 | {"nodemap", (getter)index_nodemap, NULL, "nodemap", NULL}, | |
2753 | {NULL} /* Sentinel */ |
|
2763 | {NULL} /* Sentinel */ | |
2754 | }; |
|
2764 | }; | |
2755 |
|
2765 | |||
2756 | PyTypeObject HgRevlogIndex_Type = { |
|
2766 | PyTypeObject HgRevlogIndex_Type = { | |
2757 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ |
|
2767 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ | |
2758 | "parsers.index", /* tp_name */ |
|
2768 | "parsers.index", /* tp_name */ | |
2759 | sizeof(indexObject), /* tp_basicsize */ |
|
2769 | sizeof(indexObject), /* tp_basicsize */ | |
2760 | 0, /* tp_itemsize */ |
|
2770 | 0, /* tp_itemsize */ | |
2761 | (destructor)index_dealloc, /* tp_dealloc */ |
|
2771 | (destructor)index_dealloc, /* tp_dealloc */ | |
2762 | 0, /* tp_print */ |
|
2772 | 0, /* tp_print */ | |
2763 | 0, /* tp_getattr */ |
|
2773 | 0, /* tp_getattr */ | |
2764 | 0, /* tp_setattr */ |
|
2774 | 0, /* tp_setattr */ | |
2765 | 0, /* tp_compare */ |
|
2775 | 0, /* tp_compare */ | |
2766 | 0, /* tp_repr */ |
|
2776 | 0, /* tp_repr */ | |
2767 | 0, /* tp_as_number */ |
|
2777 | 0, /* tp_as_number */ | |
2768 | &index_sequence_methods, /* tp_as_sequence */ |
|
2778 | &index_sequence_methods, /* tp_as_sequence */ | |
2769 | &index_mapping_methods, /* tp_as_mapping */ |
|
2779 | &index_mapping_methods, /* tp_as_mapping */ | |
2770 | 0, /* tp_hash */ |
|
2780 | 0, /* tp_hash */ | |
2771 | 0, /* tp_call */ |
|
2781 | 0, /* tp_call */ | |
2772 | 0, /* tp_str */ |
|
2782 | 0, /* tp_str */ | |
2773 | 0, /* tp_getattro */ |
|
2783 | 0, /* tp_getattro */ | |
2774 | 0, /* tp_setattro */ |
|
2784 | 0, /* tp_setattro */ | |
2775 | 0, /* tp_as_buffer */ |
|
2785 | 0, /* tp_as_buffer */ | |
2776 | Py_TPFLAGS_DEFAULT, /* tp_flags */ |
|
2786 | Py_TPFLAGS_DEFAULT, /* tp_flags */ | |
2777 | "revlog index", /* tp_doc */ |
|
2787 | "revlog index", /* tp_doc */ | |
2778 | 0, /* tp_traverse */ |
|
2788 | 0, /* tp_traverse */ | |
2779 | 0, /* tp_clear */ |
|
2789 | 0, /* tp_clear */ | |
2780 | 0, /* tp_richcompare */ |
|
2790 | 0, /* tp_richcompare */ | |
2781 | 0, /* tp_weaklistoffset */ |
|
2791 | 0, /* tp_weaklistoffset */ | |
2782 | 0, /* tp_iter */ |
|
2792 | 0, /* tp_iter */ | |
2783 | 0, /* tp_iternext */ |
|
2793 | 0, /* tp_iternext */ | |
2784 | index_methods, /* tp_methods */ |
|
2794 | index_methods, /* tp_methods */ | |
2785 | 0, /* tp_members */ |
|
2795 | 0, /* tp_members */ | |
2786 | index_getset, /* tp_getset */ |
|
2796 | index_getset, /* tp_getset */ | |
2787 | 0, /* tp_base */ |
|
2797 | 0, /* tp_base */ | |
2788 | 0, /* tp_dict */ |
|
2798 | 0, /* tp_dict */ | |
2789 | 0, /* tp_descr_get */ |
|
2799 | 0, /* tp_descr_get */ | |
2790 | 0, /* tp_descr_set */ |
|
2800 | 0, /* tp_descr_set */ | |
2791 | 0, /* tp_dictoffset */ |
|
2801 | 0, /* tp_dictoffset */ | |
2792 | (initproc)index_init, /* tp_init */ |
|
2802 | (initproc)index_init, /* tp_init */ | |
2793 | 0, /* tp_alloc */ |
|
2803 | 0, /* tp_alloc */ | |
2794 | }; |
|
2804 | }; | |
2795 |
|
2805 | |||
2796 | /* |
|
2806 | /* | |
2797 | * returns a tuple of the form (index, index, cache) with elements as |
|
2807 | * returns a tuple of the form (index, index, cache) with elements as | |
2798 | * follows: |
|
2808 | * follows: | |
2799 | * |
|
2809 | * | |
2800 | * index: an index object that lazily parses RevlogNG records |
|
2810 | * index: an index object that lazily parses RevlogNG records | |
2801 | * cache: if data is inlined, a tuple (0, index_file_content), else None |
|
2811 | * cache: if data is inlined, a tuple (0, index_file_content), else None | |
2802 | * index_file_content could be a string, or a buffer |
|
2812 | * index_file_content could be a string, or a buffer | |
2803 | * |
|
2813 | * | |
2804 | * added complications are for backwards compatibility |
|
2814 | * added complications are for backwards compatibility | |
2805 | */ |
|
2815 | */ | |
2806 | PyObject *parse_index2(PyObject *self, PyObject *args) |
|
2816 | PyObject *parse_index2(PyObject *self, PyObject *args) | |
2807 | { |
|
2817 | { | |
2808 | PyObject *tuple = NULL, *cache = NULL; |
|
2818 | PyObject *tuple = NULL, *cache = NULL; | |
2809 | indexObject *idx; |
|
2819 | indexObject *idx; | |
2810 | int ret; |
|
2820 | int ret; | |
2811 |
|
2821 | |||
2812 | idx = PyObject_New(indexObject, &HgRevlogIndex_Type); |
|
2822 | idx = PyObject_New(indexObject, &HgRevlogIndex_Type); | |
2813 | if (idx == NULL) |
|
2823 | if (idx == NULL) | |
2814 | goto bail; |
|
2824 | goto bail; | |
2815 |
|
2825 | |||
2816 | ret = index_init(idx, args); |
|
2826 | ret = index_init(idx, args); | |
2817 | if (ret == -1) |
|
2827 | if (ret == -1) | |
2818 | goto bail; |
|
2828 | goto bail; | |
2819 |
|
2829 | |||
2820 | if (idx->inlined) { |
|
2830 | if (idx->inlined) { | |
2821 | cache = Py_BuildValue("iO", 0, idx->data); |
|
2831 | cache = Py_BuildValue("iO", 0, idx->data); | |
2822 | if (cache == NULL) |
|
2832 | if (cache == NULL) | |
2823 | goto bail; |
|
2833 | goto bail; | |
2824 | } else { |
|
2834 | } else { | |
2825 | cache = Py_None; |
|
2835 | cache = Py_None; | |
2826 | Py_INCREF(cache); |
|
2836 | Py_INCREF(cache); | |
2827 | } |
|
2837 | } | |
2828 |
|
2838 | |||
2829 | tuple = Py_BuildValue("NN", idx, cache); |
|
2839 | tuple = Py_BuildValue("NN", idx, cache); | |
2830 | if (!tuple) |
|
2840 | if (!tuple) | |
2831 | goto bail; |
|
2841 | goto bail; | |
2832 | return tuple; |
|
2842 | return tuple; | |
2833 |
|
2843 | |||
2834 | bail: |
|
2844 | bail: | |
2835 | Py_XDECREF(idx); |
|
2845 | Py_XDECREF(idx); | |
2836 | Py_XDECREF(cache); |
|
2846 | Py_XDECREF(cache); | |
2837 | Py_XDECREF(tuple); |
|
2847 | Py_XDECREF(tuple); | |
2838 | return NULL; |
|
2848 | return NULL; | |
2839 | } |
|
2849 | } | |
2840 |
|
2850 | |||
2841 | #ifdef WITH_RUST |
|
2851 | #ifdef WITH_RUST | |
2842 |
|
2852 | |||
2843 | /* rustlazyancestors: iteration over ancestors implemented in Rust |
|
2853 | /* rustlazyancestors: iteration over ancestors implemented in Rust | |
2844 | * |
|
2854 | * | |
2845 | * This class holds a reference to an index and to the Rust iterator. |
|
2855 | * This class holds a reference to an index and to the Rust iterator. | |
2846 | */ |
|
2856 | */ | |
2847 | typedef struct rustlazyancestorsObjectStruct rustlazyancestorsObject; |
|
2857 | typedef struct rustlazyancestorsObjectStruct rustlazyancestorsObject; | |
2848 |
|
2858 | |||
2849 | struct rustlazyancestorsObjectStruct { |
|
2859 | struct rustlazyancestorsObjectStruct { | |
2850 | PyObject_HEAD |
|
2860 | PyObject_HEAD | |
2851 | /* Type-specific fields go here. */ |
|
2861 | /* Type-specific fields go here. */ | |
2852 | indexObject *index; /* Ref kept to avoid GC'ing the index */ |
|
2862 | indexObject *index; /* Ref kept to avoid GC'ing the index */ | |
2853 | void *iter; /* Rust iterator */ |
|
2863 | void *iter; /* Rust iterator */ | |
2854 | }; |
|
2864 | }; | |
2855 |
|
2865 | |||
2856 | /* FFI exposed from Rust code */ |
|
2866 | /* FFI exposed from Rust code */ | |
2857 | rustlazyancestorsObject *rustlazyancestors_init(indexObject *index, |
|
2867 | rustlazyancestorsObject *rustlazyancestors_init(indexObject *index, | |
2858 | /* intrevs vector */ |
|
2868 | /* intrevs vector */ | |
2859 | Py_ssize_t initrevslen, |
|
2869 | Py_ssize_t initrevslen, | |
2860 | long *initrevs, long stoprev, |
|
2870 | long *initrevs, long stoprev, | |
2861 | int inclusive); |
|
2871 | int inclusive); | |
2862 | void rustlazyancestors_drop(rustlazyancestorsObject *self); |
|
2872 | void rustlazyancestors_drop(rustlazyancestorsObject *self); | |
2863 | int rustlazyancestors_next(rustlazyancestorsObject *self); |
|
2873 | int rustlazyancestors_next(rustlazyancestorsObject *self); | |
2864 | int rustlazyancestors_contains(rustlazyancestorsObject *self, long rev); |
|
2874 | int rustlazyancestors_contains(rustlazyancestorsObject *self, long rev); | |
2865 |
|
2875 | |||
2866 | /* CPython instance methods */ |
|
2876 | /* CPython instance methods */ | |
2867 | static int rustla_init(rustlazyancestorsObject *self, PyObject *args) |
|
2877 | static int rustla_init(rustlazyancestorsObject *self, PyObject *args) | |
2868 | { |
|
2878 | { | |
2869 | PyObject *initrevsarg = NULL; |
|
2879 | PyObject *initrevsarg = NULL; | |
2870 | PyObject *inclusivearg = NULL; |
|
2880 | PyObject *inclusivearg = NULL; | |
2871 | long stoprev = 0; |
|
2881 | long stoprev = 0; | |
2872 | long *initrevs = NULL; |
|
2882 | long *initrevs = NULL; | |
2873 | int inclusive = 0; |
|
2883 | int inclusive = 0; | |
2874 | Py_ssize_t i; |
|
2884 | Py_ssize_t i; | |
2875 |
|
2885 | |||
2876 | indexObject *index; |
|
2886 | indexObject *index; | |
2877 | if (!PyArg_ParseTuple(args, "O!O!lO!", &HgRevlogIndex_Type, &index, |
|
2887 | if (!PyArg_ParseTuple(args, "O!O!lO!", &HgRevlogIndex_Type, &index, | |
2878 | &PyList_Type, &initrevsarg, &stoprev, |
|
2888 | &PyList_Type, &initrevsarg, &stoprev, | |
2879 | &PyBool_Type, &inclusivearg)) |
|
2889 | &PyBool_Type, &inclusivearg)) | |
2880 | return -1; |
|
2890 | return -1; | |
2881 |
|
2891 | |||
2882 | Py_INCREF(index); |
|
2892 | Py_INCREF(index); | |
2883 | self->index = index; |
|
2893 | self->index = index; | |
2884 |
|
2894 | |||
2885 | if (inclusivearg == Py_True) |
|
2895 | if (inclusivearg == Py_True) | |
2886 | inclusive = 1; |
|
2896 | inclusive = 1; | |
2887 |
|
2897 | |||
2888 | Py_ssize_t linit = PyList_GET_SIZE(initrevsarg); |
|
2898 | Py_ssize_t linit = PyList_GET_SIZE(initrevsarg); | |
2889 |
|
2899 | |||
2890 | initrevs = (long *)calloc(linit, sizeof(long)); |
|
2900 | initrevs = (long *)calloc(linit, sizeof(long)); | |
2891 |
|
2901 | |||
2892 | if (initrevs == NULL) { |
|
2902 | if (initrevs == NULL) { | |
2893 | PyErr_NoMemory(); |
|
2903 | PyErr_NoMemory(); | |
2894 | goto bail; |
|
2904 | goto bail; | |
2895 | } |
|
2905 | } | |
2896 |
|
2906 | |||
2897 | for (i = 0; i < linit; i++) { |
|
2907 | for (i = 0; i < linit; i++) { | |
2898 | initrevs[i] = PyInt_AsLong(PyList_GET_ITEM(initrevsarg, i)); |
|
2908 | initrevs[i] = PyInt_AsLong(PyList_GET_ITEM(initrevsarg, i)); | |
2899 | } |
|
2909 | } | |
2900 | if (PyErr_Occurred()) |
|
2910 | if (PyErr_Occurred()) | |
2901 | goto bail; |
|
2911 | goto bail; | |
2902 |
|
2912 | |||
2903 | self->iter = |
|
2913 | self->iter = | |
2904 | rustlazyancestors_init(index, linit, initrevs, stoprev, inclusive); |
|
2914 | rustlazyancestors_init(index, linit, initrevs, stoprev, inclusive); | |
2905 | if (self->iter == NULL) { |
|
2915 | if (self->iter == NULL) { | |
2906 | /* if this is because of GraphError::ParentOutOfRange |
|
2916 | /* if this is because of GraphError::ParentOutOfRange | |
2907 | * HgRevlogIndex_GetParents() has already set the proper |
|
2917 | * HgRevlogIndex_GetParents() has already set the proper | |
2908 | * exception */ |
|
2918 | * exception */ | |
2909 | goto bail; |
|
2919 | goto bail; | |
2910 | } |
|
2920 | } | |
2911 |
|
2921 | |||
2912 | free(initrevs); |
|
2922 | free(initrevs); | |
2913 | return 0; |
|
2923 | return 0; | |
2914 |
|
2924 | |||
2915 | bail: |
|
2925 | bail: | |
2916 | free(initrevs); |
|
2926 | free(initrevs); | |
2917 | return -1; |
|
2927 | return -1; | |
2918 | }; |
|
2928 | }; | |
2919 |
|
2929 | |||
2920 | static void rustla_dealloc(rustlazyancestorsObject *self) |
|
2930 | static void rustla_dealloc(rustlazyancestorsObject *self) | |
2921 | { |
|
2931 | { | |
2922 | Py_XDECREF(self->index); |
|
2932 | Py_XDECREF(self->index); | |
2923 | if (self->iter != NULL) { /* can happen if rustla_init failed */ |
|
2933 | if (self->iter != NULL) { /* can happen if rustla_init failed */ | |
2924 | rustlazyancestors_drop(self->iter); |
|
2934 | rustlazyancestors_drop(self->iter); | |
2925 | } |
|
2935 | } | |
2926 | PyObject_Del(self); |
|
2936 | PyObject_Del(self); | |
2927 | } |
|
2937 | } | |
2928 |
|
2938 | |||
2929 | static PyObject *rustla_next(rustlazyancestorsObject *self) |
|
2939 | static PyObject *rustla_next(rustlazyancestorsObject *self) | |
2930 | { |
|
2940 | { | |
2931 | int res = rustlazyancestors_next(self->iter); |
|
2941 | int res = rustlazyancestors_next(self->iter); | |
2932 | if (res == -1) { |
|
2942 | if (res == -1) { | |
2933 | /* Setting an explicit exception seems unnecessary |
|
2943 | /* Setting an explicit exception seems unnecessary | |
2934 | * as examples from Python source code (Objects/rangeobjets.c |
|
2944 | * as examples from Python source code (Objects/rangeobjets.c | |
2935 | * and Modules/_io/stringio.c) seem to demonstrate. |
|
2945 | * and Modules/_io/stringio.c) seem to demonstrate. | |
2936 | */ |
|
2946 | */ | |
2937 | return NULL; |
|
2947 | return NULL; | |
2938 | } |
|
2948 | } | |
2939 | return PyInt_FromLong(res); |
|
2949 | return PyInt_FromLong(res); | |
2940 | } |
|
2950 | } | |
2941 |
|
2951 | |||
2942 | static int rustla_contains(rustlazyancestorsObject *self, PyObject *rev) |
|
2952 | static int rustla_contains(rustlazyancestorsObject *self, PyObject *rev) | |
2943 | { |
|
2953 | { | |
2944 | long lrev; |
|
2954 | long lrev; | |
2945 | if (!pylong_to_long(rev, &lrev)) { |
|
2955 | if (!pylong_to_long(rev, &lrev)) { | |
2946 | PyErr_Clear(); |
|
2956 | PyErr_Clear(); | |
2947 | return 0; |
|
2957 | return 0; | |
2948 | } |
|
2958 | } | |
2949 | return rustlazyancestors_contains(self->iter, lrev); |
|
2959 | return rustlazyancestors_contains(self->iter, lrev); | |
2950 | } |
|
2960 | } | |
2951 |
|
2961 | |||
2952 | static PySequenceMethods rustla_sequence_methods = { |
|
2962 | static PySequenceMethods rustla_sequence_methods = { | |
2953 | 0, /* sq_length */ |
|
2963 | 0, /* sq_length */ | |
2954 | 0, /* sq_concat */ |
|
2964 | 0, /* sq_concat */ | |
2955 | 0, /* sq_repeat */ |
|
2965 | 0, /* sq_repeat */ | |
2956 | 0, /* sq_item */ |
|
2966 | 0, /* sq_item */ | |
2957 | 0, /* sq_slice */ |
|
2967 | 0, /* sq_slice */ | |
2958 | 0, /* sq_ass_item */ |
|
2968 | 0, /* sq_ass_item */ | |
2959 | 0, /* sq_ass_slice */ |
|
2969 | 0, /* sq_ass_slice */ | |
2960 | (objobjproc)rustla_contains, /* sq_contains */ |
|
2970 | (objobjproc)rustla_contains, /* sq_contains */ | |
2961 | }; |
|
2971 | }; | |
2962 |
|
2972 | |||
2963 | static PyTypeObject rustlazyancestorsType = { |
|
2973 | static PyTypeObject rustlazyancestorsType = { | |
2964 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ |
|
2974 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ | |
2965 | "parsers.rustlazyancestors", /* tp_name */ |
|
2975 | "parsers.rustlazyancestors", /* tp_name */ | |
2966 | sizeof(rustlazyancestorsObject), /* tp_basicsize */ |
|
2976 | sizeof(rustlazyancestorsObject), /* tp_basicsize */ | |
2967 | 0, /* tp_itemsize */ |
|
2977 | 0, /* tp_itemsize */ | |
2968 | (destructor)rustla_dealloc, /* tp_dealloc */ |
|
2978 | (destructor)rustla_dealloc, /* tp_dealloc */ | |
2969 | 0, /* tp_print */ |
|
2979 | 0, /* tp_print */ | |
2970 | 0, /* tp_getattr */ |
|
2980 | 0, /* tp_getattr */ | |
2971 | 0, /* tp_setattr */ |
|
2981 | 0, /* tp_setattr */ | |
2972 | 0, /* tp_compare */ |
|
2982 | 0, /* tp_compare */ | |
2973 | 0, /* tp_repr */ |
|
2983 | 0, /* tp_repr */ | |
2974 | 0, /* tp_as_number */ |
|
2984 | 0, /* tp_as_number */ | |
2975 | &rustla_sequence_methods, /* tp_as_sequence */ |
|
2985 | &rustla_sequence_methods, /* tp_as_sequence */ | |
2976 | 0, /* tp_as_mapping */ |
|
2986 | 0, /* tp_as_mapping */ | |
2977 | 0, /* tp_hash */ |
|
2987 | 0, /* tp_hash */ | |
2978 | 0, /* tp_call */ |
|
2988 | 0, /* tp_call */ | |
2979 | 0, /* tp_str */ |
|
2989 | 0, /* tp_str */ | |
2980 | 0, /* tp_getattro */ |
|
2990 | 0, /* tp_getattro */ | |
2981 | 0, /* tp_setattro */ |
|
2991 | 0, /* tp_setattro */ | |
2982 | 0, /* tp_as_buffer */ |
|
2992 | 0, /* tp_as_buffer */ | |
2983 | Py_TPFLAGS_DEFAULT, /* tp_flags */ |
|
2993 | Py_TPFLAGS_DEFAULT, /* tp_flags */ | |
2984 | "Iterator over ancestors, implemented in Rust", /* tp_doc */ |
|
2994 | "Iterator over ancestors, implemented in Rust", /* tp_doc */ | |
2985 | 0, /* tp_traverse */ |
|
2995 | 0, /* tp_traverse */ | |
2986 | 0, /* tp_clear */ |
|
2996 | 0, /* tp_clear */ | |
2987 | 0, /* tp_richcompare */ |
|
2997 | 0, /* tp_richcompare */ | |
2988 | 0, /* tp_weaklistoffset */ |
|
2998 | 0, /* tp_weaklistoffset */ | |
2989 | 0, /* tp_iter */ |
|
2999 | 0, /* tp_iter */ | |
2990 | (iternextfunc)rustla_next, /* tp_iternext */ |
|
3000 | (iternextfunc)rustla_next, /* tp_iternext */ | |
2991 | 0, /* tp_methods */ |
|
3001 | 0, /* tp_methods */ | |
2992 | 0, /* tp_members */ |
|
3002 | 0, /* tp_members */ | |
2993 | 0, /* tp_getset */ |
|
3003 | 0, /* tp_getset */ | |
2994 | 0, /* tp_base */ |
|
3004 | 0, /* tp_base */ | |
2995 | 0, /* tp_dict */ |
|
3005 | 0, /* tp_dict */ | |
2996 | 0, /* tp_descr_get */ |
|
3006 | 0, /* tp_descr_get */ | |
2997 | 0, /* tp_descr_set */ |
|
3007 | 0, /* tp_descr_set */ | |
2998 | 0, /* tp_dictoffset */ |
|
3008 | 0, /* tp_dictoffset */ | |
2999 | (initproc)rustla_init, /* tp_init */ |
|
3009 | (initproc)rustla_init, /* tp_init */ | |
3000 | 0, /* tp_alloc */ |
|
3010 | 0, /* tp_alloc */ | |
3001 | }; |
|
3011 | }; | |
3002 | #endif /* WITH_RUST */ |
|
3012 | #endif /* WITH_RUST */ | |
3003 |
|
3013 | |||
3004 | void revlog_module_init(PyObject *mod) |
|
3014 | void revlog_module_init(PyObject *mod) | |
3005 | { |
|
3015 | { | |
3006 | PyObject *caps = NULL; |
|
3016 | PyObject *caps = NULL; | |
3007 | HgRevlogIndex_Type.tp_new = PyType_GenericNew; |
|
3017 | HgRevlogIndex_Type.tp_new = PyType_GenericNew; | |
3008 | if (PyType_Ready(&HgRevlogIndex_Type) < 0) |
|
3018 | if (PyType_Ready(&HgRevlogIndex_Type) < 0) | |
3009 | return; |
|
3019 | return; | |
3010 | Py_INCREF(&HgRevlogIndex_Type); |
|
3020 | Py_INCREF(&HgRevlogIndex_Type); | |
3011 | PyModule_AddObject(mod, "index", (PyObject *)&HgRevlogIndex_Type); |
|
3021 | PyModule_AddObject(mod, "index", (PyObject *)&HgRevlogIndex_Type); | |
3012 |
|
3022 | |||
3013 | nodetreeType.tp_new = PyType_GenericNew; |
|
3023 | nodetreeType.tp_new = PyType_GenericNew; | |
3014 | if (PyType_Ready(&nodetreeType) < 0) |
|
3024 | if (PyType_Ready(&nodetreeType) < 0) | |
3015 | return; |
|
3025 | return; | |
3016 | Py_INCREF(&nodetreeType); |
|
3026 | Py_INCREF(&nodetreeType); | |
3017 | PyModule_AddObject(mod, "nodetree", (PyObject *)&nodetreeType); |
|
3027 | PyModule_AddObject(mod, "nodetree", (PyObject *)&nodetreeType); | |
3018 |
|
3028 | |||
3019 | if (!nullentry) { |
|
3029 | if (!nullentry) { | |
3020 | nullentry = |
|
3030 | nullentry = | |
3021 | Py_BuildValue(PY23("iiiiiiis#", "iiiiiiiy#"), 0, 0, 0, -1, |
|
3031 | Py_BuildValue(PY23("iiiiiiis#", "iiiiiiiy#"), 0, 0, 0, -1, | |
3022 | -1, -1, -1, nullid, (Py_ssize_t)20); |
|
3032 | -1, -1, -1, nullid, (Py_ssize_t)20); | |
3023 | } |
|
3033 | } | |
3024 | if (nullentry) |
|
3034 | if (nullentry) | |
3025 | PyObject_GC_UnTrack(nullentry); |
|
3035 | PyObject_GC_UnTrack(nullentry); | |
3026 |
|
3036 | |||
3027 | caps = PyCapsule_New(HgRevlogIndex_GetParents, |
|
3037 | caps = PyCapsule_New(HgRevlogIndex_GetParents, | |
3028 | "mercurial.cext.parsers.index_get_parents_CAPI", |
|
3038 | "mercurial.cext.parsers.index_get_parents_CAPI", | |
3029 | NULL); |
|
3039 | NULL); | |
3030 | if (caps != NULL) |
|
3040 | if (caps != NULL) | |
3031 | PyModule_AddObject(mod, "index_get_parents_CAPI", caps); |
|
3041 | PyModule_AddObject(mod, "index_get_parents_CAPI", caps); | |
3032 |
|
3042 | |||
3033 | #ifdef WITH_RUST |
|
3043 | #ifdef WITH_RUST | |
3034 | rustlazyancestorsType.tp_new = PyType_GenericNew; |
|
3044 | rustlazyancestorsType.tp_new = PyType_GenericNew; | |
3035 | if (PyType_Ready(&rustlazyancestorsType) < 0) |
|
3045 | if (PyType_Ready(&rustlazyancestorsType) < 0) | |
3036 | return; |
|
3046 | return; | |
3037 | Py_INCREF(&rustlazyancestorsType); |
|
3047 | Py_INCREF(&rustlazyancestorsType); | |
3038 | PyModule_AddObject(mod, "rustlazyancestors", |
|
3048 | PyModule_AddObject(mod, "rustlazyancestors", | |
3039 | (PyObject *)&rustlazyancestorsType); |
|
3049 | (PyObject *)&rustlazyancestorsType); | |
3040 | #endif |
|
3050 | #endif | |
3041 | } |
|
3051 | } |
@@ -1,157 +1,157 b'' | |||||
1 | # policy.py - module policy logic for Mercurial. |
|
1 | # policy.py - module policy logic for Mercurial. | |
2 | # |
|
2 | # | |
3 | # Copyright 2015 Gregory Szorc <gregory.szorc@gmail.com> |
|
3 | # Copyright 2015 Gregory Szorc <gregory.szorc@gmail.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import os |
|
10 | import os | |
11 | import sys |
|
11 | import sys | |
12 |
|
12 | |||
13 | from .pycompat import getattr |
|
13 | from .pycompat import getattr | |
14 |
|
14 | |||
15 | # Rules for how modules can be loaded. Values are: |
|
15 | # Rules for how modules can be loaded. Values are: | |
16 | # |
|
16 | # | |
17 | # c - require C extensions |
|
17 | # c - require C extensions | |
18 | # rust+c - require Rust and C extensions |
|
18 | # rust+c - require Rust and C extensions | |
19 | # rust+c-allow - allow Rust and C extensions with fallback to pure Python |
|
19 | # rust+c-allow - allow Rust and C extensions with fallback to pure Python | |
20 | # for each |
|
20 | # for each | |
21 | # allow - allow pure Python implementation when C loading fails |
|
21 | # allow - allow pure Python implementation when C loading fails | |
22 | # cffi - required cffi versions (implemented within pure module) |
|
22 | # cffi - required cffi versions (implemented within pure module) | |
23 | # cffi-allow - allow pure Python implementation if cffi version is missing |
|
23 | # cffi-allow - allow pure Python implementation if cffi version is missing | |
24 | # py - only load pure Python modules |
|
24 | # py - only load pure Python modules | |
25 | # |
|
25 | # | |
26 | # By default, fall back to the pure modules so the in-place build can |
|
26 | # By default, fall back to the pure modules so the in-place build can | |
27 | # run without recompiling the C extensions. This will be overridden by |
|
27 | # run without recompiling the C extensions. This will be overridden by | |
28 | # __modulepolicy__ generated by setup.py. |
|
28 | # __modulepolicy__ generated by setup.py. | |
29 | policy = b'allow' |
|
29 | policy = b'allow' | |
30 | _packageprefs = { |
|
30 | _packageprefs = { | |
31 | # policy: (versioned package, pure package) |
|
31 | # policy: (versioned package, pure package) | |
32 | b'c': ('cext', None), |
|
32 | b'c': ('cext', None), | |
33 | b'allow': ('cext', 'pure'), |
|
33 | b'allow': ('cext', 'pure'), | |
34 | b'cffi': ('cffi', None), |
|
34 | b'cffi': ('cffi', None), | |
35 | b'cffi-allow': ('cffi', 'pure'), |
|
35 | b'cffi-allow': ('cffi', 'pure'), | |
36 | b'py': (None, 'pure'), |
|
36 | b'py': (None, 'pure'), | |
37 | # For now, rust policies impact importrust only |
|
37 | # For now, rust policies impact importrust only | |
38 | b'rust+c': ('cext', None), |
|
38 | b'rust+c': ('cext', None), | |
39 | b'rust+c-allow': ('cext', 'pure'), |
|
39 | b'rust+c-allow': ('cext', 'pure'), | |
40 | } |
|
40 | } | |
41 |
|
41 | |||
42 | try: |
|
42 | try: | |
43 | from . import __modulepolicy__ |
|
43 | from . import __modulepolicy__ | |
44 |
|
44 | |||
45 | policy = __modulepolicy__.modulepolicy |
|
45 | policy = __modulepolicy__.modulepolicy | |
46 | except ImportError: |
|
46 | except ImportError: | |
47 | pass |
|
47 | pass | |
48 |
|
48 | |||
49 | # PyPy doesn't load C extensions. |
|
49 | # PyPy doesn't load C extensions. | |
50 | # |
|
50 | # | |
51 | # The canonical way to do this is to test platform.python_implementation(). |
|
51 | # The canonical way to do this is to test platform.python_implementation(). | |
52 | # But we don't import platform and don't bloat for it here. |
|
52 | # But we don't import platform and don't bloat for it here. | |
53 | if '__pypy__' in sys.builtin_module_names: |
|
53 | if '__pypy__' in sys.builtin_module_names: | |
54 | policy = b'cffi' |
|
54 | policy = b'cffi' | |
55 |
|
55 | |||
56 | # Environment variable can always force settings. |
|
56 | # Environment variable can always force settings. | |
57 | if sys.version_info[0] >= 3: |
|
57 | if sys.version_info[0] >= 3: | |
58 | if 'HGMODULEPOLICY' in os.environ: |
|
58 | if 'HGMODULEPOLICY' in os.environ: | |
59 | policy = os.environ['HGMODULEPOLICY'].encode('utf-8') |
|
59 | policy = os.environ['HGMODULEPOLICY'].encode('utf-8') | |
60 | else: |
|
60 | else: | |
61 | policy = os.environ.get('HGMODULEPOLICY', policy) |
|
61 | policy = os.environ.get('HGMODULEPOLICY', policy) | |
62 |
|
62 | |||
63 |
|
63 | |||
64 | def _importfrom(pkgname, modname): |
|
64 | def _importfrom(pkgname, modname): | |
65 | # from .<pkgname> import <modname> (where . is looked through this module) |
|
65 | # from .<pkgname> import <modname> (where . is looked through this module) | |
66 | fakelocals = {} |
|
66 | fakelocals = {} | |
67 | pkg = __import__(pkgname, globals(), fakelocals, [modname], level=1) |
|
67 | pkg = __import__(pkgname, globals(), fakelocals, [modname], level=1) | |
68 | try: |
|
68 | try: | |
69 | fakelocals[modname] = mod = getattr(pkg, modname) |
|
69 | fakelocals[modname] = mod = getattr(pkg, modname) | |
70 | except AttributeError: |
|
70 | except AttributeError: | |
71 | raise ImportError('cannot import name %s' % modname) |
|
71 | raise ImportError('cannot import name %s' % modname) | |
72 | # force import; fakelocals[modname] may be replaced with the real module |
|
72 | # force import; fakelocals[modname] may be replaced with the real module | |
73 | getattr(mod, '__doc__', None) |
|
73 | getattr(mod, '__doc__', None) | |
74 | return fakelocals[modname] |
|
74 | return fakelocals[modname] | |
75 |
|
75 | |||
76 |
|
76 | |||
77 | # keep in sync with "version" in C modules |
|
77 | # keep in sync with "version" in C modules | |
78 | _cextversions = { |
|
78 | _cextversions = { | |
79 | ('cext', 'base85'): 1, |
|
79 | ('cext', 'base85'): 1, | |
80 | ('cext', 'bdiff'): 3, |
|
80 | ('cext', 'bdiff'): 3, | |
81 | ('cext', 'mpatch'): 1, |
|
81 | ('cext', 'mpatch'): 1, | |
82 | ('cext', 'osutil'): 4, |
|
82 | ('cext', 'osutil'): 4, | |
83 |
('cext', 'parsers'): 1 |
|
83 | ('cext', 'parsers'): 14, | |
84 | } |
|
84 | } | |
85 |
|
85 | |||
86 | # map import request to other package or module |
|
86 | # map import request to other package or module | |
87 | _modredirects = { |
|
87 | _modredirects = { | |
88 | ('cext', 'charencode'): ('cext', 'parsers'), |
|
88 | ('cext', 'charencode'): ('cext', 'parsers'), | |
89 | ('cffi', 'base85'): ('pure', 'base85'), |
|
89 | ('cffi', 'base85'): ('pure', 'base85'), | |
90 | ('cffi', 'charencode'): ('pure', 'charencode'), |
|
90 | ('cffi', 'charencode'): ('pure', 'charencode'), | |
91 | ('cffi', 'parsers'): ('pure', 'parsers'), |
|
91 | ('cffi', 'parsers'): ('pure', 'parsers'), | |
92 | } |
|
92 | } | |
93 |
|
93 | |||
94 |
|
94 | |||
95 | def _checkmod(pkgname, modname, mod): |
|
95 | def _checkmod(pkgname, modname, mod): | |
96 | expected = _cextversions.get((pkgname, modname)) |
|
96 | expected = _cextversions.get((pkgname, modname)) | |
97 | actual = getattr(mod, 'version', None) |
|
97 | actual = getattr(mod, 'version', None) | |
98 | if actual != expected: |
|
98 | if actual != expected: | |
99 | raise ImportError( |
|
99 | raise ImportError( | |
100 | 'cannot import module %s.%s ' |
|
100 | 'cannot import module %s.%s ' | |
101 | '(expected version: %d, actual: %r)' |
|
101 | '(expected version: %d, actual: %r)' | |
102 | % (pkgname, modname, expected, actual) |
|
102 | % (pkgname, modname, expected, actual) | |
103 | ) |
|
103 | ) | |
104 |
|
104 | |||
105 |
|
105 | |||
106 | def importmod(modname): |
|
106 | def importmod(modname): | |
107 | """Import module according to policy and check API version""" |
|
107 | """Import module according to policy and check API version""" | |
108 | try: |
|
108 | try: | |
109 | verpkg, purepkg = _packageprefs[policy] |
|
109 | verpkg, purepkg = _packageprefs[policy] | |
110 | except KeyError: |
|
110 | except KeyError: | |
111 | raise ImportError('invalid HGMODULEPOLICY %r' % policy) |
|
111 | raise ImportError('invalid HGMODULEPOLICY %r' % policy) | |
112 | assert verpkg or purepkg |
|
112 | assert verpkg or purepkg | |
113 | if verpkg: |
|
113 | if verpkg: | |
114 | pn, mn = _modredirects.get((verpkg, modname), (verpkg, modname)) |
|
114 | pn, mn = _modredirects.get((verpkg, modname), (verpkg, modname)) | |
115 | try: |
|
115 | try: | |
116 | mod = _importfrom(pn, mn) |
|
116 | mod = _importfrom(pn, mn) | |
117 | if pn == verpkg: |
|
117 | if pn == verpkg: | |
118 | _checkmod(pn, mn, mod) |
|
118 | _checkmod(pn, mn, mod) | |
119 | return mod |
|
119 | return mod | |
120 | except ImportError: |
|
120 | except ImportError: | |
121 | if not purepkg: |
|
121 | if not purepkg: | |
122 | raise |
|
122 | raise | |
123 | pn, mn = _modredirects.get((purepkg, modname), (purepkg, modname)) |
|
123 | pn, mn = _modredirects.get((purepkg, modname), (purepkg, modname)) | |
124 | return _importfrom(pn, mn) |
|
124 | return _importfrom(pn, mn) | |
125 |
|
125 | |||
126 |
|
126 | |||
127 | def _isrustpermissive(): |
|
127 | def _isrustpermissive(): | |
128 | """Assuming the policy is a Rust one, tell if it's permissive.""" |
|
128 | """Assuming the policy is a Rust one, tell if it's permissive.""" | |
129 | return policy.endswith(b'-allow') |
|
129 | return policy.endswith(b'-allow') | |
130 |
|
130 | |||
131 |
|
131 | |||
132 | def importrust(modname, member=None, default=None): |
|
132 | def importrust(modname, member=None, default=None): | |
133 | """Import Rust module according to policy and availability. |
|
133 | """Import Rust module according to policy and availability. | |
134 |
|
134 | |||
135 | If policy isn't a Rust one, this returns `default`. |
|
135 | If policy isn't a Rust one, this returns `default`. | |
136 |
|
136 | |||
137 | If either the module or its member is not available, this returns `default` |
|
137 | If either the module or its member is not available, this returns `default` | |
138 | if policy is permissive and raises `ImportError` if not. |
|
138 | if policy is permissive and raises `ImportError` if not. | |
139 | """ |
|
139 | """ | |
140 | if not policy.startswith(b'rust'): |
|
140 | if not policy.startswith(b'rust'): | |
141 | return default |
|
141 | return default | |
142 |
|
142 | |||
143 | try: |
|
143 | try: | |
144 | mod = _importfrom('rustext', modname) |
|
144 | mod = _importfrom('rustext', modname) | |
145 | except ImportError: |
|
145 | except ImportError: | |
146 | if _isrustpermissive(): |
|
146 | if _isrustpermissive(): | |
147 | return default |
|
147 | return default | |
148 | raise |
|
148 | raise | |
149 | if member is None: |
|
149 | if member is None: | |
150 | return mod |
|
150 | return mod | |
151 |
|
151 | |||
152 | try: |
|
152 | try: | |
153 | return getattr(mod, member) |
|
153 | return getattr(mod, member) | |
154 | except AttributeError: |
|
154 | except AttributeError: | |
155 | if _isrustpermissive(): |
|
155 | if _isrustpermissive(): | |
156 | return default |
|
156 | return default | |
157 | raise ImportError("Cannot import name %s" % member) |
|
157 | raise ImportError("Cannot import name %s" % member) |
@@ -1,213 +1,217 b'' | |||||
1 | # parsers.py - Python implementation of parsers.c |
|
1 | # parsers.py - Python implementation of parsers.c | |
2 | # |
|
2 | # | |
3 | # Copyright 2009 Matt Mackall <mpm@selenic.com> and others |
|
3 | # Copyright 2009 Matt Mackall <mpm@selenic.com> and others | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import struct |
|
10 | import struct | |
11 | import zlib |
|
11 | import zlib | |
12 |
|
12 | |||
13 | from ..node import nullid, nullrev |
|
13 | from ..node import nullid, nullrev | |
14 | from .. import ( |
|
14 | from .. import ( | |
15 | pycompat, |
|
15 | pycompat, | |
16 | revlogutils, |
|
16 | revlogutils, | |
17 | util, |
|
17 | util, | |
18 | ) |
|
18 | ) | |
19 |
|
19 | |||
20 | stringio = pycompat.bytesio |
|
20 | stringio = pycompat.bytesio | |
21 |
|
21 | |||
22 |
|
22 | |||
23 | _pack = struct.pack |
|
23 | _pack = struct.pack | |
24 | _unpack = struct.unpack |
|
24 | _unpack = struct.unpack | |
25 | _compress = zlib.compress |
|
25 | _compress = zlib.compress | |
26 | _decompress = zlib.decompress |
|
26 | _decompress = zlib.decompress | |
27 |
|
27 | |||
28 | # Some code below makes tuples directly because it's more convenient. However, |
|
28 | # Some code below makes tuples directly because it's more convenient. However, | |
29 | # code outside this module should always use dirstatetuple. |
|
29 | # code outside this module should always use dirstatetuple. | |
30 | def dirstatetuple(*x): |
|
30 | def dirstatetuple(*x): | |
31 | # x is a tuple |
|
31 | # x is a tuple | |
32 | return x |
|
32 | return x | |
33 |
|
33 | |||
34 |
|
34 | |||
35 | indexformatng = b">Qiiiiii20s12x" |
|
35 | indexformatng = b">Qiiiiii20s12x" | |
36 | indexfirst = struct.calcsize(b'Q') |
|
36 | indexfirst = struct.calcsize(b'Q') | |
37 | sizeint = struct.calcsize(b'i') |
|
37 | sizeint = struct.calcsize(b'i') | |
38 | indexsize = struct.calcsize(indexformatng) |
|
38 | indexsize = struct.calcsize(indexformatng) | |
39 |
|
39 | |||
40 |
|
40 | |||
41 | def gettype(q): |
|
41 | def gettype(q): | |
42 | return int(q & 0xFFFF) |
|
42 | return int(q & 0xFFFF) | |
43 |
|
43 | |||
44 |
|
44 | |||
45 | def offset_type(offset, type): |
|
45 | def offset_type(offset, type): | |
46 | return int(int(offset) << 16 | type) |
|
46 | return int(int(offset) << 16 | type) | |
47 |
|
47 | |||
48 |
|
48 | |||
49 | class BaseIndexObject(object): |
|
49 | class BaseIndexObject(object): | |
50 | @util.propertycache |
|
50 | @util.propertycache | |
51 | def nodemap(self): |
|
51 | def nodemap(self): | |
52 | nodemap = revlogutils.NodeMap({nullid: nullrev}) |
|
52 | nodemap = revlogutils.NodeMap({nullid: nullrev}) | |
53 | for r in range(0, len(self)): |
|
53 | for r in range(0, len(self)): | |
54 | n = self[r][7] |
|
54 | n = self[r][7] | |
55 | nodemap[n] = r |
|
55 | nodemap[n] = r | |
56 | return nodemap |
|
56 | return nodemap | |
57 |
|
57 | |||
|
58 | def has_node(self, node): | |||
|
59 | """return True if the node exist in the index""" | |||
|
60 | return node in self.nodemap | |||
|
61 | ||||
58 | def _stripnodes(self, start): |
|
62 | def _stripnodes(self, start): | |
59 | if 'nodemap' in vars(self): |
|
63 | if 'nodemap' in vars(self): | |
60 | for r in range(start, len(self)): |
|
64 | for r in range(start, len(self)): | |
61 | n = self[r][7] |
|
65 | n = self[r][7] | |
62 | del self.nodemap[n] |
|
66 | del self.nodemap[n] | |
63 |
|
67 | |||
64 | def clearcaches(self): |
|
68 | def clearcaches(self): | |
65 | self.__dict__.pop('nodemap', None) |
|
69 | self.__dict__.pop('nodemap', None) | |
66 |
|
70 | |||
67 | def __len__(self): |
|
71 | def __len__(self): | |
68 | return self._lgt + len(self._extra) |
|
72 | return self._lgt + len(self._extra) | |
69 |
|
73 | |||
70 | def append(self, tup): |
|
74 | def append(self, tup): | |
71 | if 'nodemap' in vars(self): |
|
75 | if 'nodemap' in vars(self): | |
72 | self.nodemap[tup[7]] = len(self) |
|
76 | self.nodemap[tup[7]] = len(self) | |
73 | self._extra.append(tup) |
|
77 | self._extra.append(tup) | |
74 |
|
78 | |||
75 | def _check_index(self, i): |
|
79 | def _check_index(self, i): | |
76 | if not isinstance(i, int): |
|
80 | if not isinstance(i, int): | |
77 | raise TypeError(b"expecting int indexes") |
|
81 | raise TypeError(b"expecting int indexes") | |
78 | if i < 0 or i >= len(self): |
|
82 | if i < 0 or i >= len(self): | |
79 | raise IndexError |
|
83 | raise IndexError | |
80 |
|
84 | |||
81 | def __getitem__(self, i): |
|
85 | def __getitem__(self, i): | |
82 | if i == -1: |
|
86 | if i == -1: | |
83 | return (0, 0, 0, -1, -1, -1, -1, nullid) |
|
87 | return (0, 0, 0, -1, -1, -1, -1, nullid) | |
84 | self._check_index(i) |
|
88 | self._check_index(i) | |
85 | if i >= self._lgt: |
|
89 | if i >= self._lgt: | |
86 | return self._extra[i - self._lgt] |
|
90 | return self._extra[i - self._lgt] | |
87 | index = self._calculate_index(i) |
|
91 | index = self._calculate_index(i) | |
88 | r = struct.unpack(indexformatng, self._data[index : index + indexsize]) |
|
92 | r = struct.unpack(indexformatng, self._data[index : index + indexsize]) | |
89 | if i == 0: |
|
93 | if i == 0: | |
90 | e = list(r) |
|
94 | e = list(r) | |
91 | type = gettype(e[0]) |
|
95 | type = gettype(e[0]) | |
92 | e[0] = offset_type(0, type) |
|
96 | e[0] = offset_type(0, type) | |
93 | return tuple(e) |
|
97 | return tuple(e) | |
94 | return r |
|
98 | return r | |
95 |
|
99 | |||
96 |
|
100 | |||
97 | class IndexObject(BaseIndexObject): |
|
101 | class IndexObject(BaseIndexObject): | |
98 | def __init__(self, data): |
|
102 | def __init__(self, data): | |
99 | assert len(data) % indexsize == 0 |
|
103 | assert len(data) % indexsize == 0 | |
100 | self._data = data |
|
104 | self._data = data | |
101 | self._lgt = len(data) // indexsize |
|
105 | self._lgt = len(data) // indexsize | |
102 | self._extra = [] |
|
106 | self._extra = [] | |
103 |
|
107 | |||
104 | def _calculate_index(self, i): |
|
108 | def _calculate_index(self, i): | |
105 | return i * indexsize |
|
109 | return i * indexsize | |
106 |
|
110 | |||
107 | def __delitem__(self, i): |
|
111 | def __delitem__(self, i): | |
108 | if not isinstance(i, slice) or not i.stop == -1 or i.step is not None: |
|
112 | if not isinstance(i, slice) or not i.stop == -1 or i.step is not None: | |
109 | raise ValueError(b"deleting slices only supports a:-1 with step 1") |
|
113 | raise ValueError(b"deleting slices only supports a:-1 with step 1") | |
110 | i = i.start |
|
114 | i = i.start | |
111 | self._check_index(i) |
|
115 | self._check_index(i) | |
112 | self._stripnodes(i) |
|
116 | self._stripnodes(i) | |
113 | if i < self._lgt: |
|
117 | if i < self._lgt: | |
114 | self._data = self._data[: i * indexsize] |
|
118 | self._data = self._data[: i * indexsize] | |
115 | self._lgt = i |
|
119 | self._lgt = i | |
116 | self._extra = [] |
|
120 | self._extra = [] | |
117 | else: |
|
121 | else: | |
118 | self._extra = self._extra[: i - self._lgt] |
|
122 | self._extra = self._extra[: i - self._lgt] | |
119 |
|
123 | |||
120 |
|
124 | |||
121 | class InlinedIndexObject(BaseIndexObject): |
|
125 | class InlinedIndexObject(BaseIndexObject): | |
122 | def __init__(self, data, inline=0): |
|
126 | def __init__(self, data, inline=0): | |
123 | self._data = data |
|
127 | self._data = data | |
124 | self._lgt = self._inline_scan(None) |
|
128 | self._lgt = self._inline_scan(None) | |
125 | self._inline_scan(self._lgt) |
|
129 | self._inline_scan(self._lgt) | |
126 | self._extra = [] |
|
130 | self._extra = [] | |
127 |
|
131 | |||
128 | def _inline_scan(self, lgt): |
|
132 | def _inline_scan(self, lgt): | |
129 | off = 0 |
|
133 | off = 0 | |
130 | if lgt is not None: |
|
134 | if lgt is not None: | |
131 | self._offsets = [0] * lgt |
|
135 | self._offsets = [0] * lgt | |
132 | count = 0 |
|
136 | count = 0 | |
133 | while off <= len(self._data) - indexsize: |
|
137 | while off <= len(self._data) - indexsize: | |
134 | (s,) = struct.unpack( |
|
138 | (s,) = struct.unpack( | |
135 | b'>i', self._data[off + indexfirst : off + sizeint + indexfirst] |
|
139 | b'>i', self._data[off + indexfirst : off + sizeint + indexfirst] | |
136 | ) |
|
140 | ) | |
137 | if lgt is not None: |
|
141 | if lgt is not None: | |
138 | self._offsets[count] = off |
|
142 | self._offsets[count] = off | |
139 | count += 1 |
|
143 | count += 1 | |
140 | off += indexsize + s |
|
144 | off += indexsize + s | |
141 | if off != len(self._data): |
|
145 | if off != len(self._data): | |
142 | raise ValueError(b"corrupted data") |
|
146 | raise ValueError(b"corrupted data") | |
143 | return count |
|
147 | return count | |
144 |
|
148 | |||
145 | def __delitem__(self, i): |
|
149 | def __delitem__(self, i): | |
146 | if not isinstance(i, slice) or not i.stop == -1 or i.step is not None: |
|
150 | if not isinstance(i, slice) or not i.stop == -1 or i.step is not None: | |
147 | raise ValueError(b"deleting slices only supports a:-1 with step 1") |
|
151 | raise ValueError(b"deleting slices only supports a:-1 with step 1") | |
148 | i = i.start |
|
152 | i = i.start | |
149 | self._check_index(i) |
|
153 | self._check_index(i) | |
150 | self._stripnodes(i) |
|
154 | self._stripnodes(i) | |
151 | if i < self._lgt: |
|
155 | if i < self._lgt: | |
152 | self._offsets = self._offsets[:i] |
|
156 | self._offsets = self._offsets[:i] | |
153 | self._lgt = i |
|
157 | self._lgt = i | |
154 | self._extra = [] |
|
158 | self._extra = [] | |
155 | else: |
|
159 | else: | |
156 | self._extra = self._extra[: i - self._lgt] |
|
160 | self._extra = self._extra[: i - self._lgt] | |
157 |
|
161 | |||
158 | def _calculate_index(self, i): |
|
162 | def _calculate_index(self, i): | |
159 | return self._offsets[i] |
|
163 | return self._offsets[i] | |
160 |
|
164 | |||
161 |
|
165 | |||
162 | def parse_index2(data, inline): |
|
166 | def parse_index2(data, inline): | |
163 | if not inline: |
|
167 | if not inline: | |
164 | return IndexObject(data), None |
|
168 | return IndexObject(data), None | |
165 | return InlinedIndexObject(data, inline), (0, data) |
|
169 | return InlinedIndexObject(data, inline), (0, data) | |
166 |
|
170 | |||
167 |
|
171 | |||
168 | def parse_dirstate(dmap, copymap, st): |
|
172 | def parse_dirstate(dmap, copymap, st): | |
169 | parents = [st[:20], st[20:40]] |
|
173 | parents = [st[:20], st[20:40]] | |
170 | # dereference fields so they will be local in loop |
|
174 | # dereference fields so they will be local in loop | |
171 | format = b">cllll" |
|
175 | format = b">cllll" | |
172 | e_size = struct.calcsize(format) |
|
176 | e_size = struct.calcsize(format) | |
173 | pos1 = 40 |
|
177 | pos1 = 40 | |
174 | l = len(st) |
|
178 | l = len(st) | |
175 |
|
179 | |||
176 | # the inner loop |
|
180 | # the inner loop | |
177 | while pos1 < l: |
|
181 | while pos1 < l: | |
178 | pos2 = pos1 + e_size |
|
182 | pos2 = pos1 + e_size | |
179 | e = _unpack(b">cllll", st[pos1:pos2]) # a literal here is faster |
|
183 | e = _unpack(b">cllll", st[pos1:pos2]) # a literal here is faster | |
180 | pos1 = pos2 + e[4] |
|
184 | pos1 = pos2 + e[4] | |
181 | f = st[pos2:pos1] |
|
185 | f = st[pos2:pos1] | |
182 | if b'\0' in f: |
|
186 | if b'\0' in f: | |
183 | f, c = f.split(b'\0') |
|
187 | f, c = f.split(b'\0') | |
184 | copymap[f] = c |
|
188 | copymap[f] = c | |
185 | dmap[f] = e[:4] |
|
189 | dmap[f] = e[:4] | |
186 | return parents |
|
190 | return parents | |
187 |
|
191 | |||
188 |
|
192 | |||
189 | def pack_dirstate(dmap, copymap, pl, now): |
|
193 | def pack_dirstate(dmap, copymap, pl, now): | |
190 | now = int(now) |
|
194 | now = int(now) | |
191 | cs = stringio() |
|
195 | cs = stringio() | |
192 | write = cs.write |
|
196 | write = cs.write | |
193 | write(b"".join(pl)) |
|
197 | write(b"".join(pl)) | |
194 | for f, e in pycompat.iteritems(dmap): |
|
198 | for f, e in pycompat.iteritems(dmap): | |
195 | if e[0] == b'n' and e[3] == now: |
|
199 | if e[0] == b'n' and e[3] == now: | |
196 | # The file was last modified "simultaneously" with the current |
|
200 | # The file was last modified "simultaneously" with the current | |
197 | # write to dirstate (i.e. within the same second for file- |
|
201 | # write to dirstate (i.e. within the same second for file- | |
198 | # systems with a granularity of 1 sec). This commonly happens |
|
202 | # systems with a granularity of 1 sec). This commonly happens | |
199 | # for at least a couple of files on 'update'. |
|
203 | # for at least a couple of files on 'update'. | |
200 | # The user could change the file without changing its size |
|
204 | # The user could change the file without changing its size | |
201 | # within the same second. Invalidate the file's mtime in |
|
205 | # within the same second. Invalidate the file's mtime in | |
202 | # dirstate, forcing future 'status' calls to compare the |
|
206 | # dirstate, forcing future 'status' calls to compare the | |
203 | # contents of the file if the size is the same. This prevents |
|
207 | # contents of the file if the size is the same. This prevents | |
204 | # mistakenly treating such files as clean. |
|
208 | # mistakenly treating such files as clean. | |
205 | e = dirstatetuple(e[0], e[1], e[2], -1) |
|
209 | e = dirstatetuple(e[0], e[1], e[2], -1) | |
206 | dmap[f] = e |
|
210 | dmap[f] = e | |
207 |
|
211 | |||
208 | if f in copymap: |
|
212 | if f in copymap: | |
209 | f = b"%s\0%s" % (f, copymap[f]) |
|
213 | f = b"%s\0%s" % (f, copymap[f]) | |
210 | e = _pack(b">cllll", e[0], e[1], e[2], e[3], len(f)) |
|
214 | e = _pack(b">cllll", e[0], e[1], e[2], e[3], len(f)) | |
211 | write(e) |
|
215 | write(e) | |
212 | write(f) |
|
216 | write(f) | |
213 | return cs.getvalue() |
|
217 | return cs.getvalue() |
@@ -1,2960 +1,2964 b'' | |||||
1 | # revlog.py - storage back-end for mercurial |
|
1 | # revlog.py - storage back-end for mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | """Storage back-end for Mercurial. |
|
8 | """Storage back-end for Mercurial. | |
9 |
|
9 | |||
10 | This provides efficient delta storage with O(1) retrieve and append |
|
10 | This provides efficient delta storage with O(1) retrieve and append | |
11 | and O(changes) merge between branches. |
|
11 | and O(changes) merge between branches. | |
12 | """ |
|
12 | """ | |
13 |
|
13 | |||
14 | from __future__ import absolute_import |
|
14 | from __future__ import absolute_import | |
15 |
|
15 | |||
16 | import collections |
|
16 | import collections | |
17 | import contextlib |
|
17 | import contextlib | |
18 | import errno |
|
18 | import errno | |
19 | import io |
|
19 | import io | |
20 | import os |
|
20 | import os | |
21 | import struct |
|
21 | import struct | |
22 | import zlib |
|
22 | import zlib | |
23 |
|
23 | |||
24 | # import stuff from node for others to import from revlog |
|
24 | # import stuff from node for others to import from revlog | |
25 | from .node import ( |
|
25 | from .node import ( | |
26 | bin, |
|
26 | bin, | |
27 | hex, |
|
27 | hex, | |
28 | nullhex, |
|
28 | nullhex, | |
29 | nullid, |
|
29 | nullid, | |
30 | nullrev, |
|
30 | nullrev, | |
31 | short, |
|
31 | short, | |
32 | wdirfilenodeids, |
|
32 | wdirfilenodeids, | |
33 | wdirhex, |
|
33 | wdirhex, | |
34 | wdirid, |
|
34 | wdirid, | |
35 | wdirrev, |
|
35 | wdirrev, | |
36 | ) |
|
36 | ) | |
37 | from .i18n import _ |
|
37 | from .i18n import _ | |
38 | from .pycompat import getattr |
|
38 | from .pycompat import getattr | |
39 | from .revlogutils.constants import ( |
|
39 | from .revlogutils.constants import ( | |
40 | FLAG_GENERALDELTA, |
|
40 | FLAG_GENERALDELTA, | |
41 | FLAG_INLINE_DATA, |
|
41 | FLAG_INLINE_DATA, | |
42 | REVLOGV0, |
|
42 | REVLOGV0, | |
43 | REVLOGV1, |
|
43 | REVLOGV1, | |
44 | REVLOGV1_FLAGS, |
|
44 | REVLOGV1_FLAGS, | |
45 | REVLOGV2, |
|
45 | REVLOGV2, | |
46 | REVLOGV2_FLAGS, |
|
46 | REVLOGV2_FLAGS, | |
47 | REVLOG_DEFAULT_FLAGS, |
|
47 | REVLOG_DEFAULT_FLAGS, | |
48 | REVLOG_DEFAULT_FORMAT, |
|
48 | REVLOG_DEFAULT_FORMAT, | |
49 | REVLOG_DEFAULT_VERSION, |
|
49 | REVLOG_DEFAULT_VERSION, | |
50 | ) |
|
50 | ) | |
51 | from .revlogutils.flagutil import ( |
|
51 | from .revlogutils.flagutil import ( | |
52 | REVIDX_DEFAULT_FLAGS, |
|
52 | REVIDX_DEFAULT_FLAGS, | |
53 | REVIDX_ELLIPSIS, |
|
53 | REVIDX_ELLIPSIS, | |
54 | REVIDX_EXTSTORED, |
|
54 | REVIDX_EXTSTORED, | |
55 | REVIDX_FLAGS_ORDER, |
|
55 | REVIDX_FLAGS_ORDER, | |
56 | REVIDX_ISCENSORED, |
|
56 | REVIDX_ISCENSORED, | |
57 | REVIDX_RAWTEXT_CHANGING_FLAGS, |
|
57 | REVIDX_RAWTEXT_CHANGING_FLAGS, | |
58 | REVIDX_SIDEDATA, |
|
58 | REVIDX_SIDEDATA, | |
59 | ) |
|
59 | ) | |
60 | from .thirdparty import attr |
|
60 | from .thirdparty import attr | |
61 | from . import ( |
|
61 | from . import ( | |
62 | ancestor, |
|
62 | ancestor, | |
63 | dagop, |
|
63 | dagop, | |
64 | error, |
|
64 | error, | |
65 | mdiff, |
|
65 | mdiff, | |
66 | policy, |
|
66 | policy, | |
67 | pycompat, |
|
67 | pycompat, | |
68 | revlogutils, |
|
68 | revlogutils, | |
69 | templatefilters, |
|
69 | templatefilters, | |
70 | util, |
|
70 | util, | |
71 | ) |
|
71 | ) | |
72 | from .interfaces import ( |
|
72 | from .interfaces import ( | |
73 | repository, |
|
73 | repository, | |
74 | util as interfaceutil, |
|
74 | util as interfaceutil, | |
75 | ) |
|
75 | ) | |
76 | from .revlogutils import ( |
|
76 | from .revlogutils import ( | |
77 | deltas as deltautil, |
|
77 | deltas as deltautil, | |
78 | flagutil, |
|
78 | flagutil, | |
79 | sidedata as sidedatautil, |
|
79 | sidedata as sidedatautil, | |
80 | ) |
|
80 | ) | |
81 | from .utils import ( |
|
81 | from .utils import ( | |
82 | storageutil, |
|
82 | storageutil, | |
83 | stringutil, |
|
83 | stringutil, | |
84 | ) |
|
84 | ) | |
85 |
|
85 | |||
86 | # blanked usage of all the name to prevent pyflakes constraints |
|
86 | # blanked usage of all the name to prevent pyflakes constraints | |
87 | # We need these name available in the module for extensions. |
|
87 | # We need these name available in the module for extensions. | |
88 | REVLOGV0 |
|
88 | REVLOGV0 | |
89 | REVLOGV1 |
|
89 | REVLOGV1 | |
90 | REVLOGV2 |
|
90 | REVLOGV2 | |
91 | FLAG_INLINE_DATA |
|
91 | FLAG_INLINE_DATA | |
92 | FLAG_GENERALDELTA |
|
92 | FLAG_GENERALDELTA | |
93 | REVLOG_DEFAULT_FLAGS |
|
93 | REVLOG_DEFAULT_FLAGS | |
94 | REVLOG_DEFAULT_FORMAT |
|
94 | REVLOG_DEFAULT_FORMAT | |
95 | REVLOG_DEFAULT_VERSION |
|
95 | REVLOG_DEFAULT_VERSION | |
96 | REVLOGV1_FLAGS |
|
96 | REVLOGV1_FLAGS | |
97 | REVLOGV2_FLAGS |
|
97 | REVLOGV2_FLAGS | |
98 | REVIDX_ISCENSORED |
|
98 | REVIDX_ISCENSORED | |
99 | REVIDX_ELLIPSIS |
|
99 | REVIDX_ELLIPSIS | |
100 | REVIDX_SIDEDATA |
|
100 | REVIDX_SIDEDATA | |
101 | REVIDX_EXTSTORED |
|
101 | REVIDX_EXTSTORED | |
102 | REVIDX_DEFAULT_FLAGS |
|
102 | REVIDX_DEFAULT_FLAGS | |
103 | REVIDX_FLAGS_ORDER |
|
103 | REVIDX_FLAGS_ORDER | |
104 | REVIDX_RAWTEXT_CHANGING_FLAGS |
|
104 | REVIDX_RAWTEXT_CHANGING_FLAGS | |
105 |
|
105 | |||
106 | parsers = policy.importmod('parsers') |
|
106 | parsers = policy.importmod('parsers') | |
107 | rustancestor = policy.importrust('ancestor') |
|
107 | rustancestor = policy.importrust('ancestor') | |
108 | rustdagop = policy.importrust('dagop') |
|
108 | rustdagop = policy.importrust('dagop') | |
109 |
|
109 | |||
110 | # Aliased for performance. |
|
110 | # Aliased for performance. | |
111 | _zlibdecompress = zlib.decompress |
|
111 | _zlibdecompress = zlib.decompress | |
112 |
|
112 | |||
113 | # max size of revlog with inline data |
|
113 | # max size of revlog with inline data | |
114 | _maxinline = 131072 |
|
114 | _maxinline = 131072 | |
115 | _chunksize = 1048576 |
|
115 | _chunksize = 1048576 | |
116 |
|
116 | |||
117 | # Flag processors for REVIDX_ELLIPSIS. |
|
117 | # Flag processors for REVIDX_ELLIPSIS. | |
118 | def ellipsisreadprocessor(rl, text): |
|
118 | def ellipsisreadprocessor(rl, text): | |
119 | return text, False, {} |
|
119 | return text, False, {} | |
120 |
|
120 | |||
121 |
|
121 | |||
122 | def ellipsiswriteprocessor(rl, text, sidedata): |
|
122 | def ellipsiswriteprocessor(rl, text, sidedata): | |
123 | return text, False |
|
123 | return text, False | |
124 |
|
124 | |||
125 |
|
125 | |||
126 | def ellipsisrawprocessor(rl, text): |
|
126 | def ellipsisrawprocessor(rl, text): | |
127 | return False |
|
127 | return False | |
128 |
|
128 | |||
129 |
|
129 | |||
130 | ellipsisprocessor = ( |
|
130 | ellipsisprocessor = ( | |
131 | ellipsisreadprocessor, |
|
131 | ellipsisreadprocessor, | |
132 | ellipsiswriteprocessor, |
|
132 | ellipsiswriteprocessor, | |
133 | ellipsisrawprocessor, |
|
133 | ellipsisrawprocessor, | |
134 | ) |
|
134 | ) | |
135 |
|
135 | |||
136 |
|
136 | |||
137 | def getoffset(q): |
|
137 | def getoffset(q): | |
138 | return int(q >> 16) |
|
138 | return int(q >> 16) | |
139 |
|
139 | |||
140 |
|
140 | |||
141 | def gettype(q): |
|
141 | def gettype(q): | |
142 | return int(q & 0xFFFF) |
|
142 | return int(q & 0xFFFF) | |
143 |
|
143 | |||
144 |
|
144 | |||
145 | def offset_type(offset, type): |
|
145 | def offset_type(offset, type): | |
146 | if (type & ~flagutil.REVIDX_KNOWN_FLAGS) != 0: |
|
146 | if (type & ~flagutil.REVIDX_KNOWN_FLAGS) != 0: | |
147 | raise ValueError(b'unknown revlog index flags') |
|
147 | raise ValueError(b'unknown revlog index flags') | |
148 | return int(int(offset) << 16 | type) |
|
148 | return int(int(offset) << 16 | type) | |
149 |
|
149 | |||
150 |
|
150 | |||
151 | @attr.s(slots=True, frozen=True) |
|
151 | @attr.s(slots=True, frozen=True) | |
152 | class _revisioninfo(object): |
|
152 | class _revisioninfo(object): | |
153 | """Information about a revision that allows building its fulltext |
|
153 | """Information about a revision that allows building its fulltext | |
154 | node: expected hash of the revision |
|
154 | node: expected hash of the revision | |
155 | p1, p2: parent revs of the revision |
|
155 | p1, p2: parent revs of the revision | |
156 | btext: built text cache consisting of a one-element list |
|
156 | btext: built text cache consisting of a one-element list | |
157 | cachedelta: (baserev, uncompressed_delta) or None |
|
157 | cachedelta: (baserev, uncompressed_delta) or None | |
158 | flags: flags associated to the revision storage |
|
158 | flags: flags associated to the revision storage | |
159 |
|
159 | |||
160 | One of btext[0] or cachedelta must be set. |
|
160 | One of btext[0] or cachedelta must be set. | |
161 | """ |
|
161 | """ | |
162 |
|
162 | |||
163 | node = attr.ib() |
|
163 | node = attr.ib() | |
164 | p1 = attr.ib() |
|
164 | p1 = attr.ib() | |
165 | p2 = attr.ib() |
|
165 | p2 = attr.ib() | |
166 | btext = attr.ib() |
|
166 | btext = attr.ib() | |
167 | textlen = attr.ib() |
|
167 | textlen = attr.ib() | |
168 | cachedelta = attr.ib() |
|
168 | cachedelta = attr.ib() | |
169 | flags = attr.ib() |
|
169 | flags = attr.ib() | |
170 |
|
170 | |||
171 |
|
171 | |||
172 | @interfaceutil.implementer(repository.irevisiondelta) |
|
172 | @interfaceutil.implementer(repository.irevisiondelta) | |
173 | @attr.s(slots=True) |
|
173 | @attr.s(slots=True) | |
174 | class revlogrevisiondelta(object): |
|
174 | class revlogrevisiondelta(object): | |
175 | node = attr.ib() |
|
175 | node = attr.ib() | |
176 | p1node = attr.ib() |
|
176 | p1node = attr.ib() | |
177 | p2node = attr.ib() |
|
177 | p2node = attr.ib() | |
178 | basenode = attr.ib() |
|
178 | basenode = attr.ib() | |
179 | flags = attr.ib() |
|
179 | flags = attr.ib() | |
180 | baserevisionsize = attr.ib() |
|
180 | baserevisionsize = attr.ib() | |
181 | revision = attr.ib() |
|
181 | revision = attr.ib() | |
182 | delta = attr.ib() |
|
182 | delta = attr.ib() | |
183 | linknode = attr.ib(default=None) |
|
183 | linknode = attr.ib(default=None) | |
184 |
|
184 | |||
185 |
|
185 | |||
186 | @interfaceutil.implementer(repository.iverifyproblem) |
|
186 | @interfaceutil.implementer(repository.iverifyproblem) | |
187 | @attr.s(frozen=True) |
|
187 | @attr.s(frozen=True) | |
188 | class revlogproblem(object): |
|
188 | class revlogproblem(object): | |
189 | warning = attr.ib(default=None) |
|
189 | warning = attr.ib(default=None) | |
190 | error = attr.ib(default=None) |
|
190 | error = attr.ib(default=None) | |
191 | node = attr.ib(default=None) |
|
191 | node = attr.ib(default=None) | |
192 |
|
192 | |||
193 |
|
193 | |||
194 | # index v0: |
|
194 | # index v0: | |
195 | # 4 bytes: offset |
|
195 | # 4 bytes: offset | |
196 | # 4 bytes: compressed length |
|
196 | # 4 bytes: compressed length | |
197 | # 4 bytes: base rev |
|
197 | # 4 bytes: base rev | |
198 | # 4 bytes: link rev |
|
198 | # 4 bytes: link rev | |
199 | # 20 bytes: parent 1 nodeid |
|
199 | # 20 bytes: parent 1 nodeid | |
200 | # 20 bytes: parent 2 nodeid |
|
200 | # 20 bytes: parent 2 nodeid | |
201 | # 20 bytes: nodeid |
|
201 | # 20 bytes: nodeid | |
202 | indexformatv0 = struct.Struct(b">4l20s20s20s") |
|
202 | indexformatv0 = struct.Struct(b">4l20s20s20s") | |
203 | indexformatv0_pack = indexformatv0.pack |
|
203 | indexformatv0_pack = indexformatv0.pack | |
204 | indexformatv0_unpack = indexformatv0.unpack |
|
204 | indexformatv0_unpack = indexformatv0.unpack | |
205 |
|
205 | |||
206 |
|
206 | |||
207 | class revlogoldindex(list): |
|
207 | class revlogoldindex(list): | |
208 | @util.propertycache |
|
208 | @util.propertycache | |
209 | def nodemap(self): |
|
209 | def nodemap(self): | |
210 | nodemap = revlogutils.NodeMap({nullid: nullrev}) |
|
210 | nodemap = revlogutils.NodeMap({nullid: nullrev}) | |
211 | for r in range(0, len(self)): |
|
211 | for r in range(0, len(self)): | |
212 | n = self[r][7] |
|
212 | n = self[r][7] | |
213 | nodemap[n] = r |
|
213 | nodemap[n] = r | |
214 | return nodemap |
|
214 | return nodemap | |
215 |
|
215 | |||
|
216 | def has_node(self, node): | |||
|
217 | """return True if the node exist in the index""" | |||
|
218 | return node in self.nodemap | |||
|
219 | ||||
216 | def append(self, tup): |
|
220 | def append(self, tup): | |
217 | self.nodemap[tup[7]] = len(self) |
|
221 | self.nodemap[tup[7]] = len(self) | |
218 | super(revlogoldindex, self).append(tup) |
|
222 | super(revlogoldindex, self).append(tup) | |
219 |
|
223 | |||
220 | def __delitem__(self, i): |
|
224 | def __delitem__(self, i): | |
221 | if not isinstance(i, slice) or not i.stop == -1 or i.step is not None: |
|
225 | if not isinstance(i, slice) or not i.stop == -1 or i.step is not None: | |
222 | raise ValueError(b"deleting slices only supports a:-1 with step 1") |
|
226 | raise ValueError(b"deleting slices only supports a:-1 with step 1") | |
223 | for r in pycompat.xrange(i.start, len(self)): |
|
227 | for r in pycompat.xrange(i.start, len(self)): | |
224 | del self.nodemap[self[r][7]] |
|
228 | del self.nodemap[self[r][7]] | |
225 | super(revlogoldindex, self).__delitem__(i) |
|
229 | super(revlogoldindex, self).__delitem__(i) | |
226 |
|
230 | |||
227 | def clearcaches(self): |
|
231 | def clearcaches(self): | |
228 | self.__dict__.pop('nodemap', None) |
|
232 | self.__dict__.pop('nodemap', None) | |
229 |
|
233 | |||
230 | def __getitem__(self, i): |
|
234 | def __getitem__(self, i): | |
231 | if i == -1: |
|
235 | if i == -1: | |
232 | return (0, 0, 0, -1, -1, -1, -1, nullid) |
|
236 | return (0, 0, 0, -1, -1, -1, -1, nullid) | |
233 | return list.__getitem__(self, i) |
|
237 | return list.__getitem__(self, i) | |
234 |
|
238 | |||
235 |
|
239 | |||
236 | class revlogoldio(object): |
|
240 | class revlogoldio(object): | |
237 | def __init__(self): |
|
241 | def __init__(self): | |
238 | self.size = indexformatv0.size |
|
242 | self.size = indexformatv0.size | |
239 |
|
243 | |||
240 | def parseindex(self, data, inline): |
|
244 | def parseindex(self, data, inline): | |
241 | s = self.size |
|
245 | s = self.size | |
242 | index = [] |
|
246 | index = [] | |
243 | nodemap = revlogutils.NodeMap({nullid: nullrev}) |
|
247 | nodemap = revlogutils.NodeMap({nullid: nullrev}) | |
244 | n = off = 0 |
|
248 | n = off = 0 | |
245 | l = len(data) |
|
249 | l = len(data) | |
246 | while off + s <= l: |
|
250 | while off + s <= l: | |
247 | cur = data[off : off + s] |
|
251 | cur = data[off : off + s] | |
248 | off += s |
|
252 | off += s | |
249 | e = indexformatv0_unpack(cur) |
|
253 | e = indexformatv0_unpack(cur) | |
250 | # transform to revlogv1 format |
|
254 | # transform to revlogv1 format | |
251 | e2 = ( |
|
255 | e2 = ( | |
252 | offset_type(e[0], 0), |
|
256 | offset_type(e[0], 0), | |
253 | e[1], |
|
257 | e[1], | |
254 | -1, |
|
258 | -1, | |
255 | e[2], |
|
259 | e[2], | |
256 | e[3], |
|
260 | e[3], | |
257 | nodemap.get(e[4], nullrev), |
|
261 | nodemap.get(e[4], nullrev), | |
258 | nodemap.get(e[5], nullrev), |
|
262 | nodemap.get(e[5], nullrev), | |
259 | e[6], |
|
263 | e[6], | |
260 | ) |
|
264 | ) | |
261 | index.append(e2) |
|
265 | index.append(e2) | |
262 | nodemap[e[6]] = n |
|
266 | nodemap[e[6]] = n | |
263 | n += 1 |
|
267 | n += 1 | |
264 |
|
268 | |||
265 | index = revlogoldindex(index) |
|
269 | index = revlogoldindex(index) | |
266 | return index, None |
|
270 | return index, None | |
267 |
|
271 | |||
268 | def packentry(self, entry, node, version, rev): |
|
272 | def packentry(self, entry, node, version, rev): | |
269 | if gettype(entry[0]): |
|
273 | if gettype(entry[0]): | |
270 | raise error.RevlogError( |
|
274 | raise error.RevlogError( | |
271 | _(b'index entry flags need revlog version 1') |
|
275 | _(b'index entry flags need revlog version 1') | |
272 | ) |
|
276 | ) | |
273 | e2 = ( |
|
277 | e2 = ( | |
274 | getoffset(entry[0]), |
|
278 | getoffset(entry[0]), | |
275 | entry[1], |
|
279 | entry[1], | |
276 | entry[3], |
|
280 | entry[3], | |
277 | entry[4], |
|
281 | entry[4], | |
278 | node(entry[5]), |
|
282 | node(entry[5]), | |
279 | node(entry[6]), |
|
283 | node(entry[6]), | |
280 | entry[7], |
|
284 | entry[7], | |
281 | ) |
|
285 | ) | |
282 | return indexformatv0_pack(*e2) |
|
286 | return indexformatv0_pack(*e2) | |
283 |
|
287 | |||
284 |
|
288 | |||
285 | # index ng: |
|
289 | # index ng: | |
286 | # 6 bytes: offset |
|
290 | # 6 bytes: offset | |
287 | # 2 bytes: flags |
|
291 | # 2 bytes: flags | |
288 | # 4 bytes: compressed length |
|
292 | # 4 bytes: compressed length | |
289 | # 4 bytes: uncompressed length |
|
293 | # 4 bytes: uncompressed length | |
290 | # 4 bytes: base rev |
|
294 | # 4 bytes: base rev | |
291 | # 4 bytes: link rev |
|
295 | # 4 bytes: link rev | |
292 | # 4 bytes: parent 1 rev |
|
296 | # 4 bytes: parent 1 rev | |
293 | # 4 bytes: parent 2 rev |
|
297 | # 4 bytes: parent 2 rev | |
294 | # 32 bytes: nodeid |
|
298 | # 32 bytes: nodeid | |
295 | indexformatng = struct.Struct(b">Qiiiiii20s12x") |
|
299 | indexformatng = struct.Struct(b">Qiiiiii20s12x") | |
296 | indexformatng_pack = indexformatng.pack |
|
300 | indexformatng_pack = indexformatng.pack | |
297 | versionformat = struct.Struct(b">I") |
|
301 | versionformat = struct.Struct(b">I") | |
298 | versionformat_pack = versionformat.pack |
|
302 | versionformat_pack = versionformat.pack | |
299 | versionformat_unpack = versionformat.unpack |
|
303 | versionformat_unpack = versionformat.unpack | |
300 |
|
304 | |||
301 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte |
|
305 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte | |
302 | # signed integer) |
|
306 | # signed integer) | |
303 | _maxentrysize = 0x7FFFFFFF |
|
307 | _maxentrysize = 0x7FFFFFFF | |
304 |
|
308 | |||
305 |
|
309 | |||
306 | class revlogio(object): |
|
310 | class revlogio(object): | |
307 | def __init__(self): |
|
311 | def __init__(self): | |
308 | self.size = indexformatng.size |
|
312 | self.size = indexformatng.size | |
309 |
|
313 | |||
310 | def parseindex(self, data, inline): |
|
314 | def parseindex(self, data, inline): | |
311 | # call the C implementation to parse the index data |
|
315 | # call the C implementation to parse the index data | |
312 | index, cache = parsers.parse_index2(data, inline) |
|
316 | index, cache = parsers.parse_index2(data, inline) | |
313 | return index, cache |
|
317 | return index, cache | |
314 |
|
318 | |||
315 | def packentry(self, entry, node, version, rev): |
|
319 | def packentry(self, entry, node, version, rev): | |
316 | p = indexformatng_pack(*entry) |
|
320 | p = indexformatng_pack(*entry) | |
317 | if rev == 0: |
|
321 | if rev == 0: | |
318 | p = versionformat_pack(version) + p[4:] |
|
322 | p = versionformat_pack(version) + p[4:] | |
319 | return p |
|
323 | return p | |
320 |
|
324 | |||
321 |
|
325 | |||
322 | class revlog(object): |
|
326 | class revlog(object): | |
323 | """ |
|
327 | """ | |
324 | the underlying revision storage object |
|
328 | the underlying revision storage object | |
325 |
|
329 | |||
326 | A revlog consists of two parts, an index and the revision data. |
|
330 | A revlog consists of two parts, an index and the revision data. | |
327 |
|
331 | |||
328 | The index is a file with a fixed record size containing |
|
332 | The index is a file with a fixed record size containing | |
329 | information on each revision, including its nodeid (hash), the |
|
333 | information on each revision, including its nodeid (hash), the | |
330 | nodeids of its parents, the position and offset of its data within |
|
334 | nodeids of its parents, the position and offset of its data within | |
331 | the data file, and the revision it's based on. Finally, each entry |
|
335 | the data file, and the revision it's based on. Finally, each entry | |
332 | contains a linkrev entry that can serve as a pointer to external |
|
336 | contains a linkrev entry that can serve as a pointer to external | |
333 | data. |
|
337 | data. | |
334 |
|
338 | |||
335 | The revision data itself is a linear collection of data chunks. |
|
339 | The revision data itself is a linear collection of data chunks. | |
336 | Each chunk represents a revision and is usually represented as a |
|
340 | Each chunk represents a revision and is usually represented as a | |
337 | delta against the previous chunk. To bound lookup time, runs of |
|
341 | delta against the previous chunk. To bound lookup time, runs of | |
338 | deltas are limited to about 2 times the length of the original |
|
342 | deltas are limited to about 2 times the length of the original | |
339 | version data. This makes retrieval of a version proportional to |
|
343 | version data. This makes retrieval of a version proportional to | |
340 | its size, or O(1) relative to the number of revisions. |
|
344 | its size, or O(1) relative to the number of revisions. | |
341 |
|
345 | |||
342 | Both pieces of the revlog are written to in an append-only |
|
346 | Both pieces of the revlog are written to in an append-only | |
343 | fashion, which means we never need to rewrite a file to insert or |
|
347 | fashion, which means we never need to rewrite a file to insert or | |
344 | remove data, and can use some simple techniques to avoid the need |
|
348 | remove data, and can use some simple techniques to avoid the need | |
345 | for locking while reading. |
|
349 | for locking while reading. | |
346 |
|
350 | |||
347 | If checkambig, indexfile is opened with checkambig=True at |
|
351 | If checkambig, indexfile is opened with checkambig=True at | |
348 | writing, to avoid file stat ambiguity. |
|
352 | writing, to avoid file stat ambiguity. | |
349 |
|
353 | |||
350 | If mmaplargeindex is True, and an mmapindexthreshold is set, the |
|
354 | If mmaplargeindex is True, and an mmapindexthreshold is set, the | |
351 | index will be mmapped rather than read if it is larger than the |
|
355 | index will be mmapped rather than read if it is larger than the | |
352 | configured threshold. |
|
356 | configured threshold. | |
353 |
|
357 | |||
354 | If censorable is True, the revlog can have censored revisions. |
|
358 | If censorable is True, the revlog can have censored revisions. | |
355 |
|
359 | |||
356 | If `upperboundcomp` is not None, this is the expected maximal gain from |
|
360 | If `upperboundcomp` is not None, this is the expected maximal gain from | |
357 | compression for the data content. |
|
361 | compression for the data content. | |
358 | """ |
|
362 | """ | |
359 |
|
363 | |||
360 | _flagserrorclass = error.RevlogError |
|
364 | _flagserrorclass = error.RevlogError | |
361 |
|
365 | |||
362 | def __init__( |
|
366 | def __init__( | |
363 | self, |
|
367 | self, | |
364 | opener, |
|
368 | opener, | |
365 | indexfile, |
|
369 | indexfile, | |
366 | datafile=None, |
|
370 | datafile=None, | |
367 | checkambig=False, |
|
371 | checkambig=False, | |
368 | mmaplargeindex=False, |
|
372 | mmaplargeindex=False, | |
369 | censorable=False, |
|
373 | censorable=False, | |
370 | upperboundcomp=None, |
|
374 | upperboundcomp=None, | |
371 | ): |
|
375 | ): | |
372 | """ |
|
376 | """ | |
373 | create a revlog object |
|
377 | create a revlog object | |
374 |
|
378 | |||
375 | opener is a function that abstracts the file opening operation |
|
379 | opener is a function that abstracts the file opening operation | |
376 | and can be used to implement COW semantics or the like. |
|
380 | and can be used to implement COW semantics or the like. | |
377 |
|
381 | |||
378 | """ |
|
382 | """ | |
379 | self.upperboundcomp = upperboundcomp |
|
383 | self.upperboundcomp = upperboundcomp | |
380 | self.indexfile = indexfile |
|
384 | self.indexfile = indexfile | |
381 | self.datafile = datafile or (indexfile[:-2] + b".d") |
|
385 | self.datafile = datafile or (indexfile[:-2] + b".d") | |
382 | self.opener = opener |
|
386 | self.opener = opener | |
383 | # When True, indexfile is opened with checkambig=True at writing, to |
|
387 | # When True, indexfile is opened with checkambig=True at writing, to | |
384 | # avoid file stat ambiguity. |
|
388 | # avoid file stat ambiguity. | |
385 | self._checkambig = checkambig |
|
389 | self._checkambig = checkambig | |
386 | self._mmaplargeindex = mmaplargeindex |
|
390 | self._mmaplargeindex = mmaplargeindex | |
387 | self._censorable = censorable |
|
391 | self._censorable = censorable | |
388 | # 3-tuple of (node, rev, text) for a raw revision. |
|
392 | # 3-tuple of (node, rev, text) for a raw revision. | |
389 | self._revisioncache = None |
|
393 | self._revisioncache = None | |
390 | # Maps rev to chain base rev. |
|
394 | # Maps rev to chain base rev. | |
391 | self._chainbasecache = util.lrucachedict(100) |
|
395 | self._chainbasecache = util.lrucachedict(100) | |
392 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. |
|
396 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. | |
393 | self._chunkcache = (0, b'') |
|
397 | self._chunkcache = (0, b'') | |
394 | # How much data to read and cache into the raw revlog data cache. |
|
398 | # How much data to read and cache into the raw revlog data cache. | |
395 | self._chunkcachesize = 65536 |
|
399 | self._chunkcachesize = 65536 | |
396 | self._maxchainlen = None |
|
400 | self._maxchainlen = None | |
397 | self._deltabothparents = True |
|
401 | self._deltabothparents = True | |
398 | self.index = None |
|
402 | self.index = None | |
399 | # Mapping of partial identifiers to full nodes. |
|
403 | # Mapping of partial identifiers to full nodes. | |
400 | self._pcache = {} |
|
404 | self._pcache = {} | |
401 | # Mapping of revision integer to full node. |
|
405 | # Mapping of revision integer to full node. | |
402 | self._nodepos = None |
|
406 | self._nodepos = None | |
403 | self._compengine = b'zlib' |
|
407 | self._compengine = b'zlib' | |
404 | self._compengineopts = {} |
|
408 | self._compengineopts = {} | |
405 | self._maxdeltachainspan = -1 |
|
409 | self._maxdeltachainspan = -1 | |
406 | self._withsparseread = False |
|
410 | self._withsparseread = False | |
407 | self._sparserevlog = False |
|
411 | self._sparserevlog = False | |
408 | self._srdensitythreshold = 0.50 |
|
412 | self._srdensitythreshold = 0.50 | |
409 | self._srmingapsize = 262144 |
|
413 | self._srmingapsize = 262144 | |
410 |
|
414 | |||
411 | # Make copy of flag processors so each revlog instance can support |
|
415 | # Make copy of flag processors so each revlog instance can support | |
412 | # custom flags. |
|
416 | # custom flags. | |
413 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
417 | self._flagprocessors = dict(flagutil.flagprocessors) | |
414 |
|
418 | |||
415 | # 2-tuple of file handles being used for active writing. |
|
419 | # 2-tuple of file handles being used for active writing. | |
416 | self._writinghandles = None |
|
420 | self._writinghandles = None | |
417 |
|
421 | |||
418 | self._loadindex() |
|
422 | self._loadindex() | |
419 |
|
423 | |||
420 | def _loadindex(self): |
|
424 | def _loadindex(self): | |
421 | mmapindexthreshold = None |
|
425 | mmapindexthreshold = None | |
422 | opts = self.opener.options |
|
426 | opts = self.opener.options | |
423 |
|
427 | |||
424 | if b'revlogv2' in opts: |
|
428 | if b'revlogv2' in opts: | |
425 | newversionflags = REVLOGV2 | FLAG_INLINE_DATA |
|
429 | newversionflags = REVLOGV2 | FLAG_INLINE_DATA | |
426 | elif b'revlogv1' in opts: |
|
430 | elif b'revlogv1' in opts: | |
427 | newversionflags = REVLOGV1 | FLAG_INLINE_DATA |
|
431 | newversionflags = REVLOGV1 | FLAG_INLINE_DATA | |
428 | if b'generaldelta' in opts: |
|
432 | if b'generaldelta' in opts: | |
429 | newversionflags |= FLAG_GENERALDELTA |
|
433 | newversionflags |= FLAG_GENERALDELTA | |
430 | elif b'revlogv0' in self.opener.options: |
|
434 | elif b'revlogv0' in self.opener.options: | |
431 | newversionflags = REVLOGV0 |
|
435 | newversionflags = REVLOGV0 | |
432 | else: |
|
436 | else: | |
433 | newversionflags = REVLOG_DEFAULT_VERSION |
|
437 | newversionflags = REVLOG_DEFAULT_VERSION | |
434 |
|
438 | |||
435 | if b'chunkcachesize' in opts: |
|
439 | if b'chunkcachesize' in opts: | |
436 | self._chunkcachesize = opts[b'chunkcachesize'] |
|
440 | self._chunkcachesize = opts[b'chunkcachesize'] | |
437 | if b'maxchainlen' in opts: |
|
441 | if b'maxchainlen' in opts: | |
438 | self._maxchainlen = opts[b'maxchainlen'] |
|
442 | self._maxchainlen = opts[b'maxchainlen'] | |
439 | if b'deltabothparents' in opts: |
|
443 | if b'deltabothparents' in opts: | |
440 | self._deltabothparents = opts[b'deltabothparents'] |
|
444 | self._deltabothparents = opts[b'deltabothparents'] | |
441 | self._lazydelta = bool(opts.get(b'lazydelta', True)) |
|
445 | self._lazydelta = bool(opts.get(b'lazydelta', True)) | |
442 | self._lazydeltabase = False |
|
446 | self._lazydeltabase = False | |
443 | if self._lazydelta: |
|
447 | if self._lazydelta: | |
444 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) |
|
448 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) | |
445 | if b'compengine' in opts: |
|
449 | if b'compengine' in opts: | |
446 | self._compengine = opts[b'compengine'] |
|
450 | self._compengine = opts[b'compengine'] | |
447 | if b'zlib.level' in opts: |
|
451 | if b'zlib.level' in opts: | |
448 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] |
|
452 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] | |
449 | if b'zstd.level' in opts: |
|
453 | if b'zstd.level' in opts: | |
450 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] |
|
454 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] | |
451 | if b'maxdeltachainspan' in opts: |
|
455 | if b'maxdeltachainspan' in opts: | |
452 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] |
|
456 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] | |
453 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: |
|
457 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: | |
454 | mmapindexthreshold = opts[b'mmapindexthreshold'] |
|
458 | mmapindexthreshold = opts[b'mmapindexthreshold'] | |
455 | self.hassidedata = bool(opts.get(b'side-data', False)) |
|
459 | self.hassidedata = bool(opts.get(b'side-data', False)) | |
456 | if self.hassidedata: |
|
460 | if self.hassidedata: | |
457 | self._flagprocessors[REVIDX_SIDEDATA] = sidedatautil.processors |
|
461 | self._flagprocessors[REVIDX_SIDEDATA] = sidedatautil.processors | |
458 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) |
|
462 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) | |
459 | withsparseread = bool(opts.get(b'with-sparse-read', False)) |
|
463 | withsparseread = bool(opts.get(b'with-sparse-read', False)) | |
460 | # sparse-revlog forces sparse-read |
|
464 | # sparse-revlog forces sparse-read | |
461 | self._withsparseread = self._sparserevlog or withsparseread |
|
465 | self._withsparseread = self._sparserevlog or withsparseread | |
462 | if b'sparse-read-density-threshold' in opts: |
|
466 | if b'sparse-read-density-threshold' in opts: | |
463 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] |
|
467 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] | |
464 | if b'sparse-read-min-gap-size' in opts: |
|
468 | if b'sparse-read-min-gap-size' in opts: | |
465 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] |
|
469 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] | |
466 | if opts.get(b'enableellipsis'): |
|
470 | if opts.get(b'enableellipsis'): | |
467 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor |
|
471 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor | |
468 |
|
472 | |||
469 | # revlog v0 doesn't have flag processors |
|
473 | # revlog v0 doesn't have flag processors | |
470 | for flag, processor in pycompat.iteritems( |
|
474 | for flag, processor in pycompat.iteritems( | |
471 | opts.get(b'flagprocessors', {}) |
|
475 | opts.get(b'flagprocessors', {}) | |
472 | ): |
|
476 | ): | |
473 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) |
|
477 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) | |
474 |
|
478 | |||
475 | if self._chunkcachesize <= 0: |
|
479 | if self._chunkcachesize <= 0: | |
476 | raise error.RevlogError( |
|
480 | raise error.RevlogError( | |
477 | _(b'revlog chunk cache size %r is not greater than 0') |
|
481 | _(b'revlog chunk cache size %r is not greater than 0') | |
478 | % self._chunkcachesize |
|
482 | % self._chunkcachesize | |
479 | ) |
|
483 | ) | |
480 | elif self._chunkcachesize & (self._chunkcachesize - 1): |
|
484 | elif self._chunkcachesize & (self._chunkcachesize - 1): | |
481 | raise error.RevlogError( |
|
485 | raise error.RevlogError( | |
482 | _(b'revlog chunk cache size %r is not a power of 2') |
|
486 | _(b'revlog chunk cache size %r is not a power of 2') | |
483 | % self._chunkcachesize |
|
487 | % self._chunkcachesize | |
484 | ) |
|
488 | ) | |
485 |
|
489 | |||
486 | indexdata = b'' |
|
490 | indexdata = b'' | |
487 | self._initempty = True |
|
491 | self._initempty = True | |
488 | try: |
|
492 | try: | |
489 | with self._indexfp() as f: |
|
493 | with self._indexfp() as f: | |
490 | if ( |
|
494 | if ( | |
491 | mmapindexthreshold is not None |
|
495 | mmapindexthreshold is not None | |
492 | and self.opener.fstat(f).st_size >= mmapindexthreshold |
|
496 | and self.opener.fstat(f).st_size >= mmapindexthreshold | |
493 | ): |
|
497 | ): | |
494 | # TODO: should .close() to release resources without |
|
498 | # TODO: should .close() to release resources without | |
495 | # relying on Python GC |
|
499 | # relying on Python GC | |
496 | indexdata = util.buffer(util.mmapread(f)) |
|
500 | indexdata = util.buffer(util.mmapread(f)) | |
497 | else: |
|
501 | else: | |
498 | indexdata = f.read() |
|
502 | indexdata = f.read() | |
499 | if len(indexdata) > 0: |
|
503 | if len(indexdata) > 0: | |
500 | versionflags = versionformat_unpack(indexdata[:4])[0] |
|
504 | versionflags = versionformat_unpack(indexdata[:4])[0] | |
501 | self._initempty = False |
|
505 | self._initempty = False | |
502 | else: |
|
506 | else: | |
503 | versionflags = newversionflags |
|
507 | versionflags = newversionflags | |
504 | except IOError as inst: |
|
508 | except IOError as inst: | |
505 | if inst.errno != errno.ENOENT: |
|
509 | if inst.errno != errno.ENOENT: | |
506 | raise |
|
510 | raise | |
507 |
|
511 | |||
508 | versionflags = newversionflags |
|
512 | versionflags = newversionflags | |
509 |
|
513 | |||
510 | self.version = versionflags |
|
514 | self.version = versionflags | |
511 |
|
515 | |||
512 | flags = versionflags & ~0xFFFF |
|
516 | flags = versionflags & ~0xFFFF | |
513 | fmt = versionflags & 0xFFFF |
|
517 | fmt = versionflags & 0xFFFF | |
514 |
|
518 | |||
515 | if fmt == REVLOGV0: |
|
519 | if fmt == REVLOGV0: | |
516 | if flags: |
|
520 | if flags: | |
517 | raise error.RevlogError( |
|
521 | raise error.RevlogError( | |
518 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
522 | _(b'unknown flags (%#04x) in version %d revlog %s') | |
519 | % (flags >> 16, fmt, self.indexfile) |
|
523 | % (flags >> 16, fmt, self.indexfile) | |
520 | ) |
|
524 | ) | |
521 |
|
525 | |||
522 | self._inline = False |
|
526 | self._inline = False | |
523 | self._generaldelta = False |
|
527 | self._generaldelta = False | |
524 |
|
528 | |||
525 | elif fmt == REVLOGV1: |
|
529 | elif fmt == REVLOGV1: | |
526 | if flags & ~REVLOGV1_FLAGS: |
|
530 | if flags & ~REVLOGV1_FLAGS: | |
527 | raise error.RevlogError( |
|
531 | raise error.RevlogError( | |
528 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
532 | _(b'unknown flags (%#04x) in version %d revlog %s') | |
529 | % (flags >> 16, fmt, self.indexfile) |
|
533 | % (flags >> 16, fmt, self.indexfile) | |
530 | ) |
|
534 | ) | |
531 |
|
535 | |||
532 | self._inline = versionflags & FLAG_INLINE_DATA |
|
536 | self._inline = versionflags & FLAG_INLINE_DATA | |
533 | self._generaldelta = versionflags & FLAG_GENERALDELTA |
|
537 | self._generaldelta = versionflags & FLAG_GENERALDELTA | |
534 |
|
538 | |||
535 | elif fmt == REVLOGV2: |
|
539 | elif fmt == REVLOGV2: | |
536 | if flags & ~REVLOGV2_FLAGS: |
|
540 | if flags & ~REVLOGV2_FLAGS: | |
537 | raise error.RevlogError( |
|
541 | raise error.RevlogError( | |
538 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
542 | _(b'unknown flags (%#04x) in version %d revlog %s') | |
539 | % (flags >> 16, fmt, self.indexfile) |
|
543 | % (flags >> 16, fmt, self.indexfile) | |
540 | ) |
|
544 | ) | |
541 |
|
545 | |||
542 | self._inline = versionflags & FLAG_INLINE_DATA |
|
546 | self._inline = versionflags & FLAG_INLINE_DATA | |
543 | # generaldelta implied by version 2 revlogs. |
|
547 | # generaldelta implied by version 2 revlogs. | |
544 | self._generaldelta = True |
|
548 | self._generaldelta = True | |
545 |
|
549 | |||
546 | else: |
|
550 | else: | |
547 | raise error.RevlogError( |
|
551 | raise error.RevlogError( | |
548 | _(b'unknown version (%d) in revlog %s') % (fmt, self.indexfile) |
|
552 | _(b'unknown version (%d) in revlog %s') % (fmt, self.indexfile) | |
549 | ) |
|
553 | ) | |
550 | # sparse-revlog can't be on without general-delta (issue6056) |
|
554 | # sparse-revlog can't be on without general-delta (issue6056) | |
551 | if not self._generaldelta: |
|
555 | if not self._generaldelta: | |
552 | self._sparserevlog = False |
|
556 | self._sparserevlog = False | |
553 |
|
557 | |||
554 | self._storedeltachains = True |
|
558 | self._storedeltachains = True | |
555 |
|
559 | |||
556 | self._io = revlogio() |
|
560 | self._io = revlogio() | |
557 | if self.version == REVLOGV0: |
|
561 | if self.version == REVLOGV0: | |
558 | self._io = revlogoldio() |
|
562 | self._io = revlogoldio() | |
559 | try: |
|
563 | try: | |
560 | d = self._io.parseindex(indexdata, self._inline) |
|
564 | d = self._io.parseindex(indexdata, self._inline) | |
561 | except (ValueError, IndexError): |
|
565 | except (ValueError, IndexError): | |
562 | raise error.RevlogError( |
|
566 | raise error.RevlogError( | |
563 | _(b"index %s is corrupted") % self.indexfile |
|
567 | _(b"index %s is corrupted") % self.indexfile | |
564 | ) |
|
568 | ) | |
565 | self.index, self._chunkcache = d |
|
569 | self.index, self._chunkcache = d | |
566 | self.nodemap = self.index.nodemap |
|
570 | self.nodemap = self.index.nodemap | |
567 | if not self._chunkcache: |
|
571 | if not self._chunkcache: | |
568 | self._chunkclear() |
|
572 | self._chunkclear() | |
569 | # revnum -> (chain-length, sum-delta-length) |
|
573 | # revnum -> (chain-length, sum-delta-length) | |
570 | self._chaininfocache = {} |
|
574 | self._chaininfocache = {} | |
571 | # revlog header -> revlog compressor |
|
575 | # revlog header -> revlog compressor | |
572 | self._decompressors = {} |
|
576 | self._decompressors = {} | |
573 |
|
577 | |||
574 | @util.propertycache |
|
578 | @util.propertycache | |
575 | def _compressor(self): |
|
579 | def _compressor(self): | |
576 | engine = util.compengines[self._compengine] |
|
580 | engine = util.compengines[self._compengine] | |
577 | return engine.revlogcompressor(self._compengineopts) |
|
581 | return engine.revlogcompressor(self._compengineopts) | |
578 |
|
582 | |||
579 | def _indexfp(self, mode=b'r'): |
|
583 | def _indexfp(self, mode=b'r'): | |
580 | """file object for the revlog's index file""" |
|
584 | """file object for the revlog's index file""" | |
581 | args = {'mode': mode} |
|
585 | args = {'mode': mode} | |
582 | if mode != b'r': |
|
586 | if mode != b'r': | |
583 | args['checkambig'] = self._checkambig |
|
587 | args['checkambig'] = self._checkambig | |
584 | if mode == b'w': |
|
588 | if mode == b'w': | |
585 | args['atomictemp'] = True |
|
589 | args['atomictemp'] = True | |
586 | return self.opener(self.indexfile, **args) |
|
590 | return self.opener(self.indexfile, **args) | |
587 |
|
591 | |||
588 | def _datafp(self, mode=b'r'): |
|
592 | def _datafp(self, mode=b'r'): | |
589 | """file object for the revlog's data file""" |
|
593 | """file object for the revlog's data file""" | |
590 | return self.opener(self.datafile, mode=mode) |
|
594 | return self.opener(self.datafile, mode=mode) | |
591 |
|
595 | |||
592 | @contextlib.contextmanager |
|
596 | @contextlib.contextmanager | |
593 | def _datareadfp(self, existingfp=None): |
|
597 | def _datareadfp(self, existingfp=None): | |
594 | """file object suitable to read data""" |
|
598 | """file object suitable to read data""" | |
595 | # Use explicit file handle, if given. |
|
599 | # Use explicit file handle, if given. | |
596 | if existingfp is not None: |
|
600 | if existingfp is not None: | |
597 | yield existingfp |
|
601 | yield existingfp | |
598 |
|
602 | |||
599 | # Use a file handle being actively used for writes, if available. |
|
603 | # Use a file handle being actively used for writes, if available. | |
600 | # There is some danger to doing this because reads will seek the |
|
604 | # There is some danger to doing this because reads will seek the | |
601 | # file. However, _writeentry() performs a SEEK_END before all writes, |
|
605 | # file. However, _writeentry() performs a SEEK_END before all writes, | |
602 | # so we should be safe. |
|
606 | # so we should be safe. | |
603 | elif self._writinghandles: |
|
607 | elif self._writinghandles: | |
604 | if self._inline: |
|
608 | if self._inline: | |
605 | yield self._writinghandles[0] |
|
609 | yield self._writinghandles[0] | |
606 | else: |
|
610 | else: | |
607 | yield self._writinghandles[1] |
|
611 | yield self._writinghandles[1] | |
608 |
|
612 | |||
609 | # Otherwise open a new file handle. |
|
613 | # Otherwise open a new file handle. | |
610 | else: |
|
614 | else: | |
611 | if self._inline: |
|
615 | if self._inline: | |
612 | func = self._indexfp |
|
616 | func = self._indexfp | |
613 | else: |
|
617 | else: | |
614 | func = self._datafp |
|
618 | func = self._datafp | |
615 | with func() as fp: |
|
619 | with func() as fp: | |
616 | yield fp |
|
620 | yield fp | |
617 |
|
621 | |||
618 | def tiprev(self): |
|
622 | def tiprev(self): | |
619 | return len(self.index) - 1 |
|
623 | return len(self.index) - 1 | |
620 |
|
624 | |||
621 | def tip(self): |
|
625 | def tip(self): | |
622 | return self.node(self.tiprev()) |
|
626 | return self.node(self.tiprev()) | |
623 |
|
627 | |||
624 | def __contains__(self, rev): |
|
628 | def __contains__(self, rev): | |
625 | return 0 <= rev < len(self) |
|
629 | return 0 <= rev < len(self) | |
626 |
|
630 | |||
627 | def __len__(self): |
|
631 | def __len__(self): | |
628 | return len(self.index) |
|
632 | return len(self.index) | |
629 |
|
633 | |||
630 | def __iter__(self): |
|
634 | def __iter__(self): | |
631 | return iter(pycompat.xrange(len(self))) |
|
635 | return iter(pycompat.xrange(len(self))) | |
632 |
|
636 | |||
633 | def revs(self, start=0, stop=None): |
|
637 | def revs(self, start=0, stop=None): | |
634 | """iterate over all rev in this revlog (from start to stop)""" |
|
638 | """iterate over all rev in this revlog (from start to stop)""" | |
635 | return storageutil.iterrevs(len(self), start=start, stop=stop) |
|
639 | return storageutil.iterrevs(len(self), start=start, stop=stop) | |
636 |
|
640 | |||
637 | @util.propertycache |
|
641 | @util.propertycache | |
638 | def nodemap(self): |
|
642 | def nodemap(self): | |
639 | if self.index: |
|
643 | if self.index: | |
640 | # populate mapping down to the initial node |
|
644 | # populate mapping down to the initial node | |
641 | node0 = self.index[0][7] # get around changelog filtering |
|
645 | node0 = self.index[0][7] # get around changelog filtering | |
642 | self.rev(node0) |
|
646 | self.rev(node0) | |
643 | return self.index.nodemap |
|
647 | return self.index.nodemap | |
644 |
|
648 | |||
645 | @property |
|
649 | @property | |
646 | def _nodecache(self): |
|
650 | def _nodecache(self): | |
647 | msg = "revlog._nodecache is deprecated, use revlog.index.nodemap" |
|
651 | msg = "revlog._nodecache is deprecated, use revlog.index.nodemap" | |
648 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
652 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) | |
649 | return self.index.nodemap |
|
653 | return self.index.nodemap | |
650 |
|
654 | |||
651 | def hasnode(self, node): |
|
655 | def hasnode(self, node): | |
652 | try: |
|
656 | try: | |
653 | self.rev(node) |
|
657 | self.rev(node) | |
654 | return True |
|
658 | return True | |
655 | except KeyError: |
|
659 | except KeyError: | |
656 | return False |
|
660 | return False | |
657 |
|
661 | |||
658 | def candelta(self, baserev, rev): |
|
662 | def candelta(self, baserev, rev): | |
659 | """whether two revisions (baserev, rev) can be delta-ed or not""" |
|
663 | """whether two revisions (baserev, rev) can be delta-ed or not""" | |
660 | # Disable delta if either rev requires a content-changing flag |
|
664 | # Disable delta if either rev requires a content-changing flag | |
661 | # processor (ex. LFS). This is because such flag processor can alter |
|
665 | # processor (ex. LFS). This is because such flag processor can alter | |
662 | # the rawtext content that the delta will be based on, and two clients |
|
666 | # the rawtext content that the delta will be based on, and two clients | |
663 | # could have a same revlog node with different flags (i.e. different |
|
667 | # could have a same revlog node with different flags (i.e. different | |
664 | # rawtext contents) and the delta could be incompatible. |
|
668 | # rawtext contents) and the delta could be incompatible. | |
665 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( |
|
669 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( | |
666 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS |
|
670 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS | |
667 | ): |
|
671 | ): | |
668 | return False |
|
672 | return False | |
669 | return True |
|
673 | return True | |
670 |
|
674 | |||
671 | def clearcaches(self): |
|
675 | def clearcaches(self): | |
672 | self._revisioncache = None |
|
676 | self._revisioncache = None | |
673 | self._chainbasecache.clear() |
|
677 | self._chainbasecache.clear() | |
674 | self._chunkcache = (0, b'') |
|
678 | self._chunkcache = (0, b'') | |
675 | self._pcache = {} |
|
679 | self._pcache = {} | |
676 | self.index.clearcaches() |
|
680 | self.index.clearcaches() | |
677 |
|
681 | |||
678 | def rev(self, node): |
|
682 | def rev(self, node): | |
679 | try: |
|
683 | try: | |
680 | return self.index.nodemap[node] |
|
684 | return self.index.nodemap[node] | |
681 | except TypeError: |
|
685 | except TypeError: | |
682 | raise |
|
686 | raise | |
683 | except error.RevlogError: |
|
687 | except error.RevlogError: | |
684 | # parsers.c radix tree lookup failed |
|
688 | # parsers.c radix tree lookup failed | |
685 | if node == wdirid or node in wdirfilenodeids: |
|
689 | if node == wdirid or node in wdirfilenodeids: | |
686 | raise error.WdirUnsupported |
|
690 | raise error.WdirUnsupported | |
687 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
691 | raise error.LookupError(node, self.indexfile, _(b'no node')) | |
688 |
|
692 | |||
689 | # Accessors for index entries. |
|
693 | # Accessors for index entries. | |
690 |
|
694 | |||
691 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes |
|
695 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes | |
692 | # are flags. |
|
696 | # are flags. | |
693 | def start(self, rev): |
|
697 | def start(self, rev): | |
694 | return int(self.index[rev][0] >> 16) |
|
698 | return int(self.index[rev][0] >> 16) | |
695 |
|
699 | |||
696 | def flags(self, rev): |
|
700 | def flags(self, rev): | |
697 | return self.index[rev][0] & 0xFFFF |
|
701 | return self.index[rev][0] & 0xFFFF | |
698 |
|
702 | |||
699 | def length(self, rev): |
|
703 | def length(self, rev): | |
700 | return self.index[rev][1] |
|
704 | return self.index[rev][1] | |
701 |
|
705 | |||
702 | def rawsize(self, rev): |
|
706 | def rawsize(self, rev): | |
703 | """return the length of the uncompressed text for a given revision""" |
|
707 | """return the length of the uncompressed text for a given revision""" | |
704 | l = self.index[rev][2] |
|
708 | l = self.index[rev][2] | |
705 | if l >= 0: |
|
709 | if l >= 0: | |
706 | return l |
|
710 | return l | |
707 |
|
711 | |||
708 | t = self.rawdata(rev) |
|
712 | t = self.rawdata(rev) | |
709 | return len(t) |
|
713 | return len(t) | |
710 |
|
714 | |||
711 | def size(self, rev): |
|
715 | def size(self, rev): | |
712 | """length of non-raw text (processed by a "read" flag processor)""" |
|
716 | """length of non-raw text (processed by a "read" flag processor)""" | |
713 | # fast path: if no "read" flag processor could change the content, |
|
717 | # fast path: if no "read" flag processor could change the content, | |
714 | # size is rawsize. note: ELLIPSIS is known to not change the content. |
|
718 | # size is rawsize. note: ELLIPSIS is known to not change the content. | |
715 | flags = self.flags(rev) |
|
719 | flags = self.flags(rev) | |
716 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: |
|
720 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: | |
717 | return self.rawsize(rev) |
|
721 | return self.rawsize(rev) | |
718 |
|
722 | |||
719 | return len(self.revision(rev, raw=False)) |
|
723 | return len(self.revision(rev, raw=False)) | |
720 |
|
724 | |||
721 | def chainbase(self, rev): |
|
725 | def chainbase(self, rev): | |
722 | base = self._chainbasecache.get(rev) |
|
726 | base = self._chainbasecache.get(rev) | |
723 | if base is not None: |
|
727 | if base is not None: | |
724 | return base |
|
728 | return base | |
725 |
|
729 | |||
726 | index = self.index |
|
730 | index = self.index | |
727 | iterrev = rev |
|
731 | iterrev = rev | |
728 | base = index[iterrev][3] |
|
732 | base = index[iterrev][3] | |
729 | while base != iterrev: |
|
733 | while base != iterrev: | |
730 | iterrev = base |
|
734 | iterrev = base | |
731 | base = index[iterrev][3] |
|
735 | base = index[iterrev][3] | |
732 |
|
736 | |||
733 | self._chainbasecache[rev] = base |
|
737 | self._chainbasecache[rev] = base | |
734 | return base |
|
738 | return base | |
735 |
|
739 | |||
736 | def linkrev(self, rev): |
|
740 | def linkrev(self, rev): | |
737 | return self.index[rev][4] |
|
741 | return self.index[rev][4] | |
738 |
|
742 | |||
739 | def parentrevs(self, rev): |
|
743 | def parentrevs(self, rev): | |
740 | try: |
|
744 | try: | |
741 | entry = self.index[rev] |
|
745 | entry = self.index[rev] | |
742 | except IndexError: |
|
746 | except IndexError: | |
743 | if rev == wdirrev: |
|
747 | if rev == wdirrev: | |
744 | raise error.WdirUnsupported |
|
748 | raise error.WdirUnsupported | |
745 | raise |
|
749 | raise | |
746 |
|
750 | |||
747 | return entry[5], entry[6] |
|
751 | return entry[5], entry[6] | |
748 |
|
752 | |||
749 | # fast parentrevs(rev) where rev isn't filtered |
|
753 | # fast parentrevs(rev) where rev isn't filtered | |
750 | _uncheckedparentrevs = parentrevs |
|
754 | _uncheckedparentrevs = parentrevs | |
751 |
|
755 | |||
752 | def node(self, rev): |
|
756 | def node(self, rev): | |
753 | try: |
|
757 | try: | |
754 | return self.index[rev][7] |
|
758 | return self.index[rev][7] | |
755 | except IndexError: |
|
759 | except IndexError: | |
756 | if rev == wdirrev: |
|
760 | if rev == wdirrev: | |
757 | raise error.WdirUnsupported |
|
761 | raise error.WdirUnsupported | |
758 | raise |
|
762 | raise | |
759 |
|
763 | |||
760 | # Derived from index values. |
|
764 | # Derived from index values. | |
761 |
|
765 | |||
762 | def end(self, rev): |
|
766 | def end(self, rev): | |
763 | return self.start(rev) + self.length(rev) |
|
767 | return self.start(rev) + self.length(rev) | |
764 |
|
768 | |||
765 | def parents(self, node): |
|
769 | def parents(self, node): | |
766 | i = self.index |
|
770 | i = self.index | |
767 | d = i[self.rev(node)] |
|
771 | d = i[self.rev(node)] | |
768 | return i[d[5]][7], i[d[6]][7] # map revisions to nodes inline |
|
772 | return i[d[5]][7], i[d[6]][7] # map revisions to nodes inline | |
769 |
|
773 | |||
770 | def chainlen(self, rev): |
|
774 | def chainlen(self, rev): | |
771 | return self._chaininfo(rev)[0] |
|
775 | return self._chaininfo(rev)[0] | |
772 |
|
776 | |||
773 | def _chaininfo(self, rev): |
|
777 | def _chaininfo(self, rev): | |
774 | chaininfocache = self._chaininfocache |
|
778 | chaininfocache = self._chaininfocache | |
775 | if rev in chaininfocache: |
|
779 | if rev in chaininfocache: | |
776 | return chaininfocache[rev] |
|
780 | return chaininfocache[rev] | |
777 | index = self.index |
|
781 | index = self.index | |
778 | generaldelta = self._generaldelta |
|
782 | generaldelta = self._generaldelta | |
779 | iterrev = rev |
|
783 | iterrev = rev | |
780 | e = index[iterrev] |
|
784 | e = index[iterrev] | |
781 | clen = 0 |
|
785 | clen = 0 | |
782 | compresseddeltalen = 0 |
|
786 | compresseddeltalen = 0 | |
783 | while iterrev != e[3]: |
|
787 | while iterrev != e[3]: | |
784 | clen += 1 |
|
788 | clen += 1 | |
785 | compresseddeltalen += e[1] |
|
789 | compresseddeltalen += e[1] | |
786 | if generaldelta: |
|
790 | if generaldelta: | |
787 | iterrev = e[3] |
|
791 | iterrev = e[3] | |
788 | else: |
|
792 | else: | |
789 | iterrev -= 1 |
|
793 | iterrev -= 1 | |
790 | if iterrev in chaininfocache: |
|
794 | if iterrev in chaininfocache: | |
791 | t = chaininfocache[iterrev] |
|
795 | t = chaininfocache[iterrev] | |
792 | clen += t[0] |
|
796 | clen += t[0] | |
793 | compresseddeltalen += t[1] |
|
797 | compresseddeltalen += t[1] | |
794 | break |
|
798 | break | |
795 | e = index[iterrev] |
|
799 | e = index[iterrev] | |
796 | else: |
|
800 | else: | |
797 | # Add text length of base since decompressing that also takes |
|
801 | # Add text length of base since decompressing that also takes | |
798 | # work. For cache hits the length is already included. |
|
802 | # work. For cache hits the length is already included. | |
799 | compresseddeltalen += e[1] |
|
803 | compresseddeltalen += e[1] | |
800 | r = (clen, compresseddeltalen) |
|
804 | r = (clen, compresseddeltalen) | |
801 | chaininfocache[rev] = r |
|
805 | chaininfocache[rev] = r | |
802 | return r |
|
806 | return r | |
803 |
|
807 | |||
804 | def _deltachain(self, rev, stoprev=None): |
|
808 | def _deltachain(self, rev, stoprev=None): | |
805 | """Obtain the delta chain for a revision. |
|
809 | """Obtain the delta chain for a revision. | |
806 |
|
810 | |||
807 | ``stoprev`` specifies a revision to stop at. If not specified, we |
|
811 | ``stoprev`` specifies a revision to stop at. If not specified, we | |
808 | stop at the base of the chain. |
|
812 | stop at the base of the chain. | |
809 |
|
813 | |||
810 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of |
|
814 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of | |
811 | revs in ascending order and ``stopped`` is a bool indicating whether |
|
815 | revs in ascending order and ``stopped`` is a bool indicating whether | |
812 | ``stoprev`` was hit. |
|
816 | ``stoprev`` was hit. | |
813 | """ |
|
817 | """ | |
814 | # Try C implementation. |
|
818 | # Try C implementation. | |
815 | try: |
|
819 | try: | |
816 | return self.index.deltachain(rev, stoprev, self._generaldelta) |
|
820 | return self.index.deltachain(rev, stoprev, self._generaldelta) | |
817 | except AttributeError: |
|
821 | except AttributeError: | |
818 | pass |
|
822 | pass | |
819 |
|
823 | |||
820 | chain = [] |
|
824 | chain = [] | |
821 |
|
825 | |||
822 | # Alias to prevent attribute lookup in tight loop. |
|
826 | # Alias to prevent attribute lookup in tight loop. | |
823 | index = self.index |
|
827 | index = self.index | |
824 | generaldelta = self._generaldelta |
|
828 | generaldelta = self._generaldelta | |
825 |
|
829 | |||
826 | iterrev = rev |
|
830 | iterrev = rev | |
827 | e = index[iterrev] |
|
831 | e = index[iterrev] | |
828 | while iterrev != e[3] and iterrev != stoprev: |
|
832 | while iterrev != e[3] and iterrev != stoprev: | |
829 | chain.append(iterrev) |
|
833 | chain.append(iterrev) | |
830 | if generaldelta: |
|
834 | if generaldelta: | |
831 | iterrev = e[3] |
|
835 | iterrev = e[3] | |
832 | else: |
|
836 | else: | |
833 | iterrev -= 1 |
|
837 | iterrev -= 1 | |
834 | e = index[iterrev] |
|
838 | e = index[iterrev] | |
835 |
|
839 | |||
836 | if iterrev == stoprev: |
|
840 | if iterrev == stoprev: | |
837 | stopped = True |
|
841 | stopped = True | |
838 | else: |
|
842 | else: | |
839 | chain.append(iterrev) |
|
843 | chain.append(iterrev) | |
840 | stopped = False |
|
844 | stopped = False | |
841 |
|
845 | |||
842 | chain.reverse() |
|
846 | chain.reverse() | |
843 | return chain, stopped |
|
847 | return chain, stopped | |
844 |
|
848 | |||
845 | def ancestors(self, revs, stoprev=0, inclusive=False): |
|
849 | def ancestors(self, revs, stoprev=0, inclusive=False): | |
846 | """Generate the ancestors of 'revs' in reverse revision order. |
|
850 | """Generate the ancestors of 'revs' in reverse revision order. | |
847 | Does not generate revs lower than stoprev. |
|
851 | Does not generate revs lower than stoprev. | |
848 |
|
852 | |||
849 | See the documentation for ancestor.lazyancestors for more details.""" |
|
853 | See the documentation for ancestor.lazyancestors for more details.""" | |
850 |
|
854 | |||
851 | # first, make sure start revisions aren't filtered |
|
855 | # first, make sure start revisions aren't filtered | |
852 | revs = list(revs) |
|
856 | revs = list(revs) | |
853 | checkrev = self.node |
|
857 | checkrev = self.node | |
854 | for r in revs: |
|
858 | for r in revs: | |
855 | checkrev(r) |
|
859 | checkrev(r) | |
856 | # and we're sure ancestors aren't filtered as well |
|
860 | # and we're sure ancestors aren't filtered as well | |
857 |
|
861 | |||
858 | if rustancestor is not None: |
|
862 | if rustancestor is not None: | |
859 | lazyancestors = rustancestor.LazyAncestors |
|
863 | lazyancestors = rustancestor.LazyAncestors | |
860 | arg = self.index |
|
864 | arg = self.index | |
861 | elif util.safehasattr(parsers, b'rustlazyancestors'): |
|
865 | elif util.safehasattr(parsers, b'rustlazyancestors'): | |
862 | lazyancestors = ancestor.rustlazyancestors |
|
866 | lazyancestors = ancestor.rustlazyancestors | |
863 | arg = self.index |
|
867 | arg = self.index | |
864 | else: |
|
868 | else: | |
865 | lazyancestors = ancestor.lazyancestors |
|
869 | lazyancestors = ancestor.lazyancestors | |
866 | arg = self._uncheckedparentrevs |
|
870 | arg = self._uncheckedparentrevs | |
867 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) |
|
871 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) | |
868 |
|
872 | |||
869 | def descendants(self, revs): |
|
873 | def descendants(self, revs): | |
870 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) |
|
874 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) | |
871 |
|
875 | |||
872 | def findcommonmissing(self, common=None, heads=None): |
|
876 | def findcommonmissing(self, common=None, heads=None): | |
873 | """Return a tuple of the ancestors of common and the ancestors of heads |
|
877 | """Return a tuple of the ancestors of common and the ancestors of heads | |
874 | that are not ancestors of common. In revset terminology, we return the |
|
878 | that are not ancestors of common. In revset terminology, we return the | |
875 | tuple: |
|
879 | tuple: | |
876 |
|
880 | |||
877 | ::common, (::heads) - (::common) |
|
881 | ::common, (::heads) - (::common) | |
878 |
|
882 | |||
879 | The list is sorted by revision number, meaning it is |
|
883 | The list is sorted by revision number, meaning it is | |
880 | topologically sorted. |
|
884 | topologically sorted. | |
881 |
|
885 | |||
882 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
886 | 'heads' and 'common' are both lists of node IDs. If heads is | |
883 | not supplied, uses all of the revlog's heads. If common is not |
|
887 | not supplied, uses all of the revlog's heads. If common is not | |
884 | supplied, uses nullid.""" |
|
888 | supplied, uses nullid.""" | |
885 | if common is None: |
|
889 | if common is None: | |
886 | common = [nullid] |
|
890 | common = [nullid] | |
887 | if heads is None: |
|
891 | if heads is None: | |
888 | heads = self.heads() |
|
892 | heads = self.heads() | |
889 |
|
893 | |||
890 | common = [self.rev(n) for n in common] |
|
894 | common = [self.rev(n) for n in common] | |
891 | heads = [self.rev(n) for n in heads] |
|
895 | heads = [self.rev(n) for n in heads] | |
892 |
|
896 | |||
893 | # we want the ancestors, but inclusive |
|
897 | # we want the ancestors, but inclusive | |
894 | class lazyset(object): |
|
898 | class lazyset(object): | |
895 | def __init__(self, lazyvalues): |
|
899 | def __init__(self, lazyvalues): | |
896 | self.addedvalues = set() |
|
900 | self.addedvalues = set() | |
897 | self.lazyvalues = lazyvalues |
|
901 | self.lazyvalues = lazyvalues | |
898 |
|
902 | |||
899 | def __contains__(self, value): |
|
903 | def __contains__(self, value): | |
900 | return value in self.addedvalues or value in self.lazyvalues |
|
904 | return value in self.addedvalues or value in self.lazyvalues | |
901 |
|
905 | |||
902 | def __iter__(self): |
|
906 | def __iter__(self): | |
903 | added = self.addedvalues |
|
907 | added = self.addedvalues | |
904 | for r in added: |
|
908 | for r in added: | |
905 | yield r |
|
909 | yield r | |
906 | for r in self.lazyvalues: |
|
910 | for r in self.lazyvalues: | |
907 | if not r in added: |
|
911 | if not r in added: | |
908 | yield r |
|
912 | yield r | |
909 |
|
913 | |||
910 | def add(self, value): |
|
914 | def add(self, value): | |
911 | self.addedvalues.add(value) |
|
915 | self.addedvalues.add(value) | |
912 |
|
916 | |||
913 | def update(self, values): |
|
917 | def update(self, values): | |
914 | self.addedvalues.update(values) |
|
918 | self.addedvalues.update(values) | |
915 |
|
919 | |||
916 | has = lazyset(self.ancestors(common)) |
|
920 | has = lazyset(self.ancestors(common)) | |
917 | has.add(nullrev) |
|
921 | has.add(nullrev) | |
918 | has.update(common) |
|
922 | has.update(common) | |
919 |
|
923 | |||
920 | # take all ancestors from heads that aren't in has |
|
924 | # take all ancestors from heads that aren't in has | |
921 | missing = set() |
|
925 | missing = set() | |
922 | visit = collections.deque(r for r in heads if r not in has) |
|
926 | visit = collections.deque(r for r in heads if r not in has) | |
923 | while visit: |
|
927 | while visit: | |
924 | r = visit.popleft() |
|
928 | r = visit.popleft() | |
925 | if r in missing: |
|
929 | if r in missing: | |
926 | continue |
|
930 | continue | |
927 | else: |
|
931 | else: | |
928 | missing.add(r) |
|
932 | missing.add(r) | |
929 | for p in self.parentrevs(r): |
|
933 | for p in self.parentrevs(r): | |
930 | if p not in has: |
|
934 | if p not in has: | |
931 | visit.append(p) |
|
935 | visit.append(p) | |
932 | missing = list(missing) |
|
936 | missing = list(missing) | |
933 | missing.sort() |
|
937 | missing.sort() | |
934 | return has, [self.node(miss) for miss in missing] |
|
938 | return has, [self.node(miss) for miss in missing] | |
935 |
|
939 | |||
936 | def incrementalmissingrevs(self, common=None): |
|
940 | def incrementalmissingrevs(self, common=None): | |
937 | """Return an object that can be used to incrementally compute the |
|
941 | """Return an object that can be used to incrementally compute the | |
938 | revision numbers of the ancestors of arbitrary sets that are not |
|
942 | revision numbers of the ancestors of arbitrary sets that are not | |
939 | ancestors of common. This is an ancestor.incrementalmissingancestors |
|
943 | ancestors of common. This is an ancestor.incrementalmissingancestors | |
940 | object. |
|
944 | object. | |
941 |
|
945 | |||
942 | 'common' is a list of revision numbers. If common is not supplied, uses |
|
946 | 'common' is a list of revision numbers. If common is not supplied, uses | |
943 | nullrev. |
|
947 | nullrev. | |
944 | """ |
|
948 | """ | |
945 | if common is None: |
|
949 | if common is None: | |
946 | common = [nullrev] |
|
950 | common = [nullrev] | |
947 |
|
951 | |||
948 | if rustancestor is not None: |
|
952 | if rustancestor is not None: | |
949 | return rustancestor.MissingAncestors(self.index, common) |
|
953 | return rustancestor.MissingAncestors(self.index, common) | |
950 | return ancestor.incrementalmissingancestors(self.parentrevs, common) |
|
954 | return ancestor.incrementalmissingancestors(self.parentrevs, common) | |
951 |
|
955 | |||
952 | def findmissingrevs(self, common=None, heads=None): |
|
956 | def findmissingrevs(self, common=None, heads=None): | |
953 | """Return the revision numbers of the ancestors of heads that |
|
957 | """Return the revision numbers of the ancestors of heads that | |
954 | are not ancestors of common. |
|
958 | are not ancestors of common. | |
955 |
|
959 | |||
956 | More specifically, return a list of revision numbers corresponding to |
|
960 | More specifically, return a list of revision numbers corresponding to | |
957 | nodes N such that every N satisfies the following constraints: |
|
961 | nodes N such that every N satisfies the following constraints: | |
958 |
|
962 | |||
959 | 1. N is an ancestor of some node in 'heads' |
|
963 | 1. N is an ancestor of some node in 'heads' | |
960 | 2. N is not an ancestor of any node in 'common' |
|
964 | 2. N is not an ancestor of any node in 'common' | |
961 |
|
965 | |||
962 | The list is sorted by revision number, meaning it is |
|
966 | The list is sorted by revision number, meaning it is | |
963 | topologically sorted. |
|
967 | topologically sorted. | |
964 |
|
968 | |||
965 | 'heads' and 'common' are both lists of revision numbers. If heads is |
|
969 | 'heads' and 'common' are both lists of revision numbers. If heads is | |
966 | not supplied, uses all of the revlog's heads. If common is not |
|
970 | not supplied, uses all of the revlog's heads. If common is not | |
967 | supplied, uses nullid.""" |
|
971 | supplied, uses nullid.""" | |
968 | if common is None: |
|
972 | if common is None: | |
969 | common = [nullrev] |
|
973 | common = [nullrev] | |
970 | if heads is None: |
|
974 | if heads is None: | |
971 | heads = self.headrevs() |
|
975 | heads = self.headrevs() | |
972 |
|
976 | |||
973 | inc = self.incrementalmissingrevs(common=common) |
|
977 | inc = self.incrementalmissingrevs(common=common) | |
974 | return inc.missingancestors(heads) |
|
978 | return inc.missingancestors(heads) | |
975 |
|
979 | |||
976 | def findmissing(self, common=None, heads=None): |
|
980 | def findmissing(self, common=None, heads=None): | |
977 | """Return the ancestors of heads that are not ancestors of common. |
|
981 | """Return the ancestors of heads that are not ancestors of common. | |
978 |
|
982 | |||
979 | More specifically, return a list of nodes N such that every N |
|
983 | More specifically, return a list of nodes N such that every N | |
980 | satisfies the following constraints: |
|
984 | satisfies the following constraints: | |
981 |
|
985 | |||
982 | 1. N is an ancestor of some node in 'heads' |
|
986 | 1. N is an ancestor of some node in 'heads' | |
983 | 2. N is not an ancestor of any node in 'common' |
|
987 | 2. N is not an ancestor of any node in 'common' | |
984 |
|
988 | |||
985 | The list is sorted by revision number, meaning it is |
|
989 | The list is sorted by revision number, meaning it is | |
986 | topologically sorted. |
|
990 | topologically sorted. | |
987 |
|
991 | |||
988 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
992 | 'heads' and 'common' are both lists of node IDs. If heads is | |
989 | not supplied, uses all of the revlog's heads. If common is not |
|
993 | not supplied, uses all of the revlog's heads. If common is not | |
990 | supplied, uses nullid.""" |
|
994 | supplied, uses nullid.""" | |
991 | if common is None: |
|
995 | if common is None: | |
992 | common = [nullid] |
|
996 | common = [nullid] | |
993 | if heads is None: |
|
997 | if heads is None: | |
994 | heads = self.heads() |
|
998 | heads = self.heads() | |
995 |
|
999 | |||
996 | common = [self.rev(n) for n in common] |
|
1000 | common = [self.rev(n) for n in common] | |
997 | heads = [self.rev(n) for n in heads] |
|
1001 | heads = [self.rev(n) for n in heads] | |
998 |
|
1002 | |||
999 | inc = self.incrementalmissingrevs(common=common) |
|
1003 | inc = self.incrementalmissingrevs(common=common) | |
1000 | return [self.node(r) for r in inc.missingancestors(heads)] |
|
1004 | return [self.node(r) for r in inc.missingancestors(heads)] | |
1001 |
|
1005 | |||
1002 | def nodesbetween(self, roots=None, heads=None): |
|
1006 | def nodesbetween(self, roots=None, heads=None): | |
1003 | """Return a topological path from 'roots' to 'heads'. |
|
1007 | """Return a topological path from 'roots' to 'heads'. | |
1004 |
|
1008 | |||
1005 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a |
|
1009 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a | |
1006 | topologically sorted list of all nodes N that satisfy both of |
|
1010 | topologically sorted list of all nodes N that satisfy both of | |
1007 | these constraints: |
|
1011 | these constraints: | |
1008 |
|
1012 | |||
1009 | 1. N is a descendant of some node in 'roots' |
|
1013 | 1. N is a descendant of some node in 'roots' | |
1010 | 2. N is an ancestor of some node in 'heads' |
|
1014 | 2. N is an ancestor of some node in 'heads' | |
1011 |
|
1015 | |||
1012 | Every node is considered to be both a descendant and an ancestor |
|
1016 | Every node is considered to be both a descendant and an ancestor | |
1013 | of itself, so every reachable node in 'roots' and 'heads' will be |
|
1017 | of itself, so every reachable node in 'roots' and 'heads' will be | |
1014 | included in 'nodes'. |
|
1018 | included in 'nodes'. | |
1015 |
|
1019 | |||
1016 | 'outroots' is the list of reachable nodes in 'roots', i.e., the |
|
1020 | 'outroots' is the list of reachable nodes in 'roots', i.e., the | |
1017 | subset of 'roots' that is returned in 'nodes'. Likewise, |
|
1021 | subset of 'roots' that is returned in 'nodes'. Likewise, | |
1018 | 'outheads' is the subset of 'heads' that is also in 'nodes'. |
|
1022 | 'outheads' is the subset of 'heads' that is also in 'nodes'. | |
1019 |
|
1023 | |||
1020 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is |
|
1024 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is | |
1021 | unspecified, uses nullid as the only root. If 'heads' is |
|
1025 | unspecified, uses nullid as the only root. If 'heads' is | |
1022 | unspecified, uses list of all of the revlog's heads.""" |
|
1026 | unspecified, uses list of all of the revlog's heads.""" | |
1023 | nonodes = ([], [], []) |
|
1027 | nonodes = ([], [], []) | |
1024 | if roots is not None: |
|
1028 | if roots is not None: | |
1025 | roots = list(roots) |
|
1029 | roots = list(roots) | |
1026 | if not roots: |
|
1030 | if not roots: | |
1027 | return nonodes |
|
1031 | return nonodes | |
1028 | lowestrev = min([self.rev(n) for n in roots]) |
|
1032 | lowestrev = min([self.rev(n) for n in roots]) | |
1029 | else: |
|
1033 | else: | |
1030 | roots = [nullid] # Everybody's a descendant of nullid |
|
1034 | roots = [nullid] # Everybody's a descendant of nullid | |
1031 | lowestrev = nullrev |
|
1035 | lowestrev = nullrev | |
1032 | if (lowestrev == nullrev) and (heads is None): |
|
1036 | if (lowestrev == nullrev) and (heads is None): | |
1033 | # We want _all_ the nodes! |
|
1037 | # We want _all_ the nodes! | |
1034 | return ([self.node(r) for r in self], [nullid], list(self.heads())) |
|
1038 | return ([self.node(r) for r in self], [nullid], list(self.heads())) | |
1035 | if heads is None: |
|
1039 | if heads is None: | |
1036 | # All nodes are ancestors, so the latest ancestor is the last |
|
1040 | # All nodes are ancestors, so the latest ancestor is the last | |
1037 | # node. |
|
1041 | # node. | |
1038 | highestrev = len(self) - 1 |
|
1042 | highestrev = len(self) - 1 | |
1039 | # Set ancestors to None to signal that every node is an ancestor. |
|
1043 | # Set ancestors to None to signal that every node is an ancestor. | |
1040 | ancestors = None |
|
1044 | ancestors = None | |
1041 | # Set heads to an empty dictionary for later discovery of heads |
|
1045 | # Set heads to an empty dictionary for later discovery of heads | |
1042 | heads = {} |
|
1046 | heads = {} | |
1043 | else: |
|
1047 | else: | |
1044 | heads = list(heads) |
|
1048 | heads = list(heads) | |
1045 | if not heads: |
|
1049 | if not heads: | |
1046 | return nonodes |
|
1050 | return nonodes | |
1047 | ancestors = set() |
|
1051 | ancestors = set() | |
1048 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
1052 | # Turn heads into a dictionary so we can remove 'fake' heads. | |
1049 | # Also, later we will be using it to filter out the heads we can't |
|
1053 | # Also, later we will be using it to filter out the heads we can't | |
1050 | # find from roots. |
|
1054 | # find from roots. | |
1051 | heads = dict.fromkeys(heads, False) |
|
1055 | heads = dict.fromkeys(heads, False) | |
1052 | # Start at the top and keep marking parents until we're done. |
|
1056 | # Start at the top and keep marking parents until we're done. | |
1053 | nodestotag = set(heads) |
|
1057 | nodestotag = set(heads) | |
1054 | # Remember where the top was so we can use it as a limit later. |
|
1058 | # Remember where the top was so we can use it as a limit later. | |
1055 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
1059 | highestrev = max([self.rev(n) for n in nodestotag]) | |
1056 | while nodestotag: |
|
1060 | while nodestotag: | |
1057 | # grab a node to tag |
|
1061 | # grab a node to tag | |
1058 | n = nodestotag.pop() |
|
1062 | n = nodestotag.pop() | |
1059 | # Never tag nullid |
|
1063 | # Never tag nullid | |
1060 | if n == nullid: |
|
1064 | if n == nullid: | |
1061 | continue |
|
1065 | continue | |
1062 | # A node's revision number represents its place in a |
|
1066 | # A node's revision number represents its place in a | |
1063 | # topologically sorted list of nodes. |
|
1067 | # topologically sorted list of nodes. | |
1064 | r = self.rev(n) |
|
1068 | r = self.rev(n) | |
1065 | if r >= lowestrev: |
|
1069 | if r >= lowestrev: | |
1066 | if n not in ancestors: |
|
1070 | if n not in ancestors: | |
1067 | # If we are possibly a descendant of one of the roots |
|
1071 | # If we are possibly a descendant of one of the roots | |
1068 | # and we haven't already been marked as an ancestor |
|
1072 | # and we haven't already been marked as an ancestor | |
1069 | ancestors.add(n) # Mark as ancestor |
|
1073 | ancestors.add(n) # Mark as ancestor | |
1070 | # Add non-nullid parents to list of nodes to tag. |
|
1074 | # Add non-nullid parents to list of nodes to tag. | |
1071 | nodestotag.update( |
|
1075 | nodestotag.update( | |
1072 | [p for p in self.parents(n) if p != nullid] |
|
1076 | [p for p in self.parents(n) if p != nullid] | |
1073 | ) |
|
1077 | ) | |
1074 | elif n in heads: # We've seen it before, is it a fake head? |
|
1078 | elif n in heads: # We've seen it before, is it a fake head? | |
1075 | # So it is, real heads should not be the ancestors of |
|
1079 | # So it is, real heads should not be the ancestors of | |
1076 | # any other heads. |
|
1080 | # any other heads. | |
1077 | heads.pop(n) |
|
1081 | heads.pop(n) | |
1078 | if not ancestors: |
|
1082 | if not ancestors: | |
1079 | return nonodes |
|
1083 | return nonodes | |
1080 | # Now that we have our set of ancestors, we want to remove any |
|
1084 | # Now that we have our set of ancestors, we want to remove any | |
1081 | # roots that are not ancestors. |
|
1085 | # roots that are not ancestors. | |
1082 |
|
1086 | |||
1083 | # If one of the roots was nullid, everything is included anyway. |
|
1087 | # If one of the roots was nullid, everything is included anyway. | |
1084 | if lowestrev > nullrev: |
|
1088 | if lowestrev > nullrev: | |
1085 | # But, since we weren't, let's recompute the lowest rev to not |
|
1089 | # But, since we weren't, let's recompute the lowest rev to not | |
1086 | # include roots that aren't ancestors. |
|
1090 | # include roots that aren't ancestors. | |
1087 |
|
1091 | |||
1088 | # Filter out roots that aren't ancestors of heads |
|
1092 | # Filter out roots that aren't ancestors of heads | |
1089 | roots = [root for root in roots if root in ancestors] |
|
1093 | roots = [root for root in roots if root in ancestors] | |
1090 | # Recompute the lowest revision |
|
1094 | # Recompute the lowest revision | |
1091 | if roots: |
|
1095 | if roots: | |
1092 | lowestrev = min([self.rev(root) for root in roots]) |
|
1096 | lowestrev = min([self.rev(root) for root in roots]) | |
1093 | else: |
|
1097 | else: | |
1094 | # No more roots? Return empty list |
|
1098 | # No more roots? Return empty list | |
1095 | return nonodes |
|
1099 | return nonodes | |
1096 | else: |
|
1100 | else: | |
1097 | # We are descending from nullid, and don't need to care about |
|
1101 | # We are descending from nullid, and don't need to care about | |
1098 | # any other roots. |
|
1102 | # any other roots. | |
1099 | lowestrev = nullrev |
|
1103 | lowestrev = nullrev | |
1100 | roots = [nullid] |
|
1104 | roots = [nullid] | |
1101 | # Transform our roots list into a set. |
|
1105 | # Transform our roots list into a set. | |
1102 | descendants = set(roots) |
|
1106 | descendants = set(roots) | |
1103 | # Also, keep the original roots so we can filter out roots that aren't |
|
1107 | # Also, keep the original roots so we can filter out roots that aren't | |
1104 | # 'real' roots (i.e. are descended from other roots). |
|
1108 | # 'real' roots (i.e. are descended from other roots). | |
1105 | roots = descendants.copy() |
|
1109 | roots = descendants.copy() | |
1106 | # Our topologically sorted list of output nodes. |
|
1110 | # Our topologically sorted list of output nodes. | |
1107 | orderedout = [] |
|
1111 | orderedout = [] | |
1108 | # Don't start at nullid since we don't want nullid in our output list, |
|
1112 | # Don't start at nullid since we don't want nullid in our output list, | |
1109 | # and if nullid shows up in descendants, empty parents will look like |
|
1113 | # and if nullid shows up in descendants, empty parents will look like | |
1110 | # they're descendants. |
|
1114 | # they're descendants. | |
1111 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): |
|
1115 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): | |
1112 | n = self.node(r) |
|
1116 | n = self.node(r) | |
1113 | isdescendant = False |
|
1117 | isdescendant = False | |
1114 | if lowestrev == nullrev: # Everybody is a descendant of nullid |
|
1118 | if lowestrev == nullrev: # Everybody is a descendant of nullid | |
1115 | isdescendant = True |
|
1119 | isdescendant = True | |
1116 | elif n in descendants: |
|
1120 | elif n in descendants: | |
1117 | # n is already a descendant |
|
1121 | # n is already a descendant | |
1118 | isdescendant = True |
|
1122 | isdescendant = True | |
1119 | # This check only needs to be done here because all the roots |
|
1123 | # This check only needs to be done here because all the roots | |
1120 | # will start being marked is descendants before the loop. |
|
1124 | # will start being marked is descendants before the loop. | |
1121 | if n in roots: |
|
1125 | if n in roots: | |
1122 | # If n was a root, check if it's a 'real' root. |
|
1126 | # If n was a root, check if it's a 'real' root. | |
1123 | p = tuple(self.parents(n)) |
|
1127 | p = tuple(self.parents(n)) | |
1124 | # If any of its parents are descendants, it's not a root. |
|
1128 | # If any of its parents are descendants, it's not a root. | |
1125 | if (p[0] in descendants) or (p[1] in descendants): |
|
1129 | if (p[0] in descendants) or (p[1] in descendants): | |
1126 | roots.remove(n) |
|
1130 | roots.remove(n) | |
1127 | else: |
|
1131 | else: | |
1128 | p = tuple(self.parents(n)) |
|
1132 | p = tuple(self.parents(n)) | |
1129 | # A node is a descendant if either of its parents are |
|
1133 | # A node is a descendant if either of its parents are | |
1130 | # descendants. (We seeded the dependents list with the roots |
|
1134 | # descendants. (We seeded the dependents list with the roots | |
1131 | # up there, remember?) |
|
1135 | # up there, remember?) | |
1132 | if (p[0] in descendants) or (p[1] in descendants): |
|
1136 | if (p[0] in descendants) or (p[1] in descendants): | |
1133 | descendants.add(n) |
|
1137 | descendants.add(n) | |
1134 | isdescendant = True |
|
1138 | isdescendant = True | |
1135 | if isdescendant and ((ancestors is None) or (n in ancestors)): |
|
1139 | if isdescendant and ((ancestors is None) or (n in ancestors)): | |
1136 | # Only include nodes that are both descendants and ancestors. |
|
1140 | # Only include nodes that are both descendants and ancestors. | |
1137 | orderedout.append(n) |
|
1141 | orderedout.append(n) | |
1138 | if (ancestors is not None) and (n in heads): |
|
1142 | if (ancestors is not None) and (n in heads): | |
1139 | # We're trying to figure out which heads are reachable |
|
1143 | # We're trying to figure out which heads are reachable | |
1140 | # from roots. |
|
1144 | # from roots. | |
1141 | # Mark this head as having been reached |
|
1145 | # Mark this head as having been reached | |
1142 | heads[n] = True |
|
1146 | heads[n] = True | |
1143 | elif ancestors is None: |
|
1147 | elif ancestors is None: | |
1144 | # Otherwise, we're trying to discover the heads. |
|
1148 | # Otherwise, we're trying to discover the heads. | |
1145 | # Assume this is a head because if it isn't, the next step |
|
1149 | # Assume this is a head because if it isn't, the next step | |
1146 | # will eventually remove it. |
|
1150 | # will eventually remove it. | |
1147 | heads[n] = True |
|
1151 | heads[n] = True | |
1148 | # But, obviously its parents aren't. |
|
1152 | # But, obviously its parents aren't. | |
1149 | for p in self.parents(n): |
|
1153 | for p in self.parents(n): | |
1150 | heads.pop(p, None) |
|
1154 | heads.pop(p, None) | |
1151 | heads = [head for head, flag in pycompat.iteritems(heads) if flag] |
|
1155 | heads = [head for head, flag in pycompat.iteritems(heads) if flag] | |
1152 | roots = list(roots) |
|
1156 | roots = list(roots) | |
1153 | assert orderedout |
|
1157 | assert orderedout | |
1154 | assert roots |
|
1158 | assert roots | |
1155 | assert heads |
|
1159 | assert heads | |
1156 | return (orderedout, roots, heads) |
|
1160 | return (orderedout, roots, heads) | |
1157 |
|
1161 | |||
1158 | def headrevs(self, revs=None): |
|
1162 | def headrevs(self, revs=None): | |
1159 | if revs is None: |
|
1163 | if revs is None: | |
1160 | try: |
|
1164 | try: | |
1161 | return self.index.headrevs() |
|
1165 | return self.index.headrevs() | |
1162 | except AttributeError: |
|
1166 | except AttributeError: | |
1163 | return self._headrevs() |
|
1167 | return self._headrevs() | |
1164 | if rustdagop is not None: |
|
1168 | if rustdagop is not None: | |
1165 | return rustdagop.headrevs(self.index, revs) |
|
1169 | return rustdagop.headrevs(self.index, revs) | |
1166 | return dagop.headrevs(revs, self._uncheckedparentrevs) |
|
1170 | return dagop.headrevs(revs, self._uncheckedparentrevs) | |
1167 |
|
1171 | |||
1168 | def computephases(self, roots): |
|
1172 | def computephases(self, roots): | |
1169 | return self.index.computephasesmapsets(roots) |
|
1173 | return self.index.computephasesmapsets(roots) | |
1170 |
|
1174 | |||
1171 | def _headrevs(self): |
|
1175 | def _headrevs(self): | |
1172 | count = len(self) |
|
1176 | count = len(self) | |
1173 | if not count: |
|
1177 | if not count: | |
1174 | return [nullrev] |
|
1178 | return [nullrev] | |
1175 | # we won't iter over filtered rev so nobody is a head at start |
|
1179 | # we won't iter over filtered rev so nobody is a head at start | |
1176 | ishead = [0] * (count + 1) |
|
1180 | ishead = [0] * (count + 1) | |
1177 | index = self.index |
|
1181 | index = self.index | |
1178 | for r in self: |
|
1182 | for r in self: | |
1179 | ishead[r] = 1 # I may be an head |
|
1183 | ishead[r] = 1 # I may be an head | |
1180 | e = index[r] |
|
1184 | e = index[r] | |
1181 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not |
|
1185 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not | |
1182 | return [r for r, val in enumerate(ishead) if val] |
|
1186 | return [r for r, val in enumerate(ishead) if val] | |
1183 |
|
1187 | |||
1184 | def heads(self, start=None, stop=None): |
|
1188 | def heads(self, start=None, stop=None): | |
1185 | """return the list of all nodes that have no children |
|
1189 | """return the list of all nodes that have no children | |
1186 |
|
1190 | |||
1187 | if start is specified, only heads that are descendants of |
|
1191 | if start is specified, only heads that are descendants of | |
1188 | start will be returned |
|
1192 | start will be returned | |
1189 | if stop is specified, it will consider all the revs from stop |
|
1193 | if stop is specified, it will consider all the revs from stop | |
1190 | as if they had no children |
|
1194 | as if they had no children | |
1191 | """ |
|
1195 | """ | |
1192 | if start is None and stop is None: |
|
1196 | if start is None and stop is None: | |
1193 | if not len(self): |
|
1197 | if not len(self): | |
1194 | return [nullid] |
|
1198 | return [nullid] | |
1195 | return [self.node(r) for r in self.headrevs()] |
|
1199 | return [self.node(r) for r in self.headrevs()] | |
1196 |
|
1200 | |||
1197 | if start is None: |
|
1201 | if start is None: | |
1198 | start = nullrev |
|
1202 | start = nullrev | |
1199 | else: |
|
1203 | else: | |
1200 | start = self.rev(start) |
|
1204 | start = self.rev(start) | |
1201 |
|
1205 | |||
1202 | stoprevs = set(self.rev(n) for n in stop or []) |
|
1206 | stoprevs = set(self.rev(n) for n in stop or []) | |
1203 |
|
1207 | |||
1204 | revs = dagop.headrevssubset( |
|
1208 | revs = dagop.headrevssubset( | |
1205 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs |
|
1209 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs | |
1206 | ) |
|
1210 | ) | |
1207 |
|
1211 | |||
1208 | return [self.node(rev) for rev in revs] |
|
1212 | return [self.node(rev) for rev in revs] | |
1209 |
|
1213 | |||
1210 | def children(self, node): |
|
1214 | def children(self, node): | |
1211 | """find the children of a given node""" |
|
1215 | """find the children of a given node""" | |
1212 | c = [] |
|
1216 | c = [] | |
1213 | p = self.rev(node) |
|
1217 | p = self.rev(node) | |
1214 | for r in self.revs(start=p + 1): |
|
1218 | for r in self.revs(start=p + 1): | |
1215 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
1219 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] | |
1216 | if prevs: |
|
1220 | if prevs: | |
1217 | for pr in prevs: |
|
1221 | for pr in prevs: | |
1218 | if pr == p: |
|
1222 | if pr == p: | |
1219 | c.append(self.node(r)) |
|
1223 | c.append(self.node(r)) | |
1220 | elif p == nullrev: |
|
1224 | elif p == nullrev: | |
1221 | c.append(self.node(r)) |
|
1225 | c.append(self.node(r)) | |
1222 | return c |
|
1226 | return c | |
1223 |
|
1227 | |||
1224 | def commonancestorsheads(self, a, b): |
|
1228 | def commonancestorsheads(self, a, b): | |
1225 | """calculate all the heads of the common ancestors of nodes a and b""" |
|
1229 | """calculate all the heads of the common ancestors of nodes a and b""" | |
1226 | a, b = self.rev(a), self.rev(b) |
|
1230 | a, b = self.rev(a), self.rev(b) | |
1227 | ancs = self._commonancestorsheads(a, b) |
|
1231 | ancs = self._commonancestorsheads(a, b) | |
1228 | return pycompat.maplist(self.node, ancs) |
|
1232 | return pycompat.maplist(self.node, ancs) | |
1229 |
|
1233 | |||
1230 | def _commonancestorsheads(self, *revs): |
|
1234 | def _commonancestorsheads(self, *revs): | |
1231 | """calculate all the heads of the common ancestors of revs""" |
|
1235 | """calculate all the heads of the common ancestors of revs""" | |
1232 | try: |
|
1236 | try: | |
1233 | ancs = self.index.commonancestorsheads(*revs) |
|
1237 | ancs = self.index.commonancestorsheads(*revs) | |
1234 | except (AttributeError, OverflowError): # C implementation failed |
|
1238 | except (AttributeError, OverflowError): # C implementation failed | |
1235 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) |
|
1239 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) | |
1236 | return ancs |
|
1240 | return ancs | |
1237 |
|
1241 | |||
1238 | def isancestor(self, a, b): |
|
1242 | def isancestor(self, a, b): | |
1239 | """return True if node a is an ancestor of node b |
|
1243 | """return True if node a is an ancestor of node b | |
1240 |
|
1244 | |||
1241 | A revision is considered an ancestor of itself.""" |
|
1245 | A revision is considered an ancestor of itself.""" | |
1242 | a, b = self.rev(a), self.rev(b) |
|
1246 | a, b = self.rev(a), self.rev(b) | |
1243 | return self.isancestorrev(a, b) |
|
1247 | return self.isancestorrev(a, b) | |
1244 |
|
1248 | |||
1245 | def isancestorrev(self, a, b): |
|
1249 | def isancestorrev(self, a, b): | |
1246 | """return True if revision a is an ancestor of revision b |
|
1250 | """return True if revision a is an ancestor of revision b | |
1247 |
|
1251 | |||
1248 | A revision is considered an ancestor of itself. |
|
1252 | A revision is considered an ancestor of itself. | |
1249 |
|
1253 | |||
1250 | The implementation of this is trivial but the use of |
|
1254 | The implementation of this is trivial but the use of | |
1251 | reachableroots is not.""" |
|
1255 | reachableroots is not.""" | |
1252 | if a == nullrev: |
|
1256 | if a == nullrev: | |
1253 | return True |
|
1257 | return True | |
1254 | elif a == b: |
|
1258 | elif a == b: | |
1255 | return True |
|
1259 | return True | |
1256 | elif a > b: |
|
1260 | elif a > b: | |
1257 | return False |
|
1261 | return False | |
1258 | return bool(self.reachableroots(a, [b], [a], includepath=False)) |
|
1262 | return bool(self.reachableroots(a, [b], [a], includepath=False)) | |
1259 |
|
1263 | |||
1260 | def reachableroots(self, minroot, heads, roots, includepath=False): |
|
1264 | def reachableroots(self, minroot, heads, roots, includepath=False): | |
1261 | """return (heads(::<roots> and <roots>::<heads>)) |
|
1265 | """return (heads(::<roots> and <roots>::<heads>)) | |
1262 |
|
1266 | |||
1263 | If includepath is True, return (<roots>::<heads>).""" |
|
1267 | If includepath is True, return (<roots>::<heads>).""" | |
1264 | try: |
|
1268 | try: | |
1265 | return self.index.reachableroots2( |
|
1269 | return self.index.reachableroots2( | |
1266 | minroot, heads, roots, includepath |
|
1270 | minroot, heads, roots, includepath | |
1267 | ) |
|
1271 | ) | |
1268 | except AttributeError: |
|
1272 | except AttributeError: | |
1269 | return dagop._reachablerootspure( |
|
1273 | return dagop._reachablerootspure( | |
1270 | self.parentrevs, minroot, roots, heads, includepath |
|
1274 | self.parentrevs, minroot, roots, heads, includepath | |
1271 | ) |
|
1275 | ) | |
1272 |
|
1276 | |||
1273 | def ancestor(self, a, b): |
|
1277 | def ancestor(self, a, b): | |
1274 | """calculate the "best" common ancestor of nodes a and b""" |
|
1278 | """calculate the "best" common ancestor of nodes a and b""" | |
1275 |
|
1279 | |||
1276 | a, b = self.rev(a), self.rev(b) |
|
1280 | a, b = self.rev(a), self.rev(b) | |
1277 | try: |
|
1281 | try: | |
1278 | ancs = self.index.ancestors(a, b) |
|
1282 | ancs = self.index.ancestors(a, b) | |
1279 | except (AttributeError, OverflowError): |
|
1283 | except (AttributeError, OverflowError): | |
1280 | ancs = ancestor.ancestors(self.parentrevs, a, b) |
|
1284 | ancs = ancestor.ancestors(self.parentrevs, a, b) | |
1281 | if ancs: |
|
1285 | if ancs: | |
1282 | # choose a consistent winner when there's a tie |
|
1286 | # choose a consistent winner when there's a tie | |
1283 | return min(map(self.node, ancs)) |
|
1287 | return min(map(self.node, ancs)) | |
1284 | return nullid |
|
1288 | return nullid | |
1285 |
|
1289 | |||
1286 | def _match(self, id): |
|
1290 | def _match(self, id): | |
1287 | if isinstance(id, int): |
|
1291 | if isinstance(id, int): | |
1288 | # rev |
|
1292 | # rev | |
1289 | return self.node(id) |
|
1293 | return self.node(id) | |
1290 | if len(id) == 20: |
|
1294 | if len(id) == 20: | |
1291 | # possibly a binary node |
|
1295 | # possibly a binary node | |
1292 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
1296 | # odds of a binary node being all hex in ASCII are 1 in 10**25 | |
1293 | try: |
|
1297 | try: | |
1294 | node = id |
|
1298 | node = id | |
1295 | self.rev(node) # quick search the index |
|
1299 | self.rev(node) # quick search the index | |
1296 | return node |
|
1300 | return node | |
1297 | except error.LookupError: |
|
1301 | except error.LookupError: | |
1298 | pass # may be partial hex id |
|
1302 | pass # may be partial hex id | |
1299 | try: |
|
1303 | try: | |
1300 | # str(rev) |
|
1304 | # str(rev) | |
1301 | rev = int(id) |
|
1305 | rev = int(id) | |
1302 | if b"%d" % rev != id: |
|
1306 | if b"%d" % rev != id: | |
1303 | raise ValueError |
|
1307 | raise ValueError | |
1304 | if rev < 0: |
|
1308 | if rev < 0: | |
1305 | rev = len(self) + rev |
|
1309 | rev = len(self) + rev | |
1306 | if rev < 0 or rev >= len(self): |
|
1310 | if rev < 0 or rev >= len(self): | |
1307 | raise ValueError |
|
1311 | raise ValueError | |
1308 | return self.node(rev) |
|
1312 | return self.node(rev) | |
1309 | except (ValueError, OverflowError): |
|
1313 | except (ValueError, OverflowError): | |
1310 | pass |
|
1314 | pass | |
1311 | if len(id) == 40: |
|
1315 | if len(id) == 40: | |
1312 | try: |
|
1316 | try: | |
1313 | # a full hex nodeid? |
|
1317 | # a full hex nodeid? | |
1314 | node = bin(id) |
|
1318 | node = bin(id) | |
1315 | self.rev(node) |
|
1319 | self.rev(node) | |
1316 | return node |
|
1320 | return node | |
1317 | except (TypeError, error.LookupError): |
|
1321 | except (TypeError, error.LookupError): | |
1318 | pass |
|
1322 | pass | |
1319 |
|
1323 | |||
1320 | def _partialmatch(self, id): |
|
1324 | def _partialmatch(self, id): | |
1321 | # we don't care wdirfilenodeids as they should be always full hash |
|
1325 | # we don't care wdirfilenodeids as they should be always full hash | |
1322 | maybewdir = wdirhex.startswith(id) |
|
1326 | maybewdir = wdirhex.startswith(id) | |
1323 | try: |
|
1327 | try: | |
1324 | partial = self.index.partialmatch(id) |
|
1328 | partial = self.index.partialmatch(id) | |
1325 | if partial and self.hasnode(partial): |
|
1329 | if partial and self.hasnode(partial): | |
1326 | if maybewdir: |
|
1330 | if maybewdir: | |
1327 | # single 'ff...' match in radix tree, ambiguous with wdir |
|
1331 | # single 'ff...' match in radix tree, ambiguous with wdir | |
1328 | raise error.RevlogError |
|
1332 | raise error.RevlogError | |
1329 | return partial |
|
1333 | return partial | |
1330 | if maybewdir: |
|
1334 | if maybewdir: | |
1331 | # no 'ff...' match in radix tree, wdir identified |
|
1335 | # no 'ff...' match in radix tree, wdir identified | |
1332 | raise error.WdirUnsupported |
|
1336 | raise error.WdirUnsupported | |
1333 | return None |
|
1337 | return None | |
1334 | except error.RevlogError: |
|
1338 | except error.RevlogError: | |
1335 | # parsers.c radix tree lookup gave multiple matches |
|
1339 | # parsers.c radix tree lookup gave multiple matches | |
1336 | # fast path: for unfiltered changelog, radix tree is accurate |
|
1340 | # fast path: for unfiltered changelog, radix tree is accurate | |
1337 | if not getattr(self, 'filteredrevs', None): |
|
1341 | if not getattr(self, 'filteredrevs', None): | |
1338 | raise error.AmbiguousPrefixLookupError( |
|
1342 | raise error.AmbiguousPrefixLookupError( | |
1339 | id, self.indexfile, _(b'ambiguous identifier') |
|
1343 | id, self.indexfile, _(b'ambiguous identifier') | |
1340 | ) |
|
1344 | ) | |
1341 | # fall through to slow path that filters hidden revisions |
|
1345 | # fall through to slow path that filters hidden revisions | |
1342 | except (AttributeError, ValueError): |
|
1346 | except (AttributeError, ValueError): | |
1343 | # we are pure python, or key was too short to search radix tree |
|
1347 | # we are pure python, or key was too short to search radix tree | |
1344 | pass |
|
1348 | pass | |
1345 |
|
1349 | |||
1346 | if id in self._pcache: |
|
1350 | if id in self._pcache: | |
1347 | return self._pcache[id] |
|
1351 | return self._pcache[id] | |
1348 |
|
1352 | |||
1349 | if len(id) <= 40: |
|
1353 | if len(id) <= 40: | |
1350 | try: |
|
1354 | try: | |
1351 | # hex(node)[:...] |
|
1355 | # hex(node)[:...] | |
1352 | l = len(id) // 2 # grab an even number of digits |
|
1356 | l = len(id) // 2 # grab an even number of digits | |
1353 | prefix = bin(id[: l * 2]) |
|
1357 | prefix = bin(id[: l * 2]) | |
1354 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] |
|
1358 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] | |
1355 | nl = [ |
|
1359 | nl = [ | |
1356 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) |
|
1360 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) | |
1357 | ] |
|
1361 | ] | |
1358 | if nullhex.startswith(id): |
|
1362 | if nullhex.startswith(id): | |
1359 | nl.append(nullid) |
|
1363 | nl.append(nullid) | |
1360 | if len(nl) > 0: |
|
1364 | if len(nl) > 0: | |
1361 | if len(nl) == 1 and not maybewdir: |
|
1365 | if len(nl) == 1 and not maybewdir: | |
1362 | self._pcache[id] = nl[0] |
|
1366 | self._pcache[id] = nl[0] | |
1363 | return nl[0] |
|
1367 | return nl[0] | |
1364 | raise error.AmbiguousPrefixLookupError( |
|
1368 | raise error.AmbiguousPrefixLookupError( | |
1365 | id, self.indexfile, _(b'ambiguous identifier') |
|
1369 | id, self.indexfile, _(b'ambiguous identifier') | |
1366 | ) |
|
1370 | ) | |
1367 | if maybewdir: |
|
1371 | if maybewdir: | |
1368 | raise error.WdirUnsupported |
|
1372 | raise error.WdirUnsupported | |
1369 | return None |
|
1373 | return None | |
1370 | except TypeError: |
|
1374 | except TypeError: | |
1371 | pass |
|
1375 | pass | |
1372 |
|
1376 | |||
1373 | def lookup(self, id): |
|
1377 | def lookup(self, id): | |
1374 | """locate a node based on: |
|
1378 | """locate a node based on: | |
1375 | - revision number or str(revision number) |
|
1379 | - revision number or str(revision number) | |
1376 | - nodeid or subset of hex nodeid |
|
1380 | - nodeid or subset of hex nodeid | |
1377 | """ |
|
1381 | """ | |
1378 | n = self._match(id) |
|
1382 | n = self._match(id) | |
1379 | if n is not None: |
|
1383 | if n is not None: | |
1380 | return n |
|
1384 | return n | |
1381 | n = self._partialmatch(id) |
|
1385 | n = self._partialmatch(id) | |
1382 | if n: |
|
1386 | if n: | |
1383 | return n |
|
1387 | return n | |
1384 |
|
1388 | |||
1385 | raise error.LookupError(id, self.indexfile, _(b'no match found')) |
|
1389 | raise error.LookupError(id, self.indexfile, _(b'no match found')) | |
1386 |
|
1390 | |||
1387 | def shortest(self, node, minlength=1): |
|
1391 | def shortest(self, node, minlength=1): | |
1388 | """Find the shortest unambiguous prefix that matches node.""" |
|
1392 | """Find the shortest unambiguous prefix that matches node.""" | |
1389 |
|
1393 | |||
1390 | def isvalid(prefix): |
|
1394 | def isvalid(prefix): | |
1391 | try: |
|
1395 | try: | |
1392 | matchednode = self._partialmatch(prefix) |
|
1396 | matchednode = self._partialmatch(prefix) | |
1393 | except error.AmbiguousPrefixLookupError: |
|
1397 | except error.AmbiguousPrefixLookupError: | |
1394 | return False |
|
1398 | return False | |
1395 | except error.WdirUnsupported: |
|
1399 | except error.WdirUnsupported: | |
1396 | # single 'ff...' match |
|
1400 | # single 'ff...' match | |
1397 | return True |
|
1401 | return True | |
1398 | if matchednode is None: |
|
1402 | if matchednode is None: | |
1399 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
1403 | raise error.LookupError(node, self.indexfile, _(b'no node')) | |
1400 | return True |
|
1404 | return True | |
1401 |
|
1405 | |||
1402 | def maybewdir(prefix): |
|
1406 | def maybewdir(prefix): | |
1403 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) |
|
1407 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) | |
1404 |
|
1408 | |||
1405 | hexnode = hex(node) |
|
1409 | hexnode = hex(node) | |
1406 |
|
1410 | |||
1407 | def disambiguate(hexnode, minlength): |
|
1411 | def disambiguate(hexnode, minlength): | |
1408 | """Disambiguate against wdirid.""" |
|
1412 | """Disambiguate against wdirid.""" | |
1409 | for length in range(minlength, 41): |
|
1413 | for length in range(minlength, 41): | |
1410 | prefix = hexnode[:length] |
|
1414 | prefix = hexnode[:length] | |
1411 | if not maybewdir(prefix): |
|
1415 | if not maybewdir(prefix): | |
1412 | return prefix |
|
1416 | return prefix | |
1413 |
|
1417 | |||
1414 | if not getattr(self, 'filteredrevs', None): |
|
1418 | if not getattr(self, 'filteredrevs', None): | |
1415 | try: |
|
1419 | try: | |
1416 | length = max(self.index.shortest(node), minlength) |
|
1420 | length = max(self.index.shortest(node), minlength) | |
1417 | return disambiguate(hexnode, length) |
|
1421 | return disambiguate(hexnode, length) | |
1418 | except error.RevlogError: |
|
1422 | except error.RevlogError: | |
1419 | if node != wdirid: |
|
1423 | if node != wdirid: | |
1420 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
1424 | raise error.LookupError(node, self.indexfile, _(b'no node')) | |
1421 | except AttributeError: |
|
1425 | except AttributeError: | |
1422 | # Fall through to pure code |
|
1426 | # Fall through to pure code | |
1423 | pass |
|
1427 | pass | |
1424 |
|
1428 | |||
1425 | if node == wdirid: |
|
1429 | if node == wdirid: | |
1426 | for length in range(minlength, 41): |
|
1430 | for length in range(minlength, 41): | |
1427 | prefix = hexnode[:length] |
|
1431 | prefix = hexnode[:length] | |
1428 | if isvalid(prefix): |
|
1432 | if isvalid(prefix): | |
1429 | return prefix |
|
1433 | return prefix | |
1430 |
|
1434 | |||
1431 | for length in range(minlength, 41): |
|
1435 | for length in range(minlength, 41): | |
1432 | prefix = hexnode[:length] |
|
1436 | prefix = hexnode[:length] | |
1433 | if isvalid(prefix): |
|
1437 | if isvalid(prefix): | |
1434 | return disambiguate(hexnode, length) |
|
1438 | return disambiguate(hexnode, length) | |
1435 |
|
1439 | |||
1436 | def cmp(self, node, text): |
|
1440 | def cmp(self, node, text): | |
1437 | """compare text with a given file revision |
|
1441 | """compare text with a given file revision | |
1438 |
|
1442 | |||
1439 | returns True if text is different than what is stored. |
|
1443 | returns True if text is different than what is stored. | |
1440 | """ |
|
1444 | """ | |
1441 | p1, p2 = self.parents(node) |
|
1445 | p1, p2 = self.parents(node) | |
1442 | return storageutil.hashrevisionsha1(text, p1, p2) != node |
|
1446 | return storageutil.hashrevisionsha1(text, p1, p2) != node | |
1443 |
|
1447 | |||
1444 | def _cachesegment(self, offset, data): |
|
1448 | def _cachesegment(self, offset, data): | |
1445 | """Add a segment to the revlog cache. |
|
1449 | """Add a segment to the revlog cache. | |
1446 |
|
1450 | |||
1447 | Accepts an absolute offset and the data that is at that location. |
|
1451 | Accepts an absolute offset and the data that is at that location. | |
1448 | """ |
|
1452 | """ | |
1449 | o, d = self._chunkcache |
|
1453 | o, d = self._chunkcache | |
1450 | # try to add to existing cache |
|
1454 | # try to add to existing cache | |
1451 | if o + len(d) == offset and len(d) + len(data) < _chunksize: |
|
1455 | if o + len(d) == offset and len(d) + len(data) < _chunksize: | |
1452 | self._chunkcache = o, d + data |
|
1456 | self._chunkcache = o, d + data | |
1453 | else: |
|
1457 | else: | |
1454 | self._chunkcache = offset, data |
|
1458 | self._chunkcache = offset, data | |
1455 |
|
1459 | |||
1456 | def _readsegment(self, offset, length, df=None): |
|
1460 | def _readsegment(self, offset, length, df=None): | |
1457 | """Load a segment of raw data from the revlog. |
|
1461 | """Load a segment of raw data from the revlog. | |
1458 |
|
1462 | |||
1459 | Accepts an absolute offset, length to read, and an optional existing |
|
1463 | Accepts an absolute offset, length to read, and an optional existing | |
1460 | file handle to read from. |
|
1464 | file handle to read from. | |
1461 |
|
1465 | |||
1462 | If an existing file handle is passed, it will be seeked and the |
|
1466 | If an existing file handle is passed, it will be seeked and the | |
1463 | original seek position will NOT be restored. |
|
1467 | original seek position will NOT be restored. | |
1464 |
|
1468 | |||
1465 | Returns a str or buffer of raw byte data. |
|
1469 | Returns a str or buffer of raw byte data. | |
1466 |
|
1470 | |||
1467 | Raises if the requested number of bytes could not be read. |
|
1471 | Raises if the requested number of bytes could not be read. | |
1468 | """ |
|
1472 | """ | |
1469 | # Cache data both forward and backward around the requested |
|
1473 | # Cache data both forward and backward around the requested | |
1470 | # data, in a fixed size window. This helps speed up operations |
|
1474 | # data, in a fixed size window. This helps speed up operations | |
1471 | # involving reading the revlog backwards. |
|
1475 | # involving reading the revlog backwards. | |
1472 | cachesize = self._chunkcachesize |
|
1476 | cachesize = self._chunkcachesize | |
1473 | realoffset = offset & ~(cachesize - 1) |
|
1477 | realoffset = offset & ~(cachesize - 1) | |
1474 | reallength = ( |
|
1478 | reallength = ( | |
1475 | (offset + length + cachesize) & ~(cachesize - 1) |
|
1479 | (offset + length + cachesize) & ~(cachesize - 1) | |
1476 | ) - realoffset |
|
1480 | ) - realoffset | |
1477 | with self._datareadfp(df) as df: |
|
1481 | with self._datareadfp(df) as df: | |
1478 | df.seek(realoffset) |
|
1482 | df.seek(realoffset) | |
1479 | d = df.read(reallength) |
|
1483 | d = df.read(reallength) | |
1480 |
|
1484 | |||
1481 | self._cachesegment(realoffset, d) |
|
1485 | self._cachesegment(realoffset, d) | |
1482 | if offset != realoffset or reallength != length: |
|
1486 | if offset != realoffset or reallength != length: | |
1483 | startoffset = offset - realoffset |
|
1487 | startoffset = offset - realoffset | |
1484 | if len(d) - startoffset < length: |
|
1488 | if len(d) - startoffset < length: | |
1485 | raise error.RevlogError( |
|
1489 | raise error.RevlogError( | |
1486 | _( |
|
1490 | _( | |
1487 | b'partial read of revlog %s; expected %d bytes from ' |
|
1491 | b'partial read of revlog %s; expected %d bytes from ' | |
1488 | b'offset %d, got %d' |
|
1492 | b'offset %d, got %d' | |
1489 | ) |
|
1493 | ) | |
1490 | % ( |
|
1494 | % ( | |
1491 | self.indexfile if self._inline else self.datafile, |
|
1495 | self.indexfile if self._inline else self.datafile, | |
1492 | length, |
|
1496 | length, | |
1493 | realoffset, |
|
1497 | realoffset, | |
1494 | len(d) - startoffset, |
|
1498 | len(d) - startoffset, | |
1495 | ) |
|
1499 | ) | |
1496 | ) |
|
1500 | ) | |
1497 |
|
1501 | |||
1498 | return util.buffer(d, startoffset, length) |
|
1502 | return util.buffer(d, startoffset, length) | |
1499 |
|
1503 | |||
1500 | if len(d) < length: |
|
1504 | if len(d) < length: | |
1501 | raise error.RevlogError( |
|
1505 | raise error.RevlogError( | |
1502 | _( |
|
1506 | _( | |
1503 | b'partial read of revlog %s; expected %d bytes from offset ' |
|
1507 | b'partial read of revlog %s; expected %d bytes from offset ' | |
1504 | b'%d, got %d' |
|
1508 | b'%d, got %d' | |
1505 | ) |
|
1509 | ) | |
1506 | % ( |
|
1510 | % ( | |
1507 | self.indexfile if self._inline else self.datafile, |
|
1511 | self.indexfile if self._inline else self.datafile, | |
1508 | length, |
|
1512 | length, | |
1509 | offset, |
|
1513 | offset, | |
1510 | len(d), |
|
1514 | len(d), | |
1511 | ) |
|
1515 | ) | |
1512 | ) |
|
1516 | ) | |
1513 |
|
1517 | |||
1514 | return d |
|
1518 | return d | |
1515 |
|
1519 | |||
1516 | def _getsegment(self, offset, length, df=None): |
|
1520 | def _getsegment(self, offset, length, df=None): | |
1517 | """Obtain a segment of raw data from the revlog. |
|
1521 | """Obtain a segment of raw data from the revlog. | |
1518 |
|
1522 | |||
1519 | Accepts an absolute offset, length of bytes to obtain, and an |
|
1523 | Accepts an absolute offset, length of bytes to obtain, and an | |
1520 | optional file handle to the already-opened revlog. If the file |
|
1524 | optional file handle to the already-opened revlog. If the file | |
1521 | handle is used, it's original seek position will not be preserved. |
|
1525 | handle is used, it's original seek position will not be preserved. | |
1522 |
|
1526 | |||
1523 | Requests for data may be returned from a cache. |
|
1527 | Requests for data may be returned from a cache. | |
1524 |
|
1528 | |||
1525 | Returns a str or a buffer instance of raw byte data. |
|
1529 | Returns a str or a buffer instance of raw byte data. | |
1526 | """ |
|
1530 | """ | |
1527 | o, d = self._chunkcache |
|
1531 | o, d = self._chunkcache | |
1528 | l = len(d) |
|
1532 | l = len(d) | |
1529 |
|
1533 | |||
1530 | # is it in the cache? |
|
1534 | # is it in the cache? | |
1531 | cachestart = offset - o |
|
1535 | cachestart = offset - o | |
1532 | cacheend = cachestart + length |
|
1536 | cacheend = cachestart + length | |
1533 | if cachestart >= 0 and cacheend <= l: |
|
1537 | if cachestart >= 0 and cacheend <= l: | |
1534 | if cachestart == 0 and cacheend == l: |
|
1538 | if cachestart == 0 and cacheend == l: | |
1535 | return d # avoid a copy |
|
1539 | return d # avoid a copy | |
1536 | return util.buffer(d, cachestart, cacheend - cachestart) |
|
1540 | return util.buffer(d, cachestart, cacheend - cachestart) | |
1537 |
|
1541 | |||
1538 | return self._readsegment(offset, length, df=df) |
|
1542 | return self._readsegment(offset, length, df=df) | |
1539 |
|
1543 | |||
1540 | def _getsegmentforrevs(self, startrev, endrev, df=None): |
|
1544 | def _getsegmentforrevs(self, startrev, endrev, df=None): | |
1541 | """Obtain a segment of raw data corresponding to a range of revisions. |
|
1545 | """Obtain a segment of raw data corresponding to a range of revisions. | |
1542 |
|
1546 | |||
1543 | Accepts the start and end revisions and an optional already-open |
|
1547 | Accepts the start and end revisions and an optional already-open | |
1544 | file handle to be used for reading. If the file handle is read, its |
|
1548 | file handle to be used for reading. If the file handle is read, its | |
1545 | seek position will not be preserved. |
|
1549 | seek position will not be preserved. | |
1546 |
|
1550 | |||
1547 | Requests for data may be satisfied by a cache. |
|
1551 | Requests for data may be satisfied by a cache. | |
1548 |
|
1552 | |||
1549 | Returns a 2-tuple of (offset, data) for the requested range of |
|
1553 | Returns a 2-tuple of (offset, data) for the requested range of | |
1550 | revisions. Offset is the integer offset from the beginning of the |
|
1554 | revisions. Offset is the integer offset from the beginning of the | |
1551 | revlog and data is a str or buffer of the raw byte data. |
|
1555 | revlog and data is a str or buffer of the raw byte data. | |
1552 |
|
1556 | |||
1553 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` |
|
1557 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` | |
1554 | to determine where each revision's data begins and ends. |
|
1558 | to determine where each revision's data begins and ends. | |
1555 | """ |
|
1559 | """ | |
1556 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons |
|
1560 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons | |
1557 | # (functions are expensive). |
|
1561 | # (functions are expensive). | |
1558 | index = self.index |
|
1562 | index = self.index | |
1559 | istart = index[startrev] |
|
1563 | istart = index[startrev] | |
1560 | start = int(istart[0] >> 16) |
|
1564 | start = int(istart[0] >> 16) | |
1561 | if startrev == endrev: |
|
1565 | if startrev == endrev: | |
1562 | end = start + istart[1] |
|
1566 | end = start + istart[1] | |
1563 | else: |
|
1567 | else: | |
1564 | iend = index[endrev] |
|
1568 | iend = index[endrev] | |
1565 | end = int(iend[0] >> 16) + iend[1] |
|
1569 | end = int(iend[0] >> 16) + iend[1] | |
1566 |
|
1570 | |||
1567 | if self._inline: |
|
1571 | if self._inline: | |
1568 | start += (startrev + 1) * self._io.size |
|
1572 | start += (startrev + 1) * self._io.size | |
1569 | end += (endrev + 1) * self._io.size |
|
1573 | end += (endrev + 1) * self._io.size | |
1570 | length = end - start |
|
1574 | length = end - start | |
1571 |
|
1575 | |||
1572 | return start, self._getsegment(start, length, df=df) |
|
1576 | return start, self._getsegment(start, length, df=df) | |
1573 |
|
1577 | |||
1574 | def _chunk(self, rev, df=None): |
|
1578 | def _chunk(self, rev, df=None): | |
1575 | """Obtain a single decompressed chunk for a revision. |
|
1579 | """Obtain a single decompressed chunk for a revision. | |
1576 |
|
1580 | |||
1577 | Accepts an integer revision and an optional already-open file handle |
|
1581 | Accepts an integer revision and an optional already-open file handle | |
1578 | to be used for reading. If used, the seek position of the file will not |
|
1582 | to be used for reading. If used, the seek position of the file will not | |
1579 | be preserved. |
|
1583 | be preserved. | |
1580 |
|
1584 | |||
1581 | Returns a str holding uncompressed data for the requested revision. |
|
1585 | Returns a str holding uncompressed data for the requested revision. | |
1582 | """ |
|
1586 | """ | |
1583 | return self.decompress(self._getsegmentforrevs(rev, rev, df=df)[1]) |
|
1587 | return self.decompress(self._getsegmentforrevs(rev, rev, df=df)[1]) | |
1584 |
|
1588 | |||
1585 | def _chunks(self, revs, df=None, targetsize=None): |
|
1589 | def _chunks(self, revs, df=None, targetsize=None): | |
1586 | """Obtain decompressed chunks for the specified revisions. |
|
1590 | """Obtain decompressed chunks for the specified revisions. | |
1587 |
|
1591 | |||
1588 | Accepts an iterable of numeric revisions that are assumed to be in |
|
1592 | Accepts an iterable of numeric revisions that are assumed to be in | |
1589 | ascending order. Also accepts an optional already-open file handle |
|
1593 | ascending order. Also accepts an optional already-open file handle | |
1590 | to be used for reading. If used, the seek position of the file will |
|
1594 | to be used for reading. If used, the seek position of the file will | |
1591 | not be preserved. |
|
1595 | not be preserved. | |
1592 |
|
1596 | |||
1593 | This function is similar to calling ``self._chunk()`` multiple times, |
|
1597 | This function is similar to calling ``self._chunk()`` multiple times, | |
1594 | but is faster. |
|
1598 | but is faster. | |
1595 |
|
1599 | |||
1596 | Returns a list with decompressed data for each requested revision. |
|
1600 | Returns a list with decompressed data for each requested revision. | |
1597 | """ |
|
1601 | """ | |
1598 | if not revs: |
|
1602 | if not revs: | |
1599 | return [] |
|
1603 | return [] | |
1600 | start = self.start |
|
1604 | start = self.start | |
1601 | length = self.length |
|
1605 | length = self.length | |
1602 | inline = self._inline |
|
1606 | inline = self._inline | |
1603 | iosize = self._io.size |
|
1607 | iosize = self._io.size | |
1604 | buffer = util.buffer |
|
1608 | buffer = util.buffer | |
1605 |
|
1609 | |||
1606 | l = [] |
|
1610 | l = [] | |
1607 | ladd = l.append |
|
1611 | ladd = l.append | |
1608 |
|
1612 | |||
1609 | if not self._withsparseread: |
|
1613 | if not self._withsparseread: | |
1610 | slicedchunks = (revs,) |
|
1614 | slicedchunks = (revs,) | |
1611 | else: |
|
1615 | else: | |
1612 | slicedchunks = deltautil.slicechunk( |
|
1616 | slicedchunks = deltautil.slicechunk( | |
1613 | self, revs, targetsize=targetsize |
|
1617 | self, revs, targetsize=targetsize | |
1614 | ) |
|
1618 | ) | |
1615 |
|
1619 | |||
1616 | for revschunk in slicedchunks: |
|
1620 | for revschunk in slicedchunks: | |
1617 | firstrev = revschunk[0] |
|
1621 | firstrev = revschunk[0] | |
1618 | # Skip trailing revisions with empty diff |
|
1622 | # Skip trailing revisions with empty diff | |
1619 | for lastrev in revschunk[::-1]: |
|
1623 | for lastrev in revschunk[::-1]: | |
1620 | if length(lastrev) != 0: |
|
1624 | if length(lastrev) != 0: | |
1621 | break |
|
1625 | break | |
1622 |
|
1626 | |||
1623 | try: |
|
1627 | try: | |
1624 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) |
|
1628 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) | |
1625 | except OverflowError: |
|
1629 | except OverflowError: | |
1626 | # issue4215 - we can't cache a run of chunks greater than |
|
1630 | # issue4215 - we can't cache a run of chunks greater than | |
1627 | # 2G on Windows |
|
1631 | # 2G on Windows | |
1628 | return [self._chunk(rev, df=df) for rev in revschunk] |
|
1632 | return [self._chunk(rev, df=df) for rev in revschunk] | |
1629 |
|
1633 | |||
1630 | decomp = self.decompress |
|
1634 | decomp = self.decompress | |
1631 | for rev in revschunk: |
|
1635 | for rev in revschunk: | |
1632 | chunkstart = start(rev) |
|
1636 | chunkstart = start(rev) | |
1633 | if inline: |
|
1637 | if inline: | |
1634 | chunkstart += (rev + 1) * iosize |
|
1638 | chunkstart += (rev + 1) * iosize | |
1635 | chunklength = length(rev) |
|
1639 | chunklength = length(rev) | |
1636 | ladd(decomp(buffer(data, chunkstart - offset, chunklength))) |
|
1640 | ladd(decomp(buffer(data, chunkstart - offset, chunklength))) | |
1637 |
|
1641 | |||
1638 | return l |
|
1642 | return l | |
1639 |
|
1643 | |||
1640 | def _chunkclear(self): |
|
1644 | def _chunkclear(self): | |
1641 | """Clear the raw chunk cache.""" |
|
1645 | """Clear the raw chunk cache.""" | |
1642 | self._chunkcache = (0, b'') |
|
1646 | self._chunkcache = (0, b'') | |
1643 |
|
1647 | |||
1644 | def deltaparent(self, rev): |
|
1648 | def deltaparent(self, rev): | |
1645 | """return deltaparent of the given revision""" |
|
1649 | """return deltaparent of the given revision""" | |
1646 | base = self.index[rev][3] |
|
1650 | base = self.index[rev][3] | |
1647 | if base == rev: |
|
1651 | if base == rev: | |
1648 | return nullrev |
|
1652 | return nullrev | |
1649 | elif self._generaldelta: |
|
1653 | elif self._generaldelta: | |
1650 | return base |
|
1654 | return base | |
1651 | else: |
|
1655 | else: | |
1652 | return rev - 1 |
|
1656 | return rev - 1 | |
1653 |
|
1657 | |||
1654 | def issnapshot(self, rev): |
|
1658 | def issnapshot(self, rev): | |
1655 | """tells whether rev is a snapshot |
|
1659 | """tells whether rev is a snapshot | |
1656 | """ |
|
1660 | """ | |
1657 | if not self._sparserevlog: |
|
1661 | if not self._sparserevlog: | |
1658 | return self.deltaparent(rev) == nullrev |
|
1662 | return self.deltaparent(rev) == nullrev | |
1659 | elif util.safehasattr(self.index, b'issnapshot'): |
|
1663 | elif util.safehasattr(self.index, b'issnapshot'): | |
1660 | # directly assign the method to cache the testing and access |
|
1664 | # directly assign the method to cache the testing and access | |
1661 | self.issnapshot = self.index.issnapshot |
|
1665 | self.issnapshot = self.index.issnapshot | |
1662 | return self.issnapshot(rev) |
|
1666 | return self.issnapshot(rev) | |
1663 | if rev == nullrev: |
|
1667 | if rev == nullrev: | |
1664 | return True |
|
1668 | return True | |
1665 | entry = self.index[rev] |
|
1669 | entry = self.index[rev] | |
1666 | base = entry[3] |
|
1670 | base = entry[3] | |
1667 | if base == rev: |
|
1671 | if base == rev: | |
1668 | return True |
|
1672 | return True | |
1669 | if base == nullrev: |
|
1673 | if base == nullrev: | |
1670 | return True |
|
1674 | return True | |
1671 | p1 = entry[5] |
|
1675 | p1 = entry[5] | |
1672 | p2 = entry[6] |
|
1676 | p2 = entry[6] | |
1673 | if base == p1 or base == p2: |
|
1677 | if base == p1 or base == p2: | |
1674 | return False |
|
1678 | return False | |
1675 | return self.issnapshot(base) |
|
1679 | return self.issnapshot(base) | |
1676 |
|
1680 | |||
1677 | def snapshotdepth(self, rev): |
|
1681 | def snapshotdepth(self, rev): | |
1678 | """number of snapshot in the chain before this one""" |
|
1682 | """number of snapshot in the chain before this one""" | |
1679 | if not self.issnapshot(rev): |
|
1683 | if not self.issnapshot(rev): | |
1680 | raise error.ProgrammingError(b'revision %d not a snapshot') |
|
1684 | raise error.ProgrammingError(b'revision %d not a snapshot') | |
1681 | return len(self._deltachain(rev)[0]) - 1 |
|
1685 | return len(self._deltachain(rev)[0]) - 1 | |
1682 |
|
1686 | |||
1683 | def revdiff(self, rev1, rev2): |
|
1687 | def revdiff(self, rev1, rev2): | |
1684 | """return or calculate a delta between two revisions |
|
1688 | """return or calculate a delta between two revisions | |
1685 |
|
1689 | |||
1686 | The delta calculated is in binary form and is intended to be written to |
|
1690 | The delta calculated is in binary form and is intended to be written to | |
1687 | revlog data directly. So this function needs raw revision data. |
|
1691 | revlog data directly. So this function needs raw revision data. | |
1688 | """ |
|
1692 | """ | |
1689 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: |
|
1693 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: | |
1690 | return bytes(self._chunk(rev2)) |
|
1694 | return bytes(self._chunk(rev2)) | |
1691 |
|
1695 | |||
1692 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) |
|
1696 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) | |
1693 |
|
1697 | |||
1694 | def _processflags(self, text, flags, operation, raw=False): |
|
1698 | def _processflags(self, text, flags, operation, raw=False): | |
1695 | """deprecated entry point to access flag processors""" |
|
1699 | """deprecated entry point to access flag processors""" | |
1696 | msg = b'_processflag(...) use the specialized variant' |
|
1700 | msg = b'_processflag(...) use the specialized variant' | |
1697 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1701 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) | |
1698 | if raw: |
|
1702 | if raw: | |
1699 | return text, flagutil.processflagsraw(self, text, flags) |
|
1703 | return text, flagutil.processflagsraw(self, text, flags) | |
1700 | elif operation == b'read': |
|
1704 | elif operation == b'read': | |
1701 | return flagutil.processflagsread(self, text, flags) |
|
1705 | return flagutil.processflagsread(self, text, flags) | |
1702 | else: # write operation |
|
1706 | else: # write operation | |
1703 | return flagutil.processflagswrite(self, text, flags) |
|
1707 | return flagutil.processflagswrite(self, text, flags) | |
1704 |
|
1708 | |||
1705 | def revision(self, nodeorrev, _df=None, raw=False): |
|
1709 | def revision(self, nodeorrev, _df=None, raw=False): | |
1706 | """return an uncompressed revision of a given node or revision |
|
1710 | """return an uncompressed revision of a given node or revision | |
1707 | number. |
|
1711 | number. | |
1708 |
|
1712 | |||
1709 | _df - an existing file handle to read from. (internal-only) |
|
1713 | _df - an existing file handle to read from. (internal-only) | |
1710 | raw - an optional argument specifying if the revision data is to be |
|
1714 | raw - an optional argument specifying if the revision data is to be | |
1711 | treated as raw data when applying flag transforms. 'raw' should be set |
|
1715 | treated as raw data when applying flag transforms. 'raw' should be set | |
1712 | to True when generating changegroups or in debug commands. |
|
1716 | to True when generating changegroups or in debug commands. | |
1713 | """ |
|
1717 | """ | |
1714 | if raw: |
|
1718 | if raw: | |
1715 | msg = ( |
|
1719 | msg = ( | |
1716 | b'revlog.revision(..., raw=True) is deprecated, ' |
|
1720 | b'revlog.revision(..., raw=True) is deprecated, ' | |
1717 | b'use revlog.rawdata(...)' |
|
1721 | b'use revlog.rawdata(...)' | |
1718 | ) |
|
1722 | ) | |
1719 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1723 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) | |
1720 | return self._revisiondata(nodeorrev, _df, raw=raw)[0] |
|
1724 | return self._revisiondata(nodeorrev, _df, raw=raw)[0] | |
1721 |
|
1725 | |||
1722 | def sidedata(self, nodeorrev, _df=None): |
|
1726 | def sidedata(self, nodeorrev, _df=None): | |
1723 | """a map of extra data related to the changeset but not part of the hash |
|
1727 | """a map of extra data related to the changeset but not part of the hash | |
1724 |
|
1728 | |||
1725 | This function currently return a dictionary. However, more advanced |
|
1729 | This function currently return a dictionary. However, more advanced | |
1726 | mapping object will likely be used in the future for a more |
|
1730 | mapping object will likely be used in the future for a more | |
1727 | efficient/lazy code. |
|
1731 | efficient/lazy code. | |
1728 | """ |
|
1732 | """ | |
1729 | return self._revisiondata(nodeorrev, _df)[1] |
|
1733 | return self._revisiondata(nodeorrev, _df)[1] | |
1730 |
|
1734 | |||
1731 | def _revisiondata(self, nodeorrev, _df=None, raw=False): |
|
1735 | def _revisiondata(self, nodeorrev, _df=None, raw=False): | |
1732 | # deal with <nodeorrev> argument type |
|
1736 | # deal with <nodeorrev> argument type | |
1733 | if isinstance(nodeorrev, int): |
|
1737 | if isinstance(nodeorrev, int): | |
1734 | rev = nodeorrev |
|
1738 | rev = nodeorrev | |
1735 | node = self.node(rev) |
|
1739 | node = self.node(rev) | |
1736 | else: |
|
1740 | else: | |
1737 | node = nodeorrev |
|
1741 | node = nodeorrev | |
1738 | rev = None |
|
1742 | rev = None | |
1739 |
|
1743 | |||
1740 | # fast path the special `nullid` rev |
|
1744 | # fast path the special `nullid` rev | |
1741 | if node == nullid: |
|
1745 | if node == nullid: | |
1742 | return b"", {} |
|
1746 | return b"", {} | |
1743 |
|
1747 | |||
1744 | # The text as stored inside the revlog. Might be the revision or might |
|
1748 | # The text as stored inside the revlog. Might be the revision or might | |
1745 | # need to be processed to retrieve the revision. |
|
1749 | # need to be processed to retrieve the revision. | |
1746 | rawtext = None |
|
1750 | rawtext = None | |
1747 |
|
1751 | |||
1748 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) |
|
1752 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) | |
1749 |
|
1753 | |||
1750 | if raw and validated: |
|
1754 | if raw and validated: | |
1751 | # if we don't want to process the raw text and that raw |
|
1755 | # if we don't want to process the raw text and that raw | |
1752 | # text is cached, we can exit early. |
|
1756 | # text is cached, we can exit early. | |
1753 | return rawtext, {} |
|
1757 | return rawtext, {} | |
1754 | if rev is None: |
|
1758 | if rev is None: | |
1755 | rev = self.rev(node) |
|
1759 | rev = self.rev(node) | |
1756 | # the revlog's flag for this revision |
|
1760 | # the revlog's flag for this revision | |
1757 | # (usually alter its state or content) |
|
1761 | # (usually alter its state or content) | |
1758 | flags = self.flags(rev) |
|
1762 | flags = self.flags(rev) | |
1759 |
|
1763 | |||
1760 | if validated and flags == REVIDX_DEFAULT_FLAGS: |
|
1764 | if validated and flags == REVIDX_DEFAULT_FLAGS: | |
1761 | # no extra flags set, no flag processor runs, text = rawtext |
|
1765 | # no extra flags set, no flag processor runs, text = rawtext | |
1762 | return rawtext, {} |
|
1766 | return rawtext, {} | |
1763 |
|
1767 | |||
1764 | sidedata = {} |
|
1768 | sidedata = {} | |
1765 | if raw: |
|
1769 | if raw: | |
1766 | validatehash = flagutil.processflagsraw(self, rawtext, flags) |
|
1770 | validatehash = flagutil.processflagsraw(self, rawtext, flags) | |
1767 | text = rawtext |
|
1771 | text = rawtext | |
1768 | else: |
|
1772 | else: | |
1769 | try: |
|
1773 | try: | |
1770 | r = flagutil.processflagsread(self, rawtext, flags) |
|
1774 | r = flagutil.processflagsread(self, rawtext, flags) | |
1771 | except error.SidedataHashError as exc: |
|
1775 | except error.SidedataHashError as exc: | |
1772 | msg = _(b"integrity check failed on %s:%s sidedata key %d") |
|
1776 | msg = _(b"integrity check failed on %s:%s sidedata key %d") | |
1773 | msg %= (self.indexfile, pycompat.bytestr(rev), exc.sidedatakey) |
|
1777 | msg %= (self.indexfile, pycompat.bytestr(rev), exc.sidedatakey) | |
1774 | raise error.RevlogError(msg) |
|
1778 | raise error.RevlogError(msg) | |
1775 | text, validatehash, sidedata = r |
|
1779 | text, validatehash, sidedata = r | |
1776 | if validatehash: |
|
1780 | if validatehash: | |
1777 | self.checkhash(text, node, rev=rev) |
|
1781 | self.checkhash(text, node, rev=rev) | |
1778 | if not validated: |
|
1782 | if not validated: | |
1779 | self._revisioncache = (node, rev, rawtext) |
|
1783 | self._revisioncache = (node, rev, rawtext) | |
1780 |
|
1784 | |||
1781 | return text, sidedata |
|
1785 | return text, sidedata | |
1782 |
|
1786 | |||
1783 | def _rawtext(self, node, rev, _df=None): |
|
1787 | def _rawtext(self, node, rev, _df=None): | |
1784 | """return the possibly unvalidated rawtext for a revision |
|
1788 | """return the possibly unvalidated rawtext for a revision | |
1785 |
|
1789 | |||
1786 | returns (rev, rawtext, validated) |
|
1790 | returns (rev, rawtext, validated) | |
1787 | """ |
|
1791 | """ | |
1788 |
|
1792 | |||
1789 | # revision in the cache (could be useful to apply delta) |
|
1793 | # revision in the cache (could be useful to apply delta) | |
1790 | cachedrev = None |
|
1794 | cachedrev = None | |
1791 | # An intermediate text to apply deltas to |
|
1795 | # An intermediate text to apply deltas to | |
1792 | basetext = None |
|
1796 | basetext = None | |
1793 |
|
1797 | |||
1794 | # Check if we have the entry in cache |
|
1798 | # Check if we have the entry in cache | |
1795 | # The cache entry looks like (node, rev, rawtext) |
|
1799 | # The cache entry looks like (node, rev, rawtext) | |
1796 | if self._revisioncache: |
|
1800 | if self._revisioncache: | |
1797 | if self._revisioncache[0] == node: |
|
1801 | if self._revisioncache[0] == node: | |
1798 | return (rev, self._revisioncache[2], True) |
|
1802 | return (rev, self._revisioncache[2], True) | |
1799 | cachedrev = self._revisioncache[1] |
|
1803 | cachedrev = self._revisioncache[1] | |
1800 |
|
1804 | |||
1801 | if rev is None: |
|
1805 | if rev is None: | |
1802 | rev = self.rev(node) |
|
1806 | rev = self.rev(node) | |
1803 |
|
1807 | |||
1804 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) |
|
1808 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) | |
1805 | if stopped: |
|
1809 | if stopped: | |
1806 | basetext = self._revisioncache[2] |
|
1810 | basetext = self._revisioncache[2] | |
1807 |
|
1811 | |||
1808 | # drop cache to save memory, the caller is expected to |
|
1812 | # drop cache to save memory, the caller is expected to | |
1809 | # update self._revisioncache after validating the text |
|
1813 | # update self._revisioncache after validating the text | |
1810 | self._revisioncache = None |
|
1814 | self._revisioncache = None | |
1811 |
|
1815 | |||
1812 | targetsize = None |
|
1816 | targetsize = None | |
1813 | rawsize = self.index[rev][2] |
|
1817 | rawsize = self.index[rev][2] | |
1814 | if 0 <= rawsize: |
|
1818 | if 0 <= rawsize: | |
1815 | targetsize = 4 * rawsize |
|
1819 | targetsize = 4 * rawsize | |
1816 |
|
1820 | |||
1817 | bins = self._chunks(chain, df=_df, targetsize=targetsize) |
|
1821 | bins = self._chunks(chain, df=_df, targetsize=targetsize) | |
1818 | if basetext is None: |
|
1822 | if basetext is None: | |
1819 | basetext = bytes(bins[0]) |
|
1823 | basetext = bytes(bins[0]) | |
1820 | bins = bins[1:] |
|
1824 | bins = bins[1:] | |
1821 |
|
1825 | |||
1822 | rawtext = mdiff.patches(basetext, bins) |
|
1826 | rawtext = mdiff.patches(basetext, bins) | |
1823 | del basetext # let us have a chance to free memory early |
|
1827 | del basetext # let us have a chance to free memory early | |
1824 | return (rev, rawtext, False) |
|
1828 | return (rev, rawtext, False) | |
1825 |
|
1829 | |||
1826 | def rawdata(self, nodeorrev, _df=None): |
|
1830 | def rawdata(self, nodeorrev, _df=None): | |
1827 | """return an uncompressed raw data of a given node or revision number. |
|
1831 | """return an uncompressed raw data of a given node or revision number. | |
1828 |
|
1832 | |||
1829 | _df - an existing file handle to read from. (internal-only) |
|
1833 | _df - an existing file handle to read from. (internal-only) | |
1830 | """ |
|
1834 | """ | |
1831 | return self._revisiondata(nodeorrev, _df, raw=True)[0] |
|
1835 | return self._revisiondata(nodeorrev, _df, raw=True)[0] | |
1832 |
|
1836 | |||
1833 | def hash(self, text, p1, p2): |
|
1837 | def hash(self, text, p1, p2): | |
1834 | """Compute a node hash. |
|
1838 | """Compute a node hash. | |
1835 |
|
1839 | |||
1836 | Available as a function so that subclasses can replace the hash |
|
1840 | Available as a function so that subclasses can replace the hash | |
1837 | as needed. |
|
1841 | as needed. | |
1838 | """ |
|
1842 | """ | |
1839 | return storageutil.hashrevisionsha1(text, p1, p2) |
|
1843 | return storageutil.hashrevisionsha1(text, p1, p2) | |
1840 |
|
1844 | |||
1841 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
1845 | def checkhash(self, text, node, p1=None, p2=None, rev=None): | |
1842 | """Check node hash integrity. |
|
1846 | """Check node hash integrity. | |
1843 |
|
1847 | |||
1844 | Available as a function so that subclasses can extend hash mismatch |
|
1848 | Available as a function so that subclasses can extend hash mismatch | |
1845 | behaviors as needed. |
|
1849 | behaviors as needed. | |
1846 | """ |
|
1850 | """ | |
1847 | try: |
|
1851 | try: | |
1848 | if p1 is None and p2 is None: |
|
1852 | if p1 is None and p2 is None: | |
1849 | p1, p2 = self.parents(node) |
|
1853 | p1, p2 = self.parents(node) | |
1850 | if node != self.hash(text, p1, p2): |
|
1854 | if node != self.hash(text, p1, p2): | |
1851 | # Clear the revision cache on hash failure. The revision cache |
|
1855 | # Clear the revision cache on hash failure. The revision cache | |
1852 | # only stores the raw revision and clearing the cache does have |
|
1856 | # only stores the raw revision and clearing the cache does have | |
1853 | # the side-effect that we won't have a cache hit when the raw |
|
1857 | # the side-effect that we won't have a cache hit when the raw | |
1854 | # revision data is accessed. But this case should be rare and |
|
1858 | # revision data is accessed. But this case should be rare and | |
1855 | # it is extra work to teach the cache about the hash |
|
1859 | # it is extra work to teach the cache about the hash | |
1856 | # verification state. |
|
1860 | # verification state. | |
1857 | if self._revisioncache and self._revisioncache[0] == node: |
|
1861 | if self._revisioncache and self._revisioncache[0] == node: | |
1858 | self._revisioncache = None |
|
1862 | self._revisioncache = None | |
1859 |
|
1863 | |||
1860 | revornode = rev |
|
1864 | revornode = rev | |
1861 | if revornode is None: |
|
1865 | if revornode is None: | |
1862 | revornode = templatefilters.short(hex(node)) |
|
1866 | revornode = templatefilters.short(hex(node)) | |
1863 | raise error.RevlogError( |
|
1867 | raise error.RevlogError( | |
1864 | _(b"integrity check failed on %s:%s") |
|
1868 | _(b"integrity check failed on %s:%s") | |
1865 | % (self.indexfile, pycompat.bytestr(revornode)) |
|
1869 | % (self.indexfile, pycompat.bytestr(revornode)) | |
1866 | ) |
|
1870 | ) | |
1867 | except error.RevlogError: |
|
1871 | except error.RevlogError: | |
1868 | if self._censorable and storageutil.iscensoredtext(text): |
|
1872 | if self._censorable and storageutil.iscensoredtext(text): | |
1869 | raise error.CensoredNodeError(self.indexfile, node, text) |
|
1873 | raise error.CensoredNodeError(self.indexfile, node, text) | |
1870 | raise |
|
1874 | raise | |
1871 |
|
1875 | |||
1872 | def _enforceinlinesize(self, tr, fp=None): |
|
1876 | def _enforceinlinesize(self, tr, fp=None): | |
1873 | """Check if the revlog is too big for inline and convert if so. |
|
1877 | """Check if the revlog is too big for inline and convert if so. | |
1874 |
|
1878 | |||
1875 | This should be called after revisions are added to the revlog. If the |
|
1879 | This should be called after revisions are added to the revlog. If the | |
1876 | revlog has grown too large to be an inline revlog, it will convert it |
|
1880 | revlog has grown too large to be an inline revlog, it will convert it | |
1877 | to use multiple index and data files. |
|
1881 | to use multiple index and data files. | |
1878 | """ |
|
1882 | """ | |
1879 | tiprev = len(self) - 1 |
|
1883 | tiprev = len(self) - 1 | |
1880 | if ( |
|
1884 | if ( | |
1881 | not self._inline |
|
1885 | not self._inline | |
1882 | or (self.start(tiprev) + self.length(tiprev)) < _maxinline |
|
1886 | or (self.start(tiprev) + self.length(tiprev)) < _maxinline | |
1883 | ): |
|
1887 | ): | |
1884 | return |
|
1888 | return | |
1885 |
|
1889 | |||
1886 | trinfo = tr.find(self.indexfile) |
|
1890 | trinfo = tr.find(self.indexfile) | |
1887 | if trinfo is None: |
|
1891 | if trinfo is None: | |
1888 | raise error.RevlogError( |
|
1892 | raise error.RevlogError( | |
1889 | _(b"%s not found in the transaction") % self.indexfile |
|
1893 | _(b"%s not found in the transaction") % self.indexfile | |
1890 | ) |
|
1894 | ) | |
1891 |
|
1895 | |||
1892 | trindex = trinfo[2] |
|
1896 | trindex = trinfo[2] | |
1893 | if trindex is not None: |
|
1897 | if trindex is not None: | |
1894 | dataoff = self.start(trindex) |
|
1898 | dataoff = self.start(trindex) | |
1895 | else: |
|
1899 | else: | |
1896 | # revlog was stripped at start of transaction, use all leftover data |
|
1900 | # revlog was stripped at start of transaction, use all leftover data | |
1897 | trindex = len(self) - 1 |
|
1901 | trindex = len(self) - 1 | |
1898 | dataoff = self.end(tiprev) |
|
1902 | dataoff = self.end(tiprev) | |
1899 |
|
1903 | |||
1900 | tr.add(self.datafile, dataoff) |
|
1904 | tr.add(self.datafile, dataoff) | |
1901 |
|
1905 | |||
1902 | if fp: |
|
1906 | if fp: | |
1903 | fp.flush() |
|
1907 | fp.flush() | |
1904 | fp.close() |
|
1908 | fp.close() | |
1905 | # We can't use the cached file handle after close(). So prevent |
|
1909 | # We can't use the cached file handle after close(). So prevent | |
1906 | # its usage. |
|
1910 | # its usage. | |
1907 | self._writinghandles = None |
|
1911 | self._writinghandles = None | |
1908 |
|
1912 | |||
1909 | with self._indexfp(b'r') as ifh, self._datafp(b'w') as dfh: |
|
1913 | with self._indexfp(b'r') as ifh, self._datafp(b'w') as dfh: | |
1910 | for r in self: |
|
1914 | for r in self: | |
1911 | dfh.write(self._getsegmentforrevs(r, r, df=ifh)[1]) |
|
1915 | dfh.write(self._getsegmentforrevs(r, r, df=ifh)[1]) | |
1912 |
|
1916 | |||
1913 | with self._indexfp(b'w') as fp: |
|
1917 | with self._indexfp(b'w') as fp: | |
1914 | self.version &= ~FLAG_INLINE_DATA |
|
1918 | self.version &= ~FLAG_INLINE_DATA | |
1915 | self._inline = False |
|
1919 | self._inline = False | |
1916 | io = self._io |
|
1920 | io = self._io | |
1917 | for i in self: |
|
1921 | for i in self: | |
1918 | e = io.packentry(self.index[i], self.node, self.version, i) |
|
1922 | e = io.packentry(self.index[i], self.node, self.version, i) | |
1919 | fp.write(e) |
|
1923 | fp.write(e) | |
1920 |
|
1924 | |||
1921 | # the temp file replace the real index when we exit the context |
|
1925 | # the temp file replace the real index when we exit the context | |
1922 | # manager |
|
1926 | # manager | |
1923 |
|
1927 | |||
1924 | tr.replace(self.indexfile, trindex * self._io.size) |
|
1928 | tr.replace(self.indexfile, trindex * self._io.size) | |
1925 | self._chunkclear() |
|
1929 | self._chunkclear() | |
1926 |
|
1930 | |||
1927 | def _nodeduplicatecallback(self, transaction, node): |
|
1931 | def _nodeduplicatecallback(self, transaction, node): | |
1928 | """called when trying to add a node already stored. |
|
1932 | """called when trying to add a node already stored. | |
1929 | """ |
|
1933 | """ | |
1930 |
|
1934 | |||
1931 | def addrevision( |
|
1935 | def addrevision( | |
1932 | self, |
|
1936 | self, | |
1933 | text, |
|
1937 | text, | |
1934 | transaction, |
|
1938 | transaction, | |
1935 | link, |
|
1939 | link, | |
1936 | p1, |
|
1940 | p1, | |
1937 | p2, |
|
1941 | p2, | |
1938 | cachedelta=None, |
|
1942 | cachedelta=None, | |
1939 | node=None, |
|
1943 | node=None, | |
1940 | flags=REVIDX_DEFAULT_FLAGS, |
|
1944 | flags=REVIDX_DEFAULT_FLAGS, | |
1941 | deltacomputer=None, |
|
1945 | deltacomputer=None, | |
1942 | sidedata=None, |
|
1946 | sidedata=None, | |
1943 | ): |
|
1947 | ): | |
1944 | """add a revision to the log |
|
1948 | """add a revision to the log | |
1945 |
|
1949 | |||
1946 | text - the revision data to add |
|
1950 | text - the revision data to add | |
1947 | transaction - the transaction object used for rollback |
|
1951 | transaction - the transaction object used for rollback | |
1948 | link - the linkrev data to add |
|
1952 | link - the linkrev data to add | |
1949 | p1, p2 - the parent nodeids of the revision |
|
1953 | p1, p2 - the parent nodeids of the revision | |
1950 | cachedelta - an optional precomputed delta |
|
1954 | cachedelta - an optional precomputed delta | |
1951 | node - nodeid of revision; typically node is not specified, and it is |
|
1955 | node - nodeid of revision; typically node is not specified, and it is | |
1952 | computed by default as hash(text, p1, p2), however subclasses might |
|
1956 | computed by default as hash(text, p1, p2), however subclasses might | |
1953 | use different hashing method (and override checkhash() in such case) |
|
1957 | use different hashing method (and override checkhash() in such case) | |
1954 | flags - the known flags to set on the revision |
|
1958 | flags - the known flags to set on the revision | |
1955 | deltacomputer - an optional deltacomputer instance shared between |
|
1959 | deltacomputer - an optional deltacomputer instance shared between | |
1956 | multiple calls |
|
1960 | multiple calls | |
1957 | """ |
|
1961 | """ | |
1958 | if link == nullrev: |
|
1962 | if link == nullrev: | |
1959 | raise error.RevlogError( |
|
1963 | raise error.RevlogError( | |
1960 | _(b"attempted to add linkrev -1 to %s") % self.indexfile |
|
1964 | _(b"attempted to add linkrev -1 to %s") % self.indexfile | |
1961 | ) |
|
1965 | ) | |
1962 |
|
1966 | |||
1963 | if sidedata is None: |
|
1967 | if sidedata is None: | |
1964 | sidedata = {} |
|
1968 | sidedata = {} | |
1965 | flags = flags & ~REVIDX_SIDEDATA |
|
1969 | flags = flags & ~REVIDX_SIDEDATA | |
1966 | elif not self.hassidedata: |
|
1970 | elif not self.hassidedata: | |
1967 | raise error.ProgrammingError( |
|
1971 | raise error.ProgrammingError( | |
1968 | _(b"trying to add sidedata to a revlog who don't support them") |
|
1972 | _(b"trying to add sidedata to a revlog who don't support them") | |
1969 | ) |
|
1973 | ) | |
1970 | else: |
|
1974 | else: | |
1971 | flags |= REVIDX_SIDEDATA |
|
1975 | flags |= REVIDX_SIDEDATA | |
1972 |
|
1976 | |||
1973 | if flags: |
|
1977 | if flags: | |
1974 | node = node or self.hash(text, p1, p2) |
|
1978 | node = node or self.hash(text, p1, p2) | |
1975 |
|
1979 | |||
1976 | rawtext, validatehash = flagutil.processflagswrite( |
|
1980 | rawtext, validatehash = flagutil.processflagswrite( | |
1977 | self, text, flags, sidedata=sidedata |
|
1981 | self, text, flags, sidedata=sidedata | |
1978 | ) |
|
1982 | ) | |
1979 |
|
1983 | |||
1980 | # If the flag processor modifies the revision data, ignore any provided |
|
1984 | # If the flag processor modifies the revision data, ignore any provided | |
1981 | # cachedelta. |
|
1985 | # cachedelta. | |
1982 | if rawtext != text: |
|
1986 | if rawtext != text: | |
1983 | cachedelta = None |
|
1987 | cachedelta = None | |
1984 |
|
1988 | |||
1985 | if len(rawtext) > _maxentrysize: |
|
1989 | if len(rawtext) > _maxentrysize: | |
1986 | raise error.RevlogError( |
|
1990 | raise error.RevlogError( | |
1987 | _( |
|
1991 | _( | |
1988 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" |
|
1992 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" | |
1989 | ) |
|
1993 | ) | |
1990 | % (self.indexfile, len(rawtext)) |
|
1994 | % (self.indexfile, len(rawtext)) | |
1991 | ) |
|
1995 | ) | |
1992 |
|
1996 | |||
1993 | node = node or self.hash(rawtext, p1, p2) |
|
1997 | node = node or self.hash(rawtext, p1, p2) | |
1994 | if node in self.nodemap: |
|
1998 | if node in self.nodemap: | |
1995 | return node |
|
1999 | return node | |
1996 |
|
2000 | |||
1997 | if validatehash: |
|
2001 | if validatehash: | |
1998 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
2002 | self.checkhash(rawtext, node, p1=p1, p2=p2) | |
1999 |
|
2003 | |||
2000 | return self.addrawrevision( |
|
2004 | return self.addrawrevision( | |
2001 | rawtext, |
|
2005 | rawtext, | |
2002 | transaction, |
|
2006 | transaction, | |
2003 | link, |
|
2007 | link, | |
2004 | p1, |
|
2008 | p1, | |
2005 | p2, |
|
2009 | p2, | |
2006 | node, |
|
2010 | node, | |
2007 | flags, |
|
2011 | flags, | |
2008 | cachedelta=cachedelta, |
|
2012 | cachedelta=cachedelta, | |
2009 | deltacomputer=deltacomputer, |
|
2013 | deltacomputer=deltacomputer, | |
2010 | ) |
|
2014 | ) | |
2011 |
|
2015 | |||
2012 | def addrawrevision( |
|
2016 | def addrawrevision( | |
2013 | self, |
|
2017 | self, | |
2014 | rawtext, |
|
2018 | rawtext, | |
2015 | transaction, |
|
2019 | transaction, | |
2016 | link, |
|
2020 | link, | |
2017 | p1, |
|
2021 | p1, | |
2018 | p2, |
|
2022 | p2, | |
2019 | node, |
|
2023 | node, | |
2020 | flags, |
|
2024 | flags, | |
2021 | cachedelta=None, |
|
2025 | cachedelta=None, | |
2022 | deltacomputer=None, |
|
2026 | deltacomputer=None, | |
2023 | ): |
|
2027 | ): | |
2024 | """add a raw revision with known flags, node and parents |
|
2028 | """add a raw revision with known flags, node and parents | |
2025 | useful when reusing a revision not stored in this revlog (ex: received |
|
2029 | useful when reusing a revision not stored in this revlog (ex: received | |
2026 | over wire, or read from an external bundle). |
|
2030 | over wire, or read from an external bundle). | |
2027 | """ |
|
2031 | """ | |
2028 | dfh = None |
|
2032 | dfh = None | |
2029 | if not self._inline: |
|
2033 | if not self._inline: | |
2030 | dfh = self._datafp(b"a+") |
|
2034 | dfh = self._datafp(b"a+") | |
2031 | ifh = self._indexfp(b"a+") |
|
2035 | ifh = self._indexfp(b"a+") | |
2032 | try: |
|
2036 | try: | |
2033 | return self._addrevision( |
|
2037 | return self._addrevision( | |
2034 | node, |
|
2038 | node, | |
2035 | rawtext, |
|
2039 | rawtext, | |
2036 | transaction, |
|
2040 | transaction, | |
2037 | link, |
|
2041 | link, | |
2038 | p1, |
|
2042 | p1, | |
2039 | p2, |
|
2043 | p2, | |
2040 | flags, |
|
2044 | flags, | |
2041 | cachedelta, |
|
2045 | cachedelta, | |
2042 | ifh, |
|
2046 | ifh, | |
2043 | dfh, |
|
2047 | dfh, | |
2044 | deltacomputer=deltacomputer, |
|
2048 | deltacomputer=deltacomputer, | |
2045 | ) |
|
2049 | ) | |
2046 | finally: |
|
2050 | finally: | |
2047 | if dfh: |
|
2051 | if dfh: | |
2048 | dfh.close() |
|
2052 | dfh.close() | |
2049 | ifh.close() |
|
2053 | ifh.close() | |
2050 |
|
2054 | |||
2051 | def compress(self, data): |
|
2055 | def compress(self, data): | |
2052 | """Generate a possibly-compressed representation of data.""" |
|
2056 | """Generate a possibly-compressed representation of data.""" | |
2053 | if not data: |
|
2057 | if not data: | |
2054 | return b'', data |
|
2058 | return b'', data | |
2055 |
|
2059 | |||
2056 | compressed = self._compressor.compress(data) |
|
2060 | compressed = self._compressor.compress(data) | |
2057 |
|
2061 | |||
2058 | if compressed: |
|
2062 | if compressed: | |
2059 | # The revlog compressor added the header in the returned data. |
|
2063 | # The revlog compressor added the header in the returned data. | |
2060 | return b'', compressed |
|
2064 | return b'', compressed | |
2061 |
|
2065 | |||
2062 | if data[0:1] == b'\0': |
|
2066 | if data[0:1] == b'\0': | |
2063 | return b'', data |
|
2067 | return b'', data | |
2064 | return b'u', data |
|
2068 | return b'u', data | |
2065 |
|
2069 | |||
2066 | def decompress(self, data): |
|
2070 | def decompress(self, data): | |
2067 | """Decompress a revlog chunk. |
|
2071 | """Decompress a revlog chunk. | |
2068 |
|
2072 | |||
2069 | The chunk is expected to begin with a header identifying the |
|
2073 | The chunk is expected to begin with a header identifying the | |
2070 | format type so it can be routed to an appropriate decompressor. |
|
2074 | format type so it can be routed to an appropriate decompressor. | |
2071 | """ |
|
2075 | """ | |
2072 | if not data: |
|
2076 | if not data: | |
2073 | return data |
|
2077 | return data | |
2074 |
|
2078 | |||
2075 | # Revlogs are read much more frequently than they are written and many |
|
2079 | # Revlogs are read much more frequently than they are written and many | |
2076 | # chunks only take microseconds to decompress, so performance is |
|
2080 | # chunks only take microseconds to decompress, so performance is | |
2077 | # important here. |
|
2081 | # important here. | |
2078 | # |
|
2082 | # | |
2079 | # We can make a few assumptions about revlogs: |
|
2083 | # We can make a few assumptions about revlogs: | |
2080 | # |
|
2084 | # | |
2081 | # 1) the majority of chunks will be compressed (as opposed to inline |
|
2085 | # 1) the majority of chunks will be compressed (as opposed to inline | |
2082 | # raw data). |
|
2086 | # raw data). | |
2083 | # 2) decompressing *any* data will likely by at least 10x slower than |
|
2087 | # 2) decompressing *any* data will likely by at least 10x slower than | |
2084 | # returning raw inline data. |
|
2088 | # returning raw inline data. | |
2085 | # 3) we want to prioritize common and officially supported compression |
|
2089 | # 3) we want to prioritize common and officially supported compression | |
2086 | # engines |
|
2090 | # engines | |
2087 | # |
|
2091 | # | |
2088 | # It follows that we want to optimize for "decompress compressed data |
|
2092 | # It follows that we want to optimize for "decompress compressed data | |
2089 | # when encoded with common and officially supported compression engines" |
|
2093 | # when encoded with common and officially supported compression engines" | |
2090 | # case over "raw data" and "data encoded by less common or non-official |
|
2094 | # case over "raw data" and "data encoded by less common or non-official | |
2091 | # compression engines." That is why we have the inline lookup first |
|
2095 | # compression engines." That is why we have the inline lookup first | |
2092 | # followed by the compengines lookup. |
|
2096 | # followed by the compengines lookup. | |
2093 | # |
|
2097 | # | |
2094 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib |
|
2098 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib | |
2095 | # compressed chunks. And this matters for changelog and manifest reads. |
|
2099 | # compressed chunks. And this matters for changelog and manifest reads. | |
2096 | t = data[0:1] |
|
2100 | t = data[0:1] | |
2097 |
|
2101 | |||
2098 | if t == b'x': |
|
2102 | if t == b'x': | |
2099 | try: |
|
2103 | try: | |
2100 | return _zlibdecompress(data) |
|
2104 | return _zlibdecompress(data) | |
2101 | except zlib.error as e: |
|
2105 | except zlib.error as e: | |
2102 | raise error.RevlogError( |
|
2106 | raise error.RevlogError( | |
2103 | _(b'revlog decompress error: %s') |
|
2107 | _(b'revlog decompress error: %s') | |
2104 | % stringutil.forcebytestr(e) |
|
2108 | % stringutil.forcebytestr(e) | |
2105 | ) |
|
2109 | ) | |
2106 | # '\0' is more common than 'u' so it goes first. |
|
2110 | # '\0' is more common than 'u' so it goes first. | |
2107 | elif t == b'\0': |
|
2111 | elif t == b'\0': | |
2108 | return data |
|
2112 | return data | |
2109 | elif t == b'u': |
|
2113 | elif t == b'u': | |
2110 | return util.buffer(data, 1) |
|
2114 | return util.buffer(data, 1) | |
2111 |
|
2115 | |||
2112 | try: |
|
2116 | try: | |
2113 | compressor = self._decompressors[t] |
|
2117 | compressor = self._decompressors[t] | |
2114 | except KeyError: |
|
2118 | except KeyError: | |
2115 | try: |
|
2119 | try: | |
2116 | engine = util.compengines.forrevlogheader(t) |
|
2120 | engine = util.compengines.forrevlogheader(t) | |
2117 | compressor = engine.revlogcompressor(self._compengineopts) |
|
2121 | compressor = engine.revlogcompressor(self._compengineopts) | |
2118 | self._decompressors[t] = compressor |
|
2122 | self._decompressors[t] = compressor | |
2119 | except KeyError: |
|
2123 | except KeyError: | |
2120 | raise error.RevlogError(_(b'unknown compression type %r') % t) |
|
2124 | raise error.RevlogError(_(b'unknown compression type %r') % t) | |
2121 |
|
2125 | |||
2122 | return compressor.decompress(data) |
|
2126 | return compressor.decompress(data) | |
2123 |
|
2127 | |||
2124 | def _addrevision( |
|
2128 | def _addrevision( | |
2125 | self, |
|
2129 | self, | |
2126 | node, |
|
2130 | node, | |
2127 | rawtext, |
|
2131 | rawtext, | |
2128 | transaction, |
|
2132 | transaction, | |
2129 | link, |
|
2133 | link, | |
2130 | p1, |
|
2134 | p1, | |
2131 | p2, |
|
2135 | p2, | |
2132 | flags, |
|
2136 | flags, | |
2133 | cachedelta, |
|
2137 | cachedelta, | |
2134 | ifh, |
|
2138 | ifh, | |
2135 | dfh, |
|
2139 | dfh, | |
2136 | alwayscache=False, |
|
2140 | alwayscache=False, | |
2137 | deltacomputer=None, |
|
2141 | deltacomputer=None, | |
2138 | ): |
|
2142 | ): | |
2139 | """internal function to add revisions to the log |
|
2143 | """internal function to add revisions to the log | |
2140 |
|
2144 | |||
2141 | see addrevision for argument descriptions. |
|
2145 | see addrevision for argument descriptions. | |
2142 |
|
2146 | |||
2143 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. |
|
2147 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. | |
2144 |
|
2148 | |||
2145 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will |
|
2149 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will | |
2146 | be used. |
|
2150 | be used. | |
2147 |
|
2151 | |||
2148 | invariants: |
|
2152 | invariants: | |
2149 | - rawtext is optional (can be None); if not set, cachedelta must be set. |
|
2153 | - rawtext is optional (can be None); if not set, cachedelta must be set. | |
2150 | if both are set, they must correspond to each other. |
|
2154 | if both are set, they must correspond to each other. | |
2151 | """ |
|
2155 | """ | |
2152 | if node == nullid: |
|
2156 | if node == nullid: | |
2153 | raise error.RevlogError( |
|
2157 | raise error.RevlogError( | |
2154 | _(b"%s: attempt to add null revision") % self.indexfile |
|
2158 | _(b"%s: attempt to add null revision") % self.indexfile | |
2155 | ) |
|
2159 | ) | |
2156 | if node == wdirid or node in wdirfilenodeids: |
|
2160 | if node == wdirid or node in wdirfilenodeids: | |
2157 | raise error.RevlogError( |
|
2161 | raise error.RevlogError( | |
2158 | _(b"%s: attempt to add wdir revision") % self.indexfile |
|
2162 | _(b"%s: attempt to add wdir revision") % self.indexfile | |
2159 | ) |
|
2163 | ) | |
2160 |
|
2164 | |||
2161 | if self._inline: |
|
2165 | if self._inline: | |
2162 | fh = ifh |
|
2166 | fh = ifh | |
2163 | else: |
|
2167 | else: | |
2164 | fh = dfh |
|
2168 | fh = dfh | |
2165 |
|
2169 | |||
2166 | btext = [rawtext] |
|
2170 | btext = [rawtext] | |
2167 |
|
2171 | |||
2168 | curr = len(self) |
|
2172 | curr = len(self) | |
2169 | prev = curr - 1 |
|
2173 | prev = curr - 1 | |
2170 | offset = self.end(prev) |
|
2174 | offset = self.end(prev) | |
2171 | p1r, p2r = self.rev(p1), self.rev(p2) |
|
2175 | p1r, p2r = self.rev(p1), self.rev(p2) | |
2172 |
|
2176 | |||
2173 | # full versions are inserted when the needed deltas |
|
2177 | # full versions are inserted when the needed deltas | |
2174 | # become comparable to the uncompressed text |
|
2178 | # become comparable to the uncompressed text | |
2175 | if rawtext is None: |
|
2179 | if rawtext is None: | |
2176 | # need rawtext size, before changed by flag processors, which is |
|
2180 | # need rawtext size, before changed by flag processors, which is | |
2177 | # the non-raw size. use revlog explicitly to avoid filelog's extra |
|
2181 | # the non-raw size. use revlog explicitly to avoid filelog's extra | |
2178 | # logic that might remove metadata size. |
|
2182 | # logic that might remove metadata size. | |
2179 | textlen = mdiff.patchedsize( |
|
2183 | textlen = mdiff.patchedsize( | |
2180 | revlog.size(self, cachedelta[0]), cachedelta[1] |
|
2184 | revlog.size(self, cachedelta[0]), cachedelta[1] | |
2181 | ) |
|
2185 | ) | |
2182 | else: |
|
2186 | else: | |
2183 | textlen = len(rawtext) |
|
2187 | textlen = len(rawtext) | |
2184 |
|
2188 | |||
2185 | if deltacomputer is None: |
|
2189 | if deltacomputer is None: | |
2186 | deltacomputer = deltautil.deltacomputer(self) |
|
2190 | deltacomputer = deltautil.deltacomputer(self) | |
2187 |
|
2191 | |||
2188 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) |
|
2192 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) | |
2189 |
|
2193 | |||
2190 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) |
|
2194 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) | |
2191 |
|
2195 | |||
2192 | e = ( |
|
2196 | e = ( | |
2193 | offset_type(offset, flags), |
|
2197 | offset_type(offset, flags), | |
2194 | deltainfo.deltalen, |
|
2198 | deltainfo.deltalen, | |
2195 | textlen, |
|
2199 | textlen, | |
2196 | deltainfo.base, |
|
2200 | deltainfo.base, | |
2197 | link, |
|
2201 | link, | |
2198 | p1r, |
|
2202 | p1r, | |
2199 | p2r, |
|
2203 | p2r, | |
2200 | node, |
|
2204 | node, | |
2201 | ) |
|
2205 | ) | |
2202 | self.index.append(e) |
|
2206 | self.index.append(e) | |
2203 |
|
2207 | |||
2204 | # Reset the pure node cache start lookup offset to account for new |
|
2208 | # Reset the pure node cache start lookup offset to account for new | |
2205 | # revision. |
|
2209 | # revision. | |
2206 | if self._nodepos is not None: |
|
2210 | if self._nodepos is not None: | |
2207 | self._nodepos = curr |
|
2211 | self._nodepos = curr | |
2208 |
|
2212 | |||
2209 | entry = self._io.packentry(e, self.node, self.version, curr) |
|
2213 | entry = self._io.packentry(e, self.node, self.version, curr) | |
2210 | self._writeentry( |
|
2214 | self._writeentry( | |
2211 | transaction, ifh, dfh, entry, deltainfo.data, link, offset |
|
2215 | transaction, ifh, dfh, entry, deltainfo.data, link, offset | |
2212 | ) |
|
2216 | ) | |
2213 |
|
2217 | |||
2214 | rawtext = btext[0] |
|
2218 | rawtext = btext[0] | |
2215 |
|
2219 | |||
2216 | if alwayscache and rawtext is None: |
|
2220 | if alwayscache and rawtext is None: | |
2217 | rawtext = deltacomputer.buildtext(revinfo, fh) |
|
2221 | rawtext = deltacomputer.buildtext(revinfo, fh) | |
2218 |
|
2222 | |||
2219 | if type(rawtext) == bytes: # only accept immutable objects |
|
2223 | if type(rawtext) == bytes: # only accept immutable objects | |
2220 | self._revisioncache = (node, curr, rawtext) |
|
2224 | self._revisioncache = (node, curr, rawtext) | |
2221 | self._chainbasecache[curr] = deltainfo.chainbase |
|
2225 | self._chainbasecache[curr] = deltainfo.chainbase | |
2222 | return node |
|
2226 | return node | |
2223 |
|
2227 | |||
2224 | def _writeentry(self, transaction, ifh, dfh, entry, data, link, offset): |
|
2228 | def _writeentry(self, transaction, ifh, dfh, entry, data, link, offset): | |
2225 | # Files opened in a+ mode have inconsistent behavior on various |
|
2229 | # Files opened in a+ mode have inconsistent behavior on various | |
2226 | # platforms. Windows requires that a file positioning call be made |
|
2230 | # platforms. Windows requires that a file positioning call be made | |
2227 | # when the file handle transitions between reads and writes. See |
|
2231 | # when the file handle transitions between reads and writes. See | |
2228 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other |
|
2232 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other | |
2229 | # platforms, Python or the platform itself can be buggy. Some versions |
|
2233 | # platforms, Python or the platform itself can be buggy. Some versions | |
2230 | # of Solaris have been observed to not append at the end of the file |
|
2234 | # of Solaris have been observed to not append at the end of the file | |
2231 | # if the file was seeked to before the end. See issue4943 for more. |
|
2235 | # if the file was seeked to before the end. See issue4943 for more. | |
2232 | # |
|
2236 | # | |
2233 | # We work around this issue by inserting a seek() before writing. |
|
2237 | # We work around this issue by inserting a seek() before writing. | |
2234 | # Note: This is likely not necessary on Python 3. However, because |
|
2238 | # Note: This is likely not necessary on Python 3. However, because | |
2235 | # the file handle is reused for reads and may be seeked there, we need |
|
2239 | # the file handle is reused for reads and may be seeked there, we need | |
2236 | # to be careful before changing this. |
|
2240 | # to be careful before changing this. | |
2237 | ifh.seek(0, os.SEEK_END) |
|
2241 | ifh.seek(0, os.SEEK_END) | |
2238 | if dfh: |
|
2242 | if dfh: | |
2239 | dfh.seek(0, os.SEEK_END) |
|
2243 | dfh.seek(0, os.SEEK_END) | |
2240 |
|
2244 | |||
2241 | curr = len(self) - 1 |
|
2245 | curr = len(self) - 1 | |
2242 | if not self._inline: |
|
2246 | if not self._inline: | |
2243 | transaction.add(self.datafile, offset) |
|
2247 | transaction.add(self.datafile, offset) | |
2244 | transaction.add(self.indexfile, curr * len(entry)) |
|
2248 | transaction.add(self.indexfile, curr * len(entry)) | |
2245 | if data[0]: |
|
2249 | if data[0]: | |
2246 | dfh.write(data[0]) |
|
2250 | dfh.write(data[0]) | |
2247 | dfh.write(data[1]) |
|
2251 | dfh.write(data[1]) | |
2248 | ifh.write(entry) |
|
2252 | ifh.write(entry) | |
2249 | else: |
|
2253 | else: | |
2250 | offset += curr * self._io.size |
|
2254 | offset += curr * self._io.size | |
2251 | transaction.add(self.indexfile, offset, curr) |
|
2255 | transaction.add(self.indexfile, offset, curr) | |
2252 | ifh.write(entry) |
|
2256 | ifh.write(entry) | |
2253 | ifh.write(data[0]) |
|
2257 | ifh.write(data[0]) | |
2254 | ifh.write(data[1]) |
|
2258 | ifh.write(data[1]) | |
2255 | self._enforceinlinesize(transaction, ifh) |
|
2259 | self._enforceinlinesize(transaction, ifh) | |
2256 |
|
2260 | |||
2257 | def addgroup(self, deltas, linkmapper, transaction, addrevisioncb=None): |
|
2261 | def addgroup(self, deltas, linkmapper, transaction, addrevisioncb=None): | |
2258 | """ |
|
2262 | """ | |
2259 | add a delta group |
|
2263 | add a delta group | |
2260 |
|
2264 | |||
2261 | given a set of deltas, add them to the revision log. the |
|
2265 | given a set of deltas, add them to the revision log. the | |
2262 | first delta is against its parent, which should be in our |
|
2266 | first delta is against its parent, which should be in our | |
2263 | log, the rest are against the previous delta. |
|
2267 | log, the rest are against the previous delta. | |
2264 |
|
2268 | |||
2265 | If ``addrevisioncb`` is defined, it will be called with arguments of |
|
2269 | If ``addrevisioncb`` is defined, it will be called with arguments of | |
2266 | this revlog and the node that was added. |
|
2270 | this revlog and the node that was added. | |
2267 | """ |
|
2271 | """ | |
2268 |
|
2272 | |||
2269 | if self._writinghandles: |
|
2273 | if self._writinghandles: | |
2270 | raise error.ProgrammingError(b'cannot nest addgroup() calls') |
|
2274 | raise error.ProgrammingError(b'cannot nest addgroup() calls') | |
2271 |
|
2275 | |||
2272 | nodes = [] |
|
2276 | nodes = [] | |
2273 |
|
2277 | |||
2274 | r = len(self) |
|
2278 | r = len(self) | |
2275 | end = 0 |
|
2279 | end = 0 | |
2276 | if r: |
|
2280 | if r: | |
2277 | end = self.end(r - 1) |
|
2281 | end = self.end(r - 1) | |
2278 | ifh = self._indexfp(b"a+") |
|
2282 | ifh = self._indexfp(b"a+") | |
2279 | isize = r * self._io.size |
|
2283 | isize = r * self._io.size | |
2280 | if self._inline: |
|
2284 | if self._inline: | |
2281 | transaction.add(self.indexfile, end + isize, r) |
|
2285 | transaction.add(self.indexfile, end + isize, r) | |
2282 | dfh = None |
|
2286 | dfh = None | |
2283 | else: |
|
2287 | else: | |
2284 | transaction.add(self.indexfile, isize, r) |
|
2288 | transaction.add(self.indexfile, isize, r) | |
2285 | transaction.add(self.datafile, end) |
|
2289 | transaction.add(self.datafile, end) | |
2286 | dfh = self._datafp(b"a+") |
|
2290 | dfh = self._datafp(b"a+") | |
2287 |
|
2291 | |||
2288 | def flush(): |
|
2292 | def flush(): | |
2289 | if dfh: |
|
2293 | if dfh: | |
2290 | dfh.flush() |
|
2294 | dfh.flush() | |
2291 | ifh.flush() |
|
2295 | ifh.flush() | |
2292 |
|
2296 | |||
2293 | self._writinghandles = (ifh, dfh) |
|
2297 | self._writinghandles = (ifh, dfh) | |
2294 |
|
2298 | |||
2295 | try: |
|
2299 | try: | |
2296 | deltacomputer = deltautil.deltacomputer(self) |
|
2300 | deltacomputer = deltautil.deltacomputer(self) | |
2297 | # loop through our set of deltas |
|
2301 | # loop through our set of deltas | |
2298 | for data in deltas: |
|
2302 | for data in deltas: | |
2299 | node, p1, p2, linknode, deltabase, delta, flags = data |
|
2303 | node, p1, p2, linknode, deltabase, delta, flags = data | |
2300 | link = linkmapper(linknode) |
|
2304 | link = linkmapper(linknode) | |
2301 | flags = flags or REVIDX_DEFAULT_FLAGS |
|
2305 | flags = flags or REVIDX_DEFAULT_FLAGS | |
2302 |
|
2306 | |||
2303 | nodes.append(node) |
|
2307 | nodes.append(node) | |
2304 |
|
2308 | |||
2305 | if node in self.nodemap: |
|
2309 | if node in self.nodemap: | |
2306 | self._nodeduplicatecallback(transaction, node) |
|
2310 | self._nodeduplicatecallback(transaction, node) | |
2307 | # this can happen if two branches make the same change |
|
2311 | # this can happen if two branches make the same change | |
2308 | continue |
|
2312 | continue | |
2309 |
|
2313 | |||
2310 | for p in (p1, p2): |
|
2314 | for p in (p1, p2): | |
2311 | if p not in self.nodemap: |
|
2315 | if p not in self.nodemap: | |
2312 | raise error.LookupError( |
|
2316 | raise error.LookupError( | |
2313 | p, self.indexfile, _(b'unknown parent') |
|
2317 | p, self.indexfile, _(b'unknown parent') | |
2314 | ) |
|
2318 | ) | |
2315 |
|
2319 | |||
2316 | if deltabase not in self.nodemap: |
|
2320 | if deltabase not in self.nodemap: | |
2317 | raise error.LookupError( |
|
2321 | raise error.LookupError( | |
2318 | deltabase, self.indexfile, _(b'unknown delta base') |
|
2322 | deltabase, self.indexfile, _(b'unknown delta base') | |
2319 | ) |
|
2323 | ) | |
2320 |
|
2324 | |||
2321 | baserev = self.rev(deltabase) |
|
2325 | baserev = self.rev(deltabase) | |
2322 |
|
2326 | |||
2323 | if baserev != nullrev and self.iscensored(baserev): |
|
2327 | if baserev != nullrev and self.iscensored(baserev): | |
2324 | # if base is censored, delta must be full replacement in a |
|
2328 | # if base is censored, delta must be full replacement in a | |
2325 | # single patch operation |
|
2329 | # single patch operation | |
2326 | hlen = struct.calcsize(b">lll") |
|
2330 | hlen = struct.calcsize(b">lll") | |
2327 | oldlen = self.rawsize(baserev) |
|
2331 | oldlen = self.rawsize(baserev) | |
2328 | newlen = len(delta) - hlen |
|
2332 | newlen = len(delta) - hlen | |
2329 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): |
|
2333 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): | |
2330 | raise error.CensoredBaseError( |
|
2334 | raise error.CensoredBaseError( | |
2331 | self.indexfile, self.node(baserev) |
|
2335 | self.indexfile, self.node(baserev) | |
2332 | ) |
|
2336 | ) | |
2333 |
|
2337 | |||
2334 | if not flags and self._peek_iscensored(baserev, delta, flush): |
|
2338 | if not flags and self._peek_iscensored(baserev, delta, flush): | |
2335 | flags |= REVIDX_ISCENSORED |
|
2339 | flags |= REVIDX_ISCENSORED | |
2336 |
|
2340 | |||
2337 | # We assume consumers of addrevisioncb will want to retrieve |
|
2341 | # We assume consumers of addrevisioncb will want to retrieve | |
2338 | # the added revision, which will require a call to |
|
2342 | # the added revision, which will require a call to | |
2339 | # revision(). revision() will fast path if there is a cache |
|
2343 | # revision(). revision() will fast path if there is a cache | |
2340 | # hit. So, we tell _addrevision() to always cache in this case. |
|
2344 | # hit. So, we tell _addrevision() to always cache in this case. | |
2341 | # We're only using addgroup() in the context of changegroup |
|
2345 | # We're only using addgroup() in the context of changegroup | |
2342 | # generation so the revision data can always be handled as raw |
|
2346 | # generation so the revision data can always be handled as raw | |
2343 | # by the flagprocessor. |
|
2347 | # by the flagprocessor. | |
2344 | self._addrevision( |
|
2348 | self._addrevision( | |
2345 | node, |
|
2349 | node, | |
2346 | None, |
|
2350 | None, | |
2347 | transaction, |
|
2351 | transaction, | |
2348 | link, |
|
2352 | link, | |
2349 | p1, |
|
2353 | p1, | |
2350 | p2, |
|
2354 | p2, | |
2351 | flags, |
|
2355 | flags, | |
2352 | (baserev, delta), |
|
2356 | (baserev, delta), | |
2353 | ifh, |
|
2357 | ifh, | |
2354 | dfh, |
|
2358 | dfh, | |
2355 | alwayscache=bool(addrevisioncb), |
|
2359 | alwayscache=bool(addrevisioncb), | |
2356 | deltacomputer=deltacomputer, |
|
2360 | deltacomputer=deltacomputer, | |
2357 | ) |
|
2361 | ) | |
2358 |
|
2362 | |||
2359 | if addrevisioncb: |
|
2363 | if addrevisioncb: | |
2360 | addrevisioncb(self, node) |
|
2364 | addrevisioncb(self, node) | |
2361 |
|
2365 | |||
2362 | if not dfh and not self._inline: |
|
2366 | if not dfh and not self._inline: | |
2363 | # addrevision switched from inline to conventional |
|
2367 | # addrevision switched from inline to conventional | |
2364 | # reopen the index |
|
2368 | # reopen the index | |
2365 | ifh.close() |
|
2369 | ifh.close() | |
2366 | dfh = self._datafp(b"a+") |
|
2370 | dfh = self._datafp(b"a+") | |
2367 | ifh = self._indexfp(b"a+") |
|
2371 | ifh = self._indexfp(b"a+") | |
2368 | self._writinghandles = (ifh, dfh) |
|
2372 | self._writinghandles = (ifh, dfh) | |
2369 | finally: |
|
2373 | finally: | |
2370 | self._writinghandles = None |
|
2374 | self._writinghandles = None | |
2371 |
|
2375 | |||
2372 | if dfh: |
|
2376 | if dfh: | |
2373 | dfh.close() |
|
2377 | dfh.close() | |
2374 | ifh.close() |
|
2378 | ifh.close() | |
2375 |
|
2379 | |||
2376 | return nodes |
|
2380 | return nodes | |
2377 |
|
2381 | |||
2378 | def iscensored(self, rev): |
|
2382 | def iscensored(self, rev): | |
2379 | """Check if a file revision is censored.""" |
|
2383 | """Check if a file revision is censored.""" | |
2380 | if not self._censorable: |
|
2384 | if not self._censorable: | |
2381 | return False |
|
2385 | return False | |
2382 |
|
2386 | |||
2383 | return self.flags(rev) & REVIDX_ISCENSORED |
|
2387 | return self.flags(rev) & REVIDX_ISCENSORED | |
2384 |
|
2388 | |||
2385 | def _peek_iscensored(self, baserev, delta, flush): |
|
2389 | def _peek_iscensored(self, baserev, delta, flush): | |
2386 | """Quickly check if a delta produces a censored revision.""" |
|
2390 | """Quickly check if a delta produces a censored revision.""" | |
2387 | if not self._censorable: |
|
2391 | if not self._censorable: | |
2388 | return False |
|
2392 | return False | |
2389 |
|
2393 | |||
2390 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) |
|
2394 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) | |
2391 |
|
2395 | |||
2392 | def getstrippoint(self, minlink): |
|
2396 | def getstrippoint(self, minlink): | |
2393 | """find the minimum rev that must be stripped to strip the linkrev |
|
2397 | """find the minimum rev that must be stripped to strip the linkrev | |
2394 |
|
2398 | |||
2395 | Returns a tuple containing the minimum rev and a set of all revs that |
|
2399 | Returns a tuple containing the minimum rev and a set of all revs that | |
2396 | have linkrevs that will be broken by this strip. |
|
2400 | have linkrevs that will be broken by this strip. | |
2397 | """ |
|
2401 | """ | |
2398 | return storageutil.resolvestripinfo( |
|
2402 | return storageutil.resolvestripinfo( | |
2399 | minlink, |
|
2403 | minlink, | |
2400 | len(self) - 1, |
|
2404 | len(self) - 1, | |
2401 | self.headrevs(), |
|
2405 | self.headrevs(), | |
2402 | self.linkrev, |
|
2406 | self.linkrev, | |
2403 | self.parentrevs, |
|
2407 | self.parentrevs, | |
2404 | ) |
|
2408 | ) | |
2405 |
|
2409 | |||
2406 | def strip(self, minlink, transaction): |
|
2410 | def strip(self, minlink, transaction): | |
2407 | """truncate the revlog on the first revision with a linkrev >= minlink |
|
2411 | """truncate the revlog on the first revision with a linkrev >= minlink | |
2408 |
|
2412 | |||
2409 | This function is called when we're stripping revision minlink and |
|
2413 | This function is called when we're stripping revision minlink and | |
2410 | its descendants from the repository. |
|
2414 | its descendants from the repository. | |
2411 |
|
2415 | |||
2412 | We have to remove all revisions with linkrev >= minlink, because |
|
2416 | We have to remove all revisions with linkrev >= minlink, because | |
2413 | the equivalent changelog revisions will be renumbered after the |
|
2417 | the equivalent changelog revisions will be renumbered after the | |
2414 | strip. |
|
2418 | strip. | |
2415 |
|
2419 | |||
2416 | So we truncate the revlog on the first of these revisions, and |
|
2420 | So we truncate the revlog on the first of these revisions, and | |
2417 | trust that the caller has saved the revisions that shouldn't be |
|
2421 | trust that the caller has saved the revisions that shouldn't be | |
2418 | removed and that it'll re-add them after this truncation. |
|
2422 | removed and that it'll re-add them after this truncation. | |
2419 | """ |
|
2423 | """ | |
2420 | if len(self) == 0: |
|
2424 | if len(self) == 0: | |
2421 | return |
|
2425 | return | |
2422 |
|
2426 | |||
2423 | rev, _ = self.getstrippoint(minlink) |
|
2427 | rev, _ = self.getstrippoint(minlink) | |
2424 | if rev == len(self): |
|
2428 | if rev == len(self): | |
2425 | return |
|
2429 | return | |
2426 |
|
2430 | |||
2427 | # first truncate the files on disk |
|
2431 | # first truncate the files on disk | |
2428 | end = self.start(rev) |
|
2432 | end = self.start(rev) | |
2429 | if not self._inline: |
|
2433 | if not self._inline: | |
2430 | transaction.add(self.datafile, end) |
|
2434 | transaction.add(self.datafile, end) | |
2431 | end = rev * self._io.size |
|
2435 | end = rev * self._io.size | |
2432 | else: |
|
2436 | else: | |
2433 | end += rev * self._io.size |
|
2437 | end += rev * self._io.size | |
2434 |
|
2438 | |||
2435 | transaction.add(self.indexfile, end) |
|
2439 | transaction.add(self.indexfile, end) | |
2436 |
|
2440 | |||
2437 | # then reset internal state in memory to forget those revisions |
|
2441 | # then reset internal state in memory to forget those revisions | |
2438 | self._revisioncache = None |
|
2442 | self._revisioncache = None | |
2439 | self._chaininfocache = {} |
|
2443 | self._chaininfocache = {} | |
2440 | self._chunkclear() |
|
2444 | self._chunkclear() | |
2441 |
|
2445 | |||
2442 | del self.index[rev:-1] |
|
2446 | del self.index[rev:-1] | |
2443 | self._nodepos = None |
|
2447 | self._nodepos = None | |
2444 |
|
2448 | |||
2445 | def checksize(self): |
|
2449 | def checksize(self): | |
2446 | """Check size of index and data files |
|
2450 | """Check size of index and data files | |
2447 |
|
2451 | |||
2448 | return a (dd, di) tuple. |
|
2452 | return a (dd, di) tuple. | |
2449 | - dd: extra bytes for the "data" file |
|
2453 | - dd: extra bytes for the "data" file | |
2450 | - di: extra bytes for the "index" file |
|
2454 | - di: extra bytes for the "index" file | |
2451 |
|
2455 | |||
2452 | A healthy revlog will return (0, 0). |
|
2456 | A healthy revlog will return (0, 0). | |
2453 | """ |
|
2457 | """ | |
2454 | expected = 0 |
|
2458 | expected = 0 | |
2455 | if len(self): |
|
2459 | if len(self): | |
2456 | expected = max(0, self.end(len(self) - 1)) |
|
2460 | expected = max(0, self.end(len(self) - 1)) | |
2457 |
|
2461 | |||
2458 | try: |
|
2462 | try: | |
2459 | with self._datafp() as f: |
|
2463 | with self._datafp() as f: | |
2460 | f.seek(0, io.SEEK_END) |
|
2464 | f.seek(0, io.SEEK_END) | |
2461 | actual = f.tell() |
|
2465 | actual = f.tell() | |
2462 | dd = actual - expected |
|
2466 | dd = actual - expected | |
2463 | except IOError as inst: |
|
2467 | except IOError as inst: | |
2464 | if inst.errno != errno.ENOENT: |
|
2468 | if inst.errno != errno.ENOENT: | |
2465 | raise |
|
2469 | raise | |
2466 | dd = 0 |
|
2470 | dd = 0 | |
2467 |
|
2471 | |||
2468 | try: |
|
2472 | try: | |
2469 | f = self.opener(self.indexfile) |
|
2473 | f = self.opener(self.indexfile) | |
2470 | f.seek(0, io.SEEK_END) |
|
2474 | f.seek(0, io.SEEK_END) | |
2471 | actual = f.tell() |
|
2475 | actual = f.tell() | |
2472 | f.close() |
|
2476 | f.close() | |
2473 | s = self._io.size |
|
2477 | s = self._io.size | |
2474 | i = max(0, actual // s) |
|
2478 | i = max(0, actual // s) | |
2475 | di = actual - (i * s) |
|
2479 | di = actual - (i * s) | |
2476 | if self._inline: |
|
2480 | if self._inline: | |
2477 | databytes = 0 |
|
2481 | databytes = 0 | |
2478 | for r in self: |
|
2482 | for r in self: | |
2479 | databytes += max(0, self.length(r)) |
|
2483 | databytes += max(0, self.length(r)) | |
2480 | dd = 0 |
|
2484 | dd = 0 | |
2481 | di = actual - len(self) * s - databytes |
|
2485 | di = actual - len(self) * s - databytes | |
2482 | except IOError as inst: |
|
2486 | except IOError as inst: | |
2483 | if inst.errno != errno.ENOENT: |
|
2487 | if inst.errno != errno.ENOENT: | |
2484 | raise |
|
2488 | raise | |
2485 | di = 0 |
|
2489 | di = 0 | |
2486 |
|
2490 | |||
2487 | return (dd, di) |
|
2491 | return (dd, di) | |
2488 |
|
2492 | |||
2489 | def files(self): |
|
2493 | def files(self): | |
2490 | res = [self.indexfile] |
|
2494 | res = [self.indexfile] | |
2491 | if not self._inline: |
|
2495 | if not self._inline: | |
2492 | res.append(self.datafile) |
|
2496 | res.append(self.datafile) | |
2493 | return res |
|
2497 | return res | |
2494 |
|
2498 | |||
2495 | def emitrevisions( |
|
2499 | def emitrevisions( | |
2496 | self, |
|
2500 | self, | |
2497 | nodes, |
|
2501 | nodes, | |
2498 | nodesorder=None, |
|
2502 | nodesorder=None, | |
2499 | revisiondata=False, |
|
2503 | revisiondata=False, | |
2500 | assumehaveparentrevisions=False, |
|
2504 | assumehaveparentrevisions=False, | |
2501 | deltamode=repository.CG_DELTAMODE_STD, |
|
2505 | deltamode=repository.CG_DELTAMODE_STD, | |
2502 | ): |
|
2506 | ): | |
2503 | if nodesorder not in (b'nodes', b'storage', b'linear', None): |
|
2507 | if nodesorder not in (b'nodes', b'storage', b'linear', None): | |
2504 | raise error.ProgrammingError( |
|
2508 | raise error.ProgrammingError( | |
2505 | b'unhandled value for nodesorder: %s' % nodesorder |
|
2509 | b'unhandled value for nodesorder: %s' % nodesorder | |
2506 | ) |
|
2510 | ) | |
2507 |
|
2511 | |||
2508 | if nodesorder is None and not self._generaldelta: |
|
2512 | if nodesorder is None and not self._generaldelta: | |
2509 | nodesorder = b'storage' |
|
2513 | nodesorder = b'storage' | |
2510 |
|
2514 | |||
2511 | if ( |
|
2515 | if ( | |
2512 | not self._storedeltachains |
|
2516 | not self._storedeltachains | |
2513 | and deltamode != repository.CG_DELTAMODE_PREV |
|
2517 | and deltamode != repository.CG_DELTAMODE_PREV | |
2514 | ): |
|
2518 | ): | |
2515 | deltamode = repository.CG_DELTAMODE_FULL |
|
2519 | deltamode = repository.CG_DELTAMODE_FULL | |
2516 |
|
2520 | |||
2517 | return storageutil.emitrevisions( |
|
2521 | return storageutil.emitrevisions( | |
2518 | self, |
|
2522 | self, | |
2519 | nodes, |
|
2523 | nodes, | |
2520 | nodesorder, |
|
2524 | nodesorder, | |
2521 | revlogrevisiondelta, |
|
2525 | revlogrevisiondelta, | |
2522 | deltaparentfn=self.deltaparent, |
|
2526 | deltaparentfn=self.deltaparent, | |
2523 | candeltafn=self.candelta, |
|
2527 | candeltafn=self.candelta, | |
2524 | rawsizefn=self.rawsize, |
|
2528 | rawsizefn=self.rawsize, | |
2525 | revdifffn=self.revdiff, |
|
2529 | revdifffn=self.revdiff, | |
2526 | flagsfn=self.flags, |
|
2530 | flagsfn=self.flags, | |
2527 | deltamode=deltamode, |
|
2531 | deltamode=deltamode, | |
2528 | revisiondata=revisiondata, |
|
2532 | revisiondata=revisiondata, | |
2529 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
2533 | assumehaveparentrevisions=assumehaveparentrevisions, | |
2530 | ) |
|
2534 | ) | |
2531 |
|
2535 | |||
2532 | DELTAREUSEALWAYS = b'always' |
|
2536 | DELTAREUSEALWAYS = b'always' | |
2533 | DELTAREUSESAMEREVS = b'samerevs' |
|
2537 | DELTAREUSESAMEREVS = b'samerevs' | |
2534 | DELTAREUSENEVER = b'never' |
|
2538 | DELTAREUSENEVER = b'never' | |
2535 |
|
2539 | |||
2536 | DELTAREUSEFULLADD = b'fulladd' |
|
2540 | DELTAREUSEFULLADD = b'fulladd' | |
2537 |
|
2541 | |||
2538 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} |
|
2542 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} | |
2539 |
|
2543 | |||
2540 | def clone( |
|
2544 | def clone( | |
2541 | self, |
|
2545 | self, | |
2542 | tr, |
|
2546 | tr, | |
2543 | destrevlog, |
|
2547 | destrevlog, | |
2544 | addrevisioncb=None, |
|
2548 | addrevisioncb=None, | |
2545 | deltareuse=DELTAREUSESAMEREVS, |
|
2549 | deltareuse=DELTAREUSESAMEREVS, | |
2546 | forcedeltabothparents=None, |
|
2550 | forcedeltabothparents=None, | |
2547 | sidedatacompanion=None, |
|
2551 | sidedatacompanion=None, | |
2548 | ): |
|
2552 | ): | |
2549 | """Copy this revlog to another, possibly with format changes. |
|
2553 | """Copy this revlog to another, possibly with format changes. | |
2550 |
|
2554 | |||
2551 | The destination revlog will contain the same revisions and nodes. |
|
2555 | The destination revlog will contain the same revisions and nodes. | |
2552 | However, it may not be bit-for-bit identical due to e.g. delta encoding |
|
2556 | However, it may not be bit-for-bit identical due to e.g. delta encoding | |
2553 | differences. |
|
2557 | differences. | |
2554 |
|
2558 | |||
2555 | The ``deltareuse`` argument control how deltas from the existing revlog |
|
2559 | The ``deltareuse`` argument control how deltas from the existing revlog | |
2556 | are preserved in the destination revlog. The argument can have the |
|
2560 | are preserved in the destination revlog. The argument can have the | |
2557 | following values: |
|
2561 | following values: | |
2558 |
|
2562 | |||
2559 | DELTAREUSEALWAYS |
|
2563 | DELTAREUSEALWAYS | |
2560 | Deltas will always be reused (if possible), even if the destination |
|
2564 | Deltas will always be reused (if possible), even if the destination | |
2561 | revlog would not select the same revisions for the delta. This is the |
|
2565 | revlog would not select the same revisions for the delta. This is the | |
2562 | fastest mode of operation. |
|
2566 | fastest mode of operation. | |
2563 | DELTAREUSESAMEREVS |
|
2567 | DELTAREUSESAMEREVS | |
2564 | Deltas will be reused if the destination revlog would pick the same |
|
2568 | Deltas will be reused if the destination revlog would pick the same | |
2565 | revisions for the delta. This mode strikes a balance between speed |
|
2569 | revisions for the delta. This mode strikes a balance between speed | |
2566 | and optimization. |
|
2570 | and optimization. | |
2567 | DELTAREUSENEVER |
|
2571 | DELTAREUSENEVER | |
2568 | Deltas will never be reused. This is the slowest mode of execution. |
|
2572 | Deltas will never be reused. This is the slowest mode of execution. | |
2569 | This mode can be used to recompute deltas (e.g. if the diff/delta |
|
2573 | This mode can be used to recompute deltas (e.g. if the diff/delta | |
2570 | algorithm changes). |
|
2574 | algorithm changes). | |
2571 | DELTAREUSEFULLADD |
|
2575 | DELTAREUSEFULLADD | |
2572 | Revision will be re-added as if their were new content. This is |
|
2576 | Revision will be re-added as if their were new content. This is | |
2573 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. |
|
2577 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. | |
2574 | eg: large file detection and handling. |
|
2578 | eg: large file detection and handling. | |
2575 |
|
2579 | |||
2576 | Delta computation can be slow, so the choice of delta reuse policy can |
|
2580 | Delta computation can be slow, so the choice of delta reuse policy can | |
2577 | significantly affect run time. |
|
2581 | significantly affect run time. | |
2578 |
|
2582 | |||
2579 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between |
|
2583 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between | |
2580 | two extremes. Deltas will be reused if they are appropriate. But if the |
|
2584 | two extremes. Deltas will be reused if they are appropriate. But if the | |
2581 | delta could choose a better revision, it will do so. This means if you |
|
2585 | delta could choose a better revision, it will do so. This means if you | |
2582 | are converting a non-generaldelta revlog to a generaldelta revlog, |
|
2586 | are converting a non-generaldelta revlog to a generaldelta revlog, | |
2583 | deltas will be recomputed if the delta's parent isn't a parent of the |
|
2587 | deltas will be recomputed if the delta's parent isn't a parent of the | |
2584 | revision. |
|
2588 | revision. | |
2585 |
|
2589 | |||
2586 | In addition to the delta policy, the ``forcedeltabothparents`` |
|
2590 | In addition to the delta policy, the ``forcedeltabothparents`` | |
2587 | argument controls whether to force compute deltas against both parents |
|
2591 | argument controls whether to force compute deltas against both parents | |
2588 | for merges. By default, the current default is used. |
|
2592 | for merges. By default, the current default is used. | |
2589 |
|
2593 | |||
2590 | If not None, the `sidedatacompanion` is callable that accept two |
|
2594 | If not None, the `sidedatacompanion` is callable that accept two | |
2591 | arguments: |
|
2595 | arguments: | |
2592 |
|
2596 | |||
2593 | (srcrevlog, rev) |
|
2597 | (srcrevlog, rev) | |
2594 |
|
2598 | |||
2595 | and return a triplet that control changes to sidedata content from the |
|
2599 | and return a triplet that control changes to sidedata content from the | |
2596 | old revision to the new clone result: |
|
2600 | old revision to the new clone result: | |
2597 |
|
2601 | |||
2598 | (dropall, filterout, update) |
|
2602 | (dropall, filterout, update) | |
2599 |
|
2603 | |||
2600 | * if `dropall` is True, all sidedata should be dropped |
|
2604 | * if `dropall` is True, all sidedata should be dropped | |
2601 | * `filterout` is a set of sidedata keys that should be dropped |
|
2605 | * `filterout` is a set of sidedata keys that should be dropped | |
2602 | * `update` is a mapping of additionnal/new key -> value |
|
2606 | * `update` is a mapping of additionnal/new key -> value | |
2603 | """ |
|
2607 | """ | |
2604 | if deltareuse not in self.DELTAREUSEALL: |
|
2608 | if deltareuse not in self.DELTAREUSEALL: | |
2605 | raise ValueError( |
|
2609 | raise ValueError( | |
2606 | _(b'value for deltareuse invalid: %s') % deltareuse |
|
2610 | _(b'value for deltareuse invalid: %s') % deltareuse | |
2607 | ) |
|
2611 | ) | |
2608 |
|
2612 | |||
2609 | if len(destrevlog): |
|
2613 | if len(destrevlog): | |
2610 | raise ValueError(_(b'destination revlog is not empty')) |
|
2614 | raise ValueError(_(b'destination revlog is not empty')) | |
2611 |
|
2615 | |||
2612 | if getattr(self, 'filteredrevs', None): |
|
2616 | if getattr(self, 'filteredrevs', None): | |
2613 | raise ValueError(_(b'source revlog has filtered revisions')) |
|
2617 | raise ValueError(_(b'source revlog has filtered revisions')) | |
2614 | if getattr(destrevlog, 'filteredrevs', None): |
|
2618 | if getattr(destrevlog, 'filteredrevs', None): | |
2615 | raise ValueError(_(b'destination revlog has filtered revisions')) |
|
2619 | raise ValueError(_(b'destination revlog has filtered revisions')) | |
2616 |
|
2620 | |||
2617 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, |
|
2621 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, | |
2618 | # if possible. |
|
2622 | # if possible. | |
2619 | oldlazydelta = destrevlog._lazydelta |
|
2623 | oldlazydelta = destrevlog._lazydelta | |
2620 | oldlazydeltabase = destrevlog._lazydeltabase |
|
2624 | oldlazydeltabase = destrevlog._lazydeltabase | |
2621 | oldamd = destrevlog._deltabothparents |
|
2625 | oldamd = destrevlog._deltabothparents | |
2622 |
|
2626 | |||
2623 | try: |
|
2627 | try: | |
2624 | if deltareuse == self.DELTAREUSEALWAYS: |
|
2628 | if deltareuse == self.DELTAREUSEALWAYS: | |
2625 | destrevlog._lazydeltabase = True |
|
2629 | destrevlog._lazydeltabase = True | |
2626 | destrevlog._lazydelta = True |
|
2630 | destrevlog._lazydelta = True | |
2627 | elif deltareuse == self.DELTAREUSESAMEREVS: |
|
2631 | elif deltareuse == self.DELTAREUSESAMEREVS: | |
2628 | destrevlog._lazydeltabase = False |
|
2632 | destrevlog._lazydeltabase = False | |
2629 | destrevlog._lazydelta = True |
|
2633 | destrevlog._lazydelta = True | |
2630 | elif deltareuse == self.DELTAREUSENEVER: |
|
2634 | elif deltareuse == self.DELTAREUSENEVER: | |
2631 | destrevlog._lazydeltabase = False |
|
2635 | destrevlog._lazydeltabase = False | |
2632 | destrevlog._lazydelta = False |
|
2636 | destrevlog._lazydelta = False | |
2633 |
|
2637 | |||
2634 | destrevlog._deltabothparents = forcedeltabothparents or oldamd |
|
2638 | destrevlog._deltabothparents = forcedeltabothparents or oldamd | |
2635 |
|
2639 | |||
2636 | self._clone( |
|
2640 | self._clone( | |
2637 | tr, |
|
2641 | tr, | |
2638 | destrevlog, |
|
2642 | destrevlog, | |
2639 | addrevisioncb, |
|
2643 | addrevisioncb, | |
2640 | deltareuse, |
|
2644 | deltareuse, | |
2641 | forcedeltabothparents, |
|
2645 | forcedeltabothparents, | |
2642 | sidedatacompanion, |
|
2646 | sidedatacompanion, | |
2643 | ) |
|
2647 | ) | |
2644 |
|
2648 | |||
2645 | finally: |
|
2649 | finally: | |
2646 | destrevlog._lazydelta = oldlazydelta |
|
2650 | destrevlog._lazydelta = oldlazydelta | |
2647 | destrevlog._lazydeltabase = oldlazydeltabase |
|
2651 | destrevlog._lazydeltabase = oldlazydeltabase | |
2648 | destrevlog._deltabothparents = oldamd |
|
2652 | destrevlog._deltabothparents = oldamd | |
2649 |
|
2653 | |||
2650 | def _clone( |
|
2654 | def _clone( | |
2651 | self, |
|
2655 | self, | |
2652 | tr, |
|
2656 | tr, | |
2653 | destrevlog, |
|
2657 | destrevlog, | |
2654 | addrevisioncb, |
|
2658 | addrevisioncb, | |
2655 | deltareuse, |
|
2659 | deltareuse, | |
2656 | forcedeltabothparents, |
|
2660 | forcedeltabothparents, | |
2657 | sidedatacompanion, |
|
2661 | sidedatacompanion, | |
2658 | ): |
|
2662 | ): | |
2659 | """perform the core duty of `revlog.clone` after parameter processing""" |
|
2663 | """perform the core duty of `revlog.clone` after parameter processing""" | |
2660 | deltacomputer = deltautil.deltacomputer(destrevlog) |
|
2664 | deltacomputer = deltautil.deltacomputer(destrevlog) | |
2661 | index = self.index |
|
2665 | index = self.index | |
2662 | for rev in self: |
|
2666 | for rev in self: | |
2663 | entry = index[rev] |
|
2667 | entry = index[rev] | |
2664 |
|
2668 | |||
2665 | # Some classes override linkrev to take filtered revs into |
|
2669 | # Some classes override linkrev to take filtered revs into | |
2666 | # account. Use raw entry from index. |
|
2670 | # account. Use raw entry from index. | |
2667 | flags = entry[0] & 0xFFFF |
|
2671 | flags = entry[0] & 0xFFFF | |
2668 | linkrev = entry[4] |
|
2672 | linkrev = entry[4] | |
2669 | p1 = index[entry[5]][7] |
|
2673 | p1 = index[entry[5]][7] | |
2670 | p2 = index[entry[6]][7] |
|
2674 | p2 = index[entry[6]][7] | |
2671 | node = entry[7] |
|
2675 | node = entry[7] | |
2672 |
|
2676 | |||
2673 | sidedataactions = (False, [], {}) |
|
2677 | sidedataactions = (False, [], {}) | |
2674 | if sidedatacompanion is not None: |
|
2678 | if sidedatacompanion is not None: | |
2675 | sidedataactions = sidedatacompanion(self, rev) |
|
2679 | sidedataactions = sidedatacompanion(self, rev) | |
2676 |
|
2680 | |||
2677 | # (Possibly) reuse the delta from the revlog if allowed and |
|
2681 | # (Possibly) reuse the delta from the revlog if allowed and | |
2678 | # the revlog chunk is a delta. |
|
2682 | # the revlog chunk is a delta. | |
2679 | cachedelta = None |
|
2683 | cachedelta = None | |
2680 | rawtext = None |
|
2684 | rawtext = None | |
2681 | if any(sidedataactions) or deltareuse == self.DELTAREUSEFULLADD: |
|
2685 | if any(sidedataactions) or deltareuse == self.DELTAREUSEFULLADD: | |
2682 | dropall, filterout, update = sidedataactions |
|
2686 | dropall, filterout, update = sidedataactions | |
2683 | text, sidedata = self._revisiondata(rev) |
|
2687 | text, sidedata = self._revisiondata(rev) | |
2684 | if dropall: |
|
2688 | if dropall: | |
2685 | sidedata = {} |
|
2689 | sidedata = {} | |
2686 | for key in filterout: |
|
2690 | for key in filterout: | |
2687 | sidedata.pop(key, None) |
|
2691 | sidedata.pop(key, None) | |
2688 | sidedata.update(update) |
|
2692 | sidedata.update(update) | |
2689 | if not sidedata: |
|
2693 | if not sidedata: | |
2690 | sidedata = None |
|
2694 | sidedata = None | |
2691 | destrevlog.addrevision( |
|
2695 | destrevlog.addrevision( | |
2692 | text, |
|
2696 | text, | |
2693 | tr, |
|
2697 | tr, | |
2694 | linkrev, |
|
2698 | linkrev, | |
2695 | p1, |
|
2699 | p1, | |
2696 | p2, |
|
2700 | p2, | |
2697 | cachedelta=cachedelta, |
|
2701 | cachedelta=cachedelta, | |
2698 | node=node, |
|
2702 | node=node, | |
2699 | flags=flags, |
|
2703 | flags=flags, | |
2700 | deltacomputer=deltacomputer, |
|
2704 | deltacomputer=deltacomputer, | |
2701 | sidedata=sidedata, |
|
2705 | sidedata=sidedata, | |
2702 | ) |
|
2706 | ) | |
2703 | else: |
|
2707 | else: | |
2704 | if destrevlog._lazydelta: |
|
2708 | if destrevlog._lazydelta: | |
2705 | dp = self.deltaparent(rev) |
|
2709 | dp = self.deltaparent(rev) | |
2706 | if dp != nullrev: |
|
2710 | if dp != nullrev: | |
2707 | cachedelta = (dp, bytes(self._chunk(rev))) |
|
2711 | cachedelta = (dp, bytes(self._chunk(rev))) | |
2708 |
|
2712 | |||
2709 | if not cachedelta: |
|
2713 | if not cachedelta: | |
2710 | rawtext = self.rawdata(rev) |
|
2714 | rawtext = self.rawdata(rev) | |
2711 |
|
2715 | |||
2712 | ifh = destrevlog.opener( |
|
2716 | ifh = destrevlog.opener( | |
2713 | destrevlog.indexfile, b'a+', checkambig=False |
|
2717 | destrevlog.indexfile, b'a+', checkambig=False | |
2714 | ) |
|
2718 | ) | |
2715 | dfh = None |
|
2719 | dfh = None | |
2716 | if not destrevlog._inline: |
|
2720 | if not destrevlog._inline: | |
2717 | dfh = destrevlog.opener(destrevlog.datafile, b'a+') |
|
2721 | dfh = destrevlog.opener(destrevlog.datafile, b'a+') | |
2718 | try: |
|
2722 | try: | |
2719 | destrevlog._addrevision( |
|
2723 | destrevlog._addrevision( | |
2720 | node, |
|
2724 | node, | |
2721 | rawtext, |
|
2725 | rawtext, | |
2722 | tr, |
|
2726 | tr, | |
2723 | linkrev, |
|
2727 | linkrev, | |
2724 | p1, |
|
2728 | p1, | |
2725 | p2, |
|
2729 | p2, | |
2726 | flags, |
|
2730 | flags, | |
2727 | cachedelta, |
|
2731 | cachedelta, | |
2728 | ifh, |
|
2732 | ifh, | |
2729 | dfh, |
|
2733 | dfh, | |
2730 | deltacomputer=deltacomputer, |
|
2734 | deltacomputer=deltacomputer, | |
2731 | ) |
|
2735 | ) | |
2732 | finally: |
|
2736 | finally: | |
2733 | if dfh: |
|
2737 | if dfh: | |
2734 | dfh.close() |
|
2738 | dfh.close() | |
2735 | ifh.close() |
|
2739 | ifh.close() | |
2736 |
|
2740 | |||
2737 | if addrevisioncb: |
|
2741 | if addrevisioncb: | |
2738 | addrevisioncb(self, rev, node) |
|
2742 | addrevisioncb(self, rev, node) | |
2739 |
|
2743 | |||
2740 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
2744 | def censorrevision(self, tr, censornode, tombstone=b''): | |
2741 | if (self.version & 0xFFFF) == REVLOGV0: |
|
2745 | if (self.version & 0xFFFF) == REVLOGV0: | |
2742 | raise error.RevlogError( |
|
2746 | raise error.RevlogError( | |
2743 | _(b'cannot censor with version %d revlogs') % self.version |
|
2747 | _(b'cannot censor with version %d revlogs') % self.version | |
2744 | ) |
|
2748 | ) | |
2745 |
|
2749 | |||
2746 | censorrev = self.rev(censornode) |
|
2750 | censorrev = self.rev(censornode) | |
2747 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') |
|
2751 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') | |
2748 |
|
2752 | |||
2749 | if len(tombstone) > self.rawsize(censorrev): |
|
2753 | if len(tombstone) > self.rawsize(censorrev): | |
2750 | raise error.Abort( |
|
2754 | raise error.Abort( | |
2751 | _(b'censor tombstone must be no longer than censored data') |
|
2755 | _(b'censor tombstone must be no longer than censored data') | |
2752 | ) |
|
2756 | ) | |
2753 |
|
2757 | |||
2754 | # Rewriting the revlog in place is hard. Our strategy for censoring is |
|
2758 | # Rewriting the revlog in place is hard. Our strategy for censoring is | |
2755 | # to create a new revlog, copy all revisions to it, then replace the |
|
2759 | # to create a new revlog, copy all revisions to it, then replace the | |
2756 | # revlogs on transaction close. |
|
2760 | # revlogs on transaction close. | |
2757 |
|
2761 | |||
2758 | newindexfile = self.indexfile + b'.tmpcensored' |
|
2762 | newindexfile = self.indexfile + b'.tmpcensored' | |
2759 | newdatafile = self.datafile + b'.tmpcensored' |
|
2763 | newdatafile = self.datafile + b'.tmpcensored' | |
2760 |
|
2764 | |||
2761 | # This is a bit dangerous. We could easily have a mismatch of state. |
|
2765 | # This is a bit dangerous. We could easily have a mismatch of state. | |
2762 | newrl = revlog(self.opener, newindexfile, newdatafile, censorable=True) |
|
2766 | newrl = revlog(self.opener, newindexfile, newdatafile, censorable=True) | |
2763 | newrl.version = self.version |
|
2767 | newrl.version = self.version | |
2764 | newrl._generaldelta = self._generaldelta |
|
2768 | newrl._generaldelta = self._generaldelta | |
2765 | newrl._io = self._io |
|
2769 | newrl._io = self._io | |
2766 |
|
2770 | |||
2767 | for rev in self.revs(): |
|
2771 | for rev in self.revs(): | |
2768 | node = self.node(rev) |
|
2772 | node = self.node(rev) | |
2769 | p1, p2 = self.parents(node) |
|
2773 | p1, p2 = self.parents(node) | |
2770 |
|
2774 | |||
2771 | if rev == censorrev: |
|
2775 | if rev == censorrev: | |
2772 | newrl.addrawrevision( |
|
2776 | newrl.addrawrevision( | |
2773 | tombstone, |
|
2777 | tombstone, | |
2774 | tr, |
|
2778 | tr, | |
2775 | self.linkrev(censorrev), |
|
2779 | self.linkrev(censorrev), | |
2776 | p1, |
|
2780 | p1, | |
2777 | p2, |
|
2781 | p2, | |
2778 | censornode, |
|
2782 | censornode, | |
2779 | REVIDX_ISCENSORED, |
|
2783 | REVIDX_ISCENSORED, | |
2780 | ) |
|
2784 | ) | |
2781 |
|
2785 | |||
2782 | if newrl.deltaparent(rev) != nullrev: |
|
2786 | if newrl.deltaparent(rev) != nullrev: | |
2783 | raise error.Abort( |
|
2787 | raise error.Abort( | |
2784 | _( |
|
2788 | _( | |
2785 | b'censored revision stored as delta; ' |
|
2789 | b'censored revision stored as delta; ' | |
2786 | b'cannot censor' |
|
2790 | b'cannot censor' | |
2787 | ), |
|
2791 | ), | |
2788 | hint=_( |
|
2792 | hint=_( | |
2789 | b'censoring of revlogs is not ' |
|
2793 | b'censoring of revlogs is not ' | |
2790 | b'fully implemented; please report ' |
|
2794 | b'fully implemented; please report ' | |
2791 | b'this bug' |
|
2795 | b'this bug' | |
2792 | ), |
|
2796 | ), | |
2793 | ) |
|
2797 | ) | |
2794 | continue |
|
2798 | continue | |
2795 |
|
2799 | |||
2796 | if self.iscensored(rev): |
|
2800 | if self.iscensored(rev): | |
2797 | if self.deltaparent(rev) != nullrev: |
|
2801 | if self.deltaparent(rev) != nullrev: | |
2798 | raise error.Abort( |
|
2802 | raise error.Abort( | |
2799 | _( |
|
2803 | _( | |
2800 | b'cannot censor due to censored ' |
|
2804 | b'cannot censor due to censored ' | |
2801 | b'revision having delta stored' |
|
2805 | b'revision having delta stored' | |
2802 | ) |
|
2806 | ) | |
2803 | ) |
|
2807 | ) | |
2804 | rawtext = self._chunk(rev) |
|
2808 | rawtext = self._chunk(rev) | |
2805 | else: |
|
2809 | else: | |
2806 | rawtext = self.rawdata(rev) |
|
2810 | rawtext = self.rawdata(rev) | |
2807 |
|
2811 | |||
2808 | newrl.addrawrevision( |
|
2812 | newrl.addrawrevision( | |
2809 | rawtext, tr, self.linkrev(rev), p1, p2, node, self.flags(rev) |
|
2813 | rawtext, tr, self.linkrev(rev), p1, p2, node, self.flags(rev) | |
2810 | ) |
|
2814 | ) | |
2811 |
|
2815 | |||
2812 | tr.addbackup(self.indexfile, location=b'store') |
|
2816 | tr.addbackup(self.indexfile, location=b'store') | |
2813 | if not self._inline: |
|
2817 | if not self._inline: | |
2814 | tr.addbackup(self.datafile, location=b'store') |
|
2818 | tr.addbackup(self.datafile, location=b'store') | |
2815 |
|
2819 | |||
2816 | self.opener.rename(newrl.indexfile, self.indexfile) |
|
2820 | self.opener.rename(newrl.indexfile, self.indexfile) | |
2817 | if not self._inline: |
|
2821 | if not self._inline: | |
2818 | self.opener.rename(newrl.datafile, self.datafile) |
|
2822 | self.opener.rename(newrl.datafile, self.datafile) | |
2819 |
|
2823 | |||
2820 | self.clearcaches() |
|
2824 | self.clearcaches() | |
2821 | self._loadindex() |
|
2825 | self._loadindex() | |
2822 |
|
2826 | |||
2823 | def verifyintegrity(self, state): |
|
2827 | def verifyintegrity(self, state): | |
2824 | """Verifies the integrity of the revlog. |
|
2828 | """Verifies the integrity of the revlog. | |
2825 |
|
2829 | |||
2826 | Yields ``revlogproblem`` instances describing problems that are |
|
2830 | Yields ``revlogproblem`` instances describing problems that are | |
2827 | found. |
|
2831 | found. | |
2828 | """ |
|
2832 | """ | |
2829 | dd, di = self.checksize() |
|
2833 | dd, di = self.checksize() | |
2830 | if dd: |
|
2834 | if dd: | |
2831 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) |
|
2835 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) | |
2832 | if di: |
|
2836 | if di: | |
2833 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) |
|
2837 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) | |
2834 |
|
2838 | |||
2835 | version = self.version & 0xFFFF |
|
2839 | version = self.version & 0xFFFF | |
2836 |
|
2840 | |||
2837 | # The verifier tells us what version revlog we should be. |
|
2841 | # The verifier tells us what version revlog we should be. | |
2838 | if version != state[b'expectedversion']: |
|
2842 | if version != state[b'expectedversion']: | |
2839 | yield revlogproblem( |
|
2843 | yield revlogproblem( | |
2840 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") |
|
2844 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") | |
2841 | % (self.indexfile, version, state[b'expectedversion']) |
|
2845 | % (self.indexfile, version, state[b'expectedversion']) | |
2842 | ) |
|
2846 | ) | |
2843 |
|
2847 | |||
2844 | state[b'skipread'] = set() |
|
2848 | state[b'skipread'] = set() | |
2845 |
|
2849 | |||
2846 | for rev in self: |
|
2850 | for rev in self: | |
2847 | node = self.node(rev) |
|
2851 | node = self.node(rev) | |
2848 |
|
2852 | |||
2849 | # Verify contents. 4 cases to care about: |
|
2853 | # Verify contents. 4 cases to care about: | |
2850 | # |
|
2854 | # | |
2851 | # common: the most common case |
|
2855 | # common: the most common case | |
2852 | # rename: with a rename |
|
2856 | # rename: with a rename | |
2853 | # meta: file content starts with b'\1\n', the metadata |
|
2857 | # meta: file content starts with b'\1\n', the metadata | |
2854 | # header defined in filelog.py, but without a rename |
|
2858 | # header defined in filelog.py, but without a rename | |
2855 | # ext: content stored externally |
|
2859 | # ext: content stored externally | |
2856 | # |
|
2860 | # | |
2857 | # More formally, their differences are shown below: |
|
2861 | # More formally, their differences are shown below: | |
2858 | # |
|
2862 | # | |
2859 | # | common | rename | meta | ext |
|
2863 | # | common | rename | meta | ext | |
2860 | # ------------------------------------------------------- |
|
2864 | # ------------------------------------------------------- | |
2861 | # flags() | 0 | 0 | 0 | not 0 |
|
2865 | # flags() | 0 | 0 | 0 | not 0 | |
2862 | # renamed() | False | True | False | ? |
|
2866 | # renamed() | False | True | False | ? | |
2863 | # rawtext[0:2]=='\1\n'| False | True | True | ? |
|
2867 | # rawtext[0:2]=='\1\n'| False | True | True | ? | |
2864 | # |
|
2868 | # | |
2865 | # "rawtext" means the raw text stored in revlog data, which |
|
2869 | # "rawtext" means the raw text stored in revlog data, which | |
2866 | # could be retrieved by "rawdata(rev)". "text" |
|
2870 | # could be retrieved by "rawdata(rev)". "text" | |
2867 | # mentioned below is "revision(rev)". |
|
2871 | # mentioned below is "revision(rev)". | |
2868 | # |
|
2872 | # | |
2869 | # There are 3 different lengths stored physically: |
|
2873 | # There are 3 different lengths stored physically: | |
2870 | # 1. L1: rawsize, stored in revlog index |
|
2874 | # 1. L1: rawsize, stored in revlog index | |
2871 | # 2. L2: len(rawtext), stored in revlog data |
|
2875 | # 2. L2: len(rawtext), stored in revlog data | |
2872 | # 3. L3: len(text), stored in revlog data if flags==0, or |
|
2876 | # 3. L3: len(text), stored in revlog data if flags==0, or | |
2873 | # possibly somewhere else if flags!=0 |
|
2877 | # possibly somewhere else if flags!=0 | |
2874 | # |
|
2878 | # | |
2875 | # L1 should be equal to L2. L3 could be different from them. |
|
2879 | # L1 should be equal to L2. L3 could be different from them. | |
2876 | # "text" may or may not affect commit hash depending on flag |
|
2880 | # "text" may or may not affect commit hash depending on flag | |
2877 | # processors (see flagutil.addflagprocessor). |
|
2881 | # processors (see flagutil.addflagprocessor). | |
2878 | # |
|
2882 | # | |
2879 | # | common | rename | meta | ext |
|
2883 | # | common | rename | meta | ext | |
2880 | # ------------------------------------------------- |
|
2884 | # ------------------------------------------------- | |
2881 | # rawsize() | L1 | L1 | L1 | L1 |
|
2885 | # rawsize() | L1 | L1 | L1 | L1 | |
2882 | # size() | L1 | L2-LM | L1(*) | L1 (?) |
|
2886 | # size() | L1 | L2-LM | L1(*) | L1 (?) | |
2883 | # len(rawtext) | L2 | L2 | L2 | L2 |
|
2887 | # len(rawtext) | L2 | L2 | L2 | L2 | |
2884 | # len(text) | L2 | L2 | L2 | L3 |
|
2888 | # len(text) | L2 | L2 | L2 | L3 | |
2885 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) |
|
2889 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) | |
2886 | # |
|
2890 | # | |
2887 | # LM: length of metadata, depending on rawtext |
|
2891 | # LM: length of metadata, depending on rawtext | |
2888 | # (*): not ideal, see comment in filelog.size |
|
2892 | # (*): not ideal, see comment in filelog.size | |
2889 | # (?): could be "- len(meta)" if the resolved content has |
|
2893 | # (?): could be "- len(meta)" if the resolved content has | |
2890 | # rename metadata |
|
2894 | # rename metadata | |
2891 | # |
|
2895 | # | |
2892 | # Checks needed to be done: |
|
2896 | # Checks needed to be done: | |
2893 | # 1. length check: L1 == L2, in all cases. |
|
2897 | # 1. length check: L1 == L2, in all cases. | |
2894 | # 2. hash check: depending on flag processor, we may need to |
|
2898 | # 2. hash check: depending on flag processor, we may need to | |
2895 | # use either "text" (external), or "rawtext" (in revlog). |
|
2899 | # use either "text" (external), or "rawtext" (in revlog). | |
2896 |
|
2900 | |||
2897 | try: |
|
2901 | try: | |
2898 | skipflags = state.get(b'skipflags', 0) |
|
2902 | skipflags = state.get(b'skipflags', 0) | |
2899 | if skipflags: |
|
2903 | if skipflags: | |
2900 | skipflags &= self.flags(rev) |
|
2904 | skipflags &= self.flags(rev) | |
2901 |
|
2905 | |||
2902 | if skipflags: |
|
2906 | if skipflags: | |
2903 | state[b'skipread'].add(node) |
|
2907 | state[b'skipread'].add(node) | |
2904 | else: |
|
2908 | else: | |
2905 | # Side-effect: read content and verify hash. |
|
2909 | # Side-effect: read content and verify hash. | |
2906 | self.revision(node) |
|
2910 | self.revision(node) | |
2907 |
|
2911 | |||
2908 | l1 = self.rawsize(rev) |
|
2912 | l1 = self.rawsize(rev) | |
2909 | l2 = len(self.rawdata(node)) |
|
2913 | l2 = len(self.rawdata(node)) | |
2910 |
|
2914 | |||
2911 | if l1 != l2: |
|
2915 | if l1 != l2: | |
2912 | yield revlogproblem( |
|
2916 | yield revlogproblem( | |
2913 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), |
|
2917 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), | |
2914 | node=node, |
|
2918 | node=node, | |
2915 | ) |
|
2919 | ) | |
2916 |
|
2920 | |||
2917 | except error.CensoredNodeError: |
|
2921 | except error.CensoredNodeError: | |
2918 | if state[b'erroroncensored']: |
|
2922 | if state[b'erroroncensored']: | |
2919 | yield revlogproblem( |
|
2923 | yield revlogproblem( | |
2920 | error=_(b'censored file data'), node=node |
|
2924 | error=_(b'censored file data'), node=node | |
2921 | ) |
|
2925 | ) | |
2922 | state[b'skipread'].add(node) |
|
2926 | state[b'skipread'].add(node) | |
2923 | except Exception as e: |
|
2927 | except Exception as e: | |
2924 | yield revlogproblem( |
|
2928 | yield revlogproblem( | |
2925 | error=_(b'unpacking %s: %s') |
|
2929 | error=_(b'unpacking %s: %s') | |
2926 | % (short(node), stringutil.forcebytestr(e)), |
|
2930 | % (short(node), stringutil.forcebytestr(e)), | |
2927 | node=node, |
|
2931 | node=node, | |
2928 | ) |
|
2932 | ) | |
2929 | state[b'skipread'].add(node) |
|
2933 | state[b'skipread'].add(node) | |
2930 |
|
2934 | |||
2931 | def storageinfo( |
|
2935 | def storageinfo( | |
2932 | self, |
|
2936 | self, | |
2933 | exclusivefiles=False, |
|
2937 | exclusivefiles=False, | |
2934 | sharedfiles=False, |
|
2938 | sharedfiles=False, | |
2935 | revisionscount=False, |
|
2939 | revisionscount=False, | |
2936 | trackedsize=False, |
|
2940 | trackedsize=False, | |
2937 | storedsize=False, |
|
2941 | storedsize=False, | |
2938 | ): |
|
2942 | ): | |
2939 | d = {} |
|
2943 | d = {} | |
2940 |
|
2944 | |||
2941 | if exclusivefiles: |
|
2945 | if exclusivefiles: | |
2942 | d[b'exclusivefiles'] = [(self.opener, self.indexfile)] |
|
2946 | d[b'exclusivefiles'] = [(self.opener, self.indexfile)] | |
2943 | if not self._inline: |
|
2947 | if not self._inline: | |
2944 | d[b'exclusivefiles'].append((self.opener, self.datafile)) |
|
2948 | d[b'exclusivefiles'].append((self.opener, self.datafile)) | |
2945 |
|
2949 | |||
2946 | if sharedfiles: |
|
2950 | if sharedfiles: | |
2947 | d[b'sharedfiles'] = [] |
|
2951 | d[b'sharedfiles'] = [] | |
2948 |
|
2952 | |||
2949 | if revisionscount: |
|
2953 | if revisionscount: | |
2950 | d[b'revisionscount'] = len(self) |
|
2954 | d[b'revisionscount'] = len(self) | |
2951 |
|
2955 | |||
2952 | if trackedsize: |
|
2956 | if trackedsize: | |
2953 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) |
|
2957 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) | |
2954 |
|
2958 | |||
2955 | if storedsize: |
|
2959 | if storedsize: | |
2956 | d[b'storedsize'] = sum( |
|
2960 | d[b'storedsize'] = sum( | |
2957 | self.opener.stat(path).st_size for path in self.files() |
|
2961 | self.opener.stat(path).st_size for path in self.files() | |
2958 | ) |
|
2962 | ) | |
2959 |
|
2963 | |||
2960 | return d |
|
2964 | return d |
General Comments 0
You need to be logged in to leave comments.
Login now