Show More
@@ -1,192 +1,192 b'' | |||
|
1 | 1 | /* |
|
2 | 2 | base85 codec |
|
3 | 3 | |
|
4 | 4 | Copyright 2006 Brendan Cully <brendan@kublai.com> |
|
5 | 5 | |
|
6 | 6 | This software may be used and distributed according to the terms of |
|
7 | 7 | the GNU General Public License, incorporated herein by reference. |
|
8 | 8 | |
|
9 | 9 | Largely based on git's implementation |
|
10 | 10 | */ |
|
11 | 11 | |
|
12 | 12 | #define PY_SSIZE_T_CLEAN |
|
13 | 13 | #include <Python.h> |
|
14 | 14 | |
|
15 | 15 | #include "util.h" |
|
16 | 16 | |
|
17 | 17 | static const char b85chars[] = |
|
18 | 18 | "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ" |
|
19 | 19 | "abcdefghijklmnopqrstuvwxyz!#$%&()*+-;<=>?@^_`{|}~"; |
|
20 | 20 | static char b85dec[256]; |
|
21 | 21 | |
|
22 | 22 | static void b85prep(void) |
|
23 | 23 | { |
|
24 | 24 | unsigned i; |
|
25 | 25 | |
|
26 | 26 | memset(b85dec, 0, sizeof(b85dec)); |
|
27 | 27 | for (i = 0; i < sizeof(b85chars); i++) { |
|
28 | 28 | b85dec[(int)(b85chars[i])] = i + 1; |
|
29 | 29 | } |
|
30 | 30 | } |
|
31 | 31 | |
|
32 | 32 | static PyObject *b85encode(PyObject *self, PyObject *args) |
|
33 | 33 | { |
|
34 | 34 | const unsigned char *text; |
|
35 | 35 | PyObject *out; |
|
36 | 36 | char *dst; |
|
37 | 37 | Py_ssize_t len, olen, i; |
|
38 | 38 | unsigned int acc, val, ch; |
|
39 | 39 | int pad = 0; |
|
40 | 40 | |
|
41 |
if (!PyArg_ParseTuple(args, |
|
|
41 | if (!PyArg_ParseTuple(args, "y#|i", &text, &len, &pad)) { | |
|
42 | 42 | return NULL; |
|
43 | 43 | } |
|
44 | 44 | |
|
45 | 45 | if (pad) { |
|
46 | 46 | olen = ((len + 3) / 4 * 5) - 3; |
|
47 | 47 | } else { |
|
48 | 48 | olen = len % 4; |
|
49 | 49 | if (olen) { |
|
50 | 50 | olen++; |
|
51 | 51 | } |
|
52 | 52 | olen += len / 4 * 5; |
|
53 | 53 | } |
|
54 | 54 | if (!(out = PyBytes_FromStringAndSize(NULL, olen + 3))) { |
|
55 | 55 | return NULL; |
|
56 | 56 | } |
|
57 | 57 | |
|
58 | 58 | dst = PyBytes_AsString(out); |
|
59 | 59 | |
|
60 | 60 | while (len) { |
|
61 | 61 | acc = 0; |
|
62 | 62 | for (i = 24; i >= 0; i -= 8) { |
|
63 | 63 | ch = *text++; |
|
64 | 64 | acc |= ch << i; |
|
65 | 65 | if (--len == 0) { |
|
66 | 66 | break; |
|
67 | 67 | } |
|
68 | 68 | } |
|
69 | 69 | for (i = 4; i >= 0; i--) { |
|
70 | 70 | val = acc % 85; |
|
71 | 71 | acc /= 85; |
|
72 | 72 | dst[i] = b85chars[val]; |
|
73 | 73 | } |
|
74 | 74 | dst += 5; |
|
75 | 75 | } |
|
76 | 76 | |
|
77 | 77 | if (!pad) { |
|
78 | 78 | _PyBytes_Resize(&out, olen); |
|
79 | 79 | } |
|
80 | 80 | |
|
81 | 81 | return out; |
|
82 | 82 | } |
|
83 | 83 | |
|
84 | 84 | static PyObject *b85decode(PyObject *self, PyObject *args) |
|
85 | 85 | { |
|
86 | 86 | PyObject *out = NULL; |
|
87 | 87 | const char *text; |
|
88 | 88 | char *dst; |
|
89 | 89 | Py_ssize_t len, i, j, olen, cap; |
|
90 | 90 | int c; |
|
91 | 91 | unsigned int acc; |
|
92 | 92 | |
|
93 |
if (!PyArg_ParseTuple(args, |
|
|
93 | if (!PyArg_ParseTuple(args, "y#", &text, &len)) { | |
|
94 | 94 | return NULL; |
|
95 | 95 | } |
|
96 | 96 | |
|
97 | 97 | olen = len / 5 * 4; |
|
98 | 98 | i = len % 5; |
|
99 | 99 | if (i) { |
|
100 | 100 | olen += i - 1; |
|
101 | 101 | } |
|
102 | 102 | if (!(out = PyBytes_FromStringAndSize(NULL, olen))) { |
|
103 | 103 | return NULL; |
|
104 | 104 | } |
|
105 | 105 | |
|
106 | 106 | dst = PyBytes_AsString(out); |
|
107 | 107 | |
|
108 | 108 | i = 0; |
|
109 | 109 | while (i < len) { |
|
110 | 110 | acc = 0; |
|
111 | 111 | cap = len - i - 1; |
|
112 | 112 | if (cap > 4) { |
|
113 | 113 | cap = 4; |
|
114 | 114 | } |
|
115 | 115 | for (j = 0; j < cap; i++, j++) { |
|
116 | 116 | c = b85dec[(int)*text++] - 1; |
|
117 | 117 | if (c < 0) { |
|
118 | 118 | PyErr_Format( |
|
119 | 119 | PyExc_ValueError, |
|
120 | 120 | "bad base85 character at position %d", |
|
121 | 121 | (int)i); |
|
122 | 122 | goto bail; |
|
123 | 123 | } |
|
124 | 124 | acc = acc * 85 + c; |
|
125 | 125 | } |
|
126 | 126 | if (i++ < len) { |
|
127 | 127 | c = b85dec[(int)*text++] - 1; |
|
128 | 128 | if (c < 0) { |
|
129 | 129 | PyErr_Format( |
|
130 | 130 | PyExc_ValueError, |
|
131 | 131 | "bad base85 character at position %d", |
|
132 | 132 | (int)i); |
|
133 | 133 | goto bail; |
|
134 | 134 | } |
|
135 | 135 | /* overflow detection: 0xffffffff == "|NsC0", |
|
136 | 136 | * "|NsC" == 0x03030303 */ |
|
137 | 137 | if (acc > 0x03030303 || (acc *= 85) > 0xffffffff - c) { |
|
138 | 138 | PyErr_Format( |
|
139 | 139 | PyExc_ValueError, |
|
140 | 140 | "bad base85 sequence at position %d", |
|
141 | 141 | (int)i); |
|
142 | 142 | goto bail; |
|
143 | 143 | } |
|
144 | 144 | acc += c; |
|
145 | 145 | } |
|
146 | 146 | |
|
147 | 147 | cap = olen < 4 ? olen : 4; |
|
148 | 148 | olen -= cap; |
|
149 | 149 | for (j = 0; j < 4 - cap; j++) { |
|
150 | 150 | acc *= 85; |
|
151 | 151 | } |
|
152 | 152 | if (cap && cap < 4) { |
|
153 | 153 | acc += 0xffffff >> (cap - 1) * 8; |
|
154 | 154 | } |
|
155 | 155 | for (j = 0; j < cap; j++) { |
|
156 | 156 | acc = (acc << 8) | (acc >> 24); |
|
157 | 157 | *dst++ = acc; |
|
158 | 158 | } |
|
159 | 159 | } |
|
160 | 160 | |
|
161 | 161 | return out; |
|
162 | 162 | bail: |
|
163 | 163 | Py_XDECREF(out); |
|
164 | 164 | return NULL; |
|
165 | 165 | } |
|
166 | 166 | |
|
167 | 167 | static char base85_doc[] = "Base85 Data Encoding"; |
|
168 | 168 | |
|
169 | 169 | static PyMethodDef methods[] = { |
|
170 | 170 | {"b85encode", b85encode, METH_VARARGS, |
|
171 | 171 | "Encode text in base85.\n\n" |
|
172 | 172 | "If the second parameter is true, pad the result to a multiple of " |
|
173 | 173 | "five characters.\n"}, |
|
174 | 174 | {"b85decode", b85decode, METH_VARARGS, "Decode base85 text.\n"}, |
|
175 | 175 | {NULL, NULL}, |
|
176 | 176 | }; |
|
177 | 177 | |
|
178 | 178 | static const int version = 1; |
|
179 | 179 | |
|
180 | 180 | static struct PyModuleDef base85_module = { |
|
181 | 181 | PyModuleDef_HEAD_INIT, "base85", base85_doc, -1, methods, |
|
182 | 182 | }; |
|
183 | 183 | |
|
184 | 184 | PyMODINIT_FUNC PyInit_base85(void) |
|
185 | 185 | { |
|
186 | 186 | PyObject *m; |
|
187 | 187 | b85prep(); |
|
188 | 188 | |
|
189 | 189 | m = PyModule_Create(&base85_module); |
|
190 | 190 | PyModule_AddIntConstant(m, "version", version); |
|
191 | 191 | return m; |
|
192 | 192 | } |
@@ -1,350 +1,348 b'' | |||
|
1 | 1 | /* |
|
2 | 2 | bdiff.c - efficient binary diff extension for Mercurial |
|
3 | 3 | |
|
4 | 4 | Copyright 2005, 2006 Olivia Mackall <olivia@selenic.com> |
|
5 | 5 | |
|
6 | 6 | This software may be used and distributed according to the terms of |
|
7 | 7 | the GNU General Public License, incorporated herein by reference. |
|
8 | 8 | |
|
9 | 9 | Based roughly on Python difflib |
|
10 | 10 | */ |
|
11 | 11 | |
|
12 | 12 | #define PY_SSIZE_T_CLEAN |
|
13 | 13 | #include <Python.h> |
|
14 | 14 | #include <limits.h> |
|
15 | 15 | #include <stdlib.h> |
|
16 | 16 | #include <string.h> |
|
17 | 17 | |
|
18 | 18 | #include "bdiff.h" |
|
19 | 19 | #include "bitmanipulation.h" |
|
20 | 20 | #include "thirdparty/xdiff/xdiff.h" |
|
21 | 21 | #include "util.h" |
|
22 | 22 | |
|
23 | 23 | static PyObject *blocks(PyObject *self, PyObject *args) |
|
24 | 24 | { |
|
25 | 25 | PyObject *sa, *sb, *rl = NULL, *m; |
|
26 | 26 | struct bdiff_line *a, *b; |
|
27 | 27 | struct bdiff_hunk l, *h; |
|
28 | 28 | int an, bn, count, pos = 0; |
|
29 | 29 | |
|
30 | 30 | l.next = NULL; |
|
31 | 31 | |
|
32 | 32 | if (!PyArg_ParseTuple(args, "SS:bdiff", &sa, &sb)) { |
|
33 | 33 | return NULL; |
|
34 | 34 | } |
|
35 | 35 | |
|
36 | 36 | an = bdiff_splitlines(PyBytes_AsString(sa), PyBytes_Size(sa), &a); |
|
37 | 37 | bn = bdiff_splitlines(PyBytes_AsString(sb), PyBytes_Size(sb), &b); |
|
38 | 38 | |
|
39 | 39 | if (!a || !b) { |
|
40 | 40 | goto nomem; |
|
41 | 41 | } |
|
42 | 42 | |
|
43 | 43 | count = bdiff_diff(a, an, b, bn, &l); |
|
44 | 44 | if (count < 0) { |
|
45 | 45 | goto nomem; |
|
46 | 46 | } |
|
47 | 47 | |
|
48 | 48 | rl = PyList_New(count); |
|
49 | 49 | if (!rl) { |
|
50 | 50 | goto nomem; |
|
51 | 51 | } |
|
52 | 52 | |
|
53 | 53 | for (h = l.next; h; h = h->next) { |
|
54 | 54 | m = Py_BuildValue("iiii", h->a1, h->a2, h->b1, h->b2); |
|
55 | 55 | PyList_SetItem(rl, pos, m); |
|
56 | 56 | pos++; |
|
57 | 57 | } |
|
58 | 58 | |
|
59 | 59 | nomem: |
|
60 | 60 | free(a); |
|
61 | 61 | free(b); |
|
62 | 62 | bdiff_freehunks(l.next); |
|
63 | 63 | return rl ? rl : PyErr_NoMemory(); |
|
64 | 64 | } |
|
65 | 65 | |
|
66 | 66 | static PyObject *bdiff(PyObject *self, PyObject *args) |
|
67 | 67 | { |
|
68 | 68 | Py_buffer ba, bb; |
|
69 | 69 | char *rb, *ia, *ib; |
|
70 | 70 | PyObject *result = NULL; |
|
71 | 71 | struct bdiff_line *al = NULL, *bl = NULL; |
|
72 | 72 | struct bdiff_hunk l, *h; |
|
73 | 73 | int an, bn, count; |
|
74 | 74 | Py_ssize_t len = 0, la, lb, li = 0, lcommon = 0, lmax; |
|
75 | 75 | PyThreadState *_save = NULL; |
|
76 | 76 | |
|
77 | 77 | l.next = NULL; |
|
78 | 78 | |
|
79 |
if (!PyArg_ParseTuple(args, |
|
|
80 | &bb)) { | |
|
79 | if (!PyArg_ParseTuple(args, "y*y*:bdiff", &ba, &bb)) { | |
|
81 | 80 | return NULL; |
|
82 | 81 | } |
|
83 | 82 | |
|
84 | 83 | if (!PyBuffer_IsContiguous(&ba, 'C') || ba.ndim > 1) { |
|
85 | 84 | PyErr_SetString(PyExc_ValueError, "bdiff input not contiguous"); |
|
86 | 85 | goto cleanup; |
|
87 | 86 | } |
|
88 | 87 | |
|
89 | 88 | if (!PyBuffer_IsContiguous(&bb, 'C') || bb.ndim > 1) { |
|
90 | 89 | PyErr_SetString(PyExc_ValueError, "bdiff input not contiguous"); |
|
91 | 90 | goto cleanup; |
|
92 | 91 | } |
|
93 | 92 | |
|
94 | 93 | la = ba.len; |
|
95 | 94 | lb = bb.len; |
|
96 | 95 | |
|
97 | 96 | if (la > UINT_MAX || lb > UINT_MAX) { |
|
98 | 97 | PyErr_SetString(PyExc_ValueError, "bdiff inputs too large"); |
|
99 | 98 | goto cleanup; |
|
100 | 99 | } |
|
101 | 100 | |
|
102 | 101 | _save = PyEval_SaveThread(); |
|
103 | 102 | |
|
104 | 103 | lmax = la > lb ? lb : la; |
|
105 | 104 | for (ia = ba.buf, ib = bb.buf; li < lmax && *ia == *ib; |
|
106 | 105 | ++li, ++ia, ++ib) { |
|
107 | 106 | if (*ia == '\n') { |
|
108 | 107 | lcommon = li + 1; |
|
109 | 108 | } |
|
110 | 109 | } |
|
111 | 110 | /* we can almost add: if (li == lmax) lcommon = li; */ |
|
112 | 111 | |
|
113 | 112 | an = bdiff_splitlines((char *)ba.buf + lcommon, la - lcommon, &al); |
|
114 | 113 | bn = bdiff_splitlines((char *)bb.buf + lcommon, lb - lcommon, &bl); |
|
115 | 114 | if (!al || !bl) { |
|
116 | 115 | PyErr_NoMemory(); |
|
117 | 116 | goto cleanup; |
|
118 | 117 | } |
|
119 | 118 | |
|
120 | 119 | count = bdiff_diff(al, an, bl, bn, &l); |
|
121 | 120 | if (count < 0) { |
|
122 | 121 | PyErr_NoMemory(); |
|
123 | 122 | goto cleanup; |
|
124 | 123 | } |
|
125 | 124 | |
|
126 | 125 | /* calculate length of output */ |
|
127 | 126 | la = lb = 0; |
|
128 | 127 | for (h = l.next; h; h = h->next) { |
|
129 | 128 | if (h->a1 != la || h->b1 != lb) { |
|
130 | 129 | len += 12 + bl[h->b1].l - bl[lb].l; |
|
131 | 130 | } |
|
132 | 131 | la = h->a2; |
|
133 | 132 | lb = h->b2; |
|
134 | 133 | } |
|
135 | 134 | PyEval_RestoreThread(_save); |
|
136 | 135 | _save = NULL; |
|
137 | 136 | |
|
138 | 137 | result = PyBytes_FromStringAndSize(NULL, len); |
|
139 | 138 | |
|
140 | 139 | if (!result) { |
|
141 | 140 | goto cleanup; |
|
142 | 141 | } |
|
143 | 142 | |
|
144 | 143 | /* build binary patch */ |
|
145 | 144 | rb = PyBytes_AsString(result); |
|
146 | 145 | la = lb = 0; |
|
147 | 146 | |
|
148 | 147 | for (h = l.next; h; h = h->next) { |
|
149 | 148 | if (h->a1 != la || h->b1 != lb) { |
|
150 | 149 | len = bl[h->b1].l - bl[lb].l; |
|
151 | 150 | putbe32((uint32_t)(al[la].l + lcommon - al->l), rb); |
|
152 | 151 | putbe32((uint32_t)(al[h->a1].l + lcommon - al->l), |
|
153 | 152 | rb + 4); |
|
154 | 153 | putbe32((uint32_t)len, rb + 8); |
|
155 | 154 | memcpy(rb + 12, bl[lb].l, len); |
|
156 | 155 | rb += 12 + len; |
|
157 | 156 | } |
|
158 | 157 | la = h->a2; |
|
159 | 158 | lb = h->b2; |
|
160 | 159 | } |
|
161 | 160 | |
|
162 | 161 | cleanup: |
|
163 | 162 | if (_save) { |
|
164 | 163 | PyEval_RestoreThread(_save); |
|
165 | 164 | } |
|
166 | 165 | PyBuffer_Release(&ba); |
|
167 | 166 | PyBuffer_Release(&bb); |
|
168 | 167 | free(al); |
|
169 | 168 | free(bl); |
|
170 | 169 | bdiff_freehunks(l.next); |
|
171 | 170 | return result; |
|
172 | 171 | } |
|
173 | 172 | |
|
174 | 173 | /* |
|
175 | 174 | * If allws != 0, remove all whitespace (' ', \t and \r). Otherwise, |
|
176 | 175 | * reduce whitespace sequences to a single space and trim remaining whitespace |
|
177 | 176 | * from end of lines. |
|
178 | 177 | */ |
|
179 | 178 | static PyObject *fixws(PyObject *self, PyObject *args) |
|
180 | 179 | { |
|
181 | 180 | PyObject *s, *result = NULL; |
|
182 | 181 | char allws, c; |
|
183 | 182 | const char *r; |
|
184 | 183 | Py_ssize_t i, rlen, wlen = 0; |
|
185 | 184 | char *w; |
|
186 | 185 | |
|
187 | 186 | if (!PyArg_ParseTuple(args, "Sb:fixws", &s, &allws)) { |
|
188 | 187 | return NULL; |
|
189 | 188 | } |
|
190 | 189 | r = PyBytes_AsString(s); |
|
191 | 190 | rlen = PyBytes_Size(s); |
|
192 | 191 | |
|
193 | 192 | w = (char *)PyMem_Malloc(rlen ? rlen : 1); |
|
194 | 193 | if (!w) { |
|
195 | 194 | goto nomem; |
|
196 | 195 | } |
|
197 | 196 | |
|
198 | 197 | for (i = 0; i != rlen; i++) { |
|
199 | 198 | c = r[i]; |
|
200 | 199 | if (c == ' ' || c == '\t' || c == '\r') { |
|
201 | 200 | if (!allws && (wlen == 0 || w[wlen - 1] != ' ')) { |
|
202 | 201 | w[wlen++] = ' '; |
|
203 | 202 | } |
|
204 | 203 | } else if (c == '\n' && !allws && wlen > 0 && |
|
205 | 204 | w[wlen - 1] == ' ') { |
|
206 | 205 | w[wlen - 1] = '\n'; |
|
207 | 206 | } else { |
|
208 | 207 | w[wlen++] = c; |
|
209 | 208 | } |
|
210 | 209 | } |
|
211 | 210 | |
|
212 | 211 | result = PyBytes_FromStringAndSize(w, wlen); |
|
213 | 212 | |
|
214 | 213 | nomem: |
|
215 | 214 | PyMem_Free(w); |
|
216 | 215 | return result ? result : PyErr_NoMemory(); |
|
217 | 216 | } |
|
218 | 217 | |
|
219 | 218 | static bool sliceintolist(PyObject *list, Py_ssize_t destidx, |
|
220 | 219 | const char *source, Py_ssize_t len) |
|
221 | 220 | { |
|
222 | 221 | PyObject *sliced = PyBytes_FromStringAndSize(source, len); |
|
223 | 222 | if (sliced == NULL) { |
|
224 | 223 | return false; |
|
225 | 224 | } |
|
226 | 225 | PyList_SET_ITEM(list, destidx, sliced); |
|
227 | 226 | return true; |
|
228 | 227 | } |
|
229 | 228 | |
|
230 | 229 | static PyObject *splitnewlines(PyObject *self, PyObject *args) |
|
231 | 230 | { |
|
232 | 231 | const char *text; |
|
233 | 232 | Py_ssize_t nelts = 0, size, i, start = 0; |
|
234 | 233 | PyObject *result = NULL; |
|
235 | 234 | |
|
236 |
if (!PyArg_ParseTuple(args, |
|
|
235 | if (!PyArg_ParseTuple(args, "y#", &text, &size)) { | |
|
237 | 236 | goto abort; |
|
238 | 237 | } |
|
239 | 238 | if (!size) { |
|
240 | 239 | return PyList_New(0); |
|
241 | 240 | } |
|
242 | 241 | /* This loops to size-1 because if the last byte is a newline, |
|
243 | 242 | * we don't want to perform a split there. */ |
|
244 | 243 | for (i = 0; i < size - 1; ++i) { |
|
245 | 244 | if (text[i] == '\n') { |
|
246 | 245 | ++nelts; |
|
247 | 246 | } |
|
248 | 247 | } |
|
249 | 248 | if ((result = PyList_New(nelts + 1)) == NULL) { |
|
250 | 249 | goto abort; |
|
251 | 250 | } |
|
252 | 251 | nelts = 0; |
|
253 | 252 | for (i = 0; i < size - 1; ++i) { |
|
254 | 253 | if (text[i] == '\n') { |
|
255 | 254 | if (!sliceintolist(result, nelts++, text + start, |
|
256 | 255 | i - start + 1)) { |
|
257 | 256 | goto abort; |
|
258 | 257 | } |
|
259 | 258 | start = i + 1; |
|
260 | 259 | } |
|
261 | 260 | } |
|
262 | 261 | if (!sliceintolist(result, nelts++, text + start, size - start)) { |
|
263 | 262 | goto abort; |
|
264 | 263 | } |
|
265 | 264 | return result; |
|
266 | 265 | abort: |
|
267 | 266 | Py_XDECREF(result); |
|
268 | 267 | return NULL; |
|
269 | 268 | } |
|
270 | 269 | |
|
271 | 270 | static int hunk_consumer(int64_t a1, int64_t a2, int64_t b1, int64_t b2, |
|
272 | 271 | void *priv) |
|
273 | 272 | { |
|
274 | 273 | PyObject *rl = (PyObject *)priv; |
|
275 | 274 | PyObject *m = Py_BuildValue("LLLL", a1, a2, b1, b2); |
|
276 | 275 | int r; |
|
277 | 276 | if (!m) { |
|
278 | 277 | return -1; |
|
279 | 278 | } |
|
280 | 279 | r = PyList_Append(rl, m); |
|
281 | 280 | Py_DECREF(m); |
|
282 | 281 | return r; |
|
283 | 282 | } |
|
284 | 283 | |
|
285 | 284 | static PyObject *xdiffblocks(PyObject *self, PyObject *args) |
|
286 | 285 | { |
|
287 | 286 | Py_ssize_t la, lb; |
|
288 | 287 | mmfile_t a, b; |
|
289 | 288 | PyObject *rl; |
|
290 | 289 | |
|
291 | 290 | xpparam_t xpp = { |
|
292 | 291 | XDF_INDENT_HEURISTIC, /* flags */ |
|
293 | 292 | }; |
|
294 | 293 | xdemitconf_t xecfg = { |
|
295 | 294 | XDL_EMIT_BDIFFHUNK, /* flags */ |
|
296 | 295 | hunk_consumer, /* hunk_consume_func */ |
|
297 | 296 | }; |
|
298 | 297 | xdemitcb_t ecb = { |
|
299 | 298 | NULL, /* priv */ |
|
300 | 299 | }; |
|
301 | 300 | |
|
302 |
if (!PyArg_ParseTuple(args, |
|
|
303 | &lb)) { | |
|
301 | if (!PyArg_ParseTuple(args, "y#y#", &a.ptr, &la, &b.ptr, &lb)) { | |
|
304 | 302 | return NULL; |
|
305 | 303 | } |
|
306 | 304 | |
|
307 | 305 | a.size = la; |
|
308 | 306 | b.size = lb; |
|
309 | 307 | |
|
310 | 308 | rl = PyList_New(0); |
|
311 | 309 | if (!rl) { |
|
312 | 310 | return PyErr_NoMemory(); |
|
313 | 311 | } |
|
314 | 312 | |
|
315 | 313 | ecb.priv = rl; |
|
316 | 314 | |
|
317 | 315 | if (xdl_diff(&a, &b, &xpp, &xecfg, &ecb) != 0) { |
|
318 | 316 | Py_DECREF(rl); |
|
319 | 317 | return PyErr_NoMemory(); |
|
320 | 318 | } |
|
321 | 319 | |
|
322 | 320 | return rl; |
|
323 | 321 | } |
|
324 | 322 | |
|
325 | 323 | static char mdiff_doc[] = "Efficient binary diff."; |
|
326 | 324 | |
|
327 | 325 | static PyMethodDef methods[] = { |
|
328 | 326 | {"bdiff", bdiff, METH_VARARGS, "calculate a binary diff\n"}, |
|
329 | 327 | {"blocks", blocks, METH_VARARGS, "find a list of matching lines\n"}, |
|
330 | 328 | {"fixws", fixws, METH_VARARGS, "normalize diff whitespaces\n"}, |
|
331 | 329 | {"splitnewlines", splitnewlines, METH_VARARGS, |
|
332 | 330 | "like str.splitlines, but only split on newlines\n"}, |
|
333 | 331 | {"xdiffblocks", xdiffblocks, METH_VARARGS, |
|
334 | 332 | "find a list of matching lines using xdiff algorithm\n"}, |
|
335 | 333 | {NULL, NULL}, |
|
336 | 334 | }; |
|
337 | 335 | |
|
338 | 336 | static const int version = 3; |
|
339 | 337 | |
|
340 | 338 | static struct PyModuleDef bdiff_module = { |
|
341 | 339 | PyModuleDef_HEAD_INIT, "bdiff", mdiff_doc, -1, methods, |
|
342 | 340 | }; |
|
343 | 341 | |
|
344 | 342 | PyMODINIT_FUNC PyInit_bdiff(void) |
|
345 | 343 | { |
|
346 | 344 | PyObject *m; |
|
347 | 345 | m = PyModule_Create(&bdiff_module); |
|
348 | 346 | PyModule_AddIntConstant(m, "version", version); |
|
349 | 347 | return m; |
|
350 | 348 | } |
@@ -1,405 +1,404 b'' | |||
|
1 | 1 | /* |
|
2 | 2 | charencode.c - miscellaneous character encoding |
|
3 | 3 | |
|
4 | 4 | Copyright 2008 Olivia Mackall <olivia@selenic.com> and others |
|
5 | 5 | |
|
6 | 6 | This software may be used and distributed according to the terms of |
|
7 | 7 | the GNU General Public License, incorporated herein by reference. |
|
8 | 8 | */ |
|
9 | 9 | |
|
10 | 10 | #define PY_SSIZE_T_CLEAN |
|
11 | 11 | #include <Python.h> |
|
12 | 12 | #include <assert.h> |
|
13 | 13 | |
|
14 | 14 | #include "charencode.h" |
|
15 | 15 | #include "compat.h" |
|
16 | 16 | #include "util.h" |
|
17 | 17 | |
|
18 | 18 | /* clang-format off */ |
|
19 | 19 | static const char lowertable[128] = { |
|
20 | 20 | '\x00', '\x01', '\x02', '\x03', '\x04', '\x05', '\x06', '\x07', |
|
21 | 21 | '\x08', '\x09', '\x0a', '\x0b', '\x0c', '\x0d', '\x0e', '\x0f', |
|
22 | 22 | '\x10', '\x11', '\x12', '\x13', '\x14', '\x15', '\x16', '\x17', |
|
23 | 23 | '\x18', '\x19', '\x1a', '\x1b', '\x1c', '\x1d', '\x1e', '\x1f', |
|
24 | 24 | '\x20', '\x21', '\x22', '\x23', '\x24', '\x25', '\x26', '\x27', |
|
25 | 25 | '\x28', '\x29', '\x2a', '\x2b', '\x2c', '\x2d', '\x2e', '\x2f', |
|
26 | 26 | '\x30', '\x31', '\x32', '\x33', '\x34', '\x35', '\x36', '\x37', |
|
27 | 27 | '\x38', '\x39', '\x3a', '\x3b', '\x3c', '\x3d', '\x3e', '\x3f', |
|
28 | 28 | '\x40', |
|
29 | 29 | '\x61', '\x62', '\x63', '\x64', '\x65', '\x66', '\x67', /* A-G */ |
|
30 | 30 | '\x68', '\x69', '\x6a', '\x6b', '\x6c', '\x6d', '\x6e', '\x6f', /* H-O */ |
|
31 | 31 | '\x70', '\x71', '\x72', '\x73', '\x74', '\x75', '\x76', '\x77', /* P-W */ |
|
32 | 32 | '\x78', '\x79', '\x7a', /* X-Z */ |
|
33 | 33 | '\x5b', '\x5c', '\x5d', '\x5e', '\x5f', |
|
34 | 34 | '\x60', '\x61', '\x62', '\x63', '\x64', '\x65', '\x66', '\x67', |
|
35 | 35 | '\x68', '\x69', '\x6a', '\x6b', '\x6c', '\x6d', '\x6e', '\x6f', |
|
36 | 36 | '\x70', '\x71', '\x72', '\x73', '\x74', '\x75', '\x76', '\x77', |
|
37 | 37 | '\x78', '\x79', '\x7a', '\x7b', '\x7c', '\x7d', '\x7e', '\x7f' |
|
38 | 38 | }; |
|
39 | 39 | |
|
40 | 40 | static const char uppertable[128] = { |
|
41 | 41 | '\x00', '\x01', '\x02', '\x03', '\x04', '\x05', '\x06', '\x07', |
|
42 | 42 | '\x08', '\x09', '\x0a', '\x0b', '\x0c', '\x0d', '\x0e', '\x0f', |
|
43 | 43 | '\x10', '\x11', '\x12', '\x13', '\x14', '\x15', '\x16', '\x17', |
|
44 | 44 | '\x18', '\x19', '\x1a', '\x1b', '\x1c', '\x1d', '\x1e', '\x1f', |
|
45 | 45 | '\x20', '\x21', '\x22', '\x23', '\x24', '\x25', '\x26', '\x27', |
|
46 | 46 | '\x28', '\x29', '\x2a', '\x2b', '\x2c', '\x2d', '\x2e', '\x2f', |
|
47 | 47 | '\x30', '\x31', '\x32', '\x33', '\x34', '\x35', '\x36', '\x37', |
|
48 | 48 | '\x38', '\x39', '\x3a', '\x3b', '\x3c', '\x3d', '\x3e', '\x3f', |
|
49 | 49 | '\x40', '\x41', '\x42', '\x43', '\x44', '\x45', '\x46', '\x47', |
|
50 | 50 | '\x48', '\x49', '\x4a', '\x4b', '\x4c', '\x4d', '\x4e', '\x4f', |
|
51 | 51 | '\x50', '\x51', '\x52', '\x53', '\x54', '\x55', '\x56', '\x57', |
|
52 | 52 | '\x58', '\x59', '\x5a', '\x5b', '\x5c', '\x5d', '\x5e', '\x5f', |
|
53 | 53 | '\x60', |
|
54 | 54 | '\x41', '\x42', '\x43', '\x44', '\x45', '\x46', '\x47', /* a-g */ |
|
55 | 55 | '\x48', '\x49', '\x4a', '\x4b', '\x4c', '\x4d', '\x4e', '\x4f', /* h-o */ |
|
56 | 56 | '\x50', '\x51', '\x52', '\x53', '\x54', '\x55', '\x56', '\x57', /* p-w */ |
|
57 | 57 | '\x58', '\x59', '\x5a', /* x-z */ |
|
58 | 58 | '\x7b', '\x7c', '\x7d', '\x7e', '\x7f' |
|
59 | 59 | }; |
|
60 | 60 | |
|
61 | 61 | /* 1: no escape, 2: \<c>, 6: \u<x> */ |
|
62 | 62 | static const uint8_t jsonlentable[256] = { |
|
63 | 63 | 6, 6, 6, 6, 6, 6, 6, 6, 2, 2, 2, 6, 2, 2, 6, 6, /* b, t, n, f, r */ |
|
64 | 64 | 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, |
|
65 | 65 | 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* " */ |
|
66 | 66 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
67 | 67 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
68 | 68 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, /* \\ */ |
|
69 | 69 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
70 | 70 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 6, /* DEL */ |
|
71 | 71 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
72 | 72 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
73 | 73 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
74 | 74 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
75 | 75 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
76 | 76 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
77 | 77 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
78 | 78 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
79 | 79 | }; |
|
80 | 80 | |
|
81 | 81 | static const uint8_t jsonparanoidlentable[128] = { |
|
82 | 82 | 6, 6, 6, 6, 6, 6, 6, 6, 2, 2, 2, 6, 2, 2, 6, 6, /* b, t, n, f, r */ |
|
83 | 83 | 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, |
|
84 | 84 | 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* " */ |
|
85 | 85 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 6, 1, 6, 1, /* <, > */ |
|
86 | 86 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
87 | 87 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, /* \\ */ |
|
88 | 88 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, |
|
89 | 89 | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 6, /* DEL */ |
|
90 | 90 | }; |
|
91 | 91 | |
|
92 | 92 | static const char hexchartable[16] = { |
|
93 | 93 | '0', '1', '2', '3', '4', '5', '6', '7', |
|
94 | 94 | '8', '9', 'a', 'b', 'c', 'd', 'e', 'f', |
|
95 | 95 | }; |
|
96 | 96 | /* clang-format on */ |
|
97 | 97 | |
|
98 | 98 | /* |
|
99 | 99 | * Turn a hex-encoded string into binary. |
|
100 | 100 | */ |
|
101 | 101 | PyObject *unhexlify(const char *str, Py_ssize_t len) |
|
102 | 102 | { |
|
103 | 103 | PyObject *ret; |
|
104 | 104 | char *d; |
|
105 | 105 | Py_ssize_t i; |
|
106 | 106 | |
|
107 | 107 | ret = PyBytes_FromStringAndSize(NULL, len / 2); |
|
108 | 108 | |
|
109 | 109 | if (!ret) { |
|
110 | 110 | return NULL; |
|
111 | 111 | } |
|
112 | 112 | |
|
113 | 113 | d = PyBytes_AsString(ret); |
|
114 | 114 | |
|
115 | 115 | for (i = 0; i < len;) { |
|
116 | 116 | int hi = hexdigit(str, i++); |
|
117 | 117 | int lo = hexdigit(str, i++); |
|
118 | 118 | *d++ = (hi << 4) | lo; |
|
119 | 119 | } |
|
120 | 120 | |
|
121 | 121 | return ret; |
|
122 | 122 | } |
|
123 | 123 | |
|
124 | 124 | PyObject *isasciistr(PyObject *self, PyObject *args) |
|
125 | 125 | { |
|
126 | 126 | const char *buf; |
|
127 | 127 | Py_ssize_t i, len; |
|
128 |
if (!PyArg_ParseTuple(args, |
|
|
129 | &buf, &len)) { | |
|
128 | if (!PyArg_ParseTuple(args, "y#:isasciistr", &buf, &len)) { | |
|
130 | 129 | return NULL; |
|
131 | 130 | } |
|
132 | 131 | i = 0; |
|
133 | 132 | /* char array in PyStringObject should be at least 4-byte aligned */ |
|
134 | 133 | if (((uintptr_t)buf & 3) == 0) { |
|
135 | 134 | const uint32_t *p = (const uint32_t *)buf; |
|
136 | 135 | for (; i < len / 4; i++) { |
|
137 | 136 | if (p[i] & 0x80808080U) { |
|
138 | 137 | Py_RETURN_FALSE; |
|
139 | 138 | } |
|
140 | 139 | } |
|
141 | 140 | i *= 4; |
|
142 | 141 | } |
|
143 | 142 | for (; i < len; i++) { |
|
144 | 143 | if (buf[i] & 0x80) { |
|
145 | 144 | Py_RETURN_FALSE; |
|
146 | 145 | } |
|
147 | 146 | } |
|
148 | 147 | Py_RETURN_TRUE; |
|
149 | 148 | } |
|
150 | 149 | |
|
151 | 150 | static inline PyObject * |
|
152 | 151 | _asciitransform(PyObject *str_obj, const char table[128], PyObject *fallback_fn) |
|
153 | 152 | { |
|
154 | 153 | char *str, *newstr; |
|
155 | 154 | Py_ssize_t i, len; |
|
156 | 155 | PyObject *newobj = NULL; |
|
157 | 156 | PyObject *ret = NULL; |
|
158 | 157 | |
|
159 | 158 | str = PyBytes_AS_STRING(str_obj); |
|
160 | 159 | len = PyBytes_GET_SIZE(str_obj); |
|
161 | 160 | |
|
162 | 161 | newobj = PyBytes_FromStringAndSize(NULL, len); |
|
163 | 162 | if (!newobj) { |
|
164 | 163 | goto quit; |
|
165 | 164 | } |
|
166 | 165 | |
|
167 | 166 | newstr = PyBytes_AS_STRING(newobj); |
|
168 | 167 | |
|
169 | 168 | for (i = 0; i < len; i++) { |
|
170 | 169 | char c = str[i]; |
|
171 | 170 | if (c & 0x80) { |
|
172 | 171 | if (fallback_fn != NULL) { |
|
173 | 172 | ret = PyObject_CallFunctionObjArgs( |
|
174 | 173 | fallback_fn, str_obj, NULL); |
|
175 | 174 | } else { |
|
176 | 175 | PyObject *err = PyUnicodeDecodeError_Create( |
|
177 | 176 | "ascii", str, len, i, (i + 1), |
|
178 | 177 | "unexpected code byte"); |
|
179 | 178 | PyErr_SetObject(PyExc_UnicodeDecodeError, err); |
|
180 | 179 | Py_XDECREF(err); |
|
181 | 180 | } |
|
182 | 181 | goto quit; |
|
183 | 182 | } |
|
184 | 183 | newstr[i] = table[(unsigned char)c]; |
|
185 | 184 | } |
|
186 | 185 | |
|
187 | 186 | ret = newobj; |
|
188 | 187 | Py_INCREF(ret); |
|
189 | 188 | quit: |
|
190 | 189 | Py_XDECREF(newobj); |
|
191 | 190 | return ret; |
|
192 | 191 | } |
|
193 | 192 | |
|
194 | 193 | PyObject *asciilower(PyObject *self, PyObject *args) |
|
195 | 194 | { |
|
196 | 195 | PyObject *str_obj; |
|
197 | 196 | if (!PyArg_ParseTuple(args, "O!:asciilower", &PyBytes_Type, &str_obj)) { |
|
198 | 197 | return NULL; |
|
199 | 198 | } |
|
200 | 199 | return _asciitransform(str_obj, lowertable, NULL); |
|
201 | 200 | } |
|
202 | 201 | |
|
203 | 202 | PyObject *asciiupper(PyObject *self, PyObject *args) |
|
204 | 203 | { |
|
205 | 204 | PyObject *str_obj; |
|
206 | 205 | if (!PyArg_ParseTuple(args, "O!:asciiupper", &PyBytes_Type, &str_obj)) { |
|
207 | 206 | return NULL; |
|
208 | 207 | } |
|
209 | 208 | return _asciitransform(str_obj, uppertable, NULL); |
|
210 | 209 | } |
|
211 | 210 | |
|
212 | 211 | PyObject *make_file_foldmap(PyObject *self, PyObject *args) |
|
213 | 212 | { |
|
214 | 213 | PyObject *dmap, *spec_obj, *normcase_fallback; |
|
215 | 214 | PyObject *file_foldmap = NULL; |
|
216 | 215 | enum normcase_spec spec; |
|
217 | 216 | PyObject *k, *v; |
|
218 | 217 | dirstateItemObject *tuple; |
|
219 | 218 | Py_ssize_t pos = 0; |
|
220 | 219 | const char *table; |
|
221 | 220 | |
|
222 | 221 | if (!PyArg_ParseTuple(args, "O!O!O!:make_file_foldmap", &PyDict_Type, |
|
223 | 222 | &dmap, &PyLong_Type, &spec_obj, &PyFunction_Type, |
|
224 | 223 | &normcase_fallback)) { |
|
225 | 224 | goto quit; |
|
226 | 225 | } |
|
227 | 226 | |
|
228 | 227 | spec = (int)PyLong_AS_LONG(spec_obj); |
|
229 | 228 | switch (spec) { |
|
230 | 229 | case NORMCASE_LOWER: |
|
231 | 230 | table = lowertable; |
|
232 | 231 | break; |
|
233 | 232 | case NORMCASE_UPPER: |
|
234 | 233 | table = uppertable; |
|
235 | 234 | break; |
|
236 | 235 | case NORMCASE_OTHER: |
|
237 | 236 | table = NULL; |
|
238 | 237 | break; |
|
239 | 238 | default: |
|
240 | 239 | PyErr_SetString(PyExc_TypeError, "invalid normcasespec"); |
|
241 | 240 | goto quit; |
|
242 | 241 | } |
|
243 | 242 | |
|
244 | 243 | /* Add some more entries to deal with additions outside this |
|
245 | 244 | function. */ |
|
246 | 245 | file_foldmap = _dict_new_presized((PyDict_Size(dmap) / 10) * 11); |
|
247 | 246 | if (file_foldmap == NULL) { |
|
248 | 247 | goto quit; |
|
249 | 248 | } |
|
250 | 249 | |
|
251 | 250 | while (PyDict_Next(dmap, &pos, &k, &v)) { |
|
252 | 251 | if (!dirstate_tuple_check(v)) { |
|
253 | 252 | PyErr_SetString(PyExc_TypeError, |
|
254 | 253 | "expected a dirstate tuple"); |
|
255 | 254 | goto quit; |
|
256 | 255 | } |
|
257 | 256 | |
|
258 | 257 | tuple = (dirstateItemObject *)v; |
|
259 | 258 | if (tuple->flags | dirstate_flag_wc_tracked) { |
|
260 | 259 | PyObject *normed; |
|
261 | 260 | if (table != NULL) { |
|
262 | 261 | normed = _asciitransform(k, table, |
|
263 | 262 | normcase_fallback); |
|
264 | 263 | } else { |
|
265 | 264 | normed = PyObject_CallFunctionObjArgs( |
|
266 | 265 | normcase_fallback, k, NULL); |
|
267 | 266 | } |
|
268 | 267 | |
|
269 | 268 | if (normed == NULL) { |
|
270 | 269 | goto quit; |
|
271 | 270 | } |
|
272 | 271 | if (PyDict_SetItem(file_foldmap, normed, k) == -1) { |
|
273 | 272 | Py_DECREF(normed); |
|
274 | 273 | goto quit; |
|
275 | 274 | } |
|
276 | 275 | Py_DECREF(normed); |
|
277 | 276 | } |
|
278 | 277 | } |
|
279 | 278 | return file_foldmap; |
|
280 | 279 | quit: |
|
281 | 280 | Py_XDECREF(file_foldmap); |
|
282 | 281 | return NULL; |
|
283 | 282 | } |
|
284 | 283 | |
|
285 | 284 | /* calculate length of JSON-escaped string; returns -1 if unsupported */ |
|
286 | 285 | static Py_ssize_t jsonescapelen(const char *buf, Py_ssize_t len, bool paranoid) |
|
287 | 286 | { |
|
288 | 287 | Py_ssize_t i, esclen = 0; |
|
289 | 288 | |
|
290 | 289 | if (paranoid) { |
|
291 | 290 | /* don't want to process multi-byte escapes in C */ |
|
292 | 291 | for (i = 0; i < len; i++) { |
|
293 | 292 | char c = buf[i]; |
|
294 | 293 | if (c & 0x80) { |
|
295 | 294 | PyErr_SetString(PyExc_ValueError, |
|
296 | 295 | "cannot process non-ascii str"); |
|
297 | 296 | return -1; |
|
298 | 297 | } |
|
299 | 298 | esclen += jsonparanoidlentable[(unsigned char)c]; |
|
300 | 299 | if (esclen < 0) { |
|
301 | 300 | PyErr_SetString(PyExc_MemoryError, |
|
302 | 301 | "overflow in jsonescapelen"); |
|
303 | 302 | return -1; |
|
304 | 303 | } |
|
305 | 304 | } |
|
306 | 305 | } else { |
|
307 | 306 | for (i = 0; i < len; i++) { |
|
308 | 307 | char c = buf[i]; |
|
309 | 308 | esclen += jsonlentable[(unsigned char)c]; |
|
310 | 309 | if (esclen < 0) { |
|
311 | 310 | PyErr_SetString(PyExc_MemoryError, |
|
312 | 311 | "overflow in jsonescapelen"); |
|
313 | 312 | return -1; |
|
314 | 313 | } |
|
315 | 314 | } |
|
316 | 315 | } |
|
317 | 316 | |
|
318 | 317 | return esclen; |
|
319 | 318 | } |
|
320 | 319 | |
|
321 | 320 | /* map '\<c>' escape character */ |
|
322 | 321 | static char jsonescapechar2(char c) |
|
323 | 322 | { |
|
324 | 323 | switch (c) { |
|
325 | 324 | case '\b': |
|
326 | 325 | return 'b'; |
|
327 | 326 | case '\t': |
|
328 | 327 | return 't'; |
|
329 | 328 | case '\n': |
|
330 | 329 | return 'n'; |
|
331 | 330 | case '\f': |
|
332 | 331 | return 'f'; |
|
333 | 332 | case '\r': |
|
334 | 333 | return 'r'; |
|
335 | 334 | case '"': |
|
336 | 335 | return '"'; |
|
337 | 336 | case '\\': |
|
338 | 337 | return '\\'; |
|
339 | 338 | } |
|
340 | 339 | return '\0'; /* should not happen */ |
|
341 | 340 | } |
|
342 | 341 | |
|
343 | 342 | /* convert 'origbuf' to JSON-escaped form 'escbuf'; 'origbuf' should only |
|
344 | 343 | include characters mappable by json(paranoid)lentable */ |
|
345 | 344 | static void encodejsonescape(char *escbuf, Py_ssize_t esclen, |
|
346 | 345 | const char *origbuf, Py_ssize_t origlen, |
|
347 | 346 | bool paranoid) |
|
348 | 347 | { |
|
349 | 348 | const uint8_t *lentable = |
|
350 | 349 | (paranoid) ? jsonparanoidlentable : jsonlentable; |
|
351 | 350 | Py_ssize_t i, j; |
|
352 | 351 | |
|
353 | 352 | for (i = 0, j = 0; i < origlen; i++) { |
|
354 | 353 | char c = origbuf[i]; |
|
355 | 354 | uint8_t l = lentable[(unsigned char)c]; |
|
356 | 355 | assert(j + l <= esclen); |
|
357 | 356 | switch (l) { |
|
358 | 357 | case 1: |
|
359 | 358 | escbuf[j] = c; |
|
360 | 359 | break; |
|
361 | 360 | case 2: |
|
362 | 361 | escbuf[j] = '\\'; |
|
363 | 362 | escbuf[j + 1] = jsonescapechar2(c); |
|
364 | 363 | break; |
|
365 | 364 | case 6: |
|
366 | 365 | memcpy(escbuf + j, "\\u00", 4); |
|
367 | 366 | escbuf[j + 4] = hexchartable[(unsigned char)c >> 4]; |
|
368 | 367 | escbuf[j + 5] = hexchartable[(unsigned char)c & 0xf]; |
|
369 | 368 | break; |
|
370 | 369 | } |
|
371 | 370 | j += l; |
|
372 | 371 | } |
|
373 | 372 | } |
|
374 | 373 | |
|
375 | 374 | PyObject *jsonescapeu8fast(PyObject *self, PyObject *args) |
|
376 | 375 | { |
|
377 | 376 | PyObject *origstr, *escstr; |
|
378 | 377 | const char *origbuf; |
|
379 | 378 | Py_ssize_t origlen, esclen; |
|
380 | 379 | int paranoid; |
|
381 | 380 | if (!PyArg_ParseTuple(args, "O!i:jsonescapeu8fast", &PyBytes_Type, |
|
382 | 381 | &origstr, ¶noid)) { |
|
383 | 382 | return NULL; |
|
384 | 383 | } |
|
385 | 384 | |
|
386 | 385 | origbuf = PyBytes_AS_STRING(origstr); |
|
387 | 386 | origlen = PyBytes_GET_SIZE(origstr); |
|
388 | 387 | esclen = jsonescapelen(origbuf, origlen, paranoid); |
|
389 | 388 | if (esclen < 0) { |
|
390 | 389 | return NULL; /* unsupported char found or overflow */ |
|
391 | 390 | } |
|
392 | 391 | if (origlen == esclen) { |
|
393 | 392 | Py_INCREF(origstr); |
|
394 | 393 | return origstr; |
|
395 | 394 | } |
|
396 | 395 | |
|
397 | 396 | escstr = PyBytes_FromStringAndSize(NULL, esclen); |
|
398 | 397 | if (!escstr) { |
|
399 | 398 | return NULL; |
|
400 | 399 | } |
|
401 | 400 | encodejsonescape(PyBytes_AS_STRING(escstr), esclen, origbuf, origlen, |
|
402 | 401 | paranoid); |
|
403 | 402 | |
|
404 | 403 | return escstr; |
|
405 | 404 | } |
@@ -1,999 +1,999 b'' | |||
|
1 | 1 | /* |
|
2 | 2 | * manifest.c - manifest type that does on-demand parsing. |
|
3 | 3 | * |
|
4 | 4 | * Copyright 2015, Google Inc. |
|
5 | 5 | * |
|
6 | 6 | * This software may be used and distributed according to the terms of |
|
7 | 7 | * the GNU General Public License, incorporated herein by reference. |
|
8 | 8 | */ |
|
9 | 9 | #include <Python.h> |
|
10 | 10 | |
|
11 | 11 | #include <assert.h> |
|
12 | 12 | #include <stdlib.h> |
|
13 | 13 | #include <string.h> |
|
14 | 14 | |
|
15 | 15 | #include "charencode.h" |
|
16 | 16 | #include "util.h" |
|
17 | 17 | |
|
18 | 18 | #define DEFAULT_LINES 100000 |
|
19 | 19 | |
|
20 | 20 | typedef struct { |
|
21 | 21 | char *start; |
|
22 | 22 | Py_ssize_t len; /* length of line including terminal newline */ |
|
23 | 23 | char hash_suffix; |
|
24 | 24 | bool from_malloc; |
|
25 | 25 | bool deleted; |
|
26 | 26 | } line; |
|
27 | 27 | |
|
28 | 28 | typedef struct { |
|
29 | 29 | PyObject_HEAD |
|
30 | 30 | PyObject *pydata; |
|
31 | 31 | Py_ssize_t nodelen; |
|
32 | 32 | line *lines; |
|
33 | 33 | int numlines; /* number of line entries */ |
|
34 | 34 | int livelines; /* number of non-deleted lines */ |
|
35 | 35 | int maxlines; /* allocated number of lines */ |
|
36 | 36 | bool dirty; |
|
37 | 37 | } lazymanifest; |
|
38 | 38 | |
|
39 | 39 | #define MANIFEST_OOM -1 |
|
40 | 40 | #define MANIFEST_NOT_SORTED -2 |
|
41 | 41 | #define MANIFEST_MALFORMED -3 |
|
42 | 42 | #define MANIFEST_BOGUS_FILENAME -4 |
|
43 | 43 | #define MANIFEST_TOO_SHORT_LINE -5 |
|
44 | 44 | |
|
45 | 45 | /* get the length of the path for a line */ |
|
46 | 46 | static Py_ssize_t pathlen(line *l) |
|
47 | 47 | { |
|
48 | 48 | const char *end = memchr(l->start, '\0', l->len); |
|
49 | 49 | return (end) ? (Py_ssize_t)(end - l->start) : l->len; |
|
50 | 50 | } |
|
51 | 51 | |
|
52 | 52 | /* get the node value of a single line */ |
|
53 | 53 | static PyObject *nodeof(Py_ssize_t nodelen, line *l, char *flag) |
|
54 | 54 | { |
|
55 | 55 | char *s = l->start; |
|
56 | 56 | Py_ssize_t llen = pathlen(l); |
|
57 | 57 | Py_ssize_t hlen = l->len - llen - 2; |
|
58 | 58 | PyObject *hash; |
|
59 | 59 | if (llen + 1 + 40 + 1 > l->len) { /* path '\0' hash '\n' */ |
|
60 | 60 | PyErr_SetString(PyExc_ValueError, "manifest line too short"); |
|
61 | 61 | return NULL; |
|
62 | 62 | } |
|
63 | 63 | /* Detect flags after the hash first. */ |
|
64 | 64 | switch (s[llen + hlen]) { |
|
65 | 65 | case 'l': |
|
66 | 66 | case 't': |
|
67 | 67 | case 'x': |
|
68 | 68 | *flag = s[llen + hlen]; |
|
69 | 69 | --hlen; |
|
70 | 70 | break; |
|
71 | 71 | default: |
|
72 | 72 | *flag = '\0'; |
|
73 | 73 | break; |
|
74 | 74 | } |
|
75 | 75 | |
|
76 | 76 | if (hlen != 2 * nodelen) { |
|
77 | 77 | PyErr_SetString(PyExc_ValueError, "invalid node length in manifest"); |
|
78 | 78 | return NULL; |
|
79 | 79 | } |
|
80 | 80 | hash = unhexlify(s + llen + 1, nodelen * 2); |
|
81 | 81 | if (!hash) { |
|
82 | 82 | return NULL; |
|
83 | 83 | } |
|
84 | 84 | if (l->hash_suffix != '\0') { |
|
85 | 85 | char newhash[33]; |
|
86 | 86 | memcpy(newhash, PyBytes_AsString(hash), nodelen); |
|
87 | 87 | Py_DECREF(hash); |
|
88 | 88 | newhash[nodelen] = l->hash_suffix; |
|
89 | 89 | hash = PyBytes_FromStringAndSize(newhash, nodelen + 1); |
|
90 | 90 | } |
|
91 | 91 | return hash; |
|
92 | 92 | } |
|
93 | 93 | |
|
94 | 94 | /* get the node hash and flags of a line as a tuple */ |
|
95 | 95 | static PyObject *hashflags(Py_ssize_t nodelen, line *l) |
|
96 | 96 | { |
|
97 | 97 | char flag; |
|
98 | 98 | PyObject *hash = nodeof(nodelen, l, &flag); |
|
99 | 99 | PyObject *flags; |
|
100 | 100 | PyObject *tup; |
|
101 | 101 | |
|
102 | 102 | if (!hash) |
|
103 | 103 | return NULL; |
|
104 | 104 | flags = PyBytes_FromStringAndSize(&flag, flag ? 1 : 0); |
|
105 | 105 | if (!flags) { |
|
106 | 106 | Py_DECREF(hash); |
|
107 | 107 | return NULL; |
|
108 | 108 | } |
|
109 | 109 | tup = PyTuple_Pack(2, hash, flags); |
|
110 | 110 | Py_DECREF(flags); |
|
111 | 111 | Py_DECREF(hash); |
|
112 | 112 | return tup; |
|
113 | 113 | } |
|
114 | 114 | |
|
115 | 115 | /* if we're about to run out of space in the line index, add more */ |
|
116 | 116 | static bool realloc_if_full(lazymanifest *self) |
|
117 | 117 | { |
|
118 | 118 | if (self->numlines == self->maxlines) { |
|
119 | 119 | self->maxlines *= 2; |
|
120 | 120 | self->lines = realloc(self->lines, self->maxlines * sizeof(line)); |
|
121 | 121 | } |
|
122 | 122 | return !!self->lines; |
|
123 | 123 | } |
|
124 | 124 | |
|
125 | 125 | /* |
|
126 | 126 | * Find the line boundaries in the manifest that 'data' points to and store |
|
127 | 127 | * information about each line in 'self'. |
|
128 | 128 | */ |
|
129 | 129 | static int find_lines(lazymanifest *self, char *data, Py_ssize_t len) |
|
130 | 130 | { |
|
131 | 131 | char *prev = NULL; |
|
132 | 132 | while (len > 0) { |
|
133 | 133 | line *l; |
|
134 | 134 | char *next; |
|
135 | 135 | if (*data == '\0') { |
|
136 | 136 | /* It's implausible there's no filename, don't |
|
137 | 137 | * even bother looking for the newline. */ |
|
138 | 138 | return MANIFEST_BOGUS_FILENAME; |
|
139 | 139 | } |
|
140 | 140 | next = memchr(data, '\n', len); |
|
141 | 141 | if (!next) { |
|
142 | 142 | return MANIFEST_MALFORMED; |
|
143 | 143 | } |
|
144 | 144 | if ((next - data) < 42) { |
|
145 | 145 | /* We should have at least 42 bytes in a line: |
|
146 | 146 | 1 byte filename |
|
147 | 147 | 1 NUL |
|
148 | 148 | 40 bytes of hash |
|
149 | 149 | so we can give up here. |
|
150 | 150 | */ |
|
151 | 151 | return MANIFEST_TOO_SHORT_LINE; |
|
152 | 152 | } |
|
153 | 153 | next++; /* advance past newline */ |
|
154 | 154 | if (prev && strcmp(prev, data) > -1) { |
|
155 | 155 | /* This data isn't sorted, so we have to abort. */ |
|
156 | 156 | return MANIFEST_NOT_SORTED; |
|
157 | 157 | } |
|
158 | 158 | if (!realloc_if_full(self)) { |
|
159 | 159 | return MANIFEST_OOM; /* no memory */ |
|
160 | 160 | } |
|
161 | 161 | l = self->lines + ((self->numlines)++); |
|
162 | 162 | l->start = data; |
|
163 | 163 | l->len = next - data; |
|
164 | 164 | l->hash_suffix = '\0'; |
|
165 | 165 | l->from_malloc = false; |
|
166 | 166 | l->deleted = false; |
|
167 | 167 | len = len - l->len; |
|
168 | 168 | prev = data; |
|
169 | 169 | data = next; |
|
170 | 170 | } |
|
171 | 171 | self->livelines = self->numlines; |
|
172 | 172 | return 0; |
|
173 | 173 | } |
|
174 | 174 | |
|
175 | 175 | static void lazymanifest_init_early(lazymanifest *self) |
|
176 | 176 | { |
|
177 | 177 | self->pydata = NULL; |
|
178 | 178 | self->lines = NULL; |
|
179 | 179 | self->numlines = 0; |
|
180 | 180 | self->maxlines = 0; |
|
181 | 181 | } |
|
182 | 182 | |
|
183 | 183 | static int lazymanifest_init(lazymanifest *self, PyObject *args) |
|
184 | 184 | { |
|
185 | 185 | char *data; |
|
186 | 186 | Py_ssize_t nodelen, len; |
|
187 | 187 | int err, ret; |
|
188 | 188 | PyObject *pydata; |
|
189 | 189 | |
|
190 | 190 | lazymanifest_init_early(self); |
|
191 | 191 | if (!PyArg_ParseTuple(args, "nS", &nodelen, &pydata)) { |
|
192 | 192 | return -1; |
|
193 | 193 | } |
|
194 | 194 | if (nodelen != 20 && nodelen != 32) { |
|
195 | 195 | /* See fixed buffer in nodeof */ |
|
196 | 196 | PyErr_Format(PyExc_ValueError, "Unsupported node length"); |
|
197 | 197 | return -1; |
|
198 | 198 | } |
|
199 | 199 | self->nodelen = nodelen; |
|
200 | 200 | self->dirty = false; |
|
201 | 201 | |
|
202 | 202 | err = PyBytes_AsStringAndSize(pydata, &data, &len); |
|
203 | 203 | if (err == -1) |
|
204 | 204 | return -1; |
|
205 | 205 | self->pydata = pydata; |
|
206 | 206 | Py_INCREF(self->pydata); |
|
207 | 207 | Py_BEGIN_ALLOW_THREADS |
|
208 | 208 | self->lines = malloc(DEFAULT_LINES * sizeof(line)); |
|
209 | 209 | self->maxlines = DEFAULT_LINES; |
|
210 | 210 | self->numlines = 0; |
|
211 | 211 | if (!self->lines) |
|
212 | 212 | ret = MANIFEST_OOM; |
|
213 | 213 | else |
|
214 | 214 | ret = find_lines(self, data, len); |
|
215 | 215 | Py_END_ALLOW_THREADS |
|
216 | 216 | switch (ret) { |
|
217 | 217 | case 0: |
|
218 | 218 | break; |
|
219 | 219 | case MANIFEST_OOM: |
|
220 | 220 | PyErr_NoMemory(); |
|
221 | 221 | break; |
|
222 | 222 | case MANIFEST_NOT_SORTED: |
|
223 | 223 | PyErr_Format(PyExc_ValueError, |
|
224 | 224 | "Manifest lines not in sorted order."); |
|
225 | 225 | break; |
|
226 | 226 | case MANIFEST_MALFORMED: |
|
227 | 227 | PyErr_Format(PyExc_ValueError, |
|
228 | 228 | "Manifest did not end in a newline."); |
|
229 | 229 | break; |
|
230 | 230 | case MANIFEST_BOGUS_FILENAME: |
|
231 | 231 | PyErr_Format( |
|
232 | 232 | PyExc_ValueError, |
|
233 | 233 | "Manifest had an entry with a zero-length filename."); |
|
234 | 234 | break; |
|
235 | 235 | case MANIFEST_TOO_SHORT_LINE: |
|
236 | 236 | PyErr_Format( |
|
237 | 237 | PyExc_ValueError, |
|
238 | 238 | "Manifest had implausibly-short line."); |
|
239 | 239 | break; |
|
240 | 240 | default: |
|
241 | 241 | PyErr_Format(PyExc_ValueError, |
|
242 | 242 | "Unknown problem parsing manifest."); |
|
243 | 243 | } |
|
244 | 244 | return ret == 0 ? 0 : -1; |
|
245 | 245 | } |
|
246 | 246 | |
|
247 | 247 | static void lazymanifest_dealloc(lazymanifest *self) |
|
248 | 248 | { |
|
249 | 249 | /* free any extra lines we had to allocate */ |
|
250 | 250 | int i; |
|
251 | 251 | for (i = 0; self->lines && (i < self->numlines); i++) { |
|
252 | 252 | if (self->lines[i].from_malloc) { |
|
253 | 253 | free(self->lines[i].start); |
|
254 | 254 | } |
|
255 | 255 | } |
|
256 | 256 | free(self->lines); |
|
257 | 257 | self->lines = NULL; |
|
258 | 258 | if (self->pydata) { |
|
259 | 259 | Py_DECREF(self->pydata); |
|
260 | 260 | self->pydata = NULL; |
|
261 | 261 | } |
|
262 | 262 | PyObject_Del(self); |
|
263 | 263 | } |
|
264 | 264 | |
|
265 | 265 | /* iteration support */ |
|
266 | 266 | |
|
267 | 267 | typedef struct { |
|
268 | 268 | PyObject_HEAD lazymanifest *m; |
|
269 | 269 | Py_ssize_t pos; |
|
270 | 270 | } lmIter; |
|
271 | 271 | |
|
272 | 272 | static void lmiter_dealloc(PyObject *o) |
|
273 | 273 | { |
|
274 | 274 | lmIter *self = (lmIter *)o; |
|
275 | 275 | Py_DECREF(self->m); |
|
276 | 276 | PyObject_Del(self); |
|
277 | 277 | } |
|
278 | 278 | |
|
279 | 279 | static line *lmiter_nextline(lmIter *self) |
|
280 | 280 | { |
|
281 | 281 | do { |
|
282 | 282 | self->pos++; |
|
283 | 283 | if (self->pos >= self->m->numlines) { |
|
284 | 284 | return NULL; |
|
285 | 285 | } |
|
286 | 286 | /* skip over deleted manifest entries */ |
|
287 | 287 | } while (self->m->lines[self->pos].deleted); |
|
288 | 288 | return self->m->lines + self->pos; |
|
289 | 289 | } |
|
290 | 290 | |
|
291 | 291 | static PyObject *lmiter_iterentriesnext(PyObject *o) |
|
292 | 292 | { |
|
293 | 293 | lmIter *self = (lmIter *)o; |
|
294 | 294 | Py_ssize_t pl; |
|
295 | 295 | line *l; |
|
296 | 296 | char flag; |
|
297 | 297 | PyObject *ret = NULL, *path = NULL, *hash = NULL, *flags = NULL; |
|
298 | 298 | l = lmiter_nextline(self); |
|
299 | 299 | if (!l) { |
|
300 | 300 | goto done; |
|
301 | 301 | } |
|
302 | 302 | pl = pathlen(l); |
|
303 | 303 | path = PyBytes_FromStringAndSize(l->start, pl); |
|
304 | 304 | hash = nodeof(self->m->nodelen, l, &flag); |
|
305 | 305 | if (!path || !hash) { |
|
306 | 306 | goto done; |
|
307 | 307 | } |
|
308 | 308 | flags = PyBytes_FromStringAndSize(&flag, flag ? 1 : 0); |
|
309 | 309 | if (!flags) { |
|
310 | 310 | goto done; |
|
311 | 311 | } |
|
312 | 312 | ret = PyTuple_Pack(3, path, hash, flags); |
|
313 | 313 | done: |
|
314 | 314 | Py_XDECREF(path); |
|
315 | 315 | Py_XDECREF(hash); |
|
316 | 316 | Py_XDECREF(flags); |
|
317 | 317 | return ret; |
|
318 | 318 | } |
|
319 | 319 | |
|
320 | 320 | #define LAZYMANIFESTENTRIESITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT |
|
321 | 321 | |
|
322 | 322 | static PyTypeObject lazymanifestEntriesIterator = { |
|
323 | 323 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ |
|
324 | 324 | "parsers.lazymanifest.entriesiterator", /*tp_name */ |
|
325 | 325 | sizeof(lmIter), /*tp_basicsize */ |
|
326 | 326 | 0, /*tp_itemsize */ |
|
327 | 327 | lmiter_dealloc, /*tp_dealloc */ |
|
328 | 328 | 0, /*tp_print */ |
|
329 | 329 | 0, /*tp_getattr */ |
|
330 | 330 | 0, /*tp_setattr */ |
|
331 | 331 | 0, /*tp_compare */ |
|
332 | 332 | 0, /*tp_repr */ |
|
333 | 333 | 0, /*tp_as_number */ |
|
334 | 334 | 0, /*tp_as_sequence */ |
|
335 | 335 | 0, /*tp_as_mapping */ |
|
336 | 336 | 0, /*tp_hash */ |
|
337 | 337 | 0, /*tp_call */ |
|
338 | 338 | 0, /*tp_str */ |
|
339 | 339 | 0, /*tp_getattro */ |
|
340 | 340 | 0, /*tp_setattro */ |
|
341 | 341 | 0, /*tp_as_buffer */ |
|
342 | 342 | LAZYMANIFESTENTRIESITERATOR_TPFLAGS, /* tp_flags */ |
|
343 | 343 | "Iterator for 3-tuples in a lazymanifest.", /* tp_doc */ |
|
344 | 344 | 0, /* tp_traverse */ |
|
345 | 345 | 0, /* tp_clear */ |
|
346 | 346 | 0, /* tp_richcompare */ |
|
347 | 347 | 0, /* tp_weaklistoffset */ |
|
348 | 348 | PyObject_SelfIter, /* tp_iter: __iter__() method */ |
|
349 | 349 | lmiter_iterentriesnext, /* tp_iternext: next() method */ |
|
350 | 350 | }; |
|
351 | 351 | |
|
352 | 352 | static PyObject *lmiter_iterkeysnext(PyObject *o) |
|
353 | 353 | { |
|
354 | 354 | Py_ssize_t pl; |
|
355 | 355 | line *l = lmiter_nextline((lmIter *)o); |
|
356 | 356 | if (!l) { |
|
357 | 357 | return NULL; |
|
358 | 358 | } |
|
359 | 359 | pl = pathlen(l); |
|
360 | 360 | return PyBytes_FromStringAndSize(l->start, pl); |
|
361 | 361 | } |
|
362 | 362 | |
|
363 | 363 | #define LAZYMANIFESTKEYSITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT |
|
364 | 364 | |
|
365 | 365 | static PyTypeObject lazymanifestKeysIterator = { |
|
366 | 366 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ |
|
367 | 367 | "parsers.lazymanifest.keysiterator", /*tp_name */ |
|
368 | 368 | sizeof(lmIter), /*tp_basicsize */ |
|
369 | 369 | 0, /*tp_itemsize */ |
|
370 | 370 | lmiter_dealloc, /*tp_dealloc */ |
|
371 | 371 | 0, /*tp_print */ |
|
372 | 372 | 0, /*tp_getattr */ |
|
373 | 373 | 0, /*tp_setattr */ |
|
374 | 374 | 0, /*tp_compare */ |
|
375 | 375 | 0, /*tp_repr */ |
|
376 | 376 | 0, /*tp_as_number */ |
|
377 | 377 | 0, /*tp_as_sequence */ |
|
378 | 378 | 0, /*tp_as_mapping */ |
|
379 | 379 | 0, /*tp_hash */ |
|
380 | 380 | 0, /*tp_call */ |
|
381 | 381 | 0, /*tp_str */ |
|
382 | 382 | 0, /*tp_getattro */ |
|
383 | 383 | 0, /*tp_setattro */ |
|
384 | 384 | 0, /*tp_as_buffer */ |
|
385 | 385 | LAZYMANIFESTKEYSITERATOR_TPFLAGS, /* tp_flags */ |
|
386 | 386 | "Keys iterator for a lazymanifest.", /* tp_doc */ |
|
387 | 387 | 0, /* tp_traverse */ |
|
388 | 388 | 0, /* tp_clear */ |
|
389 | 389 | 0, /* tp_richcompare */ |
|
390 | 390 | 0, /* tp_weaklistoffset */ |
|
391 | 391 | PyObject_SelfIter, /* tp_iter: __iter__() method */ |
|
392 | 392 | lmiter_iterkeysnext, /* tp_iternext: next() method */ |
|
393 | 393 | }; |
|
394 | 394 | |
|
395 | 395 | static lazymanifest *lazymanifest_copy(lazymanifest *self); |
|
396 | 396 | |
|
397 | 397 | static PyObject *lazymanifest_getentriesiter(lazymanifest *self) |
|
398 | 398 | { |
|
399 | 399 | lmIter *i = NULL; |
|
400 | 400 | lazymanifest *t = lazymanifest_copy(self); |
|
401 | 401 | if (!t) { |
|
402 | 402 | PyErr_NoMemory(); |
|
403 | 403 | return NULL; |
|
404 | 404 | } |
|
405 | 405 | i = PyObject_New(lmIter, &lazymanifestEntriesIterator); |
|
406 | 406 | if (i) { |
|
407 | 407 | i->m = t; |
|
408 | 408 | i->pos = -1; |
|
409 | 409 | } else { |
|
410 | 410 | Py_DECREF(t); |
|
411 | 411 | PyErr_NoMemory(); |
|
412 | 412 | } |
|
413 | 413 | return (PyObject *)i; |
|
414 | 414 | } |
|
415 | 415 | |
|
416 | 416 | static PyObject *lazymanifest_getkeysiter(lazymanifest *self) |
|
417 | 417 | { |
|
418 | 418 | lmIter *i = NULL; |
|
419 | 419 | lazymanifest *t = lazymanifest_copy(self); |
|
420 | 420 | if (!t) { |
|
421 | 421 | PyErr_NoMemory(); |
|
422 | 422 | return NULL; |
|
423 | 423 | } |
|
424 | 424 | i = PyObject_New(lmIter, &lazymanifestKeysIterator); |
|
425 | 425 | if (i) { |
|
426 | 426 | i->m = t; |
|
427 | 427 | i->pos = -1; |
|
428 | 428 | } else { |
|
429 | 429 | Py_DECREF(t); |
|
430 | 430 | PyErr_NoMemory(); |
|
431 | 431 | } |
|
432 | 432 | return (PyObject *)i; |
|
433 | 433 | } |
|
434 | 434 | |
|
435 | 435 | /* __getitem__ and __setitem__ support */ |
|
436 | 436 | |
|
437 | 437 | static Py_ssize_t lazymanifest_size(lazymanifest *self) |
|
438 | 438 | { |
|
439 | 439 | return self->livelines; |
|
440 | 440 | } |
|
441 | 441 | |
|
442 | 442 | static int linecmp(const void *left, const void *right) |
|
443 | 443 | { |
|
444 | 444 | return strcmp(((const line *)left)->start, |
|
445 | 445 | ((const line *)right)->start); |
|
446 | 446 | } |
|
447 | 447 | |
|
448 | 448 | static PyObject *lazymanifest_getitem(lazymanifest *self, PyObject *key) |
|
449 | 449 | { |
|
450 | 450 | line needle; |
|
451 | 451 | line *hit; |
|
452 | 452 | if (!PyBytes_Check(key)) { |
|
453 | 453 | PyErr_Format(PyExc_TypeError, |
|
454 | 454 | "getitem: manifest keys must be a string."); |
|
455 | 455 | return NULL; |
|
456 | 456 | } |
|
457 | 457 | needle.start = PyBytes_AsString(key); |
|
458 | 458 | hit = bsearch(&needle, self->lines, self->numlines, sizeof(line), |
|
459 | 459 | &linecmp); |
|
460 | 460 | if (!hit || hit->deleted) { |
|
461 | 461 | PyErr_Format(PyExc_KeyError, "No such manifest entry."); |
|
462 | 462 | return NULL; |
|
463 | 463 | } |
|
464 | 464 | return hashflags(self->nodelen, hit); |
|
465 | 465 | } |
|
466 | 466 | |
|
467 | 467 | static int lazymanifest_delitem(lazymanifest *self, PyObject *key) |
|
468 | 468 | { |
|
469 | 469 | line needle; |
|
470 | 470 | line *hit; |
|
471 | 471 | if (!PyBytes_Check(key)) { |
|
472 | 472 | PyErr_Format(PyExc_TypeError, |
|
473 | 473 | "delitem: manifest keys must be a string."); |
|
474 | 474 | return -1; |
|
475 | 475 | } |
|
476 | 476 | needle.start = PyBytes_AsString(key); |
|
477 | 477 | hit = bsearch(&needle, self->lines, self->numlines, sizeof(line), |
|
478 | 478 | &linecmp); |
|
479 | 479 | if (!hit || hit->deleted) { |
|
480 | 480 | PyErr_Format(PyExc_KeyError, |
|
481 | 481 | "Tried to delete nonexistent manifest entry."); |
|
482 | 482 | return -1; |
|
483 | 483 | } |
|
484 | 484 | self->dirty = true; |
|
485 | 485 | hit->deleted = true; |
|
486 | 486 | self->livelines--; |
|
487 | 487 | return 0; |
|
488 | 488 | } |
|
489 | 489 | |
|
490 | 490 | /* Do a binary search for the insertion point for new, creating the |
|
491 | 491 | * new entry if needed. */ |
|
492 | 492 | static int internalsetitem(lazymanifest *self, line *new) |
|
493 | 493 | { |
|
494 | 494 | int start = 0, end = self->numlines; |
|
495 | 495 | while (start < end) { |
|
496 | 496 | int pos = start + (end - start) / 2; |
|
497 | 497 | int c = linecmp(new, self->lines + pos); |
|
498 | 498 | if (c < 0) |
|
499 | 499 | end = pos; |
|
500 | 500 | else if (c > 0) |
|
501 | 501 | start = pos + 1; |
|
502 | 502 | else { |
|
503 | 503 | if (self->lines[pos].deleted) |
|
504 | 504 | self->livelines++; |
|
505 | 505 | if (self->lines[pos].from_malloc) |
|
506 | 506 | free(self->lines[pos].start); |
|
507 | 507 | start = pos; |
|
508 | 508 | goto finish; |
|
509 | 509 | } |
|
510 | 510 | } |
|
511 | 511 | /* being here means we need to do an insert */ |
|
512 | 512 | if (!realloc_if_full(self)) { |
|
513 | 513 | PyErr_NoMemory(); |
|
514 | 514 | return -1; |
|
515 | 515 | } |
|
516 | 516 | memmove(self->lines + start + 1, self->lines + start, |
|
517 | 517 | (self->numlines - start) * sizeof(line)); |
|
518 | 518 | self->numlines++; |
|
519 | 519 | self->livelines++; |
|
520 | 520 | finish: |
|
521 | 521 | self->lines[start] = *new; |
|
522 | 522 | self->dirty = true; |
|
523 | 523 | return 0; |
|
524 | 524 | } |
|
525 | 525 | |
|
526 | 526 | static int lazymanifest_setitem( |
|
527 | 527 | lazymanifest *self, PyObject *key, PyObject *value) |
|
528 | 528 | { |
|
529 | 529 | char *path; |
|
530 | 530 | Py_ssize_t plen; |
|
531 | 531 | PyObject *pyhash; |
|
532 | 532 | Py_ssize_t hlen; |
|
533 | 533 | char *hash; |
|
534 | 534 | PyObject *pyflags; |
|
535 | 535 | char *flags; |
|
536 | 536 | Py_ssize_t flen; |
|
537 | 537 | Py_ssize_t dlen; |
|
538 | 538 | char *dest; |
|
539 | 539 | int i; |
|
540 | 540 | line new; |
|
541 | 541 | if (!PyBytes_Check(key)) { |
|
542 | 542 | PyErr_Format(PyExc_TypeError, |
|
543 | 543 | "setitem: manifest keys must be a string."); |
|
544 | 544 | return -1; |
|
545 | 545 | } |
|
546 | 546 | if (!value) { |
|
547 | 547 | return lazymanifest_delitem(self, key); |
|
548 | 548 | } |
|
549 | 549 | if (!PyTuple_Check(value) || PyTuple_Size(value) != 2) { |
|
550 | 550 | PyErr_Format(PyExc_TypeError, |
|
551 | 551 | "Manifest values must be a tuple of (node, flags)."); |
|
552 | 552 | return -1; |
|
553 | 553 | } |
|
554 | 554 | if (PyBytes_AsStringAndSize(key, &path, &plen) == -1) { |
|
555 | 555 | return -1; |
|
556 | 556 | } |
|
557 | 557 | |
|
558 | 558 | pyhash = PyTuple_GetItem(value, 0); |
|
559 | 559 | if (!PyBytes_Check(pyhash)) { |
|
560 | 560 | PyErr_Format(PyExc_TypeError, |
|
561 | 561 | "node must be a %zi bytes string", self->nodelen); |
|
562 | 562 | return -1; |
|
563 | 563 | } |
|
564 | 564 | hlen = PyBytes_Size(pyhash); |
|
565 | 565 | if (hlen != self->nodelen) { |
|
566 | 566 | PyErr_Format(PyExc_TypeError, |
|
567 | 567 | "node must be a %zi bytes string", self->nodelen); |
|
568 | 568 | return -1; |
|
569 | 569 | } |
|
570 | 570 | hash = PyBytes_AsString(pyhash); |
|
571 | 571 | |
|
572 | 572 | pyflags = PyTuple_GetItem(value, 1); |
|
573 | 573 | if (!PyBytes_Check(pyflags) || PyBytes_Size(pyflags) > 1) { |
|
574 | 574 | PyErr_Format(PyExc_TypeError, |
|
575 | 575 | "flags must a 0 or 1 bytes string"); |
|
576 | 576 | return -1; |
|
577 | 577 | } |
|
578 | 578 | if (PyBytes_AsStringAndSize(pyflags, &flags, &flen) == -1) { |
|
579 | 579 | return -1; |
|
580 | 580 | } |
|
581 | 581 | if (flen == 1) { |
|
582 | 582 | switch (*flags) { |
|
583 | 583 | case 'l': |
|
584 | 584 | case 't': |
|
585 | 585 | case 'x': |
|
586 | 586 | break; |
|
587 | 587 | default: |
|
588 | 588 | PyErr_Format(PyExc_TypeError, "invalid manifest flag"); |
|
589 | 589 | return -1; |
|
590 | 590 | } |
|
591 | 591 | } |
|
592 | 592 | /* one null byte and one newline */ |
|
593 | 593 | dlen = plen + hlen * 2 + 1 + flen + 1; |
|
594 | 594 | dest = malloc(dlen); |
|
595 | 595 | if (!dest) { |
|
596 | 596 | PyErr_NoMemory(); |
|
597 | 597 | return -1; |
|
598 | 598 | } |
|
599 | 599 | memcpy(dest, path, plen + 1); |
|
600 | 600 | for (i = 0; i < hlen; i++) { |
|
601 | 601 | /* Cast to unsigned, so it will not get sign-extended when promoted |
|
602 | 602 | * to int (as is done when passing to a variadic function) |
|
603 | 603 | */ |
|
604 | 604 | sprintf(dest + plen + 1 + (i * 2), "%02x", (unsigned char)hash[i]); |
|
605 | 605 | } |
|
606 | 606 | memcpy(dest + plen + 2 * hlen + 1, flags, flen); |
|
607 | 607 | dest[plen + 2 * hlen + 1 + flen] = '\n'; |
|
608 | 608 | new.start = dest; |
|
609 | 609 | new.len = dlen; |
|
610 | 610 | new.hash_suffix = '\0'; |
|
611 | 611 | if (hlen > 20) { |
|
612 | 612 | new.hash_suffix = hash[20]; |
|
613 | 613 | } |
|
614 | 614 | new.from_malloc = true; /* is `start` a pointer we allocated? */ |
|
615 | 615 | new.deleted = false; /* is this entry deleted? */ |
|
616 | 616 | if (internalsetitem(self, &new)) { |
|
617 | 617 | return -1; |
|
618 | 618 | } |
|
619 | 619 | return 0; |
|
620 | 620 | } |
|
621 | 621 | |
|
622 | 622 | static PyMappingMethods lazymanifest_mapping_methods = { |
|
623 | 623 | (lenfunc)lazymanifest_size, /* mp_length */ |
|
624 | 624 | (binaryfunc)lazymanifest_getitem, /* mp_subscript */ |
|
625 | 625 | (objobjargproc)lazymanifest_setitem, /* mp_ass_subscript */ |
|
626 | 626 | }; |
|
627 | 627 | |
|
628 | 628 | /* sequence methods (important or __contains__ builds an iterator) */ |
|
629 | 629 | |
|
630 | 630 | static int lazymanifest_contains(lazymanifest *self, PyObject *key) |
|
631 | 631 | { |
|
632 | 632 | line needle; |
|
633 | 633 | line *hit; |
|
634 | 634 | if (!PyBytes_Check(key)) { |
|
635 | 635 | /* Our keys are always strings, so if the contains |
|
636 | 636 | * check is for a non-string, just return false. */ |
|
637 | 637 | return 0; |
|
638 | 638 | } |
|
639 | 639 | needle.start = PyBytes_AsString(key); |
|
640 | 640 | hit = bsearch(&needle, self->lines, self->numlines, sizeof(line), |
|
641 | 641 | &linecmp); |
|
642 | 642 | if (!hit || hit->deleted) { |
|
643 | 643 | return 0; |
|
644 | 644 | } |
|
645 | 645 | return 1; |
|
646 | 646 | } |
|
647 | 647 | |
|
648 | 648 | static PySequenceMethods lazymanifest_seq_meths = { |
|
649 | 649 | (lenfunc)lazymanifest_size, /* sq_length */ |
|
650 | 650 | 0, /* sq_concat */ |
|
651 | 651 | 0, /* sq_repeat */ |
|
652 | 652 | 0, /* sq_item */ |
|
653 | 653 | 0, /* sq_slice */ |
|
654 | 654 | 0, /* sq_ass_item */ |
|
655 | 655 | 0, /* sq_ass_slice */ |
|
656 | 656 | (objobjproc)lazymanifest_contains, /* sq_contains */ |
|
657 | 657 | 0, /* sq_inplace_concat */ |
|
658 | 658 | 0, /* sq_inplace_repeat */ |
|
659 | 659 | }; |
|
660 | 660 | |
|
661 | 661 | |
|
662 | 662 | /* Other methods (copy, diff, etc) */ |
|
663 | 663 | static PyTypeObject lazymanifestType; |
|
664 | 664 | |
|
665 | 665 | /* If the manifest has changes, build the new manifest text and reindex it. */ |
|
666 | 666 | static int compact(lazymanifest *self) |
|
667 | 667 | { |
|
668 | 668 | int i; |
|
669 | 669 | ssize_t need = 0; |
|
670 | 670 | char *data; |
|
671 | 671 | line *src, *dst; |
|
672 | 672 | PyObject *pydata; |
|
673 | 673 | if (!self->dirty) |
|
674 | 674 | return 0; |
|
675 | 675 | for (i = 0; i < self->numlines; i++) { |
|
676 | 676 | if (!self->lines[i].deleted) { |
|
677 | 677 | need += self->lines[i].len; |
|
678 | 678 | } |
|
679 | 679 | } |
|
680 | 680 | pydata = PyBytes_FromStringAndSize(NULL, need); |
|
681 | 681 | if (!pydata) |
|
682 | 682 | return -1; |
|
683 | 683 | data = PyBytes_AsString(pydata); |
|
684 | 684 | if (!data) { |
|
685 | 685 | return -1; |
|
686 | 686 | } |
|
687 | 687 | src = self->lines; |
|
688 | 688 | dst = self->lines; |
|
689 | 689 | for (i = 0; i < self->numlines; i++, src++) { |
|
690 | 690 | char *tofree = NULL; |
|
691 | 691 | if (src->from_malloc) { |
|
692 | 692 | tofree = src->start; |
|
693 | 693 | } |
|
694 | 694 | if (!src->deleted) { |
|
695 | 695 | memcpy(data, src->start, src->len); |
|
696 | 696 | *dst = *src; |
|
697 | 697 | dst->start = data; |
|
698 | 698 | dst->from_malloc = false; |
|
699 | 699 | data += dst->len; |
|
700 | 700 | dst++; |
|
701 | 701 | } |
|
702 | 702 | free(tofree); |
|
703 | 703 | } |
|
704 | 704 | Py_DECREF(self->pydata); |
|
705 | 705 | self->pydata = pydata; |
|
706 | 706 | self->numlines = self->livelines; |
|
707 | 707 | self->dirty = false; |
|
708 | 708 | return 0; |
|
709 | 709 | } |
|
710 | 710 | |
|
711 | 711 | static PyObject *lazymanifest_text(lazymanifest *self) |
|
712 | 712 | { |
|
713 | 713 | if (compact(self) != 0) { |
|
714 | 714 | PyErr_NoMemory(); |
|
715 | 715 | return NULL; |
|
716 | 716 | } |
|
717 | 717 | Py_INCREF(self->pydata); |
|
718 | 718 | return self->pydata; |
|
719 | 719 | } |
|
720 | 720 | |
|
721 | 721 | static lazymanifest *lazymanifest_copy(lazymanifest *self) |
|
722 | 722 | { |
|
723 | 723 | lazymanifest *copy = NULL; |
|
724 | 724 | if (compact(self) != 0) { |
|
725 | 725 | goto nomem; |
|
726 | 726 | } |
|
727 | 727 | copy = PyObject_New(lazymanifest, &lazymanifestType); |
|
728 | 728 | if (!copy) { |
|
729 | 729 | goto nomem; |
|
730 | 730 | } |
|
731 | 731 | lazymanifest_init_early(copy); |
|
732 | 732 | copy->nodelen = self->nodelen; |
|
733 | 733 | copy->numlines = self->numlines; |
|
734 | 734 | copy->livelines = self->livelines; |
|
735 | 735 | copy->dirty = false; |
|
736 | 736 | copy->lines = malloc(self->maxlines *sizeof(line)); |
|
737 | 737 | if (!copy->lines) { |
|
738 | 738 | goto nomem; |
|
739 | 739 | } |
|
740 | 740 | memcpy(copy->lines, self->lines, self->numlines * sizeof(line)); |
|
741 | 741 | copy->maxlines = self->maxlines; |
|
742 | 742 | copy->pydata = self->pydata; |
|
743 | 743 | Py_INCREF(copy->pydata); |
|
744 | 744 | return copy; |
|
745 | 745 | nomem: |
|
746 | 746 | PyErr_NoMemory(); |
|
747 | 747 | Py_XDECREF(copy); |
|
748 | 748 | return NULL; |
|
749 | 749 | } |
|
750 | 750 | |
|
751 | 751 | static lazymanifest *lazymanifest_filtercopy( |
|
752 | 752 | lazymanifest *self, PyObject *matchfn) |
|
753 | 753 | { |
|
754 | 754 | lazymanifest *copy = NULL; |
|
755 | 755 | int i; |
|
756 | 756 | if (!PyCallable_Check(matchfn)) { |
|
757 | 757 | PyErr_SetString(PyExc_TypeError, "matchfn must be callable"); |
|
758 | 758 | return NULL; |
|
759 | 759 | } |
|
760 | 760 | /* compact ourselves first to avoid double-frees later when we |
|
761 | 761 | * compact tmp so that it doesn't have random pointers to our |
|
762 | 762 | * underlying from_malloc-data (self->pydata is safe) */ |
|
763 | 763 | if (compact(self) != 0) { |
|
764 | 764 | goto nomem; |
|
765 | 765 | } |
|
766 | 766 | copy = PyObject_New(lazymanifest, &lazymanifestType); |
|
767 | 767 | if (!copy) { |
|
768 | 768 | goto nomem; |
|
769 | 769 | } |
|
770 | 770 | lazymanifest_init_early(copy); |
|
771 | 771 | copy->nodelen = self->nodelen; |
|
772 | 772 | copy->dirty = true; |
|
773 | 773 | copy->lines = malloc(self->maxlines * sizeof(line)); |
|
774 | 774 | if (!copy->lines) { |
|
775 | 775 | goto nomem; |
|
776 | 776 | } |
|
777 | 777 | copy->maxlines = self->maxlines; |
|
778 | 778 | copy->numlines = 0; |
|
779 | 779 | copy->pydata = self->pydata; |
|
780 | 780 | Py_INCREF(copy->pydata); |
|
781 | 781 | for (i = 0; i < self->numlines; i++) { |
|
782 | 782 | PyObject *arglist = NULL, *result = NULL; |
|
783 |
arglist = Py_BuildValue( |
|
|
783 | arglist = Py_BuildValue("(y)", | |
|
784 | 784 | self->lines[i].start); |
|
785 | 785 | if (!arglist) { |
|
786 | 786 | goto bail; |
|
787 | 787 | } |
|
788 | 788 | result = PyObject_CallObject(matchfn, arglist); |
|
789 | 789 | Py_DECREF(arglist); |
|
790 | 790 | /* if the callback raised an exception, just let it |
|
791 | 791 | * through and give up */ |
|
792 | 792 | if (!result) { |
|
793 | 793 | goto bail; |
|
794 | 794 | } |
|
795 | 795 | if (PyObject_IsTrue(result)) { |
|
796 | 796 | assert(!(self->lines[i].from_malloc)); |
|
797 | 797 | copy->lines[copy->numlines++] = self->lines[i]; |
|
798 | 798 | } |
|
799 | 799 | Py_DECREF(result); |
|
800 | 800 | } |
|
801 | 801 | copy->livelines = copy->numlines; |
|
802 | 802 | return copy; |
|
803 | 803 | nomem: |
|
804 | 804 | PyErr_NoMemory(); |
|
805 | 805 | bail: |
|
806 | 806 | Py_XDECREF(copy); |
|
807 | 807 | return NULL; |
|
808 | 808 | } |
|
809 | 809 | |
|
810 | 810 | static PyObject *lazymanifest_diff(lazymanifest *self, PyObject *args) |
|
811 | 811 | { |
|
812 | 812 | lazymanifest *other; |
|
813 | 813 | PyObject *pyclean = NULL; |
|
814 | 814 | bool listclean; |
|
815 | 815 | PyObject *emptyTup = NULL, *ret = NULL; |
|
816 | 816 | PyObject *es; |
|
817 | 817 | int sneedle = 0, oneedle = 0; |
|
818 | 818 | if (!PyArg_ParseTuple(args, "O!|O", &lazymanifestType, &other, &pyclean)) { |
|
819 | 819 | return NULL; |
|
820 | 820 | } |
|
821 | 821 | listclean = (!pyclean) ? false : PyObject_IsTrue(pyclean); |
|
822 | 822 | es = PyBytes_FromString(""); |
|
823 | 823 | if (!es) { |
|
824 | 824 | goto nomem; |
|
825 | 825 | } |
|
826 | 826 | emptyTup = PyTuple_Pack(2, Py_None, es); |
|
827 | 827 | Py_DECREF(es); |
|
828 | 828 | if (!emptyTup) { |
|
829 | 829 | goto nomem; |
|
830 | 830 | } |
|
831 | 831 | ret = PyDict_New(); |
|
832 | 832 | if (!ret) { |
|
833 | 833 | goto nomem; |
|
834 | 834 | } |
|
835 | 835 | while (sneedle != self->numlines || oneedle != other->numlines) { |
|
836 | 836 | line *left = self->lines + sneedle; |
|
837 | 837 | line *right = other->lines + oneedle; |
|
838 | 838 | int result; |
|
839 | 839 | PyObject *key; |
|
840 | 840 | PyObject *outer; |
|
841 | 841 | /* If we're looking at a deleted entry and it's not |
|
842 | 842 | * the end of the manifest, just skip it. */ |
|
843 | 843 | if (sneedle < self->numlines && left->deleted) { |
|
844 | 844 | sneedle++; |
|
845 | 845 | continue; |
|
846 | 846 | } |
|
847 | 847 | if (oneedle < other->numlines && right->deleted) { |
|
848 | 848 | oneedle++; |
|
849 | 849 | continue; |
|
850 | 850 | } |
|
851 | 851 | /* if we're at the end of either manifest, then we |
|
852 | 852 | * know the remaining items are adds so we can skip |
|
853 | 853 | * the strcmp. */ |
|
854 | 854 | if (sneedle == self->numlines) { |
|
855 | 855 | result = 1; |
|
856 | 856 | } else if (oneedle == other->numlines) { |
|
857 | 857 | result = -1; |
|
858 | 858 | } else { |
|
859 | 859 | result = linecmp(left, right); |
|
860 | 860 | } |
|
861 | 861 | key = result <= 0 ? |
|
862 | 862 | PyBytes_FromString(left->start) : |
|
863 | 863 | PyBytes_FromString(right->start); |
|
864 | 864 | if (!key) |
|
865 | 865 | goto nomem; |
|
866 | 866 | if (result < 0) { |
|
867 | 867 | PyObject *l = hashflags(self->nodelen, left); |
|
868 | 868 | if (!l) { |
|
869 | 869 | goto nomem; |
|
870 | 870 | } |
|
871 | 871 | outer = PyTuple_Pack(2, l, emptyTup); |
|
872 | 872 | Py_DECREF(l); |
|
873 | 873 | if (!outer) { |
|
874 | 874 | goto nomem; |
|
875 | 875 | } |
|
876 | 876 | PyDict_SetItem(ret, key, outer); |
|
877 | 877 | Py_DECREF(outer); |
|
878 | 878 | sneedle++; |
|
879 | 879 | } else if (result > 0) { |
|
880 | 880 | PyObject *r = hashflags(self->nodelen, right); |
|
881 | 881 | if (!r) { |
|
882 | 882 | goto nomem; |
|
883 | 883 | } |
|
884 | 884 | outer = PyTuple_Pack(2, emptyTup, r); |
|
885 | 885 | Py_DECREF(r); |
|
886 | 886 | if (!outer) { |
|
887 | 887 | goto nomem; |
|
888 | 888 | } |
|
889 | 889 | PyDict_SetItem(ret, key, outer); |
|
890 | 890 | Py_DECREF(outer); |
|
891 | 891 | oneedle++; |
|
892 | 892 | } else { |
|
893 | 893 | /* file exists in both manifests */ |
|
894 | 894 | if (left->len != right->len |
|
895 | 895 | || memcmp(left->start, right->start, left->len) |
|
896 | 896 | || left->hash_suffix != right->hash_suffix) { |
|
897 | 897 | PyObject *l = hashflags(self->nodelen, left); |
|
898 | 898 | PyObject *r; |
|
899 | 899 | if (!l) { |
|
900 | 900 | goto nomem; |
|
901 | 901 | } |
|
902 | 902 | r = hashflags(self->nodelen, right); |
|
903 | 903 | if (!r) { |
|
904 | 904 | Py_DECREF(l); |
|
905 | 905 | goto nomem; |
|
906 | 906 | } |
|
907 | 907 | outer = PyTuple_Pack(2, l, r); |
|
908 | 908 | Py_DECREF(l); |
|
909 | 909 | Py_DECREF(r); |
|
910 | 910 | if (!outer) { |
|
911 | 911 | goto nomem; |
|
912 | 912 | } |
|
913 | 913 | PyDict_SetItem(ret, key, outer); |
|
914 | 914 | Py_DECREF(outer); |
|
915 | 915 | } else if (listclean) { |
|
916 | 916 | PyDict_SetItem(ret, key, Py_None); |
|
917 | 917 | } |
|
918 | 918 | sneedle++; |
|
919 | 919 | oneedle++; |
|
920 | 920 | } |
|
921 | 921 | Py_DECREF(key); |
|
922 | 922 | } |
|
923 | 923 | Py_DECREF(emptyTup); |
|
924 | 924 | return ret; |
|
925 | 925 | nomem: |
|
926 | 926 | PyErr_NoMemory(); |
|
927 | 927 | Py_XDECREF(ret); |
|
928 | 928 | Py_XDECREF(emptyTup); |
|
929 | 929 | return NULL; |
|
930 | 930 | } |
|
931 | 931 | |
|
932 | 932 | static PyMethodDef lazymanifest_methods[] = { |
|
933 | 933 | {"iterkeys", (PyCFunction)lazymanifest_getkeysiter, METH_NOARGS, |
|
934 | 934 | "Iterate over file names in this lazymanifest."}, |
|
935 | 935 | {"iterentries", (PyCFunction)lazymanifest_getentriesiter, METH_NOARGS, |
|
936 | 936 | "Iterate over (path, nodeid, flags) tuples in this lazymanifest."}, |
|
937 | 937 | {"copy", (PyCFunction)lazymanifest_copy, METH_NOARGS, |
|
938 | 938 | "Make a copy of this lazymanifest."}, |
|
939 | 939 | {"filtercopy", (PyCFunction)lazymanifest_filtercopy, METH_O, |
|
940 | 940 | "Make a copy of this manifest filtered by matchfn."}, |
|
941 | 941 | {"diff", (PyCFunction)lazymanifest_diff, METH_VARARGS, |
|
942 | 942 | "Compare this lazymanifest to another one."}, |
|
943 | 943 | {"text", (PyCFunction)lazymanifest_text, METH_NOARGS, |
|
944 | 944 | "Encode this manifest to text."}, |
|
945 | 945 | {NULL}, |
|
946 | 946 | }; |
|
947 | 947 | |
|
948 | 948 | #define LAZYMANIFEST_TPFLAGS Py_TPFLAGS_DEFAULT |
|
949 | 949 | |
|
950 | 950 | static PyTypeObject lazymanifestType = { |
|
951 | 951 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ |
|
952 | 952 | "parsers.lazymanifest", /* tp_name */ |
|
953 | 953 | sizeof(lazymanifest), /* tp_basicsize */ |
|
954 | 954 | 0, /* tp_itemsize */ |
|
955 | 955 | (destructor)lazymanifest_dealloc, /* tp_dealloc */ |
|
956 | 956 | 0, /* tp_print */ |
|
957 | 957 | 0, /* tp_getattr */ |
|
958 | 958 | 0, /* tp_setattr */ |
|
959 | 959 | 0, /* tp_compare */ |
|
960 | 960 | 0, /* tp_repr */ |
|
961 | 961 | 0, /* tp_as_number */ |
|
962 | 962 | &lazymanifest_seq_meths, /* tp_as_sequence */ |
|
963 | 963 | &lazymanifest_mapping_methods, /* tp_as_mapping */ |
|
964 | 964 | 0, /* tp_hash */ |
|
965 | 965 | 0, /* tp_call */ |
|
966 | 966 | 0, /* tp_str */ |
|
967 | 967 | 0, /* tp_getattro */ |
|
968 | 968 | 0, /* tp_setattro */ |
|
969 | 969 | 0, /* tp_as_buffer */ |
|
970 | 970 | LAZYMANIFEST_TPFLAGS, /* tp_flags */ |
|
971 | 971 | "TODO(augie)", /* tp_doc */ |
|
972 | 972 | 0, /* tp_traverse */ |
|
973 | 973 | 0, /* tp_clear */ |
|
974 | 974 | 0, /* tp_richcompare */ |
|
975 | 975 | 0, /* tp_weaklistoffset */ |
|
976 | 976 | (getiterfunc)lazymanifest_getkeysiter, /* tp_iter */ |
|
977 | 977 | 0, /* tp_iternext */ |
|
978 | 978 | lazymanifest_methods, /* tp_methods */ |
|
979 | 979 | 0, /* tp_members */ |
|
980 | 980 | 0, /* tp_getset */ |
|
981 | 981 | 0, /* tp_base */ |
|
982 | 982 | 0, /* tp_dict */ |
|
983 | 983 | 0, /* tp_descr_get */ |
|
984 | 984 | 0, /* tp_descr_set */ |
|
985 | 985 | 0, /* tp_dictoffset */ |
|
986 | 986 | (initproc)lazymanifest_init, /* tp_init */ |
|
987 | 987 | 0, /* tp_alloc */ |
|
988 | 988 | }; |
|
989 | 989 | |
|
990 | 990 | void manifest_module_init(PyObject * mod) |
|
991 | 991 | { |
|
992 | 992 | lazymanifestType.tp_new = PyType_GenericNew; |
|
993 | 993 | if (PyType_Ready(&lazymanifestType) < 0) |
|
994 | 994 | return; |
|
995 | 995 | Py_INCREF(&lazymanifestType); |
|
996 | 996 | |
|
997 | 997 | PyModule_AddObject(mod, "lazymanifest", |
|
998 | 998 | (PyObject *)&lazymanifestType); |
|
999 | 999 | } |
@@ -1,204 +1,203 b'' | |||
|
1 | 1 | /* |
|
2 | 2 | mpatch.c - efficient binary patching for Mercurial |
|
3 | 3 | |
|
4 | 4 | This implements a patch algorithm that's O(m + nlog n) where m is the |
|
5 | 5 | size of the output and n is the number of patches. |
|
6 | 6 | |
|
7 | 7 | Given a list of binary patches, it unpacks each into a hunk list, |
|
8 | 8 | then combines the hunk lists with a treewise recursion to form a |
|
9 | 9 | single hunk list. This hunk list is then applied to the original |
|
10 | 10 | text. |
|
11 | 11 | |
|
12 | 12 | The text (or binary) fragments are copied directly from their source |
|
13 | 13 | Python objects into a preallocated output string to avoid the |
|
14 | 14 | allocation of intermediate Python objects. Working memory is about 2x |
|
15 | 15 | the total number of hunks. |
|
16 | 16 | |
|
17 | 17 | Copyright 2005, 2006 Olivia Mackall <olivia@selenic.com> |
|
18 | 18 | |
|
19 | 19 | This software may be used and distributed according to the terms |
|
20 | 20 | of the GNU General Public License, incorporated herein by reference. |
|
21 | 21 | */ |
|
22 | 22 | |
|
23 | 23 | #define PY_SSIZE_T_CLEAN |
|
24 | 24 | #include <Python.h> |
|
25 | 25 | #include <stdlib.h> |
|
26 | 26 | #include <string.h> |
|
27 | 27 | |
|
28 | 28 | #include "bitmanipulation.h" |
|
29 | 29 | #include "compat.h" |
|
30 | 30 | #include "mpatch.h" |
|
31 | 31 | #include "util.h" |
|
32 | 32 | |
|
33 | 33 | static char mpatch_doc[] = "Efficient binary patching."; |
|
34 | 34 | static PyObject *mpatch_Error; |
|
35 | 35 | |
|
36 | 36 | static void setpyerr(int r) |
|
37 | 37 | { |
|
38 | 38 | switch (r) { |
|
39 | 39 | case MPATCH_ERR_NO_MEM: |
|
40 | 40 | PyErr_NoMemory(); |
|
41 | 41 | break; |
|
42 | 42 | case MPATCH_ERR_CANNOT_BE_DECODED: |
|
43 | 43 | PyErr_SetString(mpatch_Error, "patch cannot be decoded"); |
|
44 | 44 | break; |
|
45 | 45 | case MPATCH_ERR_INVALID_PATCH: |
|
46 | 46 | PyErr_SetString(mpatch_Error, "invalid patch"); |
|
47 | 47 | break; |
|
48 | 48 | } |
|
49 | 49 | } |
|
50 | 50 | |
|
51 | 51 | struct mpatch_flist *cpygetitem(void *bins, ssize_t pos) |
|
52 | 52 | { |
|
53 | 53 | Py_buffer buffer; |
|
54 | 54 | struct mpatch_flist *res = NULL; |
|
55 | 55 | int r; |
|
56 | 56 | |
|
57 | 57 | PyObject *tmp = PyList_GetItem((PyObject *)bins, pos); |
|
58 | 58 | if (!tmp) { |
|
59 | 59 | return NULL; |
|
60 | 60 | } |
|
61 | 61 | if (PyObject_GetBuffer(tmp, &buffer, PyBUF_CONTIG_RO)) { |
|
62 | 62 | return NULL; |
|
63 | 63 | } |
|
64 | 64 | if ((r = mpatch_decode(buffer.buf, buffer.len, &res)) < 0) { |
|
65 | 65 | if (!PyErr_Occurred()) { |
|
66 | 66 | setpyerr(r); |
|
67 | 67 | } |
|
68 | 68 | res = NULL; |
|
69 | 69 | } |
|
70 | 70 | |
|
71 | 71 | PyBuffer_Release(&buffer); |
|
72 | 72 | return res; |
|
73 | 73 | } |
|
74 | 74 | |
|
75 | 75 | static PyObject *patches(PyObject *self, PyObject *args) |
|
76 | 76 | { |
|
77 | 77 | PyObject *text, *bins, *result; |
|
78 | 78 | struct mpatch_flist *patch; |
|
79 | 79 | Py_buffer buffer; |
|
80 | 80 | int r = 0; |
|
81 | 81 | char *out; |
|
82 | 82 | Py_ssize_t len, outlen; |
|
83 | 83 | |
|
84 | 84 | if (!PyArg_ParseTuple(args, "OO:mpatch", &text, &bins)) { |
|
85 | 85 | return NULL; |
|
86 | 86 | } |
|
87 | 87 | |
|
88 | 88 | len = PyList_Size(bins); |
|
89 | 89 | if (!len) { |
|
90 | 90 | /* nothing to do */ |
|
91 | 91 | Py_INCREF(text); |
|
92 | 92 | return text; |
|
93 | 93 | } |
|
94 | 94 | |
|
95 | 95 | if (PyObject_GetBuffer(text, &buffer, PyBUF_CONTIG_RO)) { |
|
96 | 96 | return NULL; |
|
97 | 97 | } |
|
98 | 98 | |
|
99 | 99 | patch = mpatch_fold(bins, cpygetitem, 0, len); |
|
100 | 100 | if (!patch) { /* error already set or memory error */ |
|
101 | 101 | if (!PyErr_Occurred()) { |
|
102 | 102 | PyErr_NoMemory(); |
|
103 | 103 | } |
|
104 | 104 | result = NULL; |
|
105 | 105 | goto cleanup; |
|
106 | 106 | } |
|
107 | 107 | |
|
108 | 108 | outlen = mpatch_calcsize(buffer.len, patch); |
|
109 | 109 | if (outlen < 0) { |
|
110 | 110 | r = (int)outlen; |
|
111 | 111 | result = NULL; |
|
112 | 112 | goto cleanup; |
|
113 | 113 | } |
|
114 | 114 | result = PyBytes_FromStringAndSize(NULL, outlen); |
|
115 | 115 | if (!result) { |
|
116 | 116 | result = NULL; |
|
117 | 117 | goto cleanup; |
|
118 | 118 | } |
|
119 | 119 | out = PyBytes_AsString(result); |
|
120 | 120 | /* clang-format off */ |
|
121 | 121 | { |
|
122 | 122 | Py_BEGIN_ALLOW_THREADS |
|
123 | 123 | r = mpatch_apply(out, buffer.buf, buffer.len, patch); |
|
124 | 124 | Py_END_ALLOW_THREADS |
|
125 | 125 | } |
|
126 | 126 | /* clang-format on */ |
|
127 | 127 | if (r < 0) { |
|
128 | 128 | Py_DECREF(result); |
|
129 | 129 | result = NULL; |
|
130 | 130 | } |
|
131 | 131 | cleanup: |
|
132 | 132 | mpatch_lfree(patch); |
|
133 | 133 | PyBuffer_Release(&buffer); |
|
134 | 134 | if (!result && !PyErr_Occurred()) { |
|
135 | 135 | setpyerr(r); |
|
136 | 136 | } |
|
137 | 137 | return result; |
|
138 | 138 | } |
|
139 | 139 | |
|
140 | 140 | /* calculate size of a patched file directly */ |
|
141 | 141 | static PyObject *patchedsize(PyObject *self, PyObject *args) |
|
142 | 142 | { |
|
143 | 143 | long orig, start, end, len, outlen = 0, last = 0, pos = 0; |
|
144 | 144 | Py_ssize_t patchlen; |
|
145 | 145 | char *bin; |
|
146 | 146 | |
|
147 |
if (!PyArg_ParseTuple(args, |
|
|
148 | &patchlen)) { | |
|
147 | if (!PyArg_ParseTuple(args, "ly#", &orig, &bin, &patchlen)) { | |
|
149 | 148 | return NULL; |
|
150 | 149 | } |
|
151 | 150 | |
|
152 | 151 | while (pos >= 0 && pos < patchlen) { |
|
153 | 152 | start = getbe32(bin + pos); |
|
154 | 153 | end = getbe32(bin + pos + 4); |
|
155 | 154 | len = getbe32(bin + pos + 8); |
|
156 | 155 | if (start > end) { |
|
157 | 156 | break; /* sanity check */ |
|
158 | 157 | } |
|
159 | 158 | pos += 12 + len; |
|
160 | 159 | outlen += start - last; |
|
161 | 160 | last = end; |
|
162 | 161 | outlen += len; |
|
163 | 162 | } |
|
164 | 163 | |
|
165 | 164 | if (pos != patchlen) { |
|
166 | 165 | if (!PyErr_Occurred()) { |
|
167 | 166 | PyErr_SetString(mpatch_Error, |
|
168 | 167 | "patch cannot be decoded"); |
|
169 | 168 | } |
|
170 | 169 | return NULL; |
|
171 | 170 | } |
|
172 | 171 | |
|
173 | 172 | outlen += orig - last; |
|
174 | 173 | return Py_BuildValue("l", outlen); |
|
175 | 174 | } |
|
176 | 175 | |
|
177 | 176 | static PyMethodDef methods[] = { |
|
178 | 177 | {"patches", patches, METH_VARARGS, "apply a series of patches\n"}, |
|
179 | 178 | {"patchedsize", patchedsize, METH_VARARGS, "calculed patched size\n"}, |
|
180 | 179 | {NULL, NULL}, |
|
181 | 180 | }; |
|
182 | 181 | |
|
183 | 182 | static const int version = 1; |
|
184 | 183 | |
|
185 | 184 | static struct PyModuleDef mpatch_module = { |
|
186 | 185 | PyModuleDef_HEAD_INIT, "mpatch", mpatch_doc, -1, methods, |
|
187 | 186 | }; |
|
188 | 187 | |
|
189 | 188 | PyMODINIT_FUNC PyInit_mpatch(void) |
|
190 | 189 | { |
|
191 | 190 | PyObject *m; |
|
192 | 191 | |
|
193 | 192 | m = PyModule_Create(&mpatch_module); |
|
194 | 193 | if (m == NULL) |
|
195 | 194 | return NULL; |
|
196 | 195 | |
|
197 | 196 | mpatch_Error = |
|
198 | 197 | PyErr_NewException("mercurial.cext.mpatch.mpatchError", NULL, NULL); |
|
199 | 198 | Py_INCREF(mpatch_Error); |
|
200 | 199 | PyModule_AddObject(m, "mpatchError", mpatch_Error); |
|
201 | 200 | PyModule_AddIntConstant(m, "version", version); |
|
202 | 201 | |
|
203 | 202 | return m; |
|
204 | 203 | } |
@@ -1,1391 +1,1389 b'' | |||
|
1 | 1 | /* |
|
2 | 2 | osutil.c - native operating system services |
|
3 | 3 | |
|
4 | 4 | Copyright 2007 Olivia Mackall and others |
|
5 | 5 | |
|
6 | 6 | This software may be used and distributed according to the terms of |
|
7 | 7 | the GNU General Public License, incorporated herein by reference. |
|
8 | 8 | */ |
|
9 | 9 | |
|
10 | 10 | #define _ATFILE_SOURCE |
|
11 | 11 | #define PY_SSIZE_T_CLEAN |
|
12 | 12 | #include <Python.h> |
|
13 | 13 | #include <errno.h> |
|
14 | 14 | #include <fcntl.h> |
|
15 | 15 | #include <stdio.h> |
|
16 | 16 | #include <stdlib.h> |
|
17 | 17 | #include <string.h> |
|
18 | 18 | |
|
19 | 19 | #ifdef _WIN32 |
|
20 | 20 | #include <io.h> |
|
21 | 21 | #include <windows.h> |
|
22 | 22 | #else |
|
23 | 23 | #include <dirent.h> |
|
24 | 24 | #include <signal.h> |
|
25 | 25 | #include <sys/socket.h> |
|
26 | 26 | #include <sys/stat.h> |
|
27 | 27 | #include <sys/types.h> |
|
28 | 28 | #include <unistd.h> |
|
29 | 29 | #ifdef HAVE_LINUX_STATFS |
|
30 | 30 | #include <linux/magic.h> |
|
31 | 31 | #include <sys/vfs.h> |
|
32 | 32 | #endif |
|
33 | 33 | #ifdef HAVE_BSD_STATFS |
|
34 | 34 | #include <sys/mount.h> |
|
35 | 35 | #include <sys/param.h> |
|
36 | 36 | #endif |
|
37 | 37 | #endif |
|
38 | 38 | |
|
39 | 39 | #ifdef __APPLE__ |
|
40 | 40 | #include <sys/attr.h> |
|
41 | 41 | #include <sys/vnode.h> |
|
42 | 42 | #endif |
|
43 | 43 | |
|
44 | 44 | #include "util.h" |
|
45 | 45 | |
|
46 | 46 | /* some platforms lack the PATH_MAX definition (eg. GNU/Hurd) */ |
|
47 | 47 | #ifndef PATH_MAX |
|
48 | 48 | #define PATH_MAX 4096 |
|
49 | 49 | #endif |
|
50 | 50 | |
|
51 | 51 | #ifdef _WIN32 |
|
52 | 52 | /* |
|
53 | 53 | stat struct compatible with hg expectations |
|
54 | 54 | Mercurial only uses st_mode, st_size and st_mtime |
|
55 | 55 | the rest is kept to minimize changes between implementations |
|
56 | 56 | */ |
|
57 | 57 | struct hg_stat { |
|
58 | 58 | int st_dev; |
|
59 | 59 | int st_mode; |
|
60 | 60 | int st_nlink; |
|
61 | 61 | __int64 st_size; |
|
62 | 62 | int st_mtime; |
|
63 | 63 | int st_ctime; |
|
64 | 64 | }; |
|
65 | 65 | struct listdir_stat { |
|
66 | 66 | PyObject_HEAD |
|
67 | 67 | struct hg_stat st; |
|
68 | 68 | }; |
|
69 | 69 | #else |
|
70 | 70 | struct listdir_stat { |
|
71 | 71 | PyObject_HEAD |
|
72 | 72 | struct stat st; |
|
73 | 73 | }; |
|
74 | 74 | #endif |
|
75 | 75 | |
|
76 | 76 | #define listdir_slot(name) \ |
|
77 | 77 | static PyObject *listdir_stat_##name(PyObject *self, void *x) \ |
|
78 | 78 | { \ |
|
79 | 79 | return PyLong_FromLong(((struct listdir_stat *)self)->st.name); \ |
|
80 | 80 | } |
|
81 | 81 | |
|
82 | 82 | listdir_slot(st_dev) |
|
83 | 83 | listdir_slot(st_mode) |
|
84 | 84 | listdir_slot(st_nlink) |
|
85 | 85 | #ifdef _WIN32 |
|
86 | 86 | static PyObject *listdir_stat_st_size(PyObject *self, void *x) |
|
87 | 87 | { |
|
88 | 88 | return PyLong_FromLongLong( |
|
89 | 89 | (PY_LONG_LONG)((struct listdir_stat *)self)->st.st_size); |
|
90 | 90 | } |
|
91 | 91 | #else |
|
92 | 92 | listdir_slot(st_size) |
|
93 | 93 | #endif |
|
94 | 94 | listdir_slot(st_mtime) |
|
95 | 95 | listdir_slot(st_ctime) |
|
96 | 96 | |
|
97 | 97 | static struct PyGetSetDef listdir_stat_getsets[] = { |
|
98 | 98 | {"st_dev", listdir_stat_st_dev, 0, 0, 0}, |
|
99 | 99 | {"st_mode", listdir_stat_st_mode, 0, 0, 0}, |
|
100 | 100 | {"st_nlink", listdir_stat_st_nlink, 0, 0, 0}, |
|
101 | 101 | {"st_size", listdir_stat_st_size, 0, 0, 0}, |
|
102 | 102 | {"st_mtime", listdir_stat_st_mtime, 0, 0, 0}, |
|
103 | 103 | {"st_ctime", listdir_stat_st_ctime, 0, 0, 0}, |
|
104 | 104 | {0, 0, 0, 0, 0} |
|
105 | 105 | }; |
|
106 | 106 | |
|
107 | 107 | static PyObject *listdir_stat_new(PyTypeObject *t, PyObject *a, PyObject *k) |
|
108 | 108 | { |
|
109 | 109 | return t->tp_alloc(t, 0); |
|
110 | 110 | } |
|
111 | 111 | |
|
112 | 112 | static void listdir_stat_dealloc(PyObject *o) |
|
113 | 113 | { |
|
114 | 114 | Py_TYPE(o)->tp_free(o); |
|
115 | 115 | } |
|
116 | 116 | |
|
117 | 117 | static PyObject *listdir_stat_getitem(PyObject *self, PyObject *key) |
|
118 | 118 | { |
|
119 | 119 | long index = PyLong_AsLong(key); |
|
120 | 120 | if (index == -1 && PyErr_Occurred()) { |
|
121 | 121 | return NULL; |
|
122 | 122 | } |
|
123 | 123 | if (index != 8) { |
|
124 | 124 | PyErr_Format(PyExc_IndexError, "osutil.stat objects only " |
|
125 | 125 | "support stat.ST_MTIME in " |
|
126 | 126 | "__getitem__"); |
|
127 | 127 | return NULL; |
|
128 | 128 | } |
|
129 | 129 | return listdir_stat_st_mtime(self, NULL); |
|
130 | 130 | } |
|
131 | 131 | |
|
132 | 132 | static PyMappingMethods listdir_stat_type_mapping_methods = { |
|
133 | 133 | (lenfunc)NULL, /* mp_length */ |
|
134 | 134 | (binaryfunc)listdir_stat_getitem, /* mp_subscript */ |
|
135 | 135 | (objobjargproc)NULL, /* mp_ass_subscript */ |
|
136 | 136 | }; |
|
137 | 137 | |
|
138 | 138 | static PyTypeObject listdir_stat_type = { |
|
139 | 139 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ |
|
140 | 140 | "osutil.stat", /*tp_name*/ |
|
141 | 141 | sizeof(struct listdir_stat), /*tp_basicsize*/ |
|
142 | 142 | 0, /*tp_itemsize*/ |
|
143 | 143 | (destructor)listdir_stat_dealloc, /*tp_dealloc*/ |
|
144 | 144 | 0, /*tp_print*/ |
|
145 | 145 | 0, /*tp_getattr*/ |
|
146 | 146 | 0, /*tp_setattr*/ |
|
147 | 147 | 0, /*tp_compare*/ |
|
148 | 148 | 0, /*tp_repr*/ |
|
149 | 149 | 0, /*tp_as_number*/ |
|
150 | 150 | 0, /*tp_as_sequence*/ |
|
151 | 151 | &listdir_stat_type_mapping_methods, /*tp_as_mapping*/ |
|
152 | 152 | 0, /*tp_hash */ |
|
153 | 153 | 0, /*tp_call*/ |
|
154 | 154 | 0, /*tp_str*/ |
|
155 | 155 | 0, /*tp_getattro*/ |
|
156 | 156 | 0, /*tp_setattro*/ |
|
157 | 157 | 0, /*tp_as_buffer*/ |
|
158 | 158 | Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE, /*tp_flags*/ |
|
159 | 159 | "stat objects", /* tp_doc */ |
|
160 | 160 | 0, /* tp_traverse */ |
|
161 | 161 | 0, /* tp_clear */ |
|
162 | 162 | 0, /* tp_richcompare */ |
|
163 | 163 | 0, /* tp_weaklistoffset */ |
|
164 | 164 | 0, /* tp_iter */ |
|
165 | 165 | 0, /* tp_iternext */ |
|
166 | 166 | 0, /* tp_methods */ |
|
167 | 167 | 0, /* tp_members */ |
|
168 | 168 | listdir_stat_getsets, /* tp_getset */ |
|
169 | 169 | 0, /* tp_base */ |
|
170 | 170 | 0, /* tp_dict */ |
|
171 | 171 | 0, /* tp_descr_get */ |
|
172 | 172 | 0, /* tp_descr_set */ |
|
173 | 173 | 0, /* tp_dictoffset */ |
|
174 | 174 | 0, /* tp_init */ |
|
175 | 175 | 0, /* tp_alloc */ |
|
176 | 176 | listdir_stat_new, /* tp_new */ |
|
177 | 177 | }; |
|
178 | 178 | |
|
179 | 179 | #ifdef _WIN32 |
|
180 | 180 | |
|
181 | 181 | static int to_python_time(const FILETIME *tm) |
|
182 | 182 | { |
|
183 | 183 | /* number of seconds between epoch and January 1 1601 */ |
|
184 | 184 | const __int64 a0 = (__int64)134774L * (__int64)24L * (__int64)3600L; |
|
185 | 185 | /* conversion factor from 100ns to 1s */ |
|
186 | 186 | const __int64 a1 = 10000000; |
|
187 | 187 | /* explicit (int) cast to suspend compiler warnings */ |
|
188 | 188 | return (int)((((__int64)tm->dwHighDateTime << 32) |
|
189 | 189 | + tm->dwLowDateTime) / a1 - a0); |
|
190 | 190 | } |
|
191 | 191 | |
|
192 | 192 | static PyObject *make_item(const WIN32_FIND_DATAA *fd, int wantstat) |
|
193 | 193 | { |
|
194 | 194 | PyObject *py_st; |
|
195 | 195 | struct hg_stat *stp; |
|
196 | 196 | |
|
197 | 197 | int kind = (fd->dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) |
|
198 | 198 | ? _S_IFDIR : _S_IFREG; |
|
199 | 199 | |
|
200 | 200 | if (!wantstat) |
|
201 |
return Py_BuildValue( |
|
|
201 | return Py_BuildValue("yi", fd->cFileName, kind); | |
|
202 | 202 | |
|
203 | 203 | py_st = PyObject_CallObject((PyObject *)&listdir_stat_type, NULL); |
|
204 | 204 | if (!py_st) |
|
205 | 205 | return NULL; |
|
206 | 206 | |
|
207 | 207 | stp = &((struct listdir_stat *)py_st)->st; |
|
208 | 208 | /* |
|
209 | 209 | use kind as st_mode |
|
210 | 210 | rwx bits on Win32 are meaningless |
|
211 | 211 | and Hg does not use them anyway |
|
212 | 212 | */ |
|
213 | 213 | stp->st_mode = kind; |
|
214 | 214 | stp->st_mtime = to_python_time(&fd->ftLastWriteTime); |
|
215 | 215 | stp->st_ctime = to_python_time(&fd->ftCreationTime); |
|
216 | 216 | if (kind == _S_IFREG) |
|
217 | 217 | stp->st_size = ((__int64)fd->nFileSizeHigh << 32) |
|
218 | 218 | + fd->nFileSizeLow; |
|
219 |
return Py_BuildValue( |
|
|
219 | return Py_BuildValue("yiN", fd->cFileName, | |
|
220 | 220 | kind, py_st); |
|
221 | 221 | } |
|
222 | 222 | |
|
223 | 223 | static PyObject *_listdir(char *path, Py_ssize_t plen, int wantstat, char *skip) |
|
224 | 224 | { |
|
225 | 225 | PyObject *rval = NULL; /* initialize - return value */ |
|
226 | 226 | PyObject *list; |
|
227 | 227 | HANDLE fh; |
|
228 | 228 | WIN32_FIND_DATAA fd; |
|
229 | 229 | char *pattern; |
|
230 | 230 | |
|
231 | 231 | /* build the path + \* pattern string */ |
|
232 | 232 | pattern = PyMem_Malloc(plen + 3); /* path + \* + \0 */ |
|
233 | 233 | if (!pattern) { |
|
234 | 234 | PyErr_NoMemory(); |
|
235 | 235 | goto error_nomem; |
|
236 | 236 | } |
|
237 | 237 | memcpy(pattern, path, plen); |
|
238 | 238 | |
|
239 | 239 | if (plen > 0) { |
|
240 | 240 | char c = path[plen-1]; |
|
241 | 241 | if (c != ':' && c != '/' && c != '\\') |
|
242 | 242 | pattern[plen++] = '\\'; |
|
243 | 243 | } |
|
244 | 244 | pattern[plen++] = '*'; |
|
245 | 245 | pattern[plen] = '\0'; |
|
246 | 246 | |
|
247 | 247 | fh = FindFirstFileA(pattern, &fd); |
|
248 | 248 | if (fh == INVALID_HANDLE_VALUE) { |
|
249 | 249 | PyErr_SetFromWindowsErrWithFilename(GetLastError(), path); |
|
250 | 250 | goto error_file; |
|
251 | 251 | } |
|
252 | 252 | |
|
253 | 253 | list = PyList_New(0); |
|
254 | 254 | if (!list) |
|
255 | 255 | goto error_list; |
|
256 | 256 | |
|
257 | 257 | do { |
|
258 | 258 | PyObject *item; |
|
259 | 259 | |
|
260 | 260 | if (fd.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) { |
|
261 | 261 | if (!strcmp(fd.cFileName, ".") |
|
262 | 262 | || !strcmp(fd.cFileName, "..")) |
|
263 | 263 | continue; |
|
264 | 264 | |
|
265 | 265 | if (skip && !strcmp(fd.cFileName, skip)) { |
|
266 | 266 | rval = PyList_New(0); |
|
267 | 267 | goto error; |
|
268 | 268 | } |
|
269 | 269 | } |
|
270 | 270 | |
|
271 | 271 | item = make_item(&fd, wantstat); |
|
272 | 272 | if (!item) |
|
273 | 273 | goto error; |
|
274 | 274 | |
|
275 | 275 | if (PyList_Append(list, item)) { |
|
276 | 276 | Py_XDECREF(item); |
|
277 | 277 | goto error; |
|
278 | 278 | } |
|
279 | 279 | |
|
280 | 280 | Py_XDECREF(item); |
|
281 | 281 | } while (FindNextFileA(fh, &fd)); |
|
282 | 282 | |
|
283 | 283 | if (GetLastError() != ERROR_NO_MORE_FILES) { |
|
284 | 284 | PyErr_SetFromWindowsErrWithFilename(GetLastError(), path); |
|
285 | 285 | goto error; |
|
286 | 286 | } |
|
287 | 287 | |
|
288 | 288 | rval = list; |
|
289 | 289 | Py_XINCREF(rval); |
|
290 | 290 | error: |
|
291 | 291 | Py_XDECREF(list); |
|
292 | 292 | error_list: |
|
293 | 293 | FindClose(fh); |
|
294 | 294 | error_file: |
|
295 | 295 | PyMem_Free(pattern); |
|
296 | 296 | error_nomem: |
|
297 | 297 | return rval; |
|
298 | 298 | } |
|
299 | 299 | |
|
300 | 300 | #else |
|
301 | 301 | |
|
302 | 302 | int entkind(struct dirent *ent) |
|
303 | 303 | { |
|
304 | 304 | #ifdef DT_REG |
|
305 | 305 | switch (ent->d_type) { |
|
306 | 306 | case DT_REG: return S_IFREG; |
|
307 | 307 | case DT_DIR: return S_IFDIR; |
|
308 | 308 | case DT_LNK: return S_IFLNK; |
|
309 | 309 | case DT_BLK: return S_IFBLK; |
|
310 | 310 | case DT_CHR: return S_IFCHR; |
|
311 | 311 | case DT_FIFO: return S_IFIFO; |
|
312 | 312 | case DT_SOCK: return S_IFSOCK; |
|
313 | 313 | } |
|
314 | 314 | #endif |
|
315 | 315 | return -1; |
|
316 | 316 | } |
|
317 | 317 | |
|
318 | 318 | static PyObject *makestat(const struct stat *st) |
|
319 | 319 | { |
|
320 | 320 | PyObject *stat; |
|
321 | 321 | |
|
322 | 322 | stat = PyObject_CallObject((PyObject *)&listdir_stat_type, NULL); |
|
323 | 323 | if (stat) |
|
324 | 324 | memcpy(&((struct listdir_stat *)stat)->st, st, sizeof(*st)); |
|
325 | 325 | return stat; |
|
326 | 326 | } |
|
327 | 327 | |
|
328 | 328 | static PyObject *_listdir_stat(char *path, int pathlen, int keepstat, |
|
329 | 329 | char *skip) |
|
330 | 330 | { |
|
331 | 331 | PyObject *list, *elem, *ret = NULL; |
|
332 | 332 | char fullpath[PATH_MAX + 10]; |
|
333 | 333 | int kind, err; |
|
334 | 334 | struct stat st; |
|
335 | 335 | struct dirent *ent; |
|
336 | 336 | DIR *dir; |
|
337 | 337 | #ifdef AT_SYMLINK_NOFOLLOW |
|
338 | 338 | int dfd = -1; |
|
339 | 339 | #endif |
|
340 | 340 | |
|
341 | 341 | if (pathlen >= PATH_MAX) { |
|
342 | 342 | errno = ENAMETOOLONG; |
|
343 | 343 | PyErr_SetFromErrnoWithFilename(PyExc_OSError, path); |
|
344 | 344 | goto error_value; |
|
345 | 345 | } |
|
346 | 346 | strncpy(fullpath, path, PATH_MAX); |
|
347 | 347 | fullpath[pathlen] = '/'; |
|
348 | 348 | |
|
349 | 349 | #ifdef AT_SYMLINK_NOFOLLOW |
|
350 | 350 | dfd = open(path, O_RDONLY); |
|
351 | 351 | if (dfd == -1) { |
|
352 | 352 | PyErr_SetFromErrnoWithFilename(PyExc_OSError, path); |
|
353 | 353 | goto error_value; |
|
354 | 354 | } |
|
355 | 355 | dir = fdopendir(dfd); |
|
356 | 356 | #else |
|
357 | 357 | dir = opendir(path); |
|
358 | 358 | #endif |
|
359 | 359 | if (!dir) { |
|
360 | 360 | PyErr_SetFromErrnoWithFilename(PyExc_OSError, path); |
|
361 | 361 | goto error_dir; |
|
362 | 362 | } |
|
363 | 363 | |
|
364 | 364 | list = PyList_New(0); |
|
365 | 365 | if (!list) |
|
366 | 366 | goto error_list; |
|
367 | 367 | |
|
368 | 368 | while ((ent = readdir(dir))) { |
|
369 | 369 | if (!strcmp(ent->d_name, ".") || !strcmp(ent->d_name, "..")) |
|
370 | 370 | continue; |
|
371 | 371 | |
|
372 | 372 | kind = entkind(ent); |
|
373 | 373 | if (kind == -1 || keepstat) { |
|
374 | 374 | #ifdef AT_SYMLINK_NOFOLLOW |
|
375 | 375 | err = fstatat(dfd, ent->d_name, &st, |
|
376 | 376 | AT_SYMLINK_NOFOLLOW); |
|
377 | 377 | #else |
|
378 | 378 | strncpy(fullpath + pathlen + 1, ent->d_name, |
|
379 | 379 | PATH_MAX - pathlen); |
|
380 | 380 | fullpath[PATH_MAX] = '\0'; |
|
381 | 381 | err = lstat(fullpath, &st); |
|
382 | 382 | #endif |
|
383 | 383 | if (err == -1) { |
|
384 | 384 | /* race with file deletion? */ |
|
385 | 385 | if (errno == ENOENT) |
|
386 | 386 | continue; |
|
387 | 387 | strncpy(fullpath + pathlen + 1, ent->d_name, |
|
388 | 388 | PATH_MAX - pathlen); |
|
389 | 389 | fullpath[PATH_MAX] = 0; |
|
390 | 390 | PyErr_SetFromErrnoWithFilename(PyExc_OSError, |
|
391 | 391 | fullpath); |
|
392 | 392 | goto error; |
|
393 | 393 | } |
|
394 | 394 | kind = st.st_mode & S_IFMT; |
|
395 | 395 | } |
|
396 | 396 | |
|
397 | 397 | /* quit early? */ |
|
398 | 398 | if (skip && kind == S_IFDIR && !strcmp(ent->d_name, skip)) { |
|
399 | 399 | ret = PyList_New(0); |
|
400 | 400 | goto error; |
|
401 | 401 | } |
|
402 | 402 | |
|
403 | 403 | if (keepstat) { |
|
404 | 404 | PyObject *stat = makestat(&st); |
|
405 | 405 | if (!stat) |
|
406 | 406 | goto error; |
|
407 |
elem = Py_BuildValue( |
|
|
407 | elem = Py_BuildValue("yiN", ent->d_name, | |
|
408 | 408 | kind, stat); |
|
409 | 409 | } else |
|
410 |
elem = Py_BuildValue( |
|
|
410 | elem = Py_BuildValue("yi", ent->d_name, | |
|
411 | 411 | kind); |
|
412 | 412 | if (!elem) |
|
413 | 413 | goto error; |
|
414 | 414 | |
|
415 | 415 | PyList_Append(list, elem); |
|
416 | 416 | Py_DECREF(elem); |
|
417 | 417 | } |
|
418 | 418 | |
|
419 | 419 | ret = list; |
|
420 | 420 | Py_INCREF(ret); |
|
421 | 421 | |
|
422 | 422 | error: |
|
423 | 423 | Py_DECREF(list); |
|
424 | 424 | error_list: |
|
425 | 425 | closedir(dir); |
|
426 | 426 | /* closedir also closes its dirfd */ |
|
427 | 427 | goto error_value; |
|
428 | 428 | error_dir: |
|
429 | 429 | #ifdef AT_SYMLINK_NOFOLLOW |
|
430 | 430 | close(dfd); |
|
431 | 431 | #endif |
|
432 | 432 | error_value: |
|
433 | 433 | return ret; |
|
434 | 434 | } |
|
435 | 435 | |
|
436 | 436 | #ifdef __APPLE__ |
|
437 | 437 | |
|
438 | 438 | typedef struct { |
|
439 | 439 | u_int32_t length; |
|
440 | 440 | attrreference_t name; |
|
441 | 441 | fsobj_type_t obj_type; |
|
442 | 442 | struct timespec mtime; |
|
443 | 443 | #if __LITTLE_ENDIAN__ |
|
444 | 444 | mode_t access_mask; |
|
445 | 445 | uint16_t padding; |
|
446 | 446 | #else |
|
447 | 447 | uint16_t padding; |
|
448 | 448 | mode_t access_mask; |
|
449 | 449 | #endif |
|
450 | 450 | off_t size; |
|
451 | 451 | } __attribute__((packed)) attrbuf_entry; |
|
452 | 452 | |
|
453 | 453 | int attrkind(attrbuf_entry *entry) |
|
454 | 454 | { |
|
455 | 455 | switch (entry->obj_type) { |
|
456 | 456 | case VREG: return S_IFREG; |
|
457 | 457 | case VDIR: return S_IFDIR; |
|
458 | 458 | case VLNK: return S_IFLNK; |
|
459 | 459 | case VBLK: return S_IFBLK; |
|
460 | 460 | case VCHR: return S_IFCHR; |
|
461 | 461 | case VFIFO: return S_IFIFO; |
|
462 | 462 | case VSOCK: return S_IFSOCK; |
|
463 | 463 | } |
|
464 | 464 | return -1; |
|
465 | 465 | } |
|
466 | 466 | |
|
467 | 467 | /* get these many entries at a time */ |
|
468 | 468 | #define LISTDIR_BATCH_SIZE 50 |
|
469 | 469 | |
|
470 | 470 | static PyObject *_listdir_batch(char *path, int pathlen, int keepstat, |
|
471 | 471 | char *skip, bool *fallback) |
|
472 | 472 | { |
|
473 | 473 | PyObject *list, *elem, *ret = NULL; |
|
474 | 474 | int kind, err; |
|
475 | 475 | unsigned long index; |
|
476 | 476 | unsigned int count, old_state, new_state; |
|
477 | 477 | bool state_seen = false; |
|
478 | 478 | attrbuf_entry *entry; |
|
479 | 479 | /* from the getattrlist(2) man page: a path can be no longer than |
|
480 | 480 | (NAME_MAX * 3 + 1) bytes. Also, "The getattrlist() function will |
|
481 | 481 | silently truncate attribute data if attrBufSize is too small." So |
|
482 | 482 | pass in a buffer big enough for the worst case. */ |
|
483 | 483 | char attrbuf[LISTDIR_BATCH_SIZE * (sizeof(attrbuf_entry) + NAME_MAX * 3 + 1)]; |
|
484 | 484 | unsigned int basep_unused; |
|
485 | 485 | |
|
486 | 486 | struct stat st; |
|
487 | 487 | int dfd = -1; |
|
488 | 488 | |
|
489 | 489 | /* these must match the attrbuf_entry struct, otherwise you'll end up |
|
490 | 490 | with garbage */ |
|
491 | 491 | struct attrlist requested_attr = {0}; |
|
492 | 492 | requested_attr.bitmapcount = ATTR_BIT_MAP_COUNT; |
|
493 | 493 | requested_attr.commonattr = (ATTR_CMN_NAME | ATTR_CMN_OBJTYPE | |
|
494 | 494 | ATTR_CMN_MODTIME | ATTR_CMN_ACCESSMASK); |
|
495 | 495 | requested_attr.fileattr = ATTR_FILE_DATALENGTH; |
|
496 | 496 | |
|
497 | 497 | *fallback = false; |
|
498 | 498 | |
|
499 | 499 | if (pathlen >= PATH_MAX) { |
|
500 | 500 | errno = ENAMETOOLONG; |
|
501 | 501 | PyErr_SetFromErrnoWithFilename(PyExc_OSError, path); |
|
502 | 502 | goto error_value; |
|
503 | 503 | } |
|
504 | 504 | |
|
505 | 505 | dfd = open(path, O_RDONLY); |
|
506 | 506 | if (dfd == -1) { |
|
507 | 507 | PyErr_SetFromErrnoWithFilename(PyExc_OSError, path); |
|
508 | 508 | goto error_value; |
|
509 | 509 | } |
|
510 | 510 | |
|
511 | 511 | list = PyList_New(0); |
|
512 | 512 | if (!list) |
|
513 | 513 | goto error_dir; |
|
514 | 514 | |
|
515 | 515 | do { |
|
516 | 516 | count = LISTDIR_BATCH_SIZE; |
|
517 | 517 | err = getdirentriesattr(dfd, &requested_attr, &attrbuf, |
|
518 | 518 | sizeof(attrbuf), &count, &basep_unused, |
|
519 | 519 | &new_state, 0); |
|
520 | 520 | if (err < 0) { |
|
521 | 521 | if (errno == ENOTSUP) { |
|
522 | 522 | /* We're on a filesystem that doesn't support |
|
523 | 523 | getdirentriesattr. Fall back to the |
|
524 | 524 | stat-based implementation. */ |
|
525 | 525 | *fallback = true; |
|
526 | 526 | } else |
|
527 | 527 | PyErr_SetFromErrnoWithFilename(PyExc_OSError, path); |
|
528 | 528 | goto error; |
|
529 | 529 | } |
|
530 | 530 | |
|
531 | 531 | if (!state_seen) { |
|
532 | 532 | old_state = new_state; |
|
533 | 533 | state_seen = true; |
|
534 | 534 | } else if (old_state != new_state) { |
|
535 | 535 | /* There's an edge case with getdirentriesattr. Consider |
|
536 | 536 | the following initial list of files: |
|
537 | 537 | |
|
538 | 538 | a |
|
539 | 539 | b |
|
540 | 540 | <-- |
|
541 | 541 | c |
|
542 | 542 | d |
|
543 | 543 | |
|
544 | 544 | If the iteration is paused at the arrow, and b is |
|
545 | 545 | deleted before it is resumed, getdirentriesattr will |
|
546 | 546 | not return d at all! Ordinarily we're expected to |
|
547 | 547 | restart the iteration from the beginning. To avoid |
|
548 | 548 | getting stuck in a retry loop here, fall back to |
|
549 | 549 | stat. */ |
|
550 | 550 | *fallback = true; |
|
551 | 551 | goto error; |
|
552 | 552 | } |
|
553 | 553 | |
|
554 | 554 | entry = (attrbuf_entry *)attrbuf; |
|
555 | 555 | |
|
556 | 556 | for (index = 0; index < count; index++) { |
|
557 | 557 | char *filename = ((char *)&entry->name) + |
|
558 | 558 | entry->name.attr_dataoffset; |
|
559 | 559 | |
|
560 | 560 | if (!strcmp(filename, ".") || !strcmp(filename, "..")) |
|
561 | 561 | continue; |
|
562 | 562 | |
|
563 | 563 | kind = attrkind(entry); |
|
564 | 564 | if (kind == -1) { |
|
565 | 565 | PyErr_Format(PyExc_OSError, |
|
566 | 566 | "unknown object type %u for file " |
|
567 | 567 | "%s%s!", |
|
568 | 568 | entry->obj_type, path, filename); |
|
569 | 569 | goto error; |
|
570 | 570 | } |
|
571 | 571 | |
|
572 | 572 | /* quit early? */ |
|
573 | 573 | if (skip && kind == S_IFDIR && !strcmp(filename, skip)) { |
|
574 | 574 | ret = PyList_New(0); |
|
575 | 575 | goto error; |
|
576 | 576 | } |
|
577 | 577 | |
|
578 | 578 | if (keepstat) { |
|
579 | 579 | PyObject *stat = NULL; |
|
580 | 580 | /* from the getattrlist(2) man page: "Only the |
|
581 | 581 | permission bits ... are valid". */ |
|
582 | 582 | st.st_mode = (entry->access_mask & ~S_IFMT) | kind; |
|
583 | 583 | st.st_mtime = entry->mtime.tv_sec; |
|
584 | 584 | st.st_size = entry->size; |
|
585 | 585 | stat = makestat(&st); |
|
586 | 586 | if (!stat) |
|
587 | 587 | goto error; |
|
588 |
elem = Py_BuildValue( |
|
|
588 | elem = Py_BuildValue("yiN", | |
|
589 | 589 | filename, kind, stat); |
|
590 | 590 | } else |
|
591 |
elem = Py_BuildValue( |
|
|
591 | elem = Py_BuildValue("yi", | |
|
592 | 592 | filename, kind); |
|
593 | 593 | if (!elem) |
|
594 | 594 | goto error; |
|
595 | 595 | |
|
596 | 596 | PyList_Append(list, elem); |
|
597 | 597 | Py_DECREF(elem); |
|
598 | 598 | |
|
599 | 599 | entry = (attrbuf_entry *)((char *)entry + entry->length); |
|
600 | 600 | } |
|
601 | 601 | } while (err == 0); |
|
602 | 602 | |
|
603 | 603 | ret = list; |
|
604 | 604 | Py_INCREF(ret); |
|
605 | 605 | |
|
606 | 606 | error: |
|
607 | 607 | Py_DECREF(list); |
|
608 | 608 | error_dir: |
|
609 | 609 | close(dfd); |
|
610 | 610 | error_value: |
|
611 | 611 | return ret; |
|
612 | 612 | } |
|
613 | 613 | |
|
614 | 614 | #endif /* __APPLE__ */ |
|
615 | 615 | |
|
616 | 616 | static PyObject *_listdir(char *path, int pathlen, int keepstat, char *skip) |
|
617 | 617 | { |
|
618 | 618 | #ifdef __APPLE__ |
|
619 | 619 | PyObject *ret; |
|
620 | 620 | bool fallback = false; |
|
621 | 621 | |
|
622 | 622 | ret = _listdir_batch(path, pathlen, keepstat, skip, &fallback); |
|
623 | 623 | if (ret != NULL || !fallback) |
|
624 | 624 | return ret; |
|
625 | 625 | #endif |
|
626 | 626 | return _listdir_stat(path, pathlen, keepstat, skip); |
|
627 | 627 | } |
|
628 | 628 | |
|
629 | 629 | static PyObject *statfiles(PyObject *self, PyObject *args) |
|
630 | 630 | { |
|
631 | 631 | PyObject *names, *stats; |
|
632 | 632 | Py_ssize_t i, count; |
|
633 | 633 | |
|
634 | 634 | if (!PyArg_ParseTuple(args, "O:statfiles", &names)) |
|
635 | 635 | return NULL; |
|
636 | 636 | |
|
637 | 637 | count = PySequence_Length(names); |
|
638 | 638 | if (count == -1) { |
|
639 | 639 | PyErr_SetString(PyExc_TypeError, "not a sequence"); |
|
640 | 640 | return NULL; |
|
641 | 641 | } |
|
642 | 642 | |
|
643 | 643 | stats = PyList_New(count); |
|
644 | 644 | if (stats == NULL) |
|
645 | 645 | return NULL; |
|
646 | 646 | |
|
647 | 647 | for (i = 0; i < count; i++) { |
|
648 | 648 | PyObject *stat, *pypath; |
|
649 | 649 | struct stat st; |
|
650 | 650 | int ret, kind; |
|
651 | 651 | char *path; |
|
652 | 652 | |
|
653 | 653 | /* With a large file count or on a slow filesystem, |
|
654 | 654 | don't block signals for long (issue4878). */ |
|
655 | 655 | if ((i % 1000) == 999 && PyErr_CheckSignals() == -1) |
|
656 | 656 | goto bail; |
|
657 | 657 | |
|
658 | 658 | pypath = PySequence_GetItem(names, i); |
|
659 | 659 | if (!pypath) |
|
660 | 660 | goto bail; |
|
661 | 661 | path = PyBytes_AsString(pypath); |
|
662 | 662 | if (path == NULL) { |
|
663 | 663 | Py_DECREF(pypath); |
|
664 | 664 | PyErr_SetString(PyExc_TypeError, "not a string"); |
|
665 | 665 | goto bail; |
|
666 | 666 | } |
|
667 | 667 | ret = lstat(path, &st); |
|
668 | 668 | Py_DECREF(pypath); |
|
669 | 669 | kind = st.st_mode & S_IFMT; |
|
670 | 670 | if (ret != -1 && (kind == S_IFREG || kind == S_IFLNK)) { |
|
671 | 671 | stat = makestat(&st); |
|
672 | 672 | if (stat == NULL) |
|
673 | 673 | goto bail; |
|
674 | 674 | PyList_SET_ITEM(stats, i, stat); |
|
675 | 675 | } else { |
|
676 | 676 | Py_INCREF(Py_None); |
|
677 | 677 | PyList_SET_ITEM(stats, i, Py_None); |
|
678 | 678 | } |
|
679 | 679 | } |
|
680 | 680 | |
|
681 | 681 | return stats; |
|
682 | 682 | |
|
683 | 683 | bail: |
|
684 | 684 | Py_DECREF(stats); |
|
685 | 685 | return NULL; |
|
686 | 686 | } |
|
687 | 687 | |
|
688 | 688 | /* |
|
689 | 689 | * recvfds() simply does not release GIL during blocking io operation because |
|
690 | 690 | * command server is known to be single-threaded. |
|
691 | 691 | * |
|
692 | 692 | * Old systems such as Solaris don't provide CMSG_LEN, msg_control, etc. |
|
693 | 693 | * Currently, recvfds() is not supported on these platforms. |
|
694 | 694 | */ |
|
695 | 695 | #ifdef CMSG_LEN |
|
696 | 696 | |
|
697 | 697 | static ssize_t recvfdstobuf(int sockfd, int **rfds, void *cbuf, size_t cbufsize) |
|
698 | 698 | { |
|
699 | 699 | char dummy[1]; |
|
700 | 700 | struct iovec iov = {dummy, sizeof(dummy)}; |
|
701 | 701 | struct msghdr msgh = {0}; |
|
702 | 702 | struct cmsghdr *cmsg; |
|
703 | 703 | |
|
704 | 704 | msgh.msg_iov = &iov; |
|
705 | 705 | msgh.msg_iovlen = 1; |
|
706 | 706 | msgh.msg_control = cbuf; |
|
707 | 707 | msgh.msg_controllen = (socklen_t)cbufsize; |
|
708 | 708 | if (recvmsg(sockfd, &msgh, 0) < 0) |
|
709 | 709 | return -1; |
|
710 | 710 | |
|
711 | 711 | for (cmsg = CMSG_FIRSTHDR(&msgh); cmsg; |
|
712 | 712 | cmsg = CMSG_NXTHDR(&msgh, cmsg)) { |
|
713 | 713 | if (cmsg->cmsg_level != SOL_SOCKET || |
|
714 | 714 | cmsg->cmsg_type != SCM_RIGHTS) |
|
715 | 715 | continue; |
|
716 | 716 | *rfds = (int *)CMSG_DATA(cmsg); |
|
717 | 717 | return (cmsg->cmsg_len - CMSG_LEN(0)) / sizeof(int); |
|
718 | 718 | } |
|
719 | 719 | |
|
720 | 720 | *rfds = cbuf; |
|
721 | 721 | return 0; |
|
722 | 722 | } |
|
723 | 723 | |
|
724 | 724 | static PyObject *recvfds(PyObject *self, PyObject *args) |
|
725 | 725 | { |
|
726 | 726 | int sockfd; |
|
727 | 727 | int *rfds = NULL; |
|
728 | 728 | ssize_t rfdscount, i; |
|
729 | 729 | char cbuf[256]; |
|
730 | 730 | PyObject *rfdslist = NULL; |
|
731 | 731 | |
|
732 | 732 | if (!PyArg_ParseTuple(args, "i", &sockfd)) |
|
733 | 733 | return NULL; |
|
734 | 734 | |
|
735 | 735 | rfdscount = recvfdstobuf(sockfd, &rfds, cbuf, sizeof(cbuf)); |
|
736 | 736 | if (rfdscount < 0) |
|
737 | 737 | return PyErr_SetFromErrno(PyExc_OSError); |
|
738 | 738 | |
|
739 | 739 | rfdslist = PyList_New(rfdscount); |
|
740 | 740 | if (!rfdslist) |
|
741 | 741 | goto bail; |
|
742 | 742 | for (i = 0; i < rfdscount; i++) { |
|
743 | 743 | PyObject *obj = PyLong_FromLong(rfds[i]); |
|
744 | 744 | if (!obj) |
|
745 | 745 | goto bail; |
|
746 | 746 | PyList_SET_ITEM(rfdslist, i, obj); |
|
747 | 747 | } |
|
748 | 748 | return rfdslist; |
|
749 | 749 | |
|
750 | 750 | bail: |
|
751 | 751 | Py_XDECREF(rfdslist); |
|
752 | 752 | return NULL; |
|
753 | 753 | } |
|
754 | 754 | |
|
755 | 755 | #endif /* CMSG_LEN */ |
|
756 | 756 | |
|
757 | 757 | /* allow disabling setprocname via compiler flags */ |
|
758 | 758 | #ifndef SETPROCNAME_USE_NONE |
|
759 | 759 | #if defined(HAVE_SETPROCTITLE) |
|
760 | 760 | /* setproctitle is the first choice - available in FreeBSD */ |
|
761 | 761 | #define SETPROCNAME_USE_SETPROCTITLE |
|
762 | 762 | #elif (defined(__linux__) || defined(__APPLE__)) && PY_MAJOR_VERSION == 2 |
|
763 | 763 | /* rewrite the argv buffer in place - works in Linux and OS X. Py_GetArgcArgv |
|
764 | 764 | * in Python 3 returns the copied wchar_t **argv, thus unsupported. */ |
|
765 | 765 | #define SETPROCNAME_USE_ARGVREWRITE |
|
766 | 766 | #else |
|
767 | 767 | #define SETPROCNAME_USE_NONE |
|
768 | 768 | #endif |
|
769 | 769 | #endif /* ndef SETPROCNAME_USE_NONE */ |
|
770 | 770 | |
|
771 | 771 | #ifndef SETPROCNAME_USE_NONE |
|
772 | 772 | static PyObject *setprocname(PyObject *self, PyObject *args) |
|
773 | 773 | { |
|
774 | 774 | const char *name = NULL; |
|
775 |
if (!PyArg_ParseTuple(args, |
|
|
775 | if (!PyArg_ParseTuple(args, "y", &name)) | |
|
776 | 776 | return NULL; |
|
777 | 777 | |
|
778 | 778 | #if defined(SETPROCNAME_USE_SETPROCTITLE) |
|
779 | 779 | setproctitle("%s", name); |
|
780 | 780 | #elif defined(SETPROCNAME_USE_ARGVREWRITE) |
|
781 | 781 | { |
|
782 | 782 | static char *argvstart = NULL; |
|
783 | 783 | static size_t argvsize = 0; |
|
784 | 784 | if (argvstart == NULL) { |
|
785 | 785 | int argc = 0, i; |
|
786 | 786 | char **argv = NULL; |
|
787 | 787 | char *argvend; |
|
788 | 788 | extern void Py_GetArgcArgv(int *argc, char ***argv); |
|
789 | 789 | Py_GetArgcArgv(&argc, &argv); |
|
790 | 790 | /* Py_GetArgcArgv may not do much if a custom python |
|
791 | 791 | * launcher is used that doesn't record the information |
|
792 | 792 | * it needs. Let's handle this gracefully instead of |
|
793 | 793 | * segfaulting. */ |
|
794 | 794 | if (argv != NULL) |
|
795 | 795 | argvend = argvstart = argv[0]; |
|
796 | 796 | else |
|
797 | 797 | argvend = argvstart = NULL; |
|
798 | 798 | |
|
799 | 799 | /* Check the memory we can use. Typically, argv[i] and |
|
800 | 800 | * argv[i + 1] are continuous. */ |
|
801 | 801 | for (i = 0; i < argc; ++i) { |
|
802 | 802 | size_t len; |
|
803 | 803 | if (argv[i] > argvend || argv[i] < argvstart) |
|
804 | 804 | break; /* not continuous */ |
|
805 | 805 | len = strlen(argv[i]); |
|
806 | 806 | argvend = argv[i] + len + 1 /* '\0' */; |
|
807 | 807 | } |
|
808 | 808 | if (argvend > argvstart) /* sanity check */ |
|
809 | 809 | argvsize = argvend - argvstart; |
|
810 | 810 | } |
|
811 | 811 | |
|
812 | 812 | if (argvstart && argvsize > 1) { |
|
813 | 813 | int n = snprintf(argvstart, argvsize, "%s", name); |
|
814 | 814 | if (n >= 0 && (size_t)n < argvsize) |
|
815 | 815 | memset(argvstart + n, 0, argvsize - n); |
|
816 | 816 | } |
|
817 | 817 | } |
|
818 | 818 | #endif |
|
819 | 819 | |
|
820 | 820 | Py_RETURN_NONE; |
|
821 | 821 | } |
|
822 | 822 | #endif /* ndef SETPROCNAME_USE_NONE */ |
|
823 | 823 | |
|
824 | 824 | #if defined(HAVE_BSD_STATFS) |
|
825 | 825 | static const char *describefstype(const struct statfs *pbuf) |
|
826 | 826 | { |
|
827 | 827 | /* BSD or OSX provides a f_fstypename field */ |
|
828 | 828 | return pbuf->f_fstypename; |
|
829 | 829 | } |
|
830 | 830 | #elif defined(HAVE_LINUX_STATFS) |
|
831 | 831 | static const char *describefstype(const struct statfs *pbuf) |
|
832 | 832 | { |
|
833 | 833 | /* Begin of Linux filesystems */ |
|
834 | 834 | #ifdef ADFS_SUPER_MAGIC |
|
835 | 835 | if (pbuf->f_type == ADFS_SUPER_MAGIC) |
|
836 | 836 | return "adfs"; |
|
837 | 837 | #endif |
|
838 | 838 | #ifdef AFFS_SUPER_MAGIC |
|
839 | 839 | if (pbuf->f_type == AFFS_SUPER_MAGIC) |
|
840 | 840 | return "affs"; |
|
841 | 841 | #endif |
|
842 | 842 | #ifdef AUTOFS_SUPER_MAGIC |
|
843 | 843 | if (pbuf->f_type == AUTOFS_SUPER_MAGIC) |
|
844 | 844 | return "autofs"; |
|
845 | 845 | #endif |
|
846 | 846 | #ifdef BDEVFS_MAGIC |
|
847 | 847 | if (pbuf->f_type == BDEVFS_MAGIC) |
|
848 | 848 | return "bdevfs"; |
|
849 | 849 | #endif |
|
850 | 850 | #ifdef BEFS_SUPER_MAGIC |
|
851 | 851 | if (pbuf->f_type == BEFS_SUPER_MAGIC) |
|
852 | 852 | return "befs"; |
|
853 | 853 | #endif |
|
854 | 854 | #ifdef BFS_MAGIC |
|
855 | 855 | if (pbuf->f_type == BFS_MAGIC) |
|
856 | 856 | return "bfs"; |
|
857 | 857 | #endif |
|
858 | 858 | #ifdef BINFMTFS_MAGIC |
|
859 | 859 | if (pbuf->f_type == BINFMTFS_MAGIC) |
|
860 | 860 | return "binfmtfs"; |
|
861 | 861 | #endif |
|
862 | 862 | #ifdef BTRFS_SUPER_MAGIC |
|
863 | 863 | if (pbuf->f_type == BTRFS_SUPER_MAGIC) |
|
864 | 864 | return "btrfs"; |
|
865 | 865 | #endif |
|
866 | 866 | #ifdef CGROUP_SUPER_MAGIC |
|
867 | 867 | if (pbuf->f_type == CGROUP_SUPER_MAGIC) |
|
868 | 868 | return "cgroup"; |
|
869 | 869 | #endif |
|
870 | 870 | #ifdef CIFS_MAGIC_NUMBER |
|
871 | 871 | if (pbuf->f_type == CIFS_MAGIC_NUMBER) |
|
872 | 872 | return "cifs"; |
|
873 | 873 | #endif |
|
874 | 874 | #ifdef CODA_SUPER_MAGIC |
|
875 | 875 | if (pbuf->f_type == CODA_SUPER_MAGIC) |
|
876 | 876 | return "coda"; |
|
877 | 877 | #endif |
|
878 | 878 | #ifdef COH_SUPER_MAGIC |
|
879 | 879 | if (pbuf->f_type == COH_SUPER_MAGIC) |
|
880 | 880 | return "coh"; |
|
881 | 881 | #endif |
|
882 | 882 | #ifdef CRAMFS_MAGIC |
|
883 | 883 | if (pbuf->f_type == CRAMFS_MAGIC) |
|
884 | 884 | return "cramfs"; |
|
885 | 885 | #endif |
|
886 | 886 | #ifdef DEBUGFS_MAGIC |
|
887 | 887 | if (pbuf->f_type == DEBUGFS_MAGIC) |
|
888 | 888 | return "debugfs"; |
|
889 | 889 | #endif |
|
890 | 890 | #ifdef DEVFS_SUPER_MAGIC |
|
891 | 891 | if (pbuf->f_type == DEVFS_SUPER_MAGIC) |
|
892 | 892 | return "devfs"; |
|
893 | 893 | #endif |
|
894 | 894 | #ifdef DEVPTS_SUPER_MAGIC |
|
895 | 895 | if (pbuf->f_type == DEVPTS_SUPER_MAGIC) |
|
896 | 896 | return "devpts"; |
|
897 | 897 | #endif |
|
898 | 898 | #ifdef EFIVARFS_MAGIC |
|
899 | 899 | if (pbuf->f_type == EFIVARFS_MAGIC) |
|
900 | 900 | return "efivarfs"; |
|
901 | 901 | #endif |
|
902 | 902 | #ifdef EFS_SUPER_MAGIC |
|
903 | 903 | if (pbuf->f_type == EFS_SUPER_MAGIC) |
|
904 | 904 | return "efs"; |
|
905 | 905 | #endif |
|
906 | 906 | #ifdef EXT_SUPER_MAGIC |
|
907 | 907 | if (pbuf->f_type == EXT_SUPER_MAGIC) |
|
908 | 908 | return "ext"; |
|
909 | 909 | #endif |
|
910 | 910 | #ifdef EXT2_OLD_SUPER_MAGIC |
|
911 | 911 | if (pbuf->f_type == EXT2_OLD_SUPER_MAGIC) |
|
912 | 912 | return "ext2"; |
|
913 | 913 | #endif |
|
914 | 914 | #ifdef EXT2_SUPER_MAGIC |
|
915 | 915 | if (pbuf->f_type == EXT2_SUPER_MAGIC) |
|
916 | 916 | return "ext2"; |
|
917 | 917 | #endif |
|
918 | 918 | #ifdef EXT3_SUPER_MAGIC |
|
919 | 919 | if (pbuf->f_type == EXT3_SUPER_MAGIC) |
|
920 | 920 | return "ext3"; |
|
921 | 921 | #endif |
|
922 | 922 | #ifdef EXT4_SUPER_MAGIC |
|
923 | 923 | if (pbuf->f_type == EXT4_SUPER_MAGIC) |
|
924 | 924 | return "ext4"; |
|
925 | 925 | #endif |
|
926 | 926 | #ifdef F2FS_SUPER_MAGIC |
|
927 | 927 | if (pbuf->f_type == F2FS_SUPER_MAGIC) |
|
928 | 928 | return "f2fs"; |
|
929 | 929 | #endif |
|
930 | 930 | #ifdef FUSE_SUPER_MAGIC |
|
931 | 931 | if (pbuf->f_type == FUSE_SUPER_MAGIC) |
|
932 | 932 | return "fuse"; |
|
933 | 933 | #endif |
|
934 | 934 | #ifdef FUTEXFS_SUPER_MAGIC |
|
935 | 935 | if (pbuf->f_type == FUTEXFS_SUPER_MAGIC) |
|
936 | 936 | return "futexfs"; |
|
937 | 937 | #endif |
|
938 | 938 | #ifdef HFS_SUPER_MAGIC |
|
939 | 939 | if (pbuf->f_type == HFS_SUPER_MAGIC) |
|
940 | 940 | return "hfs"; |
|
941 | 941 | #endif |
|
942 | 942 | #ifdef HOSTFS_SUPER_MAGIC |
|
943 | 943 | if (pbuf->f_type == HOSTFS_SUPER_MAGIC) |
|
944 | 944 | return "hostfs"; |
|
945 | 945 | #endif |
|
946 | 946 | #ifdef HPFS_SUPER_MAGIC |
|
947 | 947 | if (pbuf->f_type == HPFS_SUPER_MAGIC) |
|
948 | 948 | return "hpfs"; |
|
949 | 949 | #endif |
|
950 | 950 | #ifdef HUGETLBFS_MAGIC |
|
951 | 951 | if (pbuf->f_type == HUGETLBFS_MAGIC) |
|
952 | 952 | return "hugetlbfs"; |
|
953 | 953 | #endif |
|
954 | 954 | #ifdef ISOFS_SUPER_MAGIC |
|
955 | 955 | if (pbuf->f_type == ISOFS_SUPER_MAGIC) |
|
956 | 956 | return "isofs"; |
|
957 | 957 | #endif |
|
958 | 958 | #ifdef JFFS2_SUPER_MAGIC |
|
959 | 959 | if (pbuf->f_type == JFFS2_SUPER_MAGIC) |
|
960 | 960 | return "jffs2"; |
|
961 | 961 | #endif |
|
962 | 962 | #ifdef JFS_SUPER_MAGIC |
|
963 | 963 | if (pbuf->f_type == JFS_SUPER_MAGIC) |
|
964 | 964 | return "jfs"; |
|
965 | 965 | #endif |
|
966 | 966 | #ifdef MINIX_SUPER_MAGIC |
|
967 | 967 | if (pbuf->f_type == MINIX_SUPER_MAGIC) |
|
968 | 968 | return "minix"; |
|
969 | 969 | #endif |
|
970 | 970 | #ifdef MINIX2_SUPER_MAGIC |
|
971 | 971 | if (pbuf->f_type == MINIX2_SUPER_MAGIC) |
|
972 | 972 | return "minix2"; |
|
973 | 973 | #endif |
|
974 | 974 | #ifdef MINIX3_SUPER_MAGIC |
|
975 | 975 | if (pbuf->f_type == MINIX3_SUPER_MAGIC) |
|
976 | 976 | return "minix3"; |
|
977 | 977 | #endif |
|
978 | 978 | #ifdef MQUEUE_MAGIC |
|
979 | 979 | if (pbuf->f_type == MQUEUE_MAGIC) |
|
980 | 980 | return "mqueue"; |
|
981 | 981 | #endif |
|
982 | 982 | #ifdef MSDOS_SUPER_MAGIC |
|
983 | 983 | if (pbuf->f_type == MSDOS_SUPER_MAGIC) |
|
984 | 984 | return "msdos"; |
|
985 | 985 | #endif |
|
986 | 986 | #ifdef NCP_SUPER_MAGIC |
|
987 | 987 | if (pbuf->f_type == NCP_SUPER_MAGIC) |
|
988 | 988 | return "ncp"; |
|
989 | 989 | #endif |
|
990 | 990 | #ifdef NFS_SUPER_MAGIC |
|
991 | 991 | if (pbuf->f_type == NFS_SUPER_MAGIC) |
|
992 | 992 | return "nfs"; |
|
993 | 993 | #endif |
|
994 | 994 | #ifdef NILFS_SUPER_MAGIC |
|
995 | 995 | if (pbuf->f_type == NILFS_SUPER_MAGIC) |
|
996 | 996 | return "nilfs"; |
|
997 | 997 | #endif |
|
998 | 998 | #ifdef NTFS_SB_MAGIC |
|
999 | 999 | if (pbuf->f_type == NTFS_SB_MAGIC) |
|
1000 | 1000 | return "ntfs-sb"; |
|
1001 | 1001 | #endif |
|
1002 | 1002 | #ifdef OCFS2_SUPER_MAGIC |
|
1003 | 1003 | if (pbuf->f_type == OCFS2_SUPER_MAGIC) |
|
1004 | 1004 | return "ocfs2"; |
|
1005 | 1005 | #endif |
|
1006 | 1006 | #ifdef OPENPROM_SUPER_MAGIC |
|
1007 | 1007 | if (pbuf->f_type == OPENPROM_SUPER_MAGIC) |
|
1008 | 1008 | return "openprom"; |
|
1009 | 1009 | #endif |
|
1010 | 1010 | #ifdef OVERLAYFS_SUPER_MAGIC |
|
1011 | 1011 | if (pbuf->f_type == OVERLAYFS_SUPER_MAGIC) |
|
1012 | 1012 | return "overlay"; |
|
1013 | 1013 | #endif |
|
1014 | 1014 | #ifdef PIPEFS_MAGIC |
|
1015 | 1015 | if (pbuf->f_type == PIPEFS_MAGIC) |
|
1016 | 1016 | return "pipefs"; |
|
1017 | 1017 | #endif |
|
1018 | 1018 | #ifdef PROC_SUPER_MAGIC |
|
1019 | 1019 | if (pbuf->f_type == PROC_SUPER_MAGIC) |
|
1020 | 1020 | return "proc"; |
|
1021 | 1021 | #endif |
|
1022 | 1022 | #ifdef PSTOREFS_MAGIC |
|
1023 | 1023 | if (pbuf->f_type == PSTOREFS_MAGIC) |
|
1024 | 1024 | return "pstorefs"; |
|
1025 | 1025 | #endif |
|
1026 | 1026 | #ifdef QNX4_SUPER_MAGIC |
|
1027 | 1027 | if (pbuf->f_type == QNX4_SUPER_MAGIC) |
|
1028 | 1028 | return "qnx4"; |
|
1029 | 1029 | #endif |
|
1030 | 1030 | #ifdef QNX6_SUPER_MAGIC |
|
1031 | 1031 | if (pbuf->f_type == QNX6_SUPER_MAGIC) |
|
1032 | 1032 | return "qnx6"; |
|
1033 | 1033 | #endif |
|
1034 | 1034 | #ifdef RAMFS_MAGIC |
|
1035 | 1035 | if (pbuf->f_type == RAMFS_MAGIC) |
|
1036 | 1036 | return "ramfs"; |
|
1037 | 1037 | #endif |
|
1038 | 1038 | #ifdef REISERFS_SUPER_MAGIC |
|
1039 | 1039 | if (pbuf->f_type == REISERFS_SUPER_MAGIC) |
|
1040 | 1040 | return "reiserfs"; |
|
1041 | 1041 | #endif |
|
1042 | 1042 | #ifdef ROMFS_MAGIC |
|
1043 | 1043 | if (pbuf->f_type == ROMFS_MAGIC) |
|
1044 | 1044 | return "romfs"; |
|
1045 | 1045 | #endif |
|
1046 | 1046 | #ifdef SECURITYFS_MAGIC |
|
1047 | 1047 | if (pbuf->f_type == SECURITYFS_MAGIC) |
|
1048 | 1048 | return "securityfs"; |
|
1049 | 1049 | #endif |
|
1050 | 1050 | #ifdef SELINUX_MAGIC |
|
1051 | 1051 | if (pbuf->f_type == SELINUX_MAGIC) |
|
1052 | 1052 | return "selinux"; |
|
1053 | 1053 | #endif |
|
1054 | 1054 | #ifdef SMACK_MAGIC |
|
1055 | 1055 | if (pbuf->f_type == SMACK_MAGIC) |
|
1056 | 1056 | return "smack"; |
|
1057 | 1057 | #endif |
|
1058 | 1058 | #ifdef SMB_SUPER_MAGIC |
|
1059 | 1059 | if (pbuf->f_type == SMB_SUPER_MAGIC) |
|
1060 | 1060 | return "smb"; |
|
1061 | 1061 | #endif |
|
1062 | 1062 | #ifdef SOCKFS_MAGIC |
|
1063 | 1063 | if (pbuf->f_type == SOCKFS_MAGIC) |
|
1064 | 1064 | return "sockfs"; |
|
1065 | 1065 | #endif |
|
1066 | 1066 | #ifdef SQUASHFS_MAGIC |
|
1067 | 1067 | if (pbuf->f_type == SQUASHFS_MAGIC) |
|
1068 | 1068 | return "squashfs"; |
|
1069 | 1069 | #endif |
|
1070 | 1070 | #ifdef SYSFS_MAGIC |
|
1071 | 1071 | if (pbuf->f_type == SYSFS_MAGIC) |
|
1072 | 1072 | return "sysfs"; |
|
1073 | 1073 | #endif |
|
1074 | 1074 | #ifdef SYSV2_SUPER_MAGIC |
|
1075 | 1075 | if (pbuf->f_type == SYSV2_SUPER_MAGIC) |
|
1076 | 1076 | return "sysv2"; |
|
1077 | 1077 | #endif |
|
1078 | 1078 | #ifdef SYSV4_SUPER_MAGIC |
|
1079 | 1079 | if (pbuf->f_type == SYSV4_SUPER_MAGIC) |
|
1080 | 1080 | return "sysv4"; |
|
1081 | 1081 | #endif |
|
1082 | 1082 | #ifdef TMPFS_MAGIC |
|
1083 | 1083 | if (pbuf->f_type == TMPFS_MAGIC) |
|
1084 | 1084 | return "tmpfs"; |
|
1085 | 1085 | #endif |
|
1086 | 1086 | #ifdef UDF_SUPER_MAGIC |
|
1087 | 1087 | if (pbuf->f_type == UDF_SUPER_MAGIC) |
|
1088 | 1088 | return "udf"; |
|
1089 | 1089 | #endif |
|
1090 | 1090 | #ifdef UFS_MAGIC |
|
1091 | 1091 | if (pbuf->f_type == UFS_MAGIC) |
|
1092 | 1092 | return "ufs"; |
|
1093 | 1093 | #endif |
|
1094 | 1094 | #ifdef USBDEVICE_SUPER_MAGIC |
|
1095 | 1095 | if (pbuf->f_type == USBDEVICE_SUPER_MAGIC) |
|
1096 | 1096 | return "usbdevice"; |
|
1097 | 1097 | #endif |
|
1098 | 1098 | #ifdef V9FS_MAGIC |
|
1099 | 1099 | if (pbuf->f_type == V9FS_MAGIC) |
|
1100 | 1100 | return "v9fs"; |
|
1101 | 1101 | #endif |
|
1102 | 1102 | #ifdef VXFS_SUPER_MAGIC |
|
1103 | 1103 | if (pbuf->f_type == VXFS_SUPER_MAGIC) |
|
1104 | 1104 | return "vxfs"; |
|
1105 | 1105 | #endif |
|
1106 | 1106 | #ifdef XENFS_SUPER_MAGIC |
|
1107 | 1107 | if (pbuf->f_type == XENFS_SUPER_MAGIC) |
|
1108 | 1108 | return "xenfs"; |
|
1109 | 1109 | #endif |
|
1110 | 1110 | #ifdef XENIX_SUPER_MAGIC |
|
1111 | 1111 | if (pbuf->f_type == XENIX_SUPER_MAGIC) |
|
1112 | 1112 | return "xenix"; |
|
1113 | 1113 | #endif |
|
1114 | 1114 | #ifdef XFS_SUPER_MAGIC |
|
1115 | 1115 | if (pbuf->f_type == XFS_SUPER_MAGIC) |
|
1116 | 1116 | return "xfs"; |
|
1117 | 1117 | #endif |
|
1118 | 1118 | /* End of Linux filesystems */ |
|
1119 | 1119 | return NULL; |
|
1120 | 1120 | } |
|
1121 | 1121 | #endif /* def HAVE_LINUX_STATFS */ |
|
1122 | 1122 | |
|
1123 | 1123 | #if defined(HAVE_BSD_STATFS) || defined(HAVE_LINUX_STATFS) |
|
1124 | 1124 | /* given a directory path, return filesystem type name (best-effort) */ |
|
1125 | 1125 | static PyObject *getfstype(PyObject *self, PyObject *args) |
|
1126 | 1126 | { |
|
1127 | 1127 | const char *path = NULL; |
|
1128 | 1128 | struct statfs buf; |
|
1129 | 1129 | int r; |
|
1130 |
if (!PyArg_ParseTuple(args, |
|
|
1130 | if (!PyArg_ParseTuple(args, "y", &path)) | |
|
1131 | 1131 | return NULL; |
|
1132 | 1132 | |
|
1133 | 1133 | memset(&buf, 0, sizeof(buf)); |
|
1134 | 1134 | r = statfs(path, &buf); |
|
1135 | 1135 | if (r != 0) |
|
1136 | 1136 | return PyErr_SetFromErrno(PyExc_OSError); |
|
1137 |
return Py_BuildValue( |
|
|
1137 | return Py_BuildValue("y", describefstype(&buf)); | |
|
1138 | 1138 | } |
|
1139 | 1139 | #endif /* defined(HAVE_LINUX_STATFS) || defined(HAVE_BSD_STATFS) */ |
|
1140 | 1140 | |
|
1141 | 1141 | #if defined(HAVE_BSD_STATFS) |
|
1142 | 1142 | /* given a directory path, return filesystem mount point (best-effort) */ |
|
1143 | 1143 | static PyObject *getfsmountpoint(PyObject *self, PyObject *args) |
|
1144 | 1144 | { |
|
1145 | 1145 | const char *path = NULL; |
|
1146 | 1146 | struct statfs buf; |
|
1147 | 1147 | int r; |
|
1148 |
if (!PyArg_ParseTuple(args, |
|
|
1148 | if (!PyArg_ParseTuple(args, "y", &path)) | |
|
1149 | 1149 | return NULL; |
|
1150 | 1150 | |
|
1151 | 1151 | memset(&buf, 0, sizeof(buf)); |
|
1152 | 1152 | r = statfs(path, &buf); |
|
1153 | 1153 | if (r != 0) |
|
1154 | 1154 | return PyErr_SetFromErrno(PyExc_OSError); |
|
1155 |
return Py_BuildValue( |
|
|
1155 | return Py_BuildValue("y", buf.f_mntonname); | |
|
1156 | 1156 | } |
|
1157 | 1157 | #endif /* defined(HAVE_BSD_STATFS) */ |
|
1158 | 1158 | |
|
1159 | 1159 | static PyObject *unblocksignal(PyObject *self, PyObject *args) |
|
1160 | 1160 | { |
|
1161 | 1161 | int sig = 0; |
|
1162 | 1162 | sigset_t set; |
|
1163 | 1163 | int r; |
|
1164 | 1164 | if (!PyArg_ParseTuple(args, "i", &sig)) |
|
1165 | 1165 | return NULL; |
|
1166 | 1166 | r = sigemptyset(&set); |
|
1167 | 1167 | if (r != 0) |
|
1168 | 1168 | return PyErr_SetFromErrno(PyExc_OSError); |
|
1169 | 1169 | r = sigaddset(&set, sig); |
|
1170 | 1170 | if (r != 0) |
|
1171 | 1171 | return PyErr_SetFromErrno(PyExc_OSError); |
|
1172 | 1172 | r = sigprocmask(SIG_UNBLOCK, &set, NULL); |
|
1173 | 1173 | if (r != 0) |
|
1174 | 1174 | return PyErr_SetFromErrno(PyExc_OSError); |
|
1175 | 1175 | Py_RETURN_NONE; |
|
1176 | 1176 | } |
|
1177 | 1177 | |
|
1178 | 1178 | #endif /* ndef _WIN32 */ |
|
1179 | 1179 | |
|
1180 | 1180 | static PyObject *listdir(PyObject *self, PyObject *args, PyObject *kwargs) |
|
1181 | 1181 | { |
|
1182 | 1182 | PyObject *statobj = NULL; /* initialize - optional arg */ |
|
1183 | 1183 | PyObject *skipobj = NULL; /* initialize - optional arg */ |
|
1184 | 1184 | char *path, *skip = NULL; |
|
1185 | 1185 | Py_ssize_t plen; |
|
1186 | 1186 | int wantstat; |
|
1187 | 1187 | |
|
1188 | 1188 | static char *kwlist[] = {"path", "stat", "skip", NULL}; |
|
1189 | 1189 | |
|
1190 |
if (!PyArg_ParseTupleAndKeywords(args, kwargs, |
|
|
1191 | "y#|OO:listdir"), | |
|
1190 | if (!PyArg_ParseTupleAndKeywords(args, kwargs, "y#|OO:listdir", | |
|
1192 | 1191 | kwlist, &path, &plen, &statobj, &skipobj)) |
|
1193 | 1192 | return NULL; |
|
1194 | 1193 | |
|
1195 | 1194 | wantstat = statobj && PyObject_IsTrue(statobj); |
|
1196 | 1195 | |
|
1197 | 1196 | if (skipobj && skipobj != Py_None) { |
|
1198 | 1197 | skip = PyBytes_AsString(skipobj); |
|
1199 | 1198 | if (!skip) |
|
1200 | 1199 | return NULL; |
|
1201 | 1200 | } |
|
1202 | 1201 | |
|
1203 | 1202 | return _listdir(path, plen, wantstat, skip); |
|
1204 | 1203 | } |
|
1205 | 1204 | |
|
1206 | 1205 | #ifdef _WIN32 |
|
1207 | 1206 | static PyObject *posixfile(PyObject *self, PyObject *args, PyObject *kwds) |
|
1208 | 1207 | { |
|
1209 | 1208 | static char *kwlist[] = {"name", "mode", "buffering", NULL}; |
|
1210 | 1209 | PyObject *file_obj = NULL; |
|
1211 | 1210 | char *name = NULL; |
|
1212 | 1211 | char *mode = "rb"; |
|
1213 | 1212 | DWORD access = 0; |
|
1214 | 1213 | DWORD creation; |
|
1215 | 1214 | HANDLE handle; |
|
1216 | 1215 | int fd, flags = 0; |
|
1217 | 1216 | int bufsize = -1; |
|
1218 | 1217 | char m0, m1, m2; |
|
1219 | 1218 | char fpmode[4]; |
|
1220 | 1219 | int fppos = 0; |
|
1221 | 1220 | int plus; |
|
1222 | 1221 | FILE *fp; |
|
1223 | 1222 | |
|
1224 |
if (!PyArg_ParseTupleAndKeywords(args, kwds, |
|
|
1225 | "et|yi:posixfile"), | |
|
1223 | if (!PyArg_ParseTupleAndKeywords(args, kwds, "et|yi:posixfile", | |
|
1226 | 1224 | kwlist, |
|
1227 | 1225 | Py_FileSystemDefaultEncoding, |
|
1228 | 1226 | &name, &mode, &bufsize)) |
|
1229 | 1227 | return NULL; |
|
1230 | 1228 | |
|
1231 | 1229 | m0 = mode[0]; |
|
1232 | 1230 | m1 = m0 ? mode[1] : '\0'; |
|
1233 | 1231 | m2 = m1 ? mode[2] : '\0'; |
|
1234 | 1232 | plus = m1 == '+' || m2 == '+'; |
|
1235 | 1233 | |
|
1236 | 1234 | fpmode[fppos++] = m0; |
|
1237 | 1235 | if (m1 == 'b' || m2 == 'b') { |
|
1238 | 1236 | flags = _O_BINARY; |
|
1239 | 1237 | fpmode[fppos++] = 'b'; |
|
1240 | 1238 | } |
|
1241 | 1239 | else |
|
1242 | 1240 | flags = _O_TEXT; |
|
1243 | 1241 | if (m0 == 'r' && !plus) { |
|
1244 | 1242 | flags |= _O_RDONLY; |
|
1245 | 1243 | access = GENERIC_READ; |
|
1246 | 1244 | } else { |
|
1247 | 1245 | /* |
|
1248 | 1246 | work around http://support.microsoft.com/kb/899149 and |
|
1249 | 1247 | set _O_RDWR for 'w' and 'a', even if mode has no '+' |
|
1250 | 1248 | */ |
|
1251 | 1249 | flags |= _O_RDWR; |
|
1252 | 1250 | access = GENERIC_READ | GENERIC_WRITE; |
|
1253 | 1251 | fpmode[fppos++] = '+'; |
|
1254 | 1252 | } |
|
1255 | 1253 | fpmode[fppos++] = '\0'; |
|
1256 | 1254 | |
|
1257 | 1255 | switch (m0) { |
|
1258 | 1256 | case 'r': |
|
1259 | 1257 | creation = OPEN_EXISTING; |
|
1260 | 1258 | break; |
|
1261 | 1259 | case 'w': |
|
1262 | 1260 | creation = CREATE_ALWAYS; |
|
1263 | 1261 | break; |
|
1264 | 1262 | case 'a': |
|
1265 | 1263 | creation = OPEN_ALWAYS; |
|
1266 | 1264 | flags |= _O_APPEND; |
|
1267 | 1265 | break; |
|
1268 | 1266 | default: |
|
1269 | 1267 | PyErr_Format(PyExc_ValueError, |
|
1270 | 1268 | "mode string must begin with one of 'r', 'w', " |
|
1271 | 1269 | "or 'a', not '%c'", m0); |
|
1272 | 1270 | goto bail; |
|
1273 | 1271 | } |
|
1274 | 1272 | |
|
1275 | 1273 | handle = CreateFile(name, access, |
|
1276 | 1274 | FILE_SHARE_READ | FILE_SHARE_WRITE | |
|
1277 | 1275 | FILE_SHARE_DELETE, |
|
1278 | 1276 | NULL, |
|
1279 | 1277 | creation, |
|
1280 | 1278 | FILE_ATTRIBUTE_NORMAL, |
|
1281 | 1279 | 0); |
|
1282 | 1280 | |
|
1283 | 1281 | if (handle == INVALID_HANDLE_VALUE) { |
|
1284 | 1282 | PyErr_SetFromWindowsErrWithFilename(GetLastError(), name); |
|
1285 | 1283 | goto bail; |
|
1286 | 1284 | } |
|
1287 | 1285 | |
|
1288 | 1286 | fd = _open_osfhandle((intptr_t)handle, flags); |
|
1289 | 1287 | |
|
1290 | 1288 | if (fd == -1) { |
|
1291 | 1289 | CloseHandle(handle); |
|
1292 | 1290 | PyErr_SetFromErrnoWithFilename(PyExc_IOError, name); |
|
1293 | 1291 | goto bail; |
|
1294 | 1292 | } |
|
1295 | 1293 | fp = _fdopen(fd, fpmode); |
|
1296 | 1294 | if (fp == NULL) { |
|
1297 | 1295 | _close(fd); |
|
1298 | 1296 | PyErr_SetFromErrnoWithFilename(PyExc_IOError, name); |
|
1299 | 1297 | goto bail; |
|
1300 | 1298 | } |
|
1301 | 1299 | |
|
1302 | 1300 | file_obj = PyFile_FromFile(fp, name, mode, fclose); |
|
1303 | 1301 | if (file_obj == NULL) { |
|
1304 | 1302 | fclose(fp); |
|
1305 | 1303 | goto bail; |
|
1306 | 1304 | } |
|
1307 | 1305 | |
|
1308 | 1306 | PyFile_SetBufSize(file_obj, bufsize); |
|
1309 | 1307 | bail: |
|
1310 | 1308 | PyMem_Free(name); |
|
1311 | 1309 | return file_obj; |
|
1312 | 1310 | } |
|
1313 | 1311 | #endif |
|
1314 | 1312 | |
|
1315 | 1313 | #ifdef __APPLE__ |
|
1316 | 1314 | #include <ApplicationServices/ApplicationServices.h> |
|
1317 | 1315 | |
|
1318 | 1316 | static PyObject *isgui(PyObject *self) |
|
1319 | 1317 | { |
|
1320 | 1318 | CFDictionaryRef dict = CGSessionCopyCurrentDictionary(); |
|
1321 | 1319 | |
|
1322 | 1320 | if (dict != NULL) { |
|
1323 | 1321 | CFRelease(dict); |
|
1324 | 1322 | Py_RETURN_TRUE; |
|
1325 | 1323 | } else { |
|
1326 | 1324 | Py_RETURN_FALSE; |
|
1327 | 1325 | } |
|
1328 | 1326 | } |
|
1329 | 1327 | #endif |
|
1330 | 1328 | |
|
1331 | 1329 | static char osutil_doc[] = "Native operating system services."; |
|
1332 | 1330 | |
|
1333 | 1331 | static PyMethodDef methods[] = { |
|
1334 | 1332 | {"listdir", (PyCFunction)listdir, METH_VARARGS | METH_KEYWORDS, |
|
1335 | 1333 | "list a directory\n"}, |
|
1336 | 1334 | #ifdef _WIN32 |
|
1337 | 1335 | {"posixfile", (PyCFunction)posixfile, METH_VARARGS | METH_KEYWORDS, |
|
1338 | 1336 | "Open a file with POSIX-like semantics.\n" |
|
1339 | 1337 | "On error, this function may raise either a WindowsError or an IOError."}, |
|
1340 | 1338 | #else |
|
1341 | 1339 | {"statfiles", (PyCFunction)statfiles, METH_VARARGS | METH_KEYWORDS, |
|
1342 | 1340 | "stat a series of files or symlinks\n" |
|
1343 | 1341 | "Returns None for non-existent entries and entries of other types.\n"}, |
|
1344 | 1342 | #ifdef CMSG_LEN |
|
1345 | 1343 | {"recvfds", (PyCFunction)recvfds, METH_VARARGS, |
|
1346 | 1344 | "receive list of file descriptors via socket\n"}, |
|
1347 | 1345 | #endif |
|
1348 | 1346 | #ifndef SETPROCNAME_USE_NONE |
|
1349 | 1347 | {"setprocname", (PyCFunction)setprocname, METH_VARARGS, |
|
1350 | 1348 | "set process title (best-effort)\n"}, |
|
1351 | 1349 | #endif |
|
1352 | 1350 | #if defined(HAVE_BSD_STATFS) || defined(HAVE_LINUX_STATFS) |
|
1353 | 1351 | {"getfstype", (PyCFunction)getfstype, METH_VARARGS, |
|
1354 | 1352 | "get filesystem type (best-effort)\n"}, |
|
1355 | 1353 | #endif |
|
1356 | 1354 | #if defined(HAVE_BSD_STATFS) |
|
1357 | 1355 | {"getfsmountpoint", (PyCFunction)getfsmountpoint, METH_VARARGS, |
|
1358 | 1356 | "get filesystem mount point (best-effort)\n"}, |
|
1359 | 1357 | #endif |
|
1360 | 1358 | {"unblocksignal", (PyCFunction)unblocksignal, METH_VARARGS, |
|
1361 | 1359 | "change signal mask to unblock a given signal\n"}, |
|
1362 | 1360 | #endif /* ndef _WIN32 */ |
|
1363 | 1361 | #ifdef __APPLE__ |
|
1364 | 1362 | { |
|
1365 | 1363 | "isgui", (PyCFunction)isgui, METH_NOARGS, |
|
1366 | 1364 | "Is a CoreGraphics session available?" |
|
1367 | 1365 | }, |
|
1368 | 1366 | #endif |
|
1369 | 1367 | {NULL, NULL} |
|
1370 | 1368 | }; |
|
1371 | 1369 | |
|
1372 | 1370 | static const int version = 4; |
|
1373 | 1371 | |
|
1374 | 1372 | static struct PyModuleDef osutil_module = { |
|
1375 | 1373 | PyModuleDef_HEAD_INIT, |
|
1376 | 1374 | "osutil", |
|
1377 | 1375 | osutil_doc, |
|
1378 | 1376 | -1, |
|
1379 | 1377 | methods |
|
1380 | 1378 | }; |
|
1381 | 1379 | |
|
1382 | 1380 | PyMODINIT_FUNC PyInit_osutil(void) |
|
1383 | 1381 | { |
|
1384 | 1382 | PyObject *m; |
|
1385 | 1383 | if (PyType_Ready(&listdir_stat_type) < 0) |
|
1386 | 1384 | return NULL; |
|
1387 | 1385 | |
|
1388 | 1386 | m = PyModule_Create(&osutil_module); |
|
1389 | 1387 | PyModule_AddIntConstant(m, "version", version); |
|
1390 | 1388 | return m; |
|
1391 | 1389 | } |
@@ -1,1306 +1,1304 b'' | |||
|
1 | 1 | /* |
|
2 | 2 | parsers.c - efficient content parsing |
|
3 | 3 | |
|
4 | 4 | Copyright 2008 Olivia Mackall <olivia@selenic.com> and others |
|
5 | 5 | |
|
6 | 6 | This software may be used and distributed according to the terms of |
|
7 | 7 | the GNU General Public License, incorporated herein by reference. |
|
8 | 8 | */ |
|
9 | 9 | |
|
10 | 10 | #define PY_SSIZE_T_CLEAN |
|
11 | 11 | #include <Python.h> |
|
12 | 12 | #include <ctype.h> |
|
13 | 13 | #include <stddef.h> |
|
14 | 14 | #include <string.h> |
|
15 | 15 | |
|
16 | 16 | #include "bitmanipulation.h" |
|
17 | 17 | #include "charencode.h" |
|
18 | 18 | #include "util.h" |
|
19 | 19 | |
|
20 | 20 | static const char *const versionerrortext = "Python minor version mismatch"; |
|
21 | 21 | |
|
22 | 22 | static const int dirstate_v1_from_p2 = -2; |
|
23 | 23 | static const int dirstate_v1_nonnormal = -1; |
|
24 | 24 | static const int ambiguous_time = -1; |
|
25 | 25 | |
|
26 | 26 | static PyObject *dict_new_presized(PyObject *self, PyObject *args) |
|
27 | 27 | { |
|
28 | 28 | Py_ssize_t expected_size; |
|
29 | 29 | |
|
30 | 30 | if (!PyArg_ParseTuple(args, "n:make_presized_dict", &expected_size)) { |
|
31 | 31 | return NULL; |
|
32 | 32 | } |
|
33 | 33 | |
|
34 | 34 | return _dict_new_presized(expected_size); |
|
35 | 35 | } |
|
36 | 36 | |
|
37 | 37 | static PyObject *dirstate_item_new(PyTypeObject *subtype, PyObject *args, |
|
38 | 38 | PyObject *kwds) |
|
39 | 39 | { |
|
40 | 40 | /* We do all the initialization here and not a tp_init function because |
|
41 | 41 | * dirstate_item is immutable. */ |
|
42 | 42 | dirstateItemObject *t; |
|
43 | 43 | int wc_tracked; |
|
44 | 44 | int p1_tracked; |
|
45 | 45 | int p2_info; |
|
46 | 46 | int has_meaningful_data; |
|
47 | 47 | int has_meaningful_mtime; |
|
48 | 48 | int mtime_second_ambiguous; |
|
49 | 49 | int mode; |
|
50 | 50 | int size; |
|
51 | 51 | int mtime_s; |
|
52 | 52 | int mtime_ns; |
|
53 | 53 | PyObject *parentfiledata; |
|
54 | 54 | PyObject *mtime; |
|
55 | 55 | PyObject *fallback_exec; |
|
56 | 56 | PyObject *fallback_symlink; |
|
57 | 57 | static char *keywords_name[] = { |
|
58 | 58 | "wc_tracked", "p1_tracked", "p2_info", |
|
59 | 59 | "has_meaningful_data", "has_meaningful_mtime", "parentfiledata", |
|
60 | 60 | "fallback_exec", "fallback_symlink", NULL, |
|
61 | 61 | }; |
|
62 | 62 | wc_tracked = 0; |
|
63 | 63 | p1_tracked = 0; |
|
64 | 64 | p2_info = 0; |
|
65 | 65 | has_meaningful_mtime = 1; |
|
66 | 66 | has_meaningful_data = 1; |
|
67 | 67 | mtime_second_ambiguous = 0; |
|
68 | 68 | parentfiledata = Py_None; |
|
69 | 69 | fallback_exec = Py_None; |
|
70 | 70 | fallback_symlink = Py_None; |
|
71 | 71 | if (!PyArg_ParseTupleAndKeywords(args, kwds, "|iiiiiOOO", keywords_name, |
|
72 | 72 | &wc_tracked, &p1_tracked, &p2_info, |
|
73 | 73 | &has_meaningful_data, |
|
74 | 74 | &has_meaningful_mtime, &parentfiledata, |
|
75 | 75 | &fallback_exec, &fallback_symlink)) { |
|
76 | 76 | return NULL; |
|
77 | 77 | } |
|
78 | 78 | t = (dirstateItemObject *)subtype->tp_alloc(subtype, 1); |
|
79 | 79 | if (!t) { |
|
80 | 80 | return NULL; |
|
81 | 81 | } |
|
82 | 82 | |
|
83 | 83 | t->flags = 0; |
|
84 | 84 | if (wc_tracked) { |
|
85 | 85 | t->flags |= dirstate_flag_wc_tracked; |
|
86 | 86 | } |
|
87 | 87 | if (p1_tracked) { |
|
88 | 88 | t->flags |= dirstate_flag_p1_tracked; |
|
89 | 89 | } |
|
90 | 90 | if (p2_info) { |
|
91 | 91 | t->flags |= dirstate_flag_p2_info; |
|
92 | 92 | } |
|
93 | 93 | |
|
94 | 94 | if (fallback_exec != Py_None) { |
|
95 | 95 | t->flags |= dirstate_flag_has_fallback_exec; |
|
96 | 96 | if (PyObject_IsTrue(fallback_exec)) { |
|
97 | 97 | t->flags |= dirstate_flag_fallback_exec; |
|
98 | 98 | } |
|
99 | 99 | } |
|
100 | 100 | if (fallback_symlink != Py_None) { |
|
101 | 101 | t->flags |= dirstate_flag_has_fallback_symlink; |
|
102 | 102 | if (PyObject_IsTrue(fallback_symlink)) { |
|
103 | 103 | t->flags |= dirstate_flag_fallback_symlink; |
|
104 | 104 | } |
|
105 | 105 | } |
|
106 | 106 | |
|
107 | 107 | if (parentfiledata != Py_None) { |
|
108 | 108 | if (!PyArg_ParseTuple(parentfiledata, "iiO", &mode, &size, |
|
109 | 109 | &mtime)) { |
|
110 | 110 | return NULL; |
|
111 | 111 | } |
|
112 | 112 | if (mtime != Py_None) { |
|
113 | 113 | if (!PyArg_ParseTuple(mtime, "iii", &mtime_s, &mtime_ns, |
|
114 | 114 | &mtime_second_ambiguous)) { |
|
115 | 115 | return NULL; |
|
116 | 116 | } |
|
117 | 117 | } else { |
|
118 | 118 | has_meaningful_mtime = 0; |
|
119 | 119 | } |
|
120 | 120 | } else { |
|
121 | 121 | has_meaningful_data = 0; |
|
122 | 122 | has_meaningful_mtime = 0; |
|
123 | 123 | } |
|
124 | 124 | if (has_meaningful_data) { |
|
125 | 125 | t->flags |= dirstate_flag_has_meaningful_data; |
|
126 | 126 | t->mode = mode; |
|
127 | 127 | t->size = size; |
|
128 | 128 | if (mtime_second_ambiguous) { |
|
129 | 129 | t->flags |= dirstate_flag_mtime_second_ambiguous; |
|
130 | 130 | } |
|
131 | 131 | } else { |
|
132 | 132 | t->mode = 0; |
|
133 | 133 | t->size = 0; |
|
134 | 134 | } |
|
135 | 135 | if (has_meaningful_mtime) { |
|
136 | 136 | t->flags |= dirstate_flag_has_mtime; |
|
137 | 137 | t->mtime_s = mtime_s; |
|
138 | 138 | t->mtime_ns = mtime_ns; |
|
139 | 139 | } else { |
|
140 | 140 | t->mtime_s = 0; |
|
141 | 141 | t->mtime_ns = 0; |
|
142 | 142 | } |
|
143 | 143 | return (PyObject *)t; |
|
144 | 144 | } |
|
145 | 145 | |
|
146 | 146 | static void dirstate_item_dealloc(PyObject *o) |
|
147 | 147 | { |
|
148 | 148 | PyObject_Del(o); |
|
149 | 149 | } |
|
150 | 150 | |
|
151 | 151 | static inline bool dirstate_item_c_tracked(dirstateItemObject *self) |
|
152 | 152 | { |
|
153 | 153 | return (self->flags & dirstate_flag_wc_tracked); |
|
154 | 154 | } |
|
155 | 155 | |
|
156 | 156 | static inline bool dirstate_item_c_any_tracked(dirstateItemObject *self) |
|
157 | 157 | { |
|
158 | 158 | const int mask = dirstate_flag_wc_tracked | dirstate_flag_p1_tracked | |
|
159 | 159 | dirstate_flag_p2_info; |
|
160 | 160 | return (self->flags & mask); |
|
161 | 161 | } |
|
162 | 162 | |
|
163 | 163 | static inline bool dirstate_item_c_added(dirstateItemObject *self) |
|
164 | 164 | { |
|
165 | 165 | const int mask = (dirstate_flag_wc_tracked | dirstate_flag_p1_tracked | |
|
166 | 166 | dirstate_flag_p2_info); |
|
167 | 167 | const int target = dirstate_flag_wc_tracked; |
|
168 | 168 | return (self->flags & mask) == target; |
|
169 | 169 | } |
|
170 | 170 | |
|
171 | 171 | static inline bool dirstate_item_c_removed(dirstateItemObject *self) |
|
172 | 172 | { |
|
173 | 173 | if (self->flags & dirstate_flag_wc_tracked) { |
|
174 | 174 | return false; |
|
175 | 175 | } |
|
176 | 176 | return (self->flags & |
|
177 | 177 | (dirstate_flag_p1_tracked | dirstate_flag_p2_info)); |
|
178 | 178 | } |
|
179 | 179 | |
|
180 | 180 | static inline bool dirstate_item_c_merged(dirstateItemObject *self) |
|
181 | 181 | { |
|
182 | 182 | return ((self->flags & dirstate_flag_wc_tracked) && |
|
183 | 183 | (self->flags & dirstate_flag_p1_tracked) && |
|
184 | 184 | (self->flags & dirstate_flag_p2_info)); |
|
185 | 185 | } |
|
186 | 186 | |
|
187 | 187 | static inline bool dirstate_item_c_from_p2(dirstateItemObject *self) |
|
188 | 188 | { |
|
189 | 189 | return ((self->flags & dirstate_flag_wc_tracked) && |
|
190 | 190 | !(self->flags & dirstate_flag_p1_tracked) && |
|
191 | 191 | (self->flags & dirstate_flag_p2_info)); |
|
192 | 192 | } |
|
193 | 193 | |
|
194 | 194 | static inline char dirstate_item_c_v1_state(dirstateItemObject *self) |
|
195 | 195 | { |
|
196 | 196 | if (dirstate_item_c_removed(self)) { |
|
197 | 197 | return 'r'; |
|
198 | 198 | } else if (dirstate_item_c_merged(self)) { |
|
199 | 199 | return 'm'; |
|
200 | 200 | } else if (dirstate_item_c_added(self)) { |
|
201 | 201 | return 'a'; |
|
202 | 202 | } else { |
|
203 | 203 | return 'n'; |
|
204 | 204 | } |
|
205 | 205 | } |
|
206 | 206 | |
|
207 | 207 | static inline bool dirstate_item_c_has_fallback_exec(dirstateItemObject *self) |
|
208 | 208 | { |
|
209 | 209 | return (bool)self->flags & dirstate_flag_has_fallback_exec; |
|
210 | 210 | } |
|
211 | 211 | |
|
212 | 212 | static inline bool |
|
213 | 213 | dirstate_item_c_has_fallback_symlink(dirstateItemObject *self) |
|
214 | 214 | { |
|
215 | 215 | return (bool)self->flags & dirstate_flag_has_fallback_symlink; |
|
216 | 216 | } |
|
217 | 217 | |
|
218 | 218 | static inline int dirstate_item_c_v1_mode(dirstateItemObject *self) |
|
219 | 219 | { |
|
220 | 220 | if (self->flags & dirstate_flag_has_meaningful_data) { |
|
221 | 221 | return self->mode; |
|
222 | 222 | } else { |
|
223 | 223 | return 0; |
|
224 | 224 | } |
|
225 | 225 | } |
|
226 | 226 | |
|
227 | 227 | static inline int dirstate_item_c_v1_size(dirstateItemObject *self) |
|
228 | 228 | { |
|
229 | 229 | if (!(self->flags & dirstate_flag_wc_tracked) && |
|
230 | 230 | (self->flags & dirstate_flag_p2_info)) { |
|
231 | 231 | if (self->flags & dirstate_flag_p1_tracked) { |
|
232 | 232 | return dirstate_v1_nonnormal; |
|
233 | 233 | } else { |
|
234 | 234 | return dirstate_v1_from_p2; |
|
235 | 235 | } |
|
236 | 236 | } else if (dirstate_item_c_removed(self)) { |
|
237 | 237 | return 0; |
|
238 | 238 | } else if (self->flags & dirstate_flag_p2_info) { |
|
239 | 239 | return dirstate_v1_from_p2; |
|
240 | 240 | } else if (dirstate_item_c_added(self)) { |
|
241 | 241 | return dirstate_v1_nonnormal; |
|
242 | 242 | } else if (self->flags & dirstate_flag_has_meaningful_data) { |
|
243 | 243 | return self->size; |
|
244 | 244 | } else { |
|
245 | 245 | return dirstate_v1_nonnormal; |
|
246 | 246 | } |
|
247 | 247 | } |
|
248 | 248 | |
|
249 | 249 | static inline int dirstate_item_c_v1_mtime(dirstateItemObject *self) |
|
250 | 250 | { |
|
251 | 251 | if (dirstate_item_c_removed(self)) { |
|
252 | 252 | return 0; |
|
253 | 253 | } else if (!(self->flags & dirstate_flag_has_mtime) || |
|
254 | 254 | !(self->flags & dirstate_flag_p1_tracked) || |
|
255 | 255 | !(self->flags & dirstate_flag_wc_tracked) || |
|
256 | 256 | (self->flags & dirstate_flag_p2_info) || |
|
257 | 257 | (self->flags & dirstate_flag_mtime_second_ambiguous)) { |
|
258 | 258 | return ambiguous_time; |
|
259 | 259 | } else { |
|
260 | 260 | return self->mtime_s; |
|
261 | 261 | } |
|
262 | 262 | } |
|
263 | 263 | |
|
264 | 264 | static PyObject *dirstate_item_v2_data(dirstateItemObject *self) |
|
265 | 265 | { |
|
266 | 266 | int flags = self->flags; |
|
267 | 267 | int mode = dirstate_item_c_v1_mode(self); |
|
268 | 268 | #ifdef S_IXUSR |
|
269 | 269 | /* This is for platforms with an exec bit */ |
|
270 | 270 | if ((mode & S_IXUSR) != 0) { |
|
271 | 271 | flags |= dirstate_flag_mode_exec_perm; |
|
272 | 272 | } else { |
|
273 | 273 | flags &= ~dirstate_flag_mode_exec_perm; |
|
274 | 274 | } |
|
275 | 275 | #else |
|
276 | 276 | flags &= ~dirstate_flag_mode_exec_perm; |
|
277 | 277 | #endif |
|
278 | 278 | #ifdef S_ISLNK |
|
279 | 279 | /* This is for platforms with support for symlinks */ |
|
280 | 280 | if (S_ISLNK(mode)) { |
|
281 | 281 | flags |= dirstate_flag_mode_is_symlink; |
|
282 | 282 | } else { |
|
283 | 283 | flags &= ~dirstate_flag_mode_is_symlink; |
|
284 | 284 | } |
|
285 | 285 | #else |
|
286 | 286 | flags &= ~dirstate_flag_mode_is_symlink; |
|
287 | 287 | #endif |
|
288 | 288 | return Py_BuildValue("iiii", flags, self->size, self->mtime_s, |
|
289 | 289 | self->mtime_ns); |
|
290 | 290 | }; |
|
291 | 291 | |
|
292 | 292 | static PyObject *dirstate_item_v1_state(dirstateItemObject *self) |
|
293 | 293 | { |
|
294 | 294 | char state = dirstate_item_c_v1_state(self); |
|
295 | 295 | return PyBytes_FromStringAndSize(&state, 1); |
|
296 | 296 | }; |
|
297 | 297 | |
|
298 | 298 | static PyObject *dirstate_item_v1_mode(dirstateItemObject *self) |
|
299 | 299 | { |
|
300 | 300 | return PyLong_FromLong(dirstate_item_c_v1_mode(self)); |
|
301 | 301 | }; |
|
302 | 302 | |
|
303 | 303 | static PyObject *dirstate_item_v1_size(dirstateItemObject *self) |
|
304 | 304 | { |
|
305 | 305 | return PyLong_FromLong(dirstate_item_c_v1_size(self)); |
|
306 | 306 | }; |
|
307 | 307 | |
|
308 | 308 | static PyObject *dirstate_item_v1_mtime(dirstateItemObject *self) |
|
309 | 309 | { |
|
310 | 310 | return PyLong_FromLong(dirstate_item_c_v1_mtime(self)); |
|
311 | 311 | }; |
|
312 | 312 | |
|
313 | 313 | static PyObject *dirstate_item_mtime_likely_equal_to(dirstateItemObject *self, |
|
314 | 314 | PyObject *other) |
|
315 | 315 | { |
|
316 | 316 | int other_s; |
|
317 | 317 | int other_ns; |
|
318 | 318 | int other_second_ambiguous; |
|
319 | 319 | if (!PyArg_ParseTuple(other, "iii", &other_s, &other_ns, |
|
320 | 320 | &other_second_ambiguous)) { |
|
321 | 321 | return NULL; |
|
322 | 322 | } |
|
323 | 323 | if (!(self->flags & dirstate_flag_has_mtime)) { |
|
324 | 324 | Py_RETURN_FALSE; |
|
325 | 325 | } |
|
326 | 326 | if (self->mtime_s != other_s) { |
|
327 | 327 | Py_RETURN_FALSE; |
|
328 | 328 | } |
|
329 | 329 | if (self->mtime_ns == 0 || other_ns == 0) { |
|
330 | 330 | if (self->flags & dirstate_flag_mtime_second_ambiguous) { |
|
331 | 331 | Py_RETURN_FALSE; |
|
332 | 332 | } else { |
|
333 | 333 | Py_RETURN_TRUE; |
|
334 | 334 | } |
|
335 | 335 | } |
|
336 | 336 | if (self->mtime_ns == other_ns) { |
|
337 | 337 | Py_RETURN_TRUE; |
|
338 | 338 | } else { |
|
339 | 339 | Py_RETURN_FALSE; |
|
340 | 340 | } |
|
341 | 341 | }; |
|
342 | 342 | |
|
343 | 343 | /* This will never change since it's bound to V1 |
|
344 | 344 | */ |
|
345 | 345 | static inline dirstateItemObject * |
|
346 | 346 | dirstate_item_from_v1_data(char state, int mode, int size, int mtime) |
|
347 | 347 | { |
|
348 | 348 | dirstateItemObject *t = |
|
349 | 349 | PyObject_New(dirstateItemObject, &dirstateItemType); |
|
350 | 350 | if (!t) { |
|
351 | 351 | return NULL; |
|
352 | 352 | } |
|
353 | 353 | t->flags = 0; |
|
354 | 354 | t->mode = 0; |
|
355 | 355 | t->size = 0; |
|
356 | 356 | t->mtime_s = 0; |
|
357 | 357 | t->mtime_ns = 0; |
|
358 | 358 | |
|
359 | 359 | if (state == 'm') { |
|
360 | 360 | t->flags = (dirstate_flag_wc_tracked | |
|
361 | 361 | dirstate_flag_p1_tracked | dirstate_flag_p2_info); |
|
362 | 362 | } else if (state == 'a') { |
|
363 | 363 | t->flags = dirstate_flag_wc_tracked; |
|
364 | 364 | } else if (state == 'r') { |
|
365 | 365 | if (size == dirstate_v1_nonnormal) { |
|
366 | 366 | t->flags = |
|
367 | 367 | dirstate_flag_p1_tracked | dirstate_flag_p2_info; |
|
368 | 368 | } else if (size == dirstate_v1_from_p2) { |
|
369 | 369 | t->flags = dirstate_flag_p2_info; |
|
370 | 370 | } else { |
|
371 | 371 | t->flags = dirstate_flag_p1_tracked; |
|
372 | 372 | } |
|
373 | 373 | } else if (state == 'n') { |
|
374 | 374 | if (size == dirstate_v1_from_p2) { |
|
375 | 375 | t->flags = |
|
376 | 376 | dirstate_flag_wc_tracked | dirstate_flag_p2_info; |
|
377 | 377 | } else if (size == dirstate_v1_nonnormal) { |
|
378 | 378 | t->flags = |
|
379 | 379 | dirstate_flag_wc_tracked | dirstate_flag_p1_tracked; |
|
380 | 380 | } else if (mtime == ambiguous_time) { |
|
381 | 381 | t->flags = (dirstate_flag_wc_tracked | |
|
382 | 382 | dirstate_flag_p1_tracked | |
|
383 | 383 | dirstate_flag_has_meaningful_data); |
|
384 | 384 | t->mode = mode; |
|
385 | 385 | t->size = size; |
|
386 | 386 | } else { |
|
387 | 387 | t->flags = (dirstate_flag_wc_tracked | |
|
388 | 388 | dirstate_flag_p1_tracked | |
|
389 | 389 | dirstate_flag_has_meaningful_data | |
|
390 | 390 | dirstate_flag_has_mtime); |
|
391 | 391 | t->mode = mode; |
|
392 | 392 | t->size = size; |
|
393 | 393 | t->mtime_s = mtime; |
|
394 | 394 | } |
|
395 | 395 | } else { |
|
396 | 396 | PyErr_Format(PyExc_RuntimeError, |
|
397 | 397 | "unknown state: `%c` (%d, %d, %d)", state, mode, |
|
398 | 398 | size, mtime, NULL); |
|
399 | 399 | Py_DECREF(t); |
|
400 | 400 | return NULL; |
|
401 | 401 | } |
|
402 | 402 | |
|
403 | 403 | return t; |
|
404 | 404 | } |
|
405 | 405 | |
|
406 | 406 | /* This will never change since it's bound to V1, unlike `dirstate_item_new` */ |
|
407 | 407 | static PyObject *dirstate_item_from_v1_meth(PyTypeObject *subtype, |
|
408 | 408 | PyObject *args) |
|
409 | 409 | { |
|
410 | 410 | /* We do all the initialization here and not a tp_init function because |
|
411 | 411 | * dirstate_item is immutable. */ |
|
412 | 412 | char state; |
|
413 | 413 | int size, mode, mtime; |
|
414 | 414 | if (!PyArg_ParseTuple(args, "ciii", &state, &mode, &size, &mtime)) { |
|
415 | 415 | return NULL; |
|
416 | 416 | } |
|
417 | 417 | return (PyObject *)dirstate_item_from_v1_data(state, mode, size, mtime); |
|
418 | 418 | }; |
|
419 | 419 | |
|
420 | 420 | static PyObject *dirstate_item_from_v2_meth(PyTypeObject *subtype, |
|
421 | 421 | PyObject *args) |
|
422 | 422 | { |
|
423 | 423 | dirstateItemObject *t = |
|
424 | 424 | PyObject_New(dirstateItemObject, &dirstateItemType); |
|
425 | 425 | if (!t) { |
|
426 | 426 | return NULL; |
|
427 | 427 | } |
|
428 | 428 | if (!PyArg_ParseTuple(args, "iiii", &t->flags, &t->size, &t->mtime_s, |
|
429 | 429 | &t->mtime_ns)) { |
|
430 | 430 | return NULL; |
|
431 | 431 | } |
|
432 | 432 | if (t->flags & dirstate_flag_expected_state_is_modified) { |
|
433 | 433 | t->flags &= ~(dirstate_flag_expected_state_is_modified | |
|
434 | 434 | dirstate_flag_has_meaningful_data | |
|
435 | 435 | dirstate_flag_has_mtime); |
|
436 | 436 | } |
|
437 | 437 | t->mode = 0; |
|
438 | 438 | if (t->flags & dirstate_flag_has_meaningful_data) { |
|
439 | 439 | if (t->flags & dirstate_flag_mode_exec_perm) { |
|
440 | 440 | t->mode = 0755; |
|
441 | 441 | } else { |
|
442 | 442 | t->mode = 0644; |
|
443 | 443 | } |
|
444 | 444 | if (t->flags & dirstate_flag_mode_is_symlink) { |
|
445 | 445 | t->mode |= S_IFLNK; |
|
446 | 446 | } else { |
|
447 | 447 | t->mode |= S_IFREG; |
|
448 | 448 | } |
|
449 | 449 | } |
|
450 | 450 | return (PyObject *)t; |
|
451 | 451 | }; |
|
452 | 452 | |
|
453 | 453 | /* This means the next status call will have to actually check its content |
|
454 | 454 | to make sure it is correct. */ |
|
455 | 455 | static PyObject *dirstate_item_set_possibly_dirty(dirstateItemObject *self) |
|
456 | 456 | { |
|
457 | 457 | self->flags &= ~dirstate_flag_has_mtime; |
|
458 | 458 | Py_RETURN_NONE; |
|
459 | 459 | } |
|
460 | 460 | |
|
461 | 461 | /* See docstring of the python implementation for details */ |
|
462 | 462 | static PyObject *dirstate_item_set_clean(dirstateItemObject *self, |
|
463 | 463 | PyObject *args) |
|
464 | 464 | { |
|
465 | 465 | int size, mode, mtime_s, mtime_ns, mtime_second_ambiguous; |
|
466 | 466 | PyObject *mtime; |
|
467 | 467 | mtime_s = 0; |
|
468 | 468 | mtime_ns = 0; |
|
469 | 469 | mtime_second_ambiguous = 0; |
|
470 | 470 | if (!PyArg_ParseTuple(args, "iiO", &mode, &size, &mtime)) { |
|
471 | 471 | return NULL; |
|
472 | 472 | } |
|
473 | 473 | if (mtime != Py_None) { |
|
474 | 474 | if (!PyArg_ParseTuple(mtime, "iii", &mtime_s, &mtime_ns, |
|
475 | 475 | &mtime_second_ambiguous)) { |
|
476 | 476 | return NULL; |
|
477 | 477 | } |
|
478 | 478 | } else { |
|
479 | 479 | self->flags &= ~dirstate_flag_has_mtime; |
|
480 | 480 | } |
|
481 | 481 | self->flags = dirstate_flag_wc_tracked | dirstate_flag_p1_tracked | |
|
482 | 482 | dirstate_flag_has_meaningful_data | |
|
483 | 483 | dirstate_flag_has_mtime; |
|
484 | 484 | if (mtime_second_ambiguous) { |
|
485 | 485 | self->flags |= dirstate_flag_mtime_second_ambiguous; |
|
486 | 486 | } |
|
487 | 487 | self->mode = mode; |
|
488 | 488 | self->size = size; |
|
489 | 489 | self->mtime_s = mtime_s; |
|
490 | 490 | self->mtime_ns = mtime_ns; |
|
491 | 491 | Py_RETURN_NONE; |
|
492 | 492 | } |
|
493 | 493 | |
|
494 | 494 | static PyObject *dirstate_item_set_tracked(dirstateItemObject *self) |
|
495 | 495 | { |
|
496 | 496 | self->flags |= dirstate_flag_wc_tracked; |
|
497 | 497 | self->flags &= ~dirstate_flag_has_mtime; |
|
498 | 498 | Py_RETURN_NONE; |
|
499 | 499 | } |
|
500 | 500 | |
|
501 | 501 | static PyObject *dirstate_item_set_untracked(dirstateItemObject *self) |
|
502 | 502 | { |
|
503 | 503 | self->flags &= ~dirstate_flag_wc_tracked; |
|
504 | 504 | self->mode = 0; |
|
505 | 505 | self->size = 0; |
|
506 | 506 | self->mtime_s = 0; |
|
507 | 507 | self->mtime_ns = 0; |
|
508 | 508 | Py_RETURN_NONE; |
|
509 | 509 | } |
|
510 | 510 | |
|
511 | 511 | static PyObject *dirstate_item_drop_merge_data(dirstateItemObject *self) |
|
512 | 512 | { |
|
513 | 513 | if (self->flags & dirstate_flag_p2_info) { |
|
514 | 514 | self->flags &= ~(dirstate_flag_p2_info | |
|
515 | 515 | dirstate_flag_has_meaningful_data | |
|
516 | 516 | dirstate_flag_has_mtime); |
|
517 | 517 | self->mode = 0; |
|
518 | 518 | self->size = 0; |
|
519 | 519 | self->mtime_s = 0; |
|
520 | 520 | self->mtime_ns = 0; |
|
521 | 521 | } |
|
522 | 522 | Py_RETURN_NONE; |
|
523 | 523 | } |
|
524 | 524 | static PyMethodDef dirstate_item_methods[] = { |
|
525 | 525 | {"v2_data", (PyCFunction)dirstate_item_v2_data, METH_NOARGS, |
|
526 | 526 | "return data suitable for v2 serialization"}, |
|
527 | 527 | {"v1_state", (PyCFunction)dirstate_item_v1_state, METH_NOARGS, |
|
528 | 528 | "return a \"state\" suitable for v1 serialization"}, |
|
529 | 529 | {"v1_mode", (PyCFunction)dirstate_item_v1_mode, METH_NOARGS, |
|
530 | 530 | "return a \"mode\" suitable for v1 serialization"}, |
|
531 | 531 | {"v1_size", (PyCFunction)dirstate_item_v1_size, METH_NOARGS, |
|
532 | 532 | "return a \"size\" suitable for v1 serialization"}, |
|
533 | 533 | {"v1_mtime", (PyCFunction)dirstate_item_v1_mtime, METH_NOARGS, |
|
534 | 534 | "return a \"mtime\" suitable for v1 serialization"}, |
|
535 | 535 | {"mtime_likely_equal_to", (PyCFunction)dirstate_item_mtime_likely_equal_to, |
|
536 | 536 | METH_O, "True if the stored mtime is likely equal to the given mtime"}, |
|
537 | 537 | {"from_v1_data", (PyCFunction)dirstate_item_from_v1_meth, |
|
538 | 538 | METH_VARARGS | METH_CLASS, "build a new DirstateItem object from V1 data"}, |
|
539 | 539 | {"from_v2_data", (PyCFunction)dirstate_item_from_v2_meth, |
|
540 | 540 | METH_VARARGS | METH_CLASS, "build a new DirstateItem object from V2 data"}, |
|
541 | 541 | {"set_possibly_dirty", (PyCFunction)dirstate_item_set_possibly_dirty, |
|
542 | 542 | METH_NOARGS, "mark a file as \"possibly dirty\""}, |
|
543 | 543 | {"set_clean", (PyCFunction)dirstate_item_set_clean, METH_VARARGS, |
|
544 | 544 | "mark a file as \"clean\""}, |
|
545 | 545 | {"set_tracked", (PyCFunction)dirstate_item_set_tracked, METH_NOARGS, |
|
546 | 546 | "mark a file as \"tracked\""}, |
|
547 | 547 | {"set_untracked", (PyCFunction)dirstate_item_set_untracked, METH_NOARGS, |
|
548 | 548 | "mark a file as \"untracked\""}, |
|
549 | 549 | {"drop_merge_data", (PyCFunction)dirstate_item_drop_merge_data, METH_NOARGS, |
|
550 | 550 | "remove all \"merge-only\" from a DirstateItem"}, |
|
551 | 551 | {NULL} /* Sentinel */ |
|
552 | 552 | }; |
|
553 | 553 | |
|
554 | 554 | static PyObject *dirstate_item_get_mode(dirstateItemObject *self) |
|
555 | 555 | { |
|
556 | 556 | return PyLong_FromLong(dirstate_item_c_v1_mode(self)); |
|
557 | 557 | }; |
|
558 | 558 | |
|
559 | 559 | static PyObject *dirstate_item_get_size(dirstateItemObject *self) |
|
560 | 560 | { |
|
561 | 561 | return PyLong_FromLong(dirstate_item_c_v1_size(self)); |
|
562 | 562 | }; |
|
563 | 563 | |
|
564 | 564 | static PyObject *dirstate_item_get_mtime(dirstateItemObject *self) |
|
565 | 565 | { |
|
566 | 566 | return PyLong_FromLong(dirstate_item_c_v1_mtime(self)); |
|
567 | 567 | }; |
|
568 | 568 | |
|
569 | 569 | static PyObject *dirstate_item_get_state(dirstateItemObject *self) |
|
570 | 570 | { |
|
571 | 571 | char state = dirstate_item_c_v1_state(self); |
|
572 | 572 | return PyBytes_FromStringAndSize(&state, 1); |
|
573 | 573 | }; |
|
574 | 574 | |
|
575 | 575 | static PyObject *dirstate_item_get_has_fallback_exec(dirstateItemObject *self) |
|
576 | 576 | { |
|
577 | 577 | if (dirstate_item_c_has_fallback_exec(self)) { |
|
578 | 578 | Py_RETURN_TRUE; |
|
579 | 579 | } else { |
|
580 | 580 | Py_RETURN_FALSE; |
|
581 | 581 | } |
|
582 | 582 | }; |
|
583 | 583 | |
|
584 | 584 | static PyObject *dirstate_item_get_fallback_exec(dirstateItemObject *self) |
|
585 | 585 | { |
|
586 | 586 | if (dirstate_item_c_has_fallback_exec(self)) { |
|
587 | 587 | if (self->flags & dirstate_flag_fallback_exec) { |
|
588 | 588 | Py_RETURN_TRUE; |
|
589 | 589 | } else { |
|
590 | 590 | Py_RETURN_FALSE; |
|
591 | 591 | } |
|
592 | 592 | } else { |
|
593 | 593 | Py_RETURN_NONE; |
|
594 | 594 | } |
|
595 | 595 | }; |
|
596 | 596 | |
|
597 | 597 | static int dirstate_item_set_fallback_exec(dirstateItemObject *self, |
|
598 | 598 | PyObject *value) |
|
599 | 599 | { |
|
600 | 600 | if ((value == Py_None) || (value == NULL)) { |
|
601 | 601 | self->flags &= ~dirstate_flag_has_fallback_exec; |
|
602 | 602 | } else { |
|
603 | 603 | self->flags |= dirstate_flag_has_fallback_exec; |
|
604 | 604 | if (PyObject_IsTrue(value)) { |
|
605 | 605 | self->flags |= dirstate_flag_fallback_exec; |
|
606 | 606 | } else { |
|
607 | 607 | self->flags &= ~dirstate_flag_fallback_exec; |
|
608 | 608 | } |
|
609 | 609 | } |
|
610 | 610 | return 0; |
|
611 | 611 | }; |
|
612 | 612 | |
|
613 | 613 | static PyObject * |
|
614 | 614 | dirstate_item_get_has_fallback_symlink(dirstateItemObject *self) |
|
615 | 615 | { |
|
616 | 616 | if (dirstate_item_c_has_fallback_symlink(self)) { |
|
617 | 617 | Py_RETURN_TRUE; |
|
618 | 618 | } else { |
|
619 | 619 | Py_RETURN_FALSE; |
|
620 | 620 | } |
|
621 | 621 | }; |
|
622 | 622 | |
|
623 | 623 | static PyObject *dirstate_item_get_fallback_symlink(dirstateItemObject *self) |
|
624 | 624 | { |
|
625 | 625 | if (dirstate_item_c_has_fallback_symlink(self)) { |
|
626 | 626 | if (self->flags & dirstate_flag_fallback_symlink) { |
|
627 | 627 | Py_RETURN_TRUE; |
|
628 | 628 | } else { |
|
629 | 629 | Py_RETURN_FALSE; |
|
630 | 630 | } |
|
631 | 631 | } else { |
|
632 | 632 | Py_RETURN_NONE; |
|
633 | 633 | } |
|
634 | 634 | }; |
|
635 | 635 | |
|
636 | 636 | static int dirstate_item_set_fallback_symlink(dirstateItemObject *self, |
|
637 | 637 | PyObject *value) |
|
638 | 638 | { |
|
639 | 639 | if ((value == Py_None) || (value == NULL)) { |
|
640 | 640 | self->flags &= ~dirstate_flag_has_fallback_symlink; |
|
641 | 641 | } else { |
|
642 | 642 | self->flags |= dirstate_flag_has_fallback_symlink; |
|
643 | 643 | if (PyObject_IsTrue(value)) { |
|
644 | 644 | self->flags |= dirstate_flag_fallback_symlink; |
|
645 | 645 | } else { |
|
646 | 646 | self->flags &= ~dirstate_flag_fallback_symlink; |
|
647 | 647 | } |
|
648 | 648 | } |
|
649 | 649 | return 0; |
|
650 | 650 | }; |
|
651 | 651 | |
|
652 | 652 | static PyObject *dirstate_item_get_tracked(dirstateItemObject *self) |
|
653 | 653 | { |
|
654 | 654 | if (dirstate_item_c_tracked(self)) { |
|
655 | 655 | Py_RETURN_TRUE; |
|
656 | 656 | } else { |
|
657 | 657 | Py_RETURN_FALSE; |
|
658 | 658 | } |
|
659 | 659 | }; |
|
660 | 660 | static PyObject *dirstate_item_get_p1_tracked(dirstateItemObject *self) |
|
661 | 661 | { |
|
662 | 662 | if (self->flags & dirstate_flag_p1_tracked) { |
|
663 | 663 | Py_RETURN_TRUE; |
|
664 | 664 | } else { |
|
665 | 665 | Py_RETURN_FALSE; |
|
666 | 666 | } |
|
667 | 667 | }; |
|
668 | 668 | |
|
669 | 669 | static PyObject *dirstate_item_get_added(dirstateItemObject *self) |
|
670 | 670 | { |
|
671 | 671 | if (dirstate_item_c_added(self)) { |
|
672 | 672 | Py_RETURN_TRUE; |
|
673 | 673 | } else { |
|
674 | 674 | Py_RETURN_FALSE; |
|
675 | 675 | } |
|
676 | 676 | }; |
|
677 | 677 | |
|
678 | 678 | static PyObject *dirstate_item_get_p2_info(dirstateItemObject *self) |
|
679 | 679 | { |
|
680 | 680 | if (self->flags & dirstate_flag_wc_tracked && |
|
681 | 681 | self->flags & dirstate_flag_p2_info) { |
|
682 | 682 | Py_RETURN_TRUE; |
|
683 | 683 | } else { |
|
684 | 684 | Py_RETURN_FALSE; |
|
685 | 685 | } |
|
686 | 686 | }; |
|
687 | 687 | |
|
688 | 688 | static PyObject *dirstate_item_get_merged(dirstateItemObject *self) |
|
689 | 689 | { |
|
690 | 690 | if (dirstate_item_c_merged(self)) { |
|
691 | 691 | Py_RETURN_TRUE; |
|
692 | 692 | } else { |
|
693 | 693 | Py_RETURN_FALSE; |
|
694 | 694 | } |
|
695 | 695 | }; |
|
696 | 696 | |
|
697 | 697 | static PyObject *dirstate_item_get_from_p2(dirstateItemObject *self) |
|
698 | 698 | { |
|
699 | 699 | if (dirstate_item_c_from_p2(self)) { |
|
700 | 700 | Py_RETURN_TRUE; |
|
701 | 701 | } else { |
|
702 | 702 | Py_RETURN_FALSE; |
|
703 | 703 | } |
|
704 | 704 | }; |
|
705 | 705 | |
|
706 | 706 | static PyObject *dirstate_item_get_maybe_clean(dirstateItemObject *self) |
|
707 | 707 | { |
|
708 | 708 | if (!(self->flags & dirstate_flag_wc_tracked)) { |
|
709 | 709 | Py_RETURN_FALSE; |
|
710 | 710 | } else if (!(self->flags & dirstate_flag_p1_tracked)) { |
|
711 | 711 | Py_RETURN_FALSE; |
|
712 | 712 | } else if (self->flags & dirstate_flag_p2_info) { |
|
713 | 713 | Py_RETURN_FALSE; |
|
714 | 714 | } else { |
|
715 | 715 | Py_RETURN_TRUE; |
|
716 | 716 | } |
|
717 | 717 | }; |
|
718 | 718 | |
|
719 | 719 | static PyObject *dirstate_item_get_any_tracked(dirstateItemObject *self) |
|
720 | 720 | { |
|
721 | 721 | if (dirstate_item_c_any_tracked(self)) { |
|
722 | 722 | Py_RETURN_TRUE; |
|
723 | 723 | } else { |
|
724 | 724 | Py_RETURN_FALSE; |
|
725 | 725 | } |
|
726 | 726 | }; |
|
727 | 727 | |
|
728 | 728 | static PyObject *dirstate_item_get_removed(dirstateItemObject *self) |
|
729 | 729 | { |
|
730 | 730 | if (dirstate_item_c_removed(self)) { |
|
731 | 731 | Py_RETURN_TRUE; |
|
732 | 732 | } else { |
|
733 | 733 | Py_RETURN_FALSE; |
|
734 | 734 | } |
|
735 | 735 | }; |
|
736 | 736 | |
|
737 | 737 | static PyGetSetDef dirstate_item_getset[] = { |
|
738 | 738 | {"mode", (getter)dirstate_item_get_mode, NULL, "mode", NULL}, |
|
739 | 739 | {"size", (getter)dirstate_item_get_size, NULL, "size", NULL}, |
|
740 | 740 | {"mtime", (getter)dirstate_item_get_mtime, NULL, "mtime", NULL}, |
|
741 | 741 | {"state", (getter)dirstate_item_get_state, NULL, "state", NULL}, |
|
742 | 742 | {"has_fallback_exec", (getter)dirstate_item_get_has_fallback_exec, NULL, |
|
743 | 743 | "has_fallback_exec", NULL}, |
|
744 | 744 | {"fallback_exec", (getter)dirstate_item_get_fallback_exec, |
|
745 | 745 | (setter)dirstate_item_set_fallback_exec, "fallback_exec", NULL}, |
|
746 | 746 | {"has_fallback_symlink", (getter)dirstate_item_get_has_fallback_symlink, |
|
747 | 747 | NULL, "has_fallback_symlink", NULL}, |
|
748 | 748 | {"fallback_symlink", (getter)dirstate_item_get_fallback_symlink, |
|
749 | 749 | (setter)dirstate_item_set_fallback_symlink, "fallback_symlink", NULL}, |
|
750 | 750 | {"tracked", (getter)dirstate_item_get_tracked, NULL, "tracked", NULL}, |
|
751 | 751 | {"p1_tracked", (getter)dirstate_item_get_p1_tracked, NULL, "p1_tracked", |
|
752 | 752 | NULL}, |
|
753 | 753 | {"added", (getter)dirstate_item_get_added, NULL, "added", NULL}, |
|
754 | 754 | {"p2_info", (getter)dirstate_item_get_p2_info, NULL, "p2_info", NULL}, |
|
755 | 755 | {"merged", (getter)dirstate_item_get_merged, NULL, "merged", NULL}, |
|
756 | 756 | {"from_p2", (getter)dirstate_item_get_from_p2, NULL, "from_p2", NULL}, |
|
757 | 757 | {"maybe_clean", (getter)dirstate_item_get_maybe_clean, NULL, "maybe_clean", |
|
758 | 758 | NULL}, |
|
759 | 759 | {"any_tracked", (getter)dirstate_item_get_any_tracked, NULL, "any_tracked", |
|
760 | 760 | NULL}, |
|
761 | 761 | {"removed", (getter)dirstate_item_get_removed, NULL, "removed", NULL}, |
|
762 | 762 | {NULL} /* Sentinel */ |
|
763 | 763 | }; |
|
764 | 764 | |
|
765 | 765 | PyTypeObject dirstateItemType = { |
|
766 | 766 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ |
|
767 | 767 | "dirstate_tuple", /* tp_name */ |
|
768 | 768 | sizeof(dirstateItemObject), /* tp_basicsize */ |
|
769 | 769 | 0, /* tp_itemsize */ |
|
770 | 770 | (destructor)dirstate_item_dealloc, /* tp_dealloc */ |
|
771 | 771 | 0, /* tp_print */ |
|
772 | 772 | 0, /* tp_getattr */ |
|
773 | 773 | 0, /* tp_setattr */ |
|
774 | 774 | 0, /* tp_compare */ |
|
775 | 775 | 0, /* tp_repr */ |
|
776 | 776 | 0, /* tp_as_number */ |
|
777 | 777 | 0, /* tp_as_sequence */ |
|
778 | 778 | 0, /* tp_as_mapping */ |
|
779 | 779 | 0, /* tp_hash */ |
|
780 | 780 | 0, /* tp_call */ |
|
781 | 781 | 0, /* tp_str */ |
|
782 | 782 | 0, /* tp_getattro */ |
|
783 | 783 | 0, /* tp_setattro */ |
|
784 | 784 | 0, /* tp_as_buffer */ |
|
785 | 785 | Py_TPFLAGS_DEFAULT, /* tp_flags */ |
|
786 | 786 | "dirstate tuple", /* tp_doc */ |
|
787 | 787 | 0, /* tp_traverse */ |
|
788 | 788 | 0, /* tp_clear */ |
|
789 | 789 | 0, /* tp_richcompare */ |
|
790 | 790 | 0, /* tp_weaklistoffset */ |
|
791 | 791 | 0, /* tp_iter */ |
|
792 | 792 | 0, /* tp_iternext */ |
|
793 | 793 | dirstate_item_methods, /* tp_methods */ |
|
794 | 794 | 0, /* tp_members */ |
|
795 | 795 | dirstate_item_getset, /* tp_getset */ |
|
796 | 796 | 0, /* tp_base */ |
|
797 | 797 | 0, /* tp_dict */ |
|
798 | 798 | 0, /* tp_descr_get */ |
|
799 | 799 | 0, /* tp_descr_set */ |
|
800 | 800 | 0, /* tp_dictoffset */ |
|
801 | 801 | 0, /* tp_init */ |
|
802 | 802 | 0, /* tp_alloc */ |
|
803 | 803 | dirstate_item_new, /* tp_new */ |
|
804 | 804 | }; |
|
805 | 805 | |
|
806 | 806 | static PyObject *parse_dirstate(PyObject *self, PyObject *args) |
|
807 | 807 | { |
|
808 | 808 | PyObject *dmap, *cmap, *parents = NULL, *ret = NULL; |
|
809 | 809 | PyObject *fname = NULL, *cname = NULL, *entry = NULL; |
|
810 | 810 | char state, *cur, *str, *cpos; |
|
811 | 811 | int mode, size, mtime; |
|
812 | 812 | unsigned int flen, pos = 40; |
|
813 | 813 | Py_ssize_t len = 40; |
|
814 | 814 | Py_ssize_t readlen; |
|
815 | 815 | |
|
816 | if (!PyArg_ParseTuple( | |
|
817 | args, PY23("O!O!s#:parse_dirstate", "O!O!y#:parse_dirstate"), | |
|
818 | &PyDict_Type, &dmap, &PyDict_Type, &cmap, &str, &readlen)) { | |
|
816 | if (!PyArg_ParseTuple(args, "O!O!y#:parse_dirstate", &PyDict_Type, | |
|
817 | &dmap, &PyDict_Type, &cmap, &str, &readlen)) { | |
|
819 | 818 | goto quit; |
|
820 | 819 | } |
|
821 | 820 | |
|
822 | 821 | len = readlen; |
|
823 | 822 | |
|
824 | 823 | /* read parents */ |
|
825 | 824 | if (len < 40) { |
|
826 | 825 | PyErr_SetString(PyExc_ValueError, |
|
827 | 826 | "too little data for parents"); |
|
828 | 827 | goto quit; |
|
829 | 828 | } |
|
830 | 829 | |
|
831 |
parents = Py_BuildValue( |
|
|
832 |
|
|
|
830 | parents = Py_BuildValue("y#y#", str, (Py_ssize_t)20, str + 20, | |
|
831 | (Py_ssize_t)20); | |
|
833 | 832 | if (!parents) { |
|
834 | 833 | goto quit; |
|
835 | 834 | } |
|
836 | 835 | |
|
837 | 836 | /* read filenames */ |
|
838 | 837 | while (pos >= 40 && pos < len) { |
|
839 | 838 | if (pos + 17 > len) { |
|
840 | 839 | PyErr_SetString(PyExc_ValueError, |
|
841 | 840 | "overflow in dirstate"); |
|
842 | 841 | goto quit; |
|
843 | 842 | } |
|
844 | 843 | cur = str + pos; |
|
845 | 844 | /* unpack header */ |
|
846 | 845 | state = *cur; |
|
847 | 846 | mode = getbe32(cur + 1); |
|
848 | 847 | size = getbe32(cur + 5); |
|
849 | 848 | mtime = getbe32(cur + 9); |
|
850 | 849 | flen = getbe32(cur + 13); |
|
851 | 850 | pos += 17; |
|
852 | 851 | cur += 17; |
|
853 | 852 | if (flen > len - pos) { |
|
854 | 853 | PyErr_SetString(PyExc_ValueError, |
|
855 | 854 | "overflow in dirstate"); |
|
856 | 855 | goto quit; |
|
857 | 856 | } |
|
858 | 857 | |
|
859 | 858 | entry = (PyObject *)dirstate_item_from_v1_data(state, mode, |
|
860 | 859 | size, mtime); |
|
861 | 860 | if (!entry) |
|
862 | 861 | goto quit; |
|
863 | 862 | cpos = memchr(cur, 0, flen); |
|
864 | 863 | if (cpos) { |
|
865 | 864 | fname = PyBytes_FromStringAndSize(cur, cpos - cur); |
|
866 | 865 | cname = PyBytes_FromStringAndSize( |
|
867 | 866 | cpos + 1, flen - (cpos - cur) - 1); |
|
868 | 867 | if (!fname || !cname || |
|
869 | 868 | PyDict_SetItem(cmap, fname, cname) == -1 || |
|
870 | 869 | PyDict_SetItem(dmap, fname, entry) == -1) { |
|
871 | 870 | goto quit; |
|
872 | 871 | } |
|
873 | 872 | Py_DECREF(cname); |
|
874 | 873 | } else { |
|
875 | 874 | fname = PyBytes_FromStringAndSize(cur, flen); |
|
876 | 875 | if (!fname || |
|
877 | 876 | PyDict_SetItem(dmap, fname, entry) == -1) { |
|
878 | 877 | goto quit; |
|
879 | 878 | } |
|
880 | 879 | } |
|
881 | 880 | Py_DECREF(fname); |
|
882 | 881 | Py_DECREF(entry); |
|
883 | 882 | fname = cname = entry = NULL; |
|
884 | 883 | pos += flen; |
|
885 | 884 | } |
|
886 | 885 | |
|
887 | 886 | ret = parents; |
|
888 | 887 | Py_INCREF(ret); |
|
889 | 888 | quit: |
|
890 | 889 | Py_XDECREF(fname); |
|
891 | 890 | Py_XDECREF(cname); |
|
892 | 891 | Py_XDECREF(entry); |
|
893 | 892 | Py_XDECREF(parents); |
|
894 | 893 | return ret; |
|
895 | 894 | } |
|
896 | 895 | |
|
897 | 896 | /* |
|
898 | 897 | * Efficiently pack a dirstate object into its on-disk format. |
|
899 | 898 | */ |
|
900 | 899 | static PyObject *pack_dirstate(PyObject *self, PyObject *args) |
|
901 | 900 | { |
|
902 | 901 | PyObject *packobj = NULL; |
|
903 | 902 | PyObject *map, *copymap, *pl, *mtime_unset = NULL; |
|
904 | 903 | Py_ssize_t nbytes, pos, l; |
|
905 | 904 | PyObject *k, *v = NULL, *pn; |
|
906 | 905 | char *p, *s; |
|
907 | 906 | |
|
908 | 907 | if (!PyArg_ParseTuple(args, "O!O!O!:pack_dirstate", &PyDict_Type, &map, |
|
909 | 908 | &PyDict_Type, ©map, &PyTuple_Type, &pl)) { |
|
910 | 909 | return NULL; |
|
911 | 910 | } |
|
912 | 911 | |
|
913 | 912 | if (PyTuple_Size(pl) != 2) { |
|
914 | 913 | PyErr_SetString(PyExc_TypeError, "expected 2-element tuple"); |
|
915 | 914 | return NULL; |
|
916 | 915 | } |
|
917 | 916 | |
|
918 | 917 | /* Figure out how much we need to allocate. */ |
|
919 | 918 | for (nbytes = 40, pos = 0; PyDict_Next(map, &pos, &k, &v);) { |
|
920 | 919 | PyObject *c; |
|
921 | 920 | if (!PyBytes_Check(k)) { |
|
922 | 921 | PyErr_SetString(PyExc_TypeError, "expected string key"); |
|
923 | 922 | goto bail; |
|
924 | 923 | } |
|
925 | 924 | nbytes += PyBytes_GET_SIZE(k) + 17; |
|
926 | 925 | c = PyDict_GetItem(copymap, k); |
|
927 | 926 | if (c) { |
|
928 | 927 | if (!PyBytes_Check(c)) { |
|
929 | 928 | PyErr_SetString(PyExc_TypeError, |
|
930 | 929 | "expected string key"); |
|
931 | 930 | goto bail; |
|
932 | 931 | } |
|
933 | 932 | nbytes += PyBytes_GET_SIZE(c) + 1; |
|
934 | 933 | } |
|
935 | 934 | } |
|
936 | 935 | |
|
937 | 936 | packobj = PyBytes_FromStringAndSize(NULL, nbytes); |
|
938 | 937 | if (packobj == NULL) { |
|
939 | 938 | goto bail; |
|
940 | 939 | } |
|
941 | 940 | |
|
942 | 941 | p = PyBytes_AS_STRING(packobj); |
|
943 | 942 | |
|
944 | 943 | pn = PyTuple_GET_ITEM(pl, 0); |
|
945 | 944 | if (PyBytes_AsStringAndSize(pn, &s, &l) == -1 || l != 20) { |
|
946 | 945 | PyErr_SetString(PyExc_TypeError, "expected a 20-byte hash"); |
|
947 | 946 | goto bail; |
|
948 | 947 | } |
|
949 | 948 | memcpy(p, s, l); |
|
950 | 949 | p += 20; |
|
951 | 950 | pn = PyTuple_GET_ITEM(pl, 1); |
|
952 | 951 | if (PyBytes_AsStringAndSize(pn, &s, &l) == -1 || l != 20) { |
|
953 | 952 | PyErr_SetString(PyExc_TypeError, "expected a 20-byte hash"); |
|
954 | 953 | goto bail; |
|
955 | 954 | } |
|
956 | 955 | memcpy(p, s, l); |
|
957 | 956 | p += 20; |
|
958 | 957 | |
|
959 | 958 | for (pos = 0; PyDict_Next(map, &pos, &k, &v);) { |
|
960 | 959 | dirstateItemObject *tuple; |
|
961 | 960 | char state; |
|
962 | 961 | int mode, size, mtime; |
|
963 | 962 | Py_ssize_t len, l; |
|
964 | 963 | PyObject *o; |
|
965 | 964 | char *t; |
|
966 | 965 | |
|
967 | 966 | if (!dirstate_tuple_check(v)) { |
|
968 | 967 | PyErr_SetString(PyExc_TypeError, |
|
969 | 968 | "expected a dirstate tuple"); |
|
970 | 969 | goto bail; |
|
971 | 970 | } |
|
972 | 971 | tuple = (dirstateItemObject *)v; |
|
973 | 972 | |
|
974 | 973 | state = dirstate_item_c_v1_state(tuple); |
|
975 | 974 | mode = dirstate_item_c_v1_mode(tuple); |
|
976 | 975 | size = dirstate_item_c_v1_size(tuple); |
|
977 | 976 | mtime = dirstate_item_c_v1_mtime(tuple); |
|
978 | 977 | *p++ = state; |
|
979 | 978 | putbe32((uint32_t)mode, p); |
|
980 | 979 | putbe32((uint32_t)size, p + 4); |
|
981 | 980 | putbe32((uint32_t)mtime, p + 8); |
|
982 | 981 | t = p + 12; |
|
983 | 982 | p += 16; |
|
984 | 983 | len = PyBytes_GET_SIZE(k); |
|
985 | 984 | memcpy(p, PyBytes_AS_STRING(k), len); |
|
986 | 985 | p += len; |
|
987 | 986 | o = PyDict_GetItem(copymap, k); |
|
988 | 987 | if (o) { |
|
989 | 988 | *p++ = '\0'; |
|
990 | 989 | l = PyBytes_GET_SIZE(o); |
|
991 | 990 | memcpy(p, PyBytes_AS_STRING(o), l); |
|
992 | 991 | p += l; |
|
993 | 992 | len += l + 1; |
|
994 | 993 | } |
|
995 | 994 | putbe32((uint32_t)len, t); |
|
996 | 995 | } |
|
997 | 996 | |
|
998 | 997 | pos = p - PyBytes_AS_STRING(packobj); |
|
999 | 998 | if (pos != nbytes) { |
|
1000 | 999 | PyErr_Format(PyExc_SystemError, "bad dirstate size: %ld != %ld", |
|
1001 | 1000 | (long)pos, (long)nbytes); |
|
1002 | 1001 | goto bail; |
|
1003 | 1002 | } |
|
1004 | 1003 | |
|
1005 | 1004 | return packobj; |
|
1006 | 1005 | bail: |
|
1007 | 1006 | Py_XDECREF(mtime_unset); |
|
1008 | 1007 | Py_XDECREF(packobj); |
|
1009 | 1008 | Py_XDECREF(v); |
|
1010 | 1009 | return NULL; |
|
1011 | 1010 | } |
|
1012 | 1011 | |
|
1013 | 1012 | #define BUMPED_FIX 1 |
|
1014 | 1013 | #define USING_SHA_256 2 |
|
1015 | 1014 | #define FM1_HEADER_SIZE (4 + 8 + 2 + 2 + 1 + 1 + 1) |
|
1016 | 1015 | |
|
1017 | 1016 | static PyObject *readshas(const char *source, unsigned char num, |
|
1018 | 1017 | Py_ssize_t hashwidth) |
|
1019 | 1018 | { |
|
1020 | 1019 | int i; |
|
1021 | 1020 | PyObject *list = PyTuple_New(num); |
|
1022 | 1021 | if (list == NULL) { |
|
1023 | 1022 | return NULL; |
|
1024 | 1023 | } |
|
1025 | 1024 | for (i = 0; i < num; i++) { |
|
1026 | 1025 | PyObject *hash = PyBytes_FromStringAndSize(source, hashwidth); |
|
1027 | 1026 | if (hash == NULL) { |
|
1028 | 1027 | Py_DECREF(list); |
|
1029 | 1028 | return NULL; |
|
1030 | 1029 | } |
|
1031 | 1030 | PyTuple_SET_ITEM(list, i, hash); |
|
1032 | 1031 | source += hashwidth; |
|
1033 | 1032 | } |
|
1034 | 1033 | return list; |
|
1035 | 1034 | } |
|
1036 | 1035 | |
|
1037 | 1036 | static PyObject *fm1readmarker(const char *databegin, const char *dataend, |
|
1038 | 1037 | uint32_t *msize) |
|
1039 | 1038 | { |
|
1040 | 1039 | const char *data = databegin; |
|
1041 | 1040 | const char *meta; |
|
1042 | 1041 | |
|
1043 | 1042 | double mtime; |
|
1044 | 1043 | int16_t tz; |
|
1045 | 1044 | uint16_t flags; |
|
1046 | 1045 | unsigned char nsuccs, nparents, nmetadata; |
|
1047 | 1046 | Py_ssize_t hashwidth = 20; |
|
1048 | 1047 | |
|
1049 | 1048 | PyObject *prec = NULL, *parents = NULL, *succs = NULL; |
|
1050 | 1049 | PyObject *metadata = NULL, *ret = NULL; |
|
1051 | 1050 | int i; |
|
1052 | 1051 | |
|
1053 | 1052 | if (data + FM1_HEADER_SIZE > dataend) { |
|
1054 | 1053 | goto overflow; |
|
1055 | 1054 | } |
|
1056 | 1055 | |
|
1057 | 1056 | *msize = getbe32(data); |
|
1058 | 1057 | data += 4; |
|
1059 | 1058 | mtime = getbefloat64(data); |
|
1060 | 1059 | data += 8; |
|
1061 | 1060 | tz = getbeint16(data); |
|
1062 | 1061 | data += 2; |
|
1063 | 1062 | flags = getbeuint16(data); |
|
1064 | 1063 | data += 2; |
|
1065 | 1064 | |
|
1066 | 1065 | if (flags & USING_SHA_256) { |
|
1067 | 1066 | hashwidth = 32; |
|
1068 | 1067 | } |
|
1069 | 1068 | |
|
1070 | 1069 | nsuccs = (unsigned char)(*data++); |
|
1071 | 1070 | nparents = (unsigned char)(*data++); |
|
1072 | 1071 | nmetadata = (unsigned char)(*data++); |
|
1073 | 1072 | |
|
1074 | 1073 | if (databegin + *msize > dataend) { |
|
1075 | 1074 | goto overflow; |
|
1076 | 1075 | } |
|
1077 | 1076 | dataend = databegin + *msize; /* narrow down to marker size */ |
|
1078 | 1077 | |
|
1079 | 1078 | if (data + hashwidth > dataend) { |
|
1080 | 1079 | goto overflow; |
|
1081 | 1080 | } |
|
1082 | 1081 | prec = PyBytes_FromStringAndSize(data, hashwidth); |
|
1083 | 1082 | data += hashwidth; |
|
1084 | 1083 | if (prec == NULL) { |
|
1085 | 1084 | goto bail; |
|
1086 | 1085 | } |
|
1087 | 1086 | |
|
1088 | 1087 | if (data + nsuccs * hashwidth > dataend) { |
|
1089 | 1088 | goto overflow; |
|
1090 | 1089 | } |
|
1091 | 1090 | succs = readshas(data, nsuccs, hashwidth); |
|
1092 | 1091 | if (succs == NULL) { |
|
1093 | 1092 | goto bail; |
|
1094 | 1093 | } |
|
1095 | 1094 | data += nsuccs * hashwidth; |
|
1096 | 1095 | |
|
1097 | 1096 | if (nparents == 1 || nparents == 2) { |
|
1098 | 1097 | if (data + nparents * hashwidth > dataend) { |
|
1099 | 1098 | goto overflow; |
|
1100 | 1099 | } |
|
1101 | 1100 | parents = readshas(data, nparents, hashwidth); |
|
1102 | 1101 | if (parents == NULL) { |
|
1103 | 1102 | goto bail; |
|
1104 | 1103 | } |
|
1105 | 1104 | data += nparents * hashwidth; |
|
1106 | 1105 | } else { |
|
1107 | 1106 | parents = Py_None; |
|
1108 | 1107 | Py_INCREF(parents); |
|
1109 | 1108 | } |
|
1110 | 1109 | |
|
1111 | 1110 | if (data + 2 * nmetadata > dataend) { |
|
1112 | 1111 | goto overflow; |
|
1113 | 1112 | } |
|
1114 | 1113 | meta = data + (2 * nmetadata); |
|
1115 | 1114 | metadata = PyTuple_New(nmetadata); |
|
1116 | 1115 | if (metadata == NULL) { |
|
1117 | 1116 | goto bail; |
|
1118 | 1117 | } |
|
1119 | 1118 | for (i = 0; i < nmetadata; i++) { |
|
1120 | 1119 | PyObject *tmp, *left = NULL, *right = NULL; |
|
1121 | 1120 | Py_ssize_t leftsize = (unsigned char)(*data++); |
|
1122 | 1121 | Py_ssize_t rightsize = (unsigned char)(*data++); |
|
1123 | 1122 | if (meta + leftsize + rightsize > dataend) { |
|
1124 | 1123 | goto overflow; |
|
1125 | 1124 | } |
|
1126 | 1125 | left = PyBytes_FromStringAndSize(meta, leftsize); |
|
1127 | 1126 | meta += leftsize; |
|
1128 | 1127 | right = PyBytes_FromStringAndSize(meta, rightsize); |
|
1129 | 1128 | meta += rightsize; |
|
1130 | 1129 | tmp = PyTuple_New(2); |
|
1131 | 1130 | if (!left || !right || !tmp) { |
|
1132 | 1131 | Py_XDECREF(left); |
|
1133 | 1132 | Py_XDECREF(right); |
|
1134 | 1133 | Py_XDECREF(tmp); |
|
1135 | 1134 | goto bail; |
|
1136 | 1135 | } |
|
1137 | 1136 | PyTuple_SET_ITEM(tmp, 0, left); |
|
1138 | 1137 | PyTuple_SET_ITEM(tmp, 1, right); |
|
1139 | 1138 | PyTuple_SET_ITEM(metadata, i, tmp); |
|
1140 | 1139 | } |
|
1141 | 1140 | ret = Py_BuildValue("(OOHO(di)O)", prec, succs, flags, metadata, mtime, |
|
1142 | 1141 | (int)tz * 60, parents); |
|
1143 | 1142 | goto bail; /* return successfully */ |
|
1144 | 1143 | |
|
1145 | 1144 | overflow: |
|
1146 | 1145 | PyErr_SetString(PyExc_ValueError, "overflow in obsstore"); |
|
1147 | 1146 | bail: |
|
1148 | 1147 | Py_XDECREF(prec); |
|
1149 | 1148 | Py_XDECREF(succs); |
|
1150 | 1149 | Py_XDECREF(metadata); |
|
1151 | 1150 | Py_XDECREF(parents); |
|
1152 | 1151 | return ret; |
|
1153 | 1152 | } |
|
1154 | 1153 | |
|
1155 | 1154 | static PyObject *fm1readmarkers(PyObject *self, PyObject *args) |
|
1156 | 1155 | { |
|
1157 | 1156 | const char *data, *dataend; |
|
1158 | 1157 | Py_ssize_t datalen, offset, stop; |
|
1159 | 1158 | PyObject *markers = NULL; |
|
1160 | 1159 | |
|
1161 |
if (!PyArg_ParseTuple(args, |
|
|
1162 | &offset, &stop)) { | |
|
1160 | if (!PyArg_ParseTuple(args, "y#nn", &data, &datalen, &offset, &stop)) { | |
|
1163 | 1161 | return NULL; |
|
1164 | 1162 | } |
|
1165 | 1163 | if (offset < 0) { |
|
1166 | 1164 | PyErr_SetString(PyExc_ValueError, |
|
1167 | 1165 | "invalid negative offset in fm1readmarkers"); |
|
1168 | 1166 | return NULL; |
|
1169 | 1167 | } |
|
1170 | 1168 | if (stop > datalen) { |
|
1171 | 1169 | PyErr_SetString( |
|
1172 | 1170 | PyExc_ValueError, |
|
1173 | 1171 | "stop longer than data length in fm1readmarkers"); |
|
1174 | 1172 | return NULL; |
|
1175 | 1173 | } |
|
1176 | 1174 | dataend = data + datalen; |
|
1177 | 1175 | data += offset; |
|
1178 | 1176 | markers = PyList_New(0); |
|
1179 | 1177 | if (!markers) { |
|
1180 | 1178 | return NULL; |
|
1181 | 1179 | } |
|
1182 | 1180 | while (offset < stop) { |
|
1183 | 1181 | uint32_t msize; |
|
1184 | 1182 | int error; |
|
1185 | 1183 | PyObject *record = fm1readmarker(data, dataend, &msize); |
|
1186 | 1184 | if (!record) { |
|
1187 | 1185 | goto bail; |
|
1188 | 1186 | } |
|
1189 | 1187 | error = PyList_Append(markers, record); |
|
1190 | 1188 | Py_DECREF(record); |
|
1191 | 1189 | if (error) { |
|
1192 | 1190 | goto bail; |
|
1193 | 1191 | } |
|
1194 | 1192 | data += msize; |
|
1195 | 1193 | offset += msize; |
|
1196 | 1194 | } |
|
1197 | 1195 | return markers; |
|
1198 | 1196 | bail: |
|
1199 | 1197 | Py_DECREF(markers); |
|
1200 | 1198 | return NULL; |
|
1201 | 1199 | } |
|
1202 | 1200 | |
|
1203 | 1201 | static char parsers_doc[] = "Efficient content parsing."; |
|
1204 | 1202 | |
|
1205 | 1203 | PyObject *encodedir(PyObject *self, PyObject *args); |
|
1206 | 1204 | PyObject *pathencode(PyObject *self, PyObject *args); |
|
1207 | 1205 | PyObject *lowerencode(PyObject *self, PyObject *args); |
|
1208 | 1206 | PyObject *parse_index2(PyObject *self, PyObject *args, PyObject *kwargs); |
|
1209 | 1207 | |
|
1210 | 1208 | static PyMethodDef methods[] = { |
|
1211 | 1209 | {"pack_dirstate", pack_dirstate, METH_VARARGS, "pack a dirstate\n"}, |
|
1212 | 1210 | {"parse_dirstate", parse_dirstate, METH_VARARGS, "parse a dirstate\n"}, |
|
1213 | 1211 | {"parse_index2", (PyCFunction)parse_index2, METH_VARARGS | METH_KEYWORDS, |
|
1214 | 1212 | "parse a revlog index\n"}, |
|
1215 | 1213 | {"isasciistr", isasciistr, METH_VARARGS, "check if an ASCII string\n"}, |
|
1216 | 1214 | {"asciilower", asciilower, METH_VARARGS, "lowercase an ASCII string\n"}, |
|
1217 | 1215 | {"asciiupper", asciiupper, METH_VARARGS, "uppercase an ASCII string\n"}, |
|
1218 | 1216 | {"dict_new_presized", dict_new_presized, METH_VARARGS, |
|
1219 | 1217 | "construct a dict with an expected size\n"}, |
|
1220 | 1218 | {"make_file_foldmap", make_file_foldmap, METH_VARARGS, |
|
1221 | 1219 | "make file foldmap\n"}, |
|
1222 | 1220 | {"jsonescapeu8fast", jsonescapeu8fast, METH_VARARGS, |
|
1223 | 1221 | "escape a UTF-8 byte string to JSON (fast path)\n"}, |
|
1224 | 1222 | {"encodedir", encodedir, METH_VARARGS, "encodedir a path\n"}, |
|
1225 | 1223 | {"pathencode", pathencode, METH_VARARGS, "fncache-encode a path\n"}, |
|
1226 | 1224 | {"lowerencode", lowerencode, METH_VARARGS, "lower-encode a path\n"}, |
|
1227 | 1225 | {"fm1readmarkers", fm1readmarkers, METH_VARARGS, |
|
1228 | 1226 | "parse v1 obsolete markers\n"}, |
|
1229 | 1227 | {NULL, NULL}}; |
|
1230 | 1228 | |
|
1231 | 1229 | void dirs_module_init(PyObject *mod); |
|
1232 | 1230 | void manifest_module_init(PyObject *mod); |
|
1233 | 1231 | void revlog_module_init(PyObject *mod); |
|
1234 | 1232 | |
|
1235 | 1233 | static const int version = 20; |
|
1236 | 1234 | |
|
1237 | 1235 | static void module_init(PyObject *mod) |
|
1238 | 1236 | { |
|
1239 | 1237 | PyModule_AddIntConstant(mod, "version", version); |
|
1240 | 1238 | |
|
1241 | 1239 | /* This module constant has two purposes. First, it lets us unit test |
|
1242 | 1240 | * the ImportError raised without hard-coding any error text. This |
|
1243 | 1241 | * means we can change the text in the future without breaking tests, |
|
1244 | 1242 | * even across changesets without a recompile. Second, its presence |
|
1245 | 1243 | * can be used to determine whether the version-checking logic is |
|
1246 | 1244 | * present, which also helps in testing across changesets without a |
|
1247 | 1245 | * recompile. Note that this means the pure-Python version of parsers |
|
1248 | 1246 | * should not have this module constant. */ |
|
1249 | 1247 | PyModule_AddStringConstant(mod, "versionerrortext", versionerrortext); |
|
1250 | 1248 | |
|
1251 | 1249 | dirs_module_init(mod); |
|
1252 | 1250 | manifest_module_init(mod); |
|
1253 | 1251 | revlog_module_init(mod); |
|
1254 | 1252 | |
|
1255 | 1253 | if (PyType_Ready(&dirstateItemType) < 0) { |
|
1256 | 1254 | return; |
|
1257 | 1255 | } |
|
1258 | 1256 | Py_INCREF(&dirstateItemType); |
|
1259 | 1257 | PyModule_AddObject(mod, "DirstateItem", (PyObject *)&dirstateItemType); |
|
1260 | 1258 | } |
|
1261 | 1259 | |
|
1262 | 1260 | static int check_python_version(void) |
|
1263 | 1261 | { |
|
1264 | 1262 | PyObject *sys = PyImport_ImportModule("sys"), *ver; |
|
1265 | 1263 | long hexversion; |
|
1266 | 1264 | if (!sys) { |
|
1267 | 1265 | return -1; |
|
1268 | 1266 | } |
|
1269 | 1267 | ver = PyObject_GetAttrString(sys, "hexversion"); |
|
1270 | 1268 | Py_DECREF(sys); |
|
1271 | 1269 | if (!ver) { |
|
1272 | 1270 | return -1; |
|
1273 | 1271 | } |
|
1274 | 1272 | hexversion = PyLong_AsLong(ver); |
|
1275 | 1273 | Py_DECREF(ver); |
|
1276 | 1274 | /* sys.hexversion is a 32-bit number by default, so the -1 case |
|
1277 | 1275 | * should only occur in unusual circumstances (e.g. if sys.hexversion |
|
1278 | 1276 | * is manually set to an invalid value). */ |
|
1279 | 1277 | if ((hexversion == -1) || (hexversion >> 16 != PY_VERSION_HEX >> 16)) { |
|
1280 | 1278 | PyErr_Format(PyExc_ImportError, |
|
1281 | 1279 | "%s: The Mercurial extension " |
|
1282 | 1280 | "modules were compiled with Python " PY_VERSION |
|
1283 | 1281 | ", but " |
|
1284 | 1282 | "Mercurial is currently using Python with " |
|
1285 | 1283 | "sys.hexversion=%ld: " |
|
1286 | 1284 | "Python %s\n at: %s", |
|
1287 | 1285 | versionerrortext, hexversion, Py_GetVersion(), |
|
1288 | 1286 | Py_GetProgramFullPath()); |
|
1289 | 1287 | return -1; |
|
1290 | 1288 | } |
|
1291 | 1289 | return 0; |
|
1292 | 1290 | } |
|
1293 | 1291 | |
|
1294 | 1292 | static struct PyModuleDef parsers_module = {PyModuleDef_HEAD_INIT, "parsers", |
|
1295 | 1293 | parsers_doc, -1, methods}; |
|
1296 | 1294 | |
|
1297 | 1295 | PyMODINIT_FUNC PyInit_parsers(void) |
|
1298 | 1296 | { |
|
1299 | 1297 | PyObject *mod; |
|
1300 | 1298 | |
|
1301 | 1299 | if (check_python_version() == -1) |
|
1302 | 1300 | return NULL; |
|
1303 | 1301 | mod = PyModule_Create(&parsers_module); |
|
1304 | 1302 | module_init(mod); |
|
1305 | 1303 | return mod; |
|
1306 | 1304 | } |
@@ -1,803 +1,802 b'' | |||
|
1 | 1 | /* |
|
2 | 2 | pathencode.c - efficient path name encoding |
|
3 | 3 | |
|
4 | 4 | Copyright 2012 Facebook |
|
5 | 5 | |
|
6 | 6 | This software may be used and distributed according to the terms of |
|
7 | 7 | the GNU General Public License, incorporated herein by reference. |
|
8 | 8 | */ |
|
9 | 9 | |
|
10 | 10 | /* |
|
11 | 11 | * An implementation of the name encoding scheme used by the fncache |
|
12 | 12 | * store. The common case is of a path < 120 bytes long, which is |
|
13 | 13 | * handled either in a single pass with no allocations or two passes |
|
14 | 14 | * with a single allocation. For longer paths, multiple passes are |
|
15 | 15 | * required. |
|
16 | 16 | */ |
|
17 | 17 | |
|
18 | 18 | #define PY_SSIZE_T_CLEAN |
|
19 | 19 | #include <Python.h> |
|
20 | 20 | #include <assert.h> |
|
21 | 21 | #include <ctype.h> |
|
22 | 22 | #include <stdlib.h> |
|
23 | 23 | #include <string.h> |
|
24 | 24 | #include "pythoncapi_compat.h" |
|
25 | 25 | |
|
26 | 26 | #include "util.h" |
|
27 | 27 | |
|
28 | 28 | /* state machine for the fast path */ |
|
29 | 29 | enum path_state { |
|
30 | 30 | START, /* first byte of a path component */ |
|
31 | 31 | A, /* "AUX" */ |
|
32 | 32 | AU, |
|
33 | 33 | THIRD, /* third of a 3-byte sequence, e.g. "AUX", "NUL" */ |
|
34 | 34 | C, /* "CON" or "COMn" */ |
|
35 | 35 | CO, |
|
36 | 36 | COMLPT, /* "COM" or "LPT" */ |
|
37 | 37 | COMLPTn, |
|
38 | 38 | L, |
|
39 | 39 | LP, |
|
40 | 40 | N, |
|
41 | 41 | NU, |
|
42 | 42 | P, /* "PRN" */ |
|
43 | 43 | PR, |
|
44 | 44 | LDOT, /* leading '.' */ |
|
45 | 45 | DOT, /* '.' in a non-leading position */ |
|
46 | 46 | H, /* ".h" */ |
|
47 | 47 | HGDI, /* ".hg", ".d", or ".i" */ |
|
48 | 48 | SPACE, |
|
49 | 49 | DEFAULT, /* byte of a path component after the first */ |
|
50 | 50 | }; |
|
51 | 51 | |
|
52 | 52 | /* state machine for dir-encoding */ |
|
53 | 53 | enum dir_state { |
|
54 | 54 | DDOT, |
|
55 | 55 | DH, |
|
56 | 56 | DHGDI, |
|
57 | 57 | DDEFAULT, |
|
58 | 58 | }; |
|
59 | 59 | |
|
60 | 60 | static inline int inset(const uint32_t bitset[], char c) |
|
61 | 61 | { |
|
62 | 62 | return bitset[((uint8_t)c) >> 5] & (1 << (((uint8_t)c) & 31)); |
|
63 | 63 | } |
|
64 | 64 | |
|
65 | 65 | static inline void charcopy(char *dest, Py_ssize_t *destlen, size_t destsize, |
|
66 | 66 | char c) |
|
67 | 67 | { |
|
68 | 68 | if (dest) { |
|
69 | 69 | assert(*destlen < destsize); |
|
70 | 70 | dest[*destlen] = c; |
|
71 | 71 | } |
|
72 | 72 | (*destlen)++; |
|
73 | 73 | } |
|
74 | 74 | |
|
75 | 75 | static inline void memcopy(char *dest, Py_ssize_t *destlen, size_t destsize, |
|
76 | 76 | const void *src, Py_ssize_t len) |
|
77 | 77 | { |
|
78 | 78 | if (dest) { |
|
79 | 79 | assert(*destlen + len < destsize); |
|
80 | 80 | memcpy((void *)&dest[*destlen], src, len); |
|
81 | 81 | } |
|
82 | 82 | *destlen += len; |
|
83 | 83 | } |
|
84 | 84 | |
|
85 | 85 | static inline void hexencode(char *dest, Py_ssize_t *destlen, size_t destsize, |
|
86 | 86 | uint8_t c) |
|
87 | 87 | { |
|
88 | 88 | static const char hexdigit[] = "0123456789abcdef"; |
|
89 | 89 | |
|
90 | 90 | charcopy(dest, destlen, destsize, hexdigit[c >> 4]); |
|
91 | 91 | charcopy(dest, destlen, destsize, hexdigit[c & 15]); |
|
92 | 92 | } |
|
93 | 93 | |
|
94 | 94 | /* 3-byte escape: tilde followed by two hex digits */ |
|
95 | 95 | static inline void escape3(char *dest, Py_ssize_t *destlen, size_t destsize, |
|
96 | 96 | char c) |
|
97 | 97 | { |
|
98 | 98 | charcopy(dest, destlen, destsize, '~'); |
|
99 | 99 | hexencode(dest, destlen, destsize, c); |
|
100 | 100 | } |
|
101 | 101 | |
|
102 | 102 | static Py_ssize_t _encodedir(char *dest, size_t destsize, const char *src, |
|
103 | 103 | Py_ssize_t len) |
|
104 | 104 | { |
|
105 | 105 | enum dir_state state = DDEFAULT; |
|
106 | 106 | Py_ssize_t i = 0, destlen = 0; |
|
107 | 107 | |
|
108 | 108 | while (i < len) { |
|
109 | 109 | switch (state) { |
|
110 | 110 | case DDOT: |
|
111 | 111 | switch (src[i]) { |
|
112 | 112 | case 'd': |
|
113 | 113 | case 'i': |
|
114 | 114 | state = DHGDI; |
|
115 | 115 | charcopy(dest, &destlen, destsize, src[i++]); |
|
116 | 116 | break; |
|
117 | 117 | case 'h': |
|
118 | 118 | state = DH; |
|
119 | 119 | charcopy(dest, &destlen, destsize, src[i++]); |
|
120 | 120 | break; |
|
121 | 121 | default: |
|
122 | 122 | state = DDEFAULT; |
|
123 | 123 | break; |
|
124 | 124 | } |
|
125 | 125 | break; |
|
126 | 126 | case DH: |
|
127 | 127 | if (src[i] == 'g') { |
|
128 | 128 | state = DHGDI; |
|
129 | 129 | charcopy(dest, &destlen, destsize, src[i++]); |
|
130 | 130 | } else { |
|
131 | 131 | state = DDEFAULT; |
|
132 | 132 | } |
|
133 | 133 | break; |
|
134 | 134 | case DHGDI: |
|
135 | 135 | if (src[i] == '/') { |
|
136 | 136 | memcopy(dest, &destlen, destsize, ".hg", 3); |
|
137 | 137 | charcopy(dest, &destlen, destsize, src[i++]); |
|
138 | 138 | } |
|
139 | 139 | state = DDEFAULT; |
|
140 | 140 | break; |
|
141 | 141 | case DDEFAULT: |
|
142 | 142 | if (src[i] == '.') { |
|
143 | 143 | state = DDOT; |
|
144 | 144 | } |
|
145 | 145 | charcopy(dest, &destlen, destsize, src[i++]); |
|
146 | 146 | break; |
|
147 | 147 | } |
|
148 | 148 | } |
|
149 | 149 | |
|
150 | 150 | return destlen; |
|
151 | 151 | } |
|
152 | 152 | |
|
153 | 153 | PyObject *encodedir(PyObject *self, PyObject *args) |
|
154 | 154 | { |
|
155 | 155 | Py_ssize_t len, newlen; |
|
156 | 156 | PyObject *pathobj, *newobj; |
|
157 | 157 | char *path; |
|
158 | 158 | |
|
159 | 159 | if (!PyArg_ParseTuple(args, "O:encodedir", &pathobj)) { |
|
160 | 160 | return NULL; |
|
161 | 161 | } |
|
162 | 162 | |
|
163 | 163 | if (PyBytes_AsStringAndSize(pathobj, &path, &len) == -1) { |
|
164 | 164 | PyErr_SetString(PyExc_TypeError, "expected a string"); |
|
165 | 165 | return NULL; |
|
166 | 166 | } |
|
167 | 167 | |
|
168 | 168 | newlen = len ? _encodedir(NULL, 0, path, len + 1) : 1; |
|
169 | 169 | |
|
170 | 170 | if (newlen == len + 1) { |
|
171 | 171 | Py_INCREF(pathobj); |
|
172 | 172 | return pathobj; |
|
173 | 173 | } |
|
174 | 174 | |
|
175 | 175 | newobj = PyBytes_FromStringAndSize(NULL, newlen); |
|
176 | 176 | |
|
177 | 177 | if (newobj) { |
|
178 | 178 | assert(PyBytes_Check(newobj)); |
|
179 | 179 | Py_SET_SIZE(newobj, Py_SIZE(newobj) - 1); |
|
180 | 180 | _encodedir(PyBytes_AS_STRING(newobj), newlen, path, len + 1); |
|
181 | 181 | } |
|
182 | 182 | |
|
183 | 183 | return newobj; |
|
184 | 184 | } |
|
185 | 185 | |
|
186 | 186 | static Py_ssize_t _encode(const uint32_t twobytes[8], const uint32_t onebyte[8], |
|
187 | 187 | char *dest, Py_ssize_t destlen, size_t destsize, |
|
188 | 188 | const char *src, Py_ssize_t len, int encodedir) |
|
189 | 189 | { |
|
190 | 190 | enum path_state state = START; |
|
191 | 191 | Py_ssize_t i = 0; |
|
192 | 192 | |
|
193 | 193 | /* |
|
194 | 194 | * Python strings end with a zero byte, which we use as a |
|
195 | 195 | * terminal token as they are not valid inside path names. |
|
196 | 196 | */ |
|
197 | 197 | |
|
198 | 198 | while (i < len) { |
|
199 | 199 | switch (state) { |
|
200 | 200 | case START: |
|
201 | 201 | switch (src[i]) { |
|
202 | 202 | case '/': |
|
203 | 203 | charcopy(dest, &destlen, destsize, src[i++]); |
|
204 | 204 | break; |
|
205 | 205 | case '.': |
|
206 | 206 | state = LDOT; |
|
207 | 207 | escape3(dest, &destlen, destsize, src[i++]); |
|
208 | 208 | break; |
|
209 | 209 | case ' ': |
|
210 | 210 | state = DEFAULT; |
|
211 | 211 | escape3(dest, &destlen, destsize, src[i++]); |
|
212 | 212 | break; |
|
213 | 213 | case 'a': |
|
214 | 214 | state = A; |
|
215 | 215 | charcopy(dest, &destlen, destsize, src[i++]); |
|
216 | 216 | break; |
|
217 | 217 | case 'c': |
|
218 | 218 | state = C; |
|
219 | 219 | charcopy(dest, &destlen, destsize, src[i++]); |
|
220 | 220 | break; |
|
221 | 221 | case 'l': |
|
222 | 222 | state = L; |
|
223 | 223 | charcopy(dest, &destlen, destsize, src[i++]); |
|
224 | 224 | break; |
|
225 | 225 | case 'n': |
|
226 | 226 | state = N; |
|
227 | 227 | charcopy(dest, &destlen, destsize, src[i++]); |
|
228 | 228 | break; |
|
229 | 229 | case 'p': |
|
230 | 230 | state = P; |
|
231 | 231 | charcopy(dest, &destlen, destsize, src[i++]); |
|
232 | 232 | break; |
|
233 | 233 | default: |
|
234 | 234 | state = DEFAULT; |
|
235 | 235 | break; |
|
236 | 236 | } |
|
237 | 237 | break; |
|
238 | 238 | case A: |
|
239 | 239 | if (src[i] == 'u') { |
|
240 | 240 | state = AU; |
|
241 | 241 | charcopy(dest, &destlen, destsize, src[i++]); |
|
242 | 242 | } else { |
|
243 | 243 | state = DEFAULT; |
|
244 | 244 | } |
|
245 | 245 | break; |
|
246 | 246 | case AU: |
|
247 | 247 | if (src[i] == 'x') { |
|
248 | 248 | state = THIRD; |
|
249 | 249 | i++; |
|
250 | 250 | } else { |
|
251 | 251 | state = DEFAULT; |
|
252 | 252 | } |
|
253 | 253 | break; |
|
254 | 254 | case THIRD: |
|
255 | 255 | state = DEFAULT; |
|
256 | 256 | switch (src[i]) { |
|
257 | 257 | case '.': |
|
258 | 258 | case '/': |
|
259 | 259 | case '\0': |
|
260 | 260 | escape3(dest, &destlen, destsize, src[i - 1]); |
|
261 | 261 | break; |
|
262 | 262 | default: |
|
263 | 263 | i--; |
|
264 | 264 | break; |
|
265 | 265 | } |
|
266 | 266 | break; |
|
267 | 267 | case C: |
|
268 | 268 | if (src[i] == 'o') { |
|
269 | 269 | state = CO; |
|
270 | 270 | charcopy(dest, &destlen, destsize, src[i++]); |
|
271 | 271 | } else { |
|
272 | 272 | state = DEFAULT; |
|
273 | 273 | } |
|
274 | 274 | break; |
|
275 | 275 | case CO: |
|
276 | 276 | if (src[i] == 'm') { |
|
277 | 277 | state = COMLPT; |
|
278 | 278 | i++; |
|
279 | 279 | } else if (src[i] == 'n') { |
|
280 | 280 | state = THIRD; |
|
281 | 281 | i++; |
|
282 | 282 | } else { |
|
283 | 283 | state = DEFAULT; |
|
284 | 284 | } |
|
285 | 285 | break; |
|
286 | 286 | case COMLPT: |
|
287 | 287 | switch (src[i]) { |
|
288 | 288 | case '1': |
|
289 | 289 | case '2': |
|
290 | 290 | case '3': |
|
291 | 291 | case '4': |
|
292 | 292 | case '5': |
|
293 | 293 | case '6': |
|
294 | 294 | case '7': |
|
295 | 295 | case '8': |
|
296 | 296 | case '9': |
|
297 | 297 | state = COMLPTn; |
|
298 | 298 | i++; |
|
299 | 299 | break; |
|
300 | 300 | default: |
|
301 | 301 | state = DEFAULT; |
|
302 | 302 | charcopy(dest, &destlen, destsize, src[i - 1]); |
|
303 | 303 | break; |
|
304 | 304 | } |
|
305 | 305 | break; |
|
306 | 306 | case COMLPTn: |
|
307 | 307 | state = DEFAULT; |
|
308 | 308 | switch (src[i]) { |
|
309 | 309 | case '.': |
|
310 | 310 | case '/': |
|
311 | 311 | case '\0': |
|
312 | 312 | escape3(dest, &destlen, destsize, src[i - 2]); |
|
313 | 313 | charcopy(dest, &destlen, destsize, src[i - 1]); |
|
314 | 314 | break; |
|
315 | 315 | default: |
|
316 | 316 | memcopy(dest, &destlen, destsize, &src[i - 2], |
|
317 | 317 | 2); |
|
318 | 318 | break; |
|
319 | 319 | } |
|
320 | 320 | break; |
|
321 | 321 | case L: |
|
322 | 322 | if (src[i] == 'p') { |
|
323 | 323 | state = LP; |
|
324 | 324 | charcopy(dest, &destlen, destsize, src[i++]); |
|
325 | 325 | } else { |
|
326 | 326 | state = DEFAULT; |
|
327 | 327 | } |
|
328 | 328 | break; |
|
329 | 329 | case LP: |
|
330 | 330 | if (src[i] == 't') { |
|
331 | 331 | state = COMLPT; |
|
332 | 332 | i++; |
|
333 | 333 | } else { |
|
334 | 334 | state = DEFAULT; |
|
335 | 335 | } |
|
336 | 336 | break; |
|
337 | 337 | case N: |
|
338 | 338 | if (src[i] == 'u') { |
|
339 | 339 | state = NU; |
|
340 | 340 | charcopy(dest, &destlen, destsize, src[i++]); |
|
341 | 341 | } else { |
|
342 | 342 | state = DEFAULT; |
|
343 | 343 | } |
|
344 | 344 | break; |
|
345 | 345 | case NU: |
|
346 | 346 | if (src[i] == 'l') { |
|
347 | 347 | state = THIRD; |
|
348 | 348 | i++; |
|
349 | 349 | } else { |
|
350 | 350 | state = DEFAULT; |
|
351 | 351 | } |
|
352 | 352 | break; |
|
353 | 353 | case P: |
|
354 | 354 | if (src[i] == 'r') { |
|
355 | 355 | state = PR; |
|
356 | 356 | charcopy(dest, &destlen, destsize, src[i++]); |
|
357 | 357 | } else { |
|
358 | 358 | state = DEFAULT; |
|
359 | 359 | } |
|
360 | 360 | break; |
|
361 | 361 | case PR: |
|
362 | 362 | if (src[i] == 'n') { |
|
363 | 363 | state = THIRD; |
|
364 | 364 | i++; |
|
365 | 365 | } else { |
|
366 | 366 | state = DEFAULT; |
|
367 | 367 | } |
|
368 | 368 | break; |
|
369 | 369 | case LDOT: |
|
370 | 370 | switch (src[i]) { |
|
371 | 371 | case 'd': |
|
372 | 372 | case 'i': |
|
373 | 373 | state = HGDI; |
|
374 | 374 | charcopy(dest, &destlen, destsize, src[i++]); |
|
375 | 375 | break; |
|
376 | 376 | case 'h': |
|
377 | 377 | state = H; |
|
378 | 378 | charcopy(dest, &destlen, destsize, src[i++]); |
|
379 | 379 | break; |
|
380 | 380 | default: |
|
381 | 381 | state = DEFAULT; |
|
382 | 382 | break; |
|
383 | 383 | } |
|
384 | 384 | break; |
|
385 | 385 | case DOT: |
|
386 | 386 | switch (src[i]) { |
|
387 | 387 | case '/': |
|
388 | 388 | case '\0': |
|
389 | 389 | state = START; |
|
390 | 390 | memcopy(dest, &destlen, destsize, "~2e", 3); |
|
391 | 391 | charcopy(dest, &destlen, destsize, src[i++]); |
|
392 | 392 | break; |
|
393 | 393 | case 'd': |
|
394 | 394 | case 'i': |
|
395 | 395 | state = HGDI; |
|
396 | 396 | charcopy(dest, &destlen, destsize, '.'); |
|
397 | 397 | charcopy(dest, &destlen, destsize, src[i++]); |
|
398 | 398 | break; |
|
399 | 399 | case 'h': |
|
400 | 400 | state = H; |
|
401 | 401 | memcopy(dest, &destlen, destsize, ".h", 2); |
|
402 | 402 | i++; |
|
403 | 403 | break; |
|
404 | 404 | default: |
|
405 | 405 | state = DEFAULT; |
|
406 | 406 | charcopy(dest, &destlen, destsize, '.'); |
|
407 | 407 | break; |
|
408 | 408 | } |
|
409 | 409 | break; |
|
410 | 410 | case H: |
|
411 | 411 | if (src[i] == 'g') { |
|
412 | 412 | state = HGDI; |
|
413 | 413 | charcopy(dest, &destlen, destsize, src[i++]); |
|
414 | 414 | } else { |
|
415 | 415 | state = DEFAULT; |
|
416 | 416 | } |
|
417 | 417 | break; |
|
418 | 418 | case HGDI: |
|
419 | 419 | if (src[i] == '/') { |
|
420 | 420 | state = START; |
|
421 | 421 | if (encodedir) { |
|
422 | 422 | memcopy(dest, &destlen, destsize, ".hg", |
|
423 | 423 | 3); |
|
424 | 424 | } |
|
425 | 425 | charcopy(dest, &destlen, destsize, src[i++]); |
|
426 | 426 | } else { |
|
427 | 427 | state = DEFAULT; |
|
428 | 428 | } |
|
429 | 429 | break; |
|
430 | 430 | case SPACE: |
|
431 | 431 | switch (src[i]) { |
|
432 | 432 | case '/': |
|
433 | 433 | case '\0': |
|
434 | 434 | state = START; |
|
435 | 435 | memcopy(dest, &destlen, destsize, "~20", 3); |
|
436 | 436 | charcopy(dest, &destlen, destsize, src[i++]); |
|
437 | 437 | break; |
|
438 | 438 | default: |
|
439 | 439 | state = DEFAULT; |
|
440 | 440 | charcopy(dest, &destlen, destsize, ' '); |
|
441 | 441 | break; |
|
442 | 442 | } |
|
443 | 443 | break; |
|
444 | 444 | case DEFAULT: |
|
445 | 445 | while (inset(onebyte, src[i])) { |
|
446 | 446 | charcopy(dest, &destlen, destsize, src[i++]); |
|
447 | 447 | if (i == len) { |
|
448 | 448 | goto done; |
|
449 | 449 | } |
|
450 | 450 | } |
|
451 | 451 | switch (src[i]) { |
|
452 | 452 | case '.': |
|
453 | 453 | state = DOT; |
|
454 | 454 | i++; |
|
455 | 455 | break; |
|
456 | 456 | case ' ': |
|
457 | 457 | state = SPACE; |
|
458 | 458 | i++; |
|
459 | 459 | break; |
|
460 | 460 | case '/': |
|
461 | 461 | state = START; |
|
462 | 462 | charcopy(dest, &destlen, destsize, '/'); |
|
463 | 463 | i++; |
|
464 | 464 | break; |
|
465 | 465 | default: |
|
466 | 466 | if (inset(onebyte, src[i])) { |
|
467 | 467 | do { |
|
468 | 468 | charcopy(dest, &destlen, |
|
469 | 469 | destsize, src[i++]); |
|
470 | 470 | } while (i < len && |
|
471 | 471 | inset(onebyte, src[i])); |
|
472 | 472 | } else if (inset(twobytes, src[i])) { |
|
473 | 473 | char c = src[i++]; |
|
474 | 474 | charcopy(dest, &destlen, destsize, '_'); |
|
475 | 475 | charcopy(dest, &destlen, destsize, |
|
476 | 476 | c == '_' ? '_' : c + 32); |
|
477 | 477 | } else { |
|
478 | 478 | escape3(dest, &destlen, destsize, |
|
479 | 479 | src[i++]); |
|
480 | 480 | } |
|
481 | 481 | break; |
|
482 | 482 | } |
|
483 | 483 | break; |
|
484 | 484 | } |
|
485 | 485 | } |
|
486 | 486 | done: |
|
487 | 487 | return destlen; |
|
488 | 488 | } |
|
489 | 489 | |
|
490 | 490 | static Py_ssize_t basicencode(char *dest, size_t destsize, const char *src, |
|
491 | 491 | Py_ssize_t len) |
|
492 | 492 | { |
|
493 | 493 | static const uint32_t twobytes[8] = {0, 0, 0x87fffffe}; |
|
494 | 494 | |
|
495 | 495 | static const uint32_t onebyte[8] = { |
|
496 | 496 | 1, |
|
497 | 497 | 0x2bff3bfa, |
|
498 | 498 | 0x68000001, |
|
499 | 499 | 0x2fffffff, |
|
500 | 500 | }; |
|
501 | 501 | |
|
502 | 502 | Py_ssize_t destlen = 0; |
|
503 | 503 | |
|
504 | 504 | return _encode(twobytes, onebyte, dest, destlen, destsize, src, len, 1); |
|
505 | 505 | } |
|
506 | 506 | |
|
507 | 507 | static const Py_ssize_t maxstorepathlen = 120; |
|
508 | 508 | |
|
509 | 509 | static Py_ssize_t _lowerencode(char *dest, size_t destsize, const char *src, |
|
510 | 510 | Py_ssize_t len) |
|
511 | 511 | { |
|
512 | 512 | static const uint32_t onebyte[8] = {1, 0x2bfffbfb, 0xe8000001, |
|
513 | 513 | 0x2fffffff}; |
|
514 | 514 | |
|
515 | 515 | static const uint32_t lower[8] = {0, 0, 0x7fffffe}; |
|
516 | 516 | |
|
517 | 517 | Py_ssize_t i, destlen = 0; |
|
518 | 518 | |
|
519 | 519 | for (i = 0; i < len; i++) { |
|
520 | 520 | if (inset(onebyte, src[i])) { |
|
521 | 521 | charcopy(dest, &destlen, destsize, src[i]); |
|
522 | 522 | } else if (inset(lower, src[i])) { |
|
523 | 523 | charcopy(dest, &destlen, destsize, src[i] + 32); |
|
524 | 524 | } else { |
|
525 | 525 | escape3(dest, &destlen, destsize, src[i]); |
|
526 | 526 | } |
|
527 | 527 | } |
|
528 | 528 | |
|
529 | 529 | return destlen; |
|
530 | 530 | } |
|
531 | 531 | |
|
532 | 532 | PyObject *lowerencode(PyObject *self, PyObject *args) |
|
533 | 533 | { |
|
534 | 534 | char *path; |
|
535 | 535 | Py_ssize_t len, newlen; |
|
536 | 536 | PyObject *ret; |
|
537 | 537 | |
|
538 |
if (!PyArg_ParseTuple(args, |
|
|
539 | &path, &len)) { | |
|
538 | if (!PyArg_ParseTuple(args, "y#:lowerencode", &path, &len)) { | |
|
540 | 539 | return NULL; |
|
541 | 540 | } |
|
542 | 541 | |
|
543 | 542 | newlen = _lowerencode(NULL, 0, path, len); |
|
544 | 543 | ret = PyBytes_FromStringAndSize(NULL, newlen); |
|
545 | 544 | if (ret) { |
|
546 | 545 | _lowerencode(PyBytes_AS_STRING(ret), newlen, path, len); |
|
547 | 546 | } |
|
548 | 547 | |
|
549 | 548 | return ret; |
|
550 | 549 | } |
|
551 | 550 | |
|
552 | 551 | /* See store.py:_auxencode for a description. */ |
|
553 | 552 | static Py_ssize_t auxencode(char *dest, size_t destsize, const char *src, |
|
554 | 553 | Py_ssize_t len) |
|
555 | 554 | { |
|
556 | 555 | static const uint32_t twobytes[8]; |
|
557 | 556 | |
|
558 | 557 | static const uint32_t onebyte[8] = { |
|
559 | 558 | ~0U, 0xffff3ffe, ~0U, ~0U, ~0U, ~0U, ~0U, ~0U, |
|
560 | 559 | }; |
|
561 | 560 | |
|
562 | 561 | return _encode(twobytes, onebyte, dest, 0, destsize, src, len, 0); |
|
563 | 562 | } |
|
564 | 563 | |
|
565 | 564 | static PyObject *hashmangle(const char *src, Py_ssize_t len, const char sha[20]) |
|
566 | 565 | { |
|
567 | 566 | static const Py_ssize_t dirprefixlen = 8; |
|
568 | 567 | static const Py_ssize_t maxshortdirslen = 68; |
|
569 | 568 | char *dest; |
|
570 | 569 | PyObject *ret; |
|
571 | 570 | |
|
572 | 571 | Py_ssize_t i, d, p, lastslash = len - 1, lastdot = -1; |
|
573 | 572 | Py_ssize_t destsize, destlen = 0, slop, used; |
|
574 | 573 | |
|
575 | 574 | while (lastslash >= 0 && src[lastslash] != '/') { |
|
576 | 575 | if (src[lastslash] == '.' && lastdot == -1) { |
|
577 | 576 | lastdot = lastslash; |
|
578 | 577 | } |
|
579 | 578 | lastslash--; |
|
580 | 579 | } |
|
581 | 580 | |
|
582 | 581 | #if 0 |
|
583 | 582 | /* All paths should end in a suffix of ".i" or ".d". |
|
584 | 583 | Unfortunately, the file names in test-hybridencode.py |
|
585 | 584 | violate this rule. */ |
|
586 | 585 | if (lastdot != len - 3) { |
|
587 | 586 | PyErr_SetString(PyExc_ValueError, |
|
588 | 587 | "suffix missing or wrong length"); |
|
589 | 588 | return NULL; |
|
590 | 589 | } |
|
591 | 590 | #endif |
|
592 | 591 | |
|
593 | 592 | /* If src contains a suffix, we will append it to the end of |
|
594 | 593 | the new string, so make room. */ |
|
595 | 594 | destsize = 120; |
|
596 | 595 | if (lastdot >= 0) { |
|
597 | 596 | destsize += len - lastdot - 1; |
|
598 | 597 | } |
|
599 | 598 | |
|
600 | 599 | ret = PyBytes_FromStringAndSize(NULL, destsize); |
|
601 | 600 | if (ret == NULL) { |
|
602 | 601 | return NULL; |
|
603 | 602 | } |
|
604 | 603 | |
|
605 | 604 | dest = PyBytes_AS_STRING(ret); |
|
606 | 605 | memcopy(dest, &destlen, destsize, "dh/", 3); |
|
607 | 606 | |
|
608 | 607 | /* Copy up to dirprefixlen bytes of each path component, up to |
|
609 | 608 | a limit of maxshortdirslen bytes. */ |
|
610 | 609 | for (i = d = p = 0; i < lastslash; i++, p++) { |
|
611 | 610 | if (src[i] == '/') { |
|
612 | 611 | char d = dest[destlen - 1]; |
|
613 | 612 | /* After truncation, a directory name may end |
|
614 | 613 | in a space or dot, which are unportable. */ |
|
615 | 614 | if (d == '.' || d == ' ') { |
|
616 | 615 | dest[destlen - 1] = '_'; |
|
617 | 616 | /* The + 3 is to account for "dh/" in the |
|
618 | 617 | * beginning */ |
|
619 | 618 | } |
|
620 | 619 | if (destlen > maxshortdirslen + 3) { |
|
621 | 620 | break; |
|
622 | 621 | } |
|
623 | 622 | charcopy(dest, &destlen, destsize, src[i]); |
|
624 | 623 | p = -1; |
|
625 | 624 | } else if (p < dirprefixlen) { |
|
626 | 625 | charcopy(dest, &destlen, destsize, src[i]); |
|
627 | 626 | } |
|
628 | 627 | } |
|
629 | 628 | |
|
630 | 629 | /* Rewind to just before the last slash copied. */ |
|
631 | 630 | if (destlen > maxshortdirslen + 3) { |
|
632 | 631 | do { |
|
633 | 632 | destlen--; |
|
634 | 633 | } while (destlen > 0 && dest[destlen] != '/'); |
|
635 | 634 | } |
|
636 | 635 | |
|
637 | 636 | if (destlen > 3) { |
|
638 | 637 | if (lastslash > 0) { |
|
639 | 638 | char d = dest[destlen - 1]; |
|
640 | 639 | /* The last directory component may be |
|
641 | 640 | truncated, so make it safe. */ |
|
642 | 641 | if (d == '.' || d == ' ') { |
|
643 | 642 | dest[destlen - 1] = '_'; |
|
644 | 643 | } |
|
645 | 644 | } |
|
646 | 645 | |
|
647 | 646 | charcopy(dest, &destlen, destsize, '/'); |
|
648 | 647 | } |
|
649 | 648 | |
|
650 | 649 | /* Add a prefix of the original file's name. Its length |
|
651 | 650 | depends on the number of bytes left after accounting for |
|
652 | 651 | hash and suffix. */ |
|
653 | 652 | used = destlen + 40; |
|
654 | 653 | if (lastdot >= 0) { |
|
655 | 654 | used += len - lastdot - 1; |
|
656 | 655 | } |
|
657 | 656 | slop = maxstorepathlen - used; |
|
658 | 657 | if (slop > 0) { |
|
659 | 658 | Py_ssize_t basenamelen = |
|
660 | 659 | lastslash >= 0 ? len - lastslash - 2 : len - 1; |
|
661 | 660 | |
|
662 | 661 | if (basenamelen > slop) { |
|
663 | 662 | basenamelen = slop; |
|
664 | 663 | } |
|
665 | 664 | if (basenamelen > 0) { |
|
666 | 665 | memcopy(dest, &destlen, destsize, &src[lastslash + 1], |
|
667 | 666 | basenamelen); |
|
668 | 667 | } |
|
669 | 668 | } |
|
670 | 669 | |
|
671 | 670 | /* Add hash and suffix. */ |
|
672 | 671 | for (i = 0; i < 20; i++) { |
|
673 | 672 | hexencode(dest, &destlen, destsize, sha[i]); |
|
674 | 673 | } |
|
675 | 674 | |
|
676 | 675 | if (lastdot >= 0) { |
|
677 | 676 | memcopy(dest, &destlen, destsize, &src[lastdot], |
|
678 | 677 | len - lastdot - 1); |
|
679 | 678 | } |
|
680 | 679 | |
|
681 | 680 | assert(PyBytes_Check(ret)); |
|
682 | 681 | Py_SET_SIZE(ret, destlen); |
|
683 | 682 | |
|
684 | 683 | return ret; |
|
685 | 684 | } |
|
686 | 685 | |
|
687 | 686 | /* |
|
688 | 687 | * Avoiding a trip through Python would improve performance by 50%, |
|
689 | 688 | * but we don't encounter enough long names to be worth the code. |
|
690 | 689 | */ |
|
691 | 690 | static int sha1hash(char hash[20], const char *str, Py_ssize_t len) |
|
692 | 691 | { |
|
693 | 692 | static PyObject *shafunc; |
|
694 | 693 | PyObject *shaobj, *hashobj; |
|
695 | 694 | |
|
696 | 695 | if (shafunc == NULL) { |
|
697 | 696 | PyObject *hashlib = PyImport_ImportModule("hashlib"); |
|
698 | 697 | if (hashlib == NULL) { |
|
699 | 698 | PyErr_SetString(PyExc_ImportError, |
|
700 | 699 | "pathencode failed to find hashlib"); |
|
701 | 700 | return -1; |
|
702 | 701 | } |
|
703 | 702 | shafunc = PyObject_GetAttrString(hashlib, "sha1"); |
|
704 | 703 | Py_DECREF(hashlib); |
|
705 | 704 | |
|
706 | 705 | if (shafunc == NULL) { |
|
707 | 706 | PyErr_SetString(PyExc_AttributeError, |
|
708 | 707 | "module 'hashlib' has no " |
|
709 | 708 | "attribute 'sha1' in pathencode"); |
|
710 | 709 | return -1; |
|
711 | 710 | } |
|
712 | 711 | } |
|
713 | 712 | |
|
714 |
shaobj = PyObject_CallFunction(shafunc, |
|
|
713 | shaobj = PyObject_CallFunction(shafunc, "y#", str, len); | |
|
715 | 714 | |
|
716 | 715 | if (shaobj == NULL) { |
|
717 | 716 | return -1; |
|
718 | 717 | } |
|
719 | 718 | |
|
720 | 719 | hashobj = PyObject_CallMethod(shaobj, "digest", ""); |
|
721 | 720 | Py_DECREF(shaobj); |
|
722 | 721 | if (hashobj == NULL) { |
|
723 | 722 | return -1; |
|
724 | 723 | } |
|
725 | 724 | |
|
726 | 725 | if (!PyBytes_Check(hashobj) || PyBytes_GET_SIZE(hashobj) != 20) { |
|
727 | 726 | PyErr_SetString(PyExc_TypeError, |
|
728 | 727 | "result of digest is not a 20-byte hash"); |
|
729 | 728 | Py_DECREF(hashobj); |
|
730 | 729 | return -1; |
|
731 | 730 | } |
|
732 | 731 | |
|
733 | 732 | memcpy(hash, PyBytes_AS_STRING(hashobj), 20); |
|
734 | 733 | Py_DECREF(hashobj); |
|
735 | 734 | return 0; |
|
736 | 735 | } |
|
737 | 736 | |
|
738 | 737 | #define MAXENCODE 4096 * 4 |
|
739 | 738 | |
|
740 | 739 | static PyObject *hashencode(const char *src, Py_ssize_t len) |
|
741 | 740 | { |
|
742 | 741 | char dired[MAXENCODE]; |
|
743 | 742 | char lowered[MAXENCODE]; |
|
744 | 743 | char auxed[MAXENCODE]; |
|
745 | 744 | Py_ssize_t dirlen, lowerlen, auxlen, baselen; |
|
746 | 745 | char sha[20]; |
|
747 | 746 | |
|
748 | 747 | baselen = (len - 5) * 3; |
|
749 | 748 | if (baselen >= MAXENCODE) { |
|
750 | 749 | PyErr_SetString(PyExc_ValueError, "string too long"); |
|
751 | 750 | return NULL; |
|
752 | 751 | } |
|
753 | 752 | |
|
754 | 753 | dirlen = _encodedir(dired, baselen, src, len); |
|
755 | 754 | if (sha1hash(sha, dired, dirlen - 1) == -1) { |
|
756 | 755 | return NULL; |
|
757 | 756 | } |
|
758 | 757 | lowerlen = _lowerencode(lowered, baselen, dired + 5, dirlen - 5); |
|
759 | 758 | auxlen = auxencode(auxed, baselen, lowered, lowerlen); |
|
760 | 759 | return hashmangle(auxed, auxlen, sha); |
|
761 | 760 | } |
|
762 | 761 | |
|
763 | 762 | PyObject *pathencode(PyObject *self, PyObject *args) |
|
764 | 763 | { |
|
765 | 764 | Py_ssize_t len, newlen; |
|
766 | 765 | PyObject *pathobj, *newobj; |
|
767 | 766 | char *path; |
|
768 | 767 | |
|
769 | 768 | if (!PyArg_ParseTuple(args, "O:pathencode", &pathobj)) { |
|
770 | 769 | return NULL; |
|
771 | 770 | } |
|
772 | 771 | |
|
773 | 772 | if (PyBytes_AsStringAndSize(pathobj, &path, &len) == -1) { |
|
774 | 773 | PyErr_SetString(PyExc_TypeError, "expected a string"); |
|
775 | 774 | return NULL; |
|
776 | 775 | } |
|
777 | 776 | |
|
778 | 777 | if (len > maxstorepathlen) { |
|
779 | 778 | newlen = maxstorepathlen + 2; |
|
780 | 779 | } else { |
|
781 | 780 | newlen = len ? basicencode(NULL, 0, path, len + 1) : 1; |
|
782 | 781 | } |
|
783 | 782 | |
|
784 | 783 | if (newlen <= maxstorepathlen + 1) { |
|
785 | 784 | if (newlen == len + 1) { |
|
786 | 785 | Py_INCREF(pathobj); |
|
787 | 786 | return pathobj; |
|
788 | 787 | } |
|
789 | 788 | |
|
790 | 789 | newobj = PyBytes_FromStringAndSize(NULL, newlen); |
|
791 | 790 | |
|
792 | 791 | if (newobj) { |
|
793 | 792 | assert(PyBytes_Check(newobj)); |
|
794 | 793 | Py_SET_SIZE(newobj, Py_SIZE(newobj) - 1); |
|
795 | 794 | basicencode(PyBytes_AS_STRING(newobj), newlen, path, |
|
796 | 795 | len + 1); |
|
797 | 796 | } |
|
798 | 797 | } else { |
|
799 | 798 | newobj = hashencode(path, len + 1); |
|
800 | 799 | } |
|
801 | 800 | |
|
802 | 801 | return newobj; |
|
803 | 802 | } |
@@ -1,3278 +1,3275 b'' | |||
|
1 | 1 | /* |
|
2 | 2 | parsers.c - efficient content parsing |
|
3 | 3 | |
|
4 | 4 | Copyright 2008 Olivia Mackall <olivia@selenic.com> and others |
|
5 | 5 | |
|
6 | 6 | This software may be used and distributed according to the terms of |
|
7 | 7 | the GNU General Public License, incorporated herein by reference. |
|
8 | 8 | */ |
|
9 | 9 | |
|
10 | 10 | #define PY_SSIZE_T_CLEAN |
|
11 | 11 | #include <Python.h> |
|
12 | 12 | #include <assert.h> |
|
13 | 13 | #include <ctype.h> |
|
14 | 14 | #include <limits.h> |
|
15 | 15 | #include <stddef.h> |
|
16 | 16 | #include <stdlib.h> |
|
17 | 17 | #include <string.h> |
|
18 | 18 | #include <structmember.h> |
|
19 | 19 | |
|
20 | 20 | #include "bitmanipulation.h" |
|
21 | 21 | #include "charencode.h" |
|
22 | 22 | #include "compat.h" |
|
23 | 23 | #include "revlog.h" |
|
24 | 24 | #include "util.h" |
|
25 | 25 | |
|
26 | 26 | typedef struct indexObjectStruct indexObject; |
|
27 | 27 | |
|
28 | 28 | typedef struct { |
|
29 | 29 | int children[16]; |
|
30 | 30 | } nodetreenode; |
|
31 | 31 | |
|
32 | 32 | typedef struct { |
|
33 | 33 | int abi_version; |
|
34 | 34 | Py_ssize_t (*index_length)(const indexObject *); |
|
35 | 35 | const char *(*index_node)(indexObject *, Py_ssize_t); |
|
36 | 36 | int (*index_parents)(PyObject *, int, int *); |
|
37 | 37 | } Revlog_CAPI; |
|
38 | 38 | |
|
39 | 39 | /* |
|
40 | 40 | * A base-16 trie for fast node->rev mapping. |
|
41 | 41 | * |
|
42 | 42 | * Positive value is index of the next node in the trie |
|
43 | 43 | * Negative value is a leaf: -(rev + 2) |
|
44 | 44 | * Zero is empty |
|
45 | 45 | */ |
|
46 | 46 | typedef struct { |
|
47 | 47 | indexObject *index; |
|
48 | 48 | nodetreenode *nodes; |
|
49 | 49 | Py_ssize_t nodelen; |
|
50 | 50 | size_t length; /* # nodes in use */ |
|
51 | 51 | size_t capacity; /* # nodes allocated */ |
|
52 | 52 | int depth; /* maximum depth of tree */ |
|
53 | 53 | int splits; /* # splits performed */ |
|
54 | 54 | } nodetree; |
|
55 | 55 | |
|
56 | 56 | typedef struct { |
|
57 | 57 | PyObject_HEAD /* ; */ |
|
58 | 58 | nodetree nt; |
|
59 | 59 | } nodetreeObject; |
|
60 | 60 | |
|
61 | 61 | /* |
|
62 | 62 | * This class has two behaviors. |
|
63 | 63 | * |
|
64 | 64 | * When used in a list-like way (with integer keys), we decode an |
|
65 | 65 | * entry in a RevlogNG index file on demand. We have limited support for |
|
66 | 66 | * integer-keyed insert and delete, only at elements right before the |
|
67 | 67 | * end. |
|
68 | 68 | * |
|
69 | 69 | * With string keys, we lazily perform a reverse mapping from node to |
|
70 | 70 | * rev, using a base-16 trie. |
|
71 | 71 | */ |
|
72 | 72 | struct indexObjectStruct { |
|
73 | 73 | PyObject_HEAD |
|
74 | 74 | /* Type-specific fields go here. */ |
|
75 | 75 | PyObject *data; /* raw bytes of index */ |
|
76 | 76 | Py_ssize_t nodelen; /* digest size of the hash, 20 for SHA-1 */ |
|
77 | 77 | PyObject *nullentry; /* fast path for references to null */ |
|
78 | 78 | Py_buffer buf; /* buffer of data */ |
|
79 | 79 | const char **offsets; /* populated on demand */ |
|
80 | 80 | Py_ssize_t length; /* current on-disk number of elements */ |
|
81 | 81 | unsigned new_length; /* number of added elements */ |
|
82 | 82 | unsigned added_length; /* space reserved for added elements */ |
|
83 | 83 | char *added; /* populated on demand */ |
|
84 | 84 | PyObject *headrevs; /* cache, invalidated on changes */ |
|
85 | 85 | PyObject *filteredrevs; /* filtered revs set */ |
|
86 | 86 | nodetree nt; /* base-16 trie */ |
|
87 | 87 | int ntinitialized; /* 0 or 1 */ |
|
88 | 88 | int ntrev; /* last rev scanned */ |
|
89 | 89 | int ntlookups; /* # lookups */ |
|
90 | 90 | int ntmisses; /* # lookups that miss the cache */ |
|
91 | 91 | int inlined; |
|
92 | 92 | long entry_size; /* size of index headers. Differs in v1 v.s. v2 format |
|
93 | 93 | */ |
|
94 | 94 | long rust_ext_compat; /* compatibility with being used in rust |
|
95 | 95 | extensions */ |
|
96 | 96 | long format_version; /* format version selector (format_*) */ |
|
97 | 97 | }; |
|
98 | 98 | |
|
99 | 99 | static Py_ssize_t index_length(const indexObject *self) |
|
100 | 100 | { |
|
101 | 101 | return self->length + self->new_length; |
|
102 | 102 | } |
|
103 | 103 | |
|
104 | 104 | static const char nullid[32] = {0}; |
|
105 | 105 | static const Py_ssize_t nullrev = -1; |
|
106 | 106 | |
|
107 | 107 | static Py_ssize_t inline_scan(indexObject *self, const char **offsets); |
|
108 | 108 | |
|
109 | 109 | static int index_find_node(indexObject *self, const char *node); |
|
110 | 110 | |
|
111 | 111 | #if LONG_MAX == 0x7fffffffL |
|
112 | static const char *const tuple_format = | |
|
113 | PY23("Kiiiiiis#KiBBi", "Kiiiiiiy#KiBBi"); | |
|
112 | static const char *const tuple_format = "Kiiiiiiy#KiBBi"; | |
|
114 | 113 | #else |
|
115 | static const char *const tuple_format = | |
|
116 | PY23("kiiiiiis#kiBBi", "kiiiiiiy#kiBBi"); | |
|
114 | static const char *const tuple_format = "kiiiiiiy#kiBBi"; | |
|
117 | 115 | #endif |
|
118 | 116 | |
|
119 | 117 | /* A RevlogNG v1 index entry is 64 bytes long. */ |
|
120 | 118 | static const long v1_entry_size = 64; |
|
121 | 119 | |
|
122 | 120 | /* A Revlogv2 index entry is 96 bytes long. */ |
|
123 | 121 | static const long v2_entry_size = 96; |
|
124 | 122 | |
|
125 | 123 | /* A Changelogv2 index entry is 96 bytes long. */ |
|
126 | 124 | static const long cl2_entry_size = 96; |
|
127 | 125 | |
|
128 | 126 | /* Internal format version. |
|
129 | 127 | * Must match their counterparts in revlogutils/constants.py */ |
|
130 | 128 | static const long format_v1 = 1; /* constants.py: REVLOGV1 */ |
|
131 | 129 | static const long format_v2 = 0xDEAD; /* constants.py: REVLOGV2 */ |
|
132 | 130 | static const long format_cl2 = 0xD34D; /* constants.py: CHANGELOGV2 */ |
|
133 | 131 | |
|
134 | 132 | static const long entry_v1_offset_high = 0; |
|
135 | 133 | static const long entry_v1_offset_offset_flags = 4; |
|
136 | 134 | static const long entry_v1_offset_comp_len = 8; |
|
137 | 135 | static const long entry_v1_offset_uncomp_len = 12; |
|
138 | 136 | static const long entry_v1_offset_base_rev = 16; |
|
139 | 137 | static const long entry_v1_offset_link_rev = 20; |
|
140 | 138 | static const long entry_v1_offset_parent_1 = 24; |
|
141 | 139 | static const long entry_v1_offset_parent_2 = 28; |
|
142 | 140 | static const long entry_v1_offset_node_id = 32; |
|
143 | 141 | |
|
144 | 142 | static const long entry_v2_offset_high = 0; |
|
145 | 143 | static const long entry_v2_offset_offset_flags = 4; |
|
146 | 144 | static const long entry_v2_offset_comp_len = 8; |
|
147 | 145 | static const long entry_v2_offset_uncomp_len = 12; |
|
148 | 146 | static const long entry_v2_offset_base_rev = 16; |
|
149 | 147 | static const long entry_v2_offset_link_rev = 20; |
|
150 | 148 | static const long entry_v2_offset_parent_1 = 24; |
|
151 | 149 | static const long entry_v2_offset_parent_2 = 28; |
|
152 | 150 | static const long entry_v2_offset_node_id = 32; |
|
153 | 151 | static const long entry_v2_offset_sidedata_offset = 64; |
|
154 | 152 | static const long entry_v2_offset_sidedata_comp_len = 72; |
|
155 | 153 | static const long entry_v2_offset_all_comp_mode = 76; |
|
156 | 154 | /* next free offset: 77 */ |
|
157 | 155 | |
|
158 | 156 | static const long entry_cl2_offset_high = 0; |
|
159 | 157 | static const long entry_cl2_offset_offset_flags = 4; |
|
160 | 158 | static const long entry_cl2_offset_comp_len = 8; |
|
161 | 159 | static const long entry_cl2_offset_uncomp_len = 12; |
|
162 | 160 | static const long entry_cl2_offset_parent_1 = 16; |
|
163 | 161 | static const long entry_cl2_offset_parent_2 = 20; |
|
164 | 162 | static const long entry_cl2_offset_node_id = 24; |
|
165 | 163 | static const long entry_cl2_offset_sidedata_offset = 56; |
|
166 | 164 | static const long entry_cl2_offset_sidedata_comp_len = 64; |
|
167 | 165 | static const long entry_cl2_offset_all_comp_mode = 68; |
|
168 | 166 | static const long entry_cl2_offset_rank = 69; |
|
169 | 167 | /* next free offset: 73 */ |
|
170 | 168 | |
|
171 | 169 | static const char comp_mode_inline = 2; |
|
172 | 170 | static const char rank_unknown = -1; |
|
173 | 171 | |
|
174 | 172 | static void raise_revlog_error(void) |
|
175 | 173 | { |
|
176 | 174 | PyObject *mod = NULL, *dict = NULL, *errclass = NULL; |
|
177 | 175 | |
|
178 | 176 | mod = PyImport_ImportModule("mercurial.error"); |
|
179 | 177 | if (mod == NULL) { |
|
180 | 178 | goto cleanup; |
|
181 | 179 | } |
|
182 | 180 | |
|
183 | 181 | dict = PyModule_GetDict(mod); |
|
184 | 182 | if (dict == NULL) { |
|
185 | 183 | goto cleanup; |
|
186 | 184 | } |
|
187 | 185 | Py_INCREF(dict); |
|
188 | 186 | |
|
189 | 187 | errclass = PyDict_GetItemString(dict, "RevlogError"); |
|
190 | 188 | if (errclass == NULL) { |
|
191 | 189 | PyErr_SetString(PyExc_SystemError, |
|
192 | 190 | "could not find RevlogError"); |
|
193 | 191 | goto cleanup; |
|
194 | 192 | } |
|
195 | 193 | |
|
196 | 194 | /* value of exception is ignored by callers */ |
|
197 | 195 | PyErr_SetString(errclass, "RevlogError"); |
|
198 | 196 | |
|
199 | 197 | cleanup: |
|
200 | 198 | Py_XDECREF(dict); |
|
201 | 199 | Py_XDECREF(mod); |
|
202 | 200 | } |
|
203 | 201 | |
|
204 | 202 | /* |
|
205 | 203 | * Return a pointer to the beginning of a RevlogNG record. |
|
206 | 204 | */ |
|
207 | 205 | static const char *index_deref(indexObject *self, Py_ssize_t pos) |
|
208 | 206 | { |
|
209 | 207 | if (pos >= self->length) |
|
210 | 208 | return self->added + (pos - self->length) * self->entry_size; |
|
211 | 209 | |
|
212 | 210 | if (self->inlined && pos > 0) { |
|
213 | 211 | if (self->offsets == NULL) { |
|
214 | 212 | Py_ssize_t ret; |
|
215 | 213 | self->offsets = |
|
216 | 214 | PyMem_Malloc(self->length * sizeof(*self->offsets)); |
|
217 | 215 | if (self->offsets == NULL) |
|
218 | 216 | return (const char *)PyErr_NoMemory(); |
|
219 | 217 | ret = inline_scan(self, self->offsets); |
|
220 | 218 | if (ret == -1) { |
|
221 | 219 | return NULL; |
|
222 | 220 | }; |
|
223 | 221 | } |
|
224 | 222 | return self->offsets[pos]; |
|
225 | 223 | } |
|
226 | 224 | |
|
227 | 225 | return (const char *)(self->buf.buf) + pos * self->entry_size; |
|
228 | 226 | } |
|
229 | 227 | |
|
230 | 228 | /* |
|
231 | 229 | * Get parents of the given rev. |
|
232 | 230 | * |
|
233 | 231 | * The specified rev must be valid and must not be nullrev. A returned |
|
234 | 232 | * parent revision may be nullrev, but is guaranteed to be in valid range. |
|
235 | 233 | */ |
|
236 | 234 | static inline int index_get_parents(indexObject *self, Py_ssize_t rev, int *ps, |
|
237 | 235 | int maxrev) |
|
238 | 236 | { |
|
239 | 237 | const char *data = index_deref(self, rev); |
|
240 | 238 | |
|
241 | 239 | if (self->format_version == format_v1) { |
|
242 | 240 | ps[0] = getbe32(data + entry_v1_offset_parent_1); |
|
243 | 241 | ps[1] = getbe32(data + entry_v1_offset_parent_2); |
|
244 | 242 | } else if (self->format_version == format_v2) { |
|
245 | 243 | ps[0] = getbe32(data + entry_v2_offset_parent_1); |
|
246 | 244 | ps[1] = getbe32(data + entry_v2_offset_parent_2); |
|
247 | 245 | } else if (self->format_version == format_cl2) { |
|
248 | 246 | ps[0] = getbe32(data + entry_cl2_offset_parent_1); |
|
249 | 247 | ps[1] = getbe32(data + entry_cl2_offset_parent_2); |
|
250 | 248 | } else { |
|
251 | 249 | raise_revlog_error(); |
|
252 | 250 | return -1; |
|
253 | 251 | } |
|
254 | 252 | |
|
255 | 253 | /* If index file is corrupted, ps[] may point to invalid revisions. So |
|
256 | 254 | * there is a risk of buffer overflow to trust them unconditionally. */ |
|
257 | 255 | if (ps[0] < -1 || ps[0] > maxrev || ps[1] < -1 || ps[1] > maxrev) { |
|
258 | 256 | PyErr_SetString(PyExc_ValueError, "parent out of range"); |
|
259 | 257 | return -1; |
|
260 | 258 | } |
|
261 | 259 | return 0; |
|
262 | 260 | } |
|
263 | 261 | |
|
264 | 262 | /* |
|
265 | 263 | * Get parents of the given rev. |
|
266 | 264 | * |
|
267 | 265 | * If the specified rev is out of range, IndexError will be raised. If the |
|
268 | 266 | * revlog entry is corrupted, ValueError may be raised. |
|
269 | 267 | * |
|
270 | 268 | * Returns 0 on success or -1 on failure. |
|
271 | 269 | */ |
|
272 | 270 | static int HgRevlogIndex_GetParents(PyObject *op, int rev, int *ps) |
|
273 | 271 | { |
|
274 | 272 | int tiprev; |
|
275 | 273 | if (!op || !HgRevlogIndex_Check(op) || !ps) { |
|
276 | 274 | PyErr_BadInternalCall(); |
|
277 | 275 | return -1; |
|
278 | 276 | } |
|
279 | 277 | tiprev = (int)index_length((indexObject *)op) - 1; |
|
280 | 278 | if (rev < -1 || rev > tiprev) { |
|
281 | 279 | PyErr_Format(PyExc_IndexError, "rev out of range: %d", rev); |
|
282 | 280 | return -1; |
|
283 | 281 | } else if (rev == -1) { |
|
284 | 282 | ps[0] = ps[1] = -1; |
|
285 | 283 | return 0; |
|
286 | 284 | } else { |
|
287 | 285 | return index_get_parents((indexObject *)op, rev, ps, tiprev); |
|
288 | 286 | } |
|
289 | 287 | } |
|
290 | 288 | |
|
291 | 289 | static inline int64_t index_get_start(indexObject *self, Py_ssize_t rev) |
|
292 | 290 | { |
|
293 | 291 | const char *data; |
|
294 | 292 | uint64_t offset; |
|
295 | 293 | |
|
296 | 294 | if (rev == nullrev) |
|
297 | 295 | return 0; |
|
298 | 296 | |
|
299 | 297 | data = index_deref(self, rev); |
|
300 | 298 | |
|
301 | 299 | if (self->format_version == format_v1) { |
|
302 | 300 | offset = getbe32(data + entry_v1_offset_offset_flags); |
|
303 | 301 | if (rev == 0) { |
|
304 | 302 | /* mask out version number for the first entry */ |
|
305 | 303 | offset &= 0xFFFF; |
|
306 | 304 | } else { |
|
307 | 305 | uint32_t offset_high = |
|
308 | 306 | getbe32(data + entry_v1_offset_high); |
|
309 | 307 | offset |= ((uint64_t)offset_high) << 32; |
|
310 | 308 | } |
|
311 | 309 | } else if (self->format_version == format_v2) { |
|
312 | 310 | offset = getbe32(data + entry_v2_offset_offset_flags); |
|
313 | 311 | if (rev == 0) { |
|
314 | 312 | /* mask out version number for the first entry */ |
|
315 | 313 | offset &= 0xFFFF; |
|
316 | 314 | } else { |
|
317 | 315 | uint32_t offset_high = |
|
318 | 316 | getbe32(data + entry_v2_offset_high); |
|
319 | 317 | offset |= ((uint64_t)offset_high) << 32; |
|
320 | 318 | } |
|
321 | 319 | } else if (self->format_version == format_cl2) { |
|
322 | 320 | uint32_t offset_high = getbe32(data + entry_cl2_offset_high); |
|
323 | 321 | offset = getbe32(data + entry_cl2_offset_offset_flags); |
|
324 | 322 | offset |= ((uint64_t)offset_high) << 32; |
|
325 | 323 | } else { |
|
326 | 324 | raise_revlog_error(); |
|
327 | 325 | return -1; |
|
328 | 326 | } |
|
329 | 327 | |
|
330 | 328 | return (int64_t)(offset >> 16); |
|
331 | 329 | } |
|
332 | 330 | |
|
333 | 331 | static inline int index_get_length(indexObject *self, Py_ssize_t rev) |
|
334 | 332 | { |
|
335 | 333 | const char *data; |
|
336 | 334 | int tmp; |
|
337 | 335 | |
|
338 | 336 | if (rev == nullrev) |
|
339 | 337 | return 0; |
|
340 | 338 | |
|
341 | 339 | data = index_deref(self, rev); |
|
342 | 340 | |
|
343 | 341 | if (self->format_version == format_v1) { |
|
344 | 342 | tmp = (int)getbe32(data + entry_v1_offset_comp_len); |
|
345 | 343 | } else if (self->format_version == format_v2) { |
|
346 | 344 | tmp = (int)getbe32(data + entry_v2_offset_comp_len); |
|
347 | 345 | } else if (self->format_version == format_cl2) { |
|
348 | 346 | tmp = (int)getbe32(data + entry_cl2_offset_comp_len); |
|
349 | 347 | } else { |
|
350 | 348 | raise_revlog_error(); |
|
351 | 349 | return -1; |
|
352 | 350 | } |
|
353 | 351 | if (tmp < 0) { |
|
354 | 352 | PyErr_Format(PyExc_OverflowError, |
|
355 | 353 | "revlog entry size out of bound (%d)", tmp); |
|
356 | 354 | return -1; |
|
357 | 355 | } |
|
358 | 356 | return tmp; |
|
359 | 357 | } |
|
360 | 358 | |
|
361 | 359 | /* |
|
362 | 360 | * RevlogNG format (all in big endian, data may be inlined): |
|
363 | 361 | * 6 bytes: offset |
|
364 | 362 | * 2 bytes: flags |
|
365 | 363 | * 4 bytes: compressed length |
|
366 | 364 | * 4 bytes: uncompressed length |
|
367 | 365 | * 4 bytes: base revision |
|
368 | 366 | * 4 bytes: link revision |
|
369 | 367 | * 4 bytes: parent 1 revision |
|
370 | 368 | * 4 bytes: parent 2 revision |
|
371 | 369 | * 32 bytes: nodeid (only 20 bytes used with SHA-1) |
|
372 | 370 | */ |
|
373 | 371 | static PyObject *index_get(indexObject *self, Py_ssize_t pos) |
|
374 | 372 | { |
|
375 | 373 | uint64_t offset_flags, sidedata_offset; |
|
376 | 374 | int comp_len, uncomp_len, base_rev, link_rev, parent_1, parent_2, |
|
377 | 375 | sidedata_comp_len, rank = rank_unknown; |
|
378 | 376 | char data_comp_mode, sidedata_comp_mode; |
|
379 | 377 | const char *c_node_id; |
|
380 | 378 | const char *data; |
|
381 | 379 | Py_ssize_t length = index_length(self); |
|
382 | 380 | |
|
383 | 381 | if (pos == nullrev) { |
|
384 | 382 | Py_INCREF(self->nullentry); |
|
385 | 383 | return self->nullentry; |
|
386 | 384 | } |
|
387 | 385 | |
|
388 | 386 | if (pos < 0 || pos >= length) { |
|
389 | 387 | PyErr_SetString(PyExc_IndexError, "revlog index out of range"); |
|
390 | 388 | return NULL; |
|
391 | 389 | } |
|
392 | 390 | |
|
393 | 391 | data = index_deref(self, pos); |
|
394 | 392 | if (data == NULL) |
|
395 | 393 | return NULL; |
|
396 | 394 | |
|
397 | 395 | if (self->format_version == format_v1) { |
|
398 | 396 | offset_flags = getbe32(data + entry_v1_offset_offset_flags); |
|
399 | 397 | /* |
|
400 | 398 | * The first entry on-disk needs the version number masked out, |
|
401 | 399 | * but this doesn't apply if entries are added to an empty |
|
402 | 400 | * index. |
|
403 | 401 | */ |
|
404 | 402 | if (self->length && pos == 0) |
|
405 | 403 | offset_flags &= 0xFFFF; |
|
406 | 404 | else { |
|
407 | 405 | uint32_t offset_high = |
|
408 | 406 | getbe32(data + entry_v1_offset_high); |
|
409 | 407 | offset_flags |= ((uint64_t)offset_high) << 32; |
|
410 | 408 | } |
|
411 | 409 | |
|
412 | 410 | comp_len = getbe32(data + entry_v1_offset_comp_len); |
|
413 | 411 | uncomp_len = getbe32(data + entry_v1_offset_uncomp_len); |
|
414 | 412 | base_rev = getbe32(data + entry_v1_offset_base_rev); |
|
415 | 413 | link_rev = getbe32(data + entry_v1_offset_link_rev); |
|
416 | 414 | parent_1 = getbe32(data + entry_v1_offset_parent_1); |
|
417 | 415 | parent_2 = getbe32(data + entry_v1_offset_parent_2); |
|
418 | 416 | c_node_id = data + entry_v1_offset_node_id; |
|
419 | 417 | |
|
420 | 418 | sidedata_offset = 0; |
|
421 | 419 | sidedata_comp_len = 0; |
|
422 | 420 | data_comp_mode = comp_mode_inline; |
|
423 | 421 | sidedata_comp_mode = comp_mode_inline; |
|
424 | 422 | } else if (self->format_version == format_v2) { |
|
425 | 423 | offset_flags = getbe32(data + entry_v2_offset_offset_flags); |
|
426 | 424 | /* |
|
427 | 425 | * The first entry on-disk needs the version number masked out, |
|
428 | 426 | * but this doesn't apply if entries are added to an empty |
|
429 | 427 | * index. |
|
430 | 428 | */ |
|
431 | 429 | if (self->length && pos == 0) |
|
432 | 430 | offset_flags &= 0xFFFF; |
|
433 | 431 | else { |
|
434 | 432 | uint32_t offset_high = |
|
435 | 433 | getbe32(data + entry_v2_offset_high); |
|
436 | 434 | offset_flags |= ((uint64_t)offset_high) << 32; |
|
437 | 435 | } |
|
438 | 436 | |
|
439 | 437 | comp_len = getbe32(data + entry_v2_offset_comp_len); |
|
440 | 438 | uncomp_len = getbe32(data + entry_v2_offset_uncomp_len); |
|
441 | 439 | base_rev = getbe32(data + entry_v2_offset_base_rev); |
|
442 | 440 | link_rev = getbe32(data + entry_v2_offset_link_rev); |
|
443 | 441 | parent_1 = getbe32(data + entry_v2_offset_parent_1); |
|
444 | 442 | parent_2 = getbe32(data + entry_v2_offset_parent_2); |
|
445 | 443 | c_node_id = data + entry_v2_offset_node_id; |
|
446 | 444 | |
|
447 | 445 | sidedata_offset = |
|
448 | 446 | getbe64(data + entry_v2_offset_sidedata_offset); |
|
449 | 447 | sidedata_comp_len = |
|
450 | 448 | getbe32(data + entry_v2_offset_sidedata_comp_len); |
|
451 | 449 | data_comp_mode = data[entry_v2_offset_all_comp_mode] & 3; |
|
452 | 450 | sidedata_comp_mode = |
|
453 | 451 | ((data[entry_v2_offset_all_comp_mode] >> 2) & 3); |
|
454 | 452 | } else if (self->format_version == format_cl2) { |
|
455 | 453 | uint32_t offset_high = getbe32(data + entry_cl2_offset_high); |
|
456 | 454 | offset_flags = getbe32(data + entry_cl2_offset_offset_flags); |
|
457 | 455 | offset_flags |= ((uint64_t)offset_high) << 32; |
|
458 | 456 | comp_len = getbe32(data + entry_cl2_offset_comp_len); |
|
459 | 457 | uncomp_len = getbe32(data + entry_cl2_offset_uncomp_len); |
|
460 | 458 | /* base_rev and link_rev are not stored in changelogv2, but are |
|
461 | 459 | still used by some functions shared with the other revlogs. |
|
462 | 460 | They are supposed to contain links to other revisions, |
|
463 | 461 | but they always point to themselves in the case of a changelog. |
|
464 | 462 | */ |
|
465 | 463 | base_rev = pos; |
|
466 | 464 | link_rev = pos; |
|
467 | 465 | parent_1 = getbe32(data + entry_cl2_offset_parent_1); |
|
468 | 466 | parent_2 = getbe32(data + entry_cl2_offset_parent_2); |
|
469 | 467 | c_node_id = data + entry_cl2_offset_node_id; |
|
470 | 468 | sidedata_offset = |
|
471 | 469 | getbe64(data + entry_cl2_offset_sidedata_offset); |
|
472 | 470 | sidedata_comp_len = |
|
473 | 471 | getbe32(data + entry_cl2_offset_sidedata_comp_len); |
|
474 | 472 | data_comp_mode = data[entry_cl2_offset_all_comp_mode] & 3; |
|
475 | 473 | sidedata_comp_mode = |
|
476 | 474 | ((data[entry_cl2_offset_all_comp_mode] >> 2) & 3); |
|
477 | 475 | rank = getbe32(data + entry_cl2_offset_rank); |
|
478 | 476 | } else { |
|
479 | 477 | raise_revlog_error(); |
|
480 | 478 | return NULL; |
|
481 | 479 | } |
|
482 | 480 | |
|
483 | 481 | return Py_BuildValue(tuple_format, offset_flags, comp_len, uncomp_len, |
|
484 | 482 | base_rev, link_rev, parent_1, parent_2, c_node_id, |
|
485 | 483 | self->nodelen, sidedata_offset, sidedata_comp_len, |
|
486 | 484 | data_comp_mode, sidedata_comp_mode, rank); |
|
487 | 485 | } |
|
488 | 486 | /* |
|
489 | 487 | * Pack header information in binary |
|
490 | 488 | */ |
|
491 | 489 | static PyObject *index_pack_header(indexObject *self, PyObject *args) |
|
492 | 490 | { |
|
493 | 491 | int header; |
|
494 | 492 | char out[4]; |
|
495 | 493 | if (!PyArg_ParseTuple(args, "I", &header)) { |
|
496 | 494 | return NULL; |
|
497 | 495 | } |
|
498 | 496 | if (self->format_version != format_v1) { |
|
499 | 497 | PyErr_Format(PyExc_RuntimeError, |
|
500 | 498 | "version header should go in the docket, not the " |
|
501 | 499 | "index: %lu", |
|
502 | 500 | header); |
|
503 | 501 | return NULL; |
|
504 | 502 | } |
|
505 | 503 | putbe32(header, out); |
|
506 | 504 | return PyBytes_FromStringAndSize(out, 4); |
|
507 | 505 | } |
|
508 | 506 | /* |
|
509 | 507 | * Return the raw binary string representing a revision |
|
510 | 508 | */ |
|
511 | 509 | static PyObject *index_entry_binary(indexObject *self, PyObject *value) |
|
512 | 510 | { |
|
513 | 511 | long rev; |
|
514 | 512 | const char *data; |
|
515 | 513 | Py_ssize_t length = index_length(self); |
|
516 | 514 | |
|
517 | 515 | if (!pylong_to_long(value, &rev)) { |
|
518 | 516 | return NULL; |
|
519 | 517 | } |
|
520 | 518 | if (rev < 0 || rev >= length) { |
|
521 | 519 | PyErr_Format(PyExc_ValueError, "revlog index out of range: %ld", |
|
522 | 520 | rev); |
|
523 | 521 | return NULL; |
|
524 | 522 | }; |
|
525 | 523 | |
|
526 | 524 | data = index_deref(self, rev); |
|
527 | 525 | if (data == NULL) |
|
528 | 526 | return NULL; |
|
529 | 527 | if (rev == 0 && self->format_version == format_v1) { |
|
530 | 528 | /* the header is eating the start of the first entry */ |
|
531 | 529 | return PyBytes_FromStringAndSize(data + 4, |
|
532 | 530 | self->entry_size - 4); |
|
533 | 531 | } |
|
534 | 532 | return PyBytes_FromStringAndSize(data, self->entry_size); |
|
535 | 533 | } |
|
536 | 534 | |
|
537 | 535 | /* |
|
538 | 536 | * Return the hash of node corresponding to the given rev. |
|
539 | 537 | */ |
|
540 | 538 | static const char *index_node(indexObject *self, Py_ssize_t pos) |
|
541 | 539 | { |
|
542 | 540 | Py_ssize_t length = index_length(self); |
|
543 | 541 | const char *data; |
|
544 | 542 | const char *node_id; |
|
545 | 543 | |
|
546 | 544 | if (pos == nullrev) |
|
547 | 545 | return nullid; |
|
548 | 546 | |
|
549 | 547 | if (pos >= length) |
|
550 | 548 | return NULL; |
|
551 | 549 | |
|
552 | 550 | data = index_deref(self, pos); |
|
553 | 551 | |
|
554 | 552 | if (self->format_version == format_v1) { |
|
555 | 553 | node_id = data + entry_v1_offset_node_id; |
|
556 | 554 | } else if (self->format_version == format_v2) { |
|
557 | 555 | node_id = data + entry_v2_offset_node_id; |
|
558 | 556 | } else if (self->format_version == format_cl2) { |
|
559 | 557 | node_id = data + entry_cl2_offset_node_id; |
|
560 | 558 | } else { |
|
561 | 559 | raise_revlog_error(); |
|
562 | 560 | return NULL; |
|
563 | 561 | } |
|
564 | 562 | |
|
565 | 563 | return data ? node_id : NULL; |
|
566 | 564 | } |
|
567 | 565 | |
|
568 | 566 | /* |
|
569 | 567 | * Return the hash of the node corresponding to the given rev. The |
|
570 | 568 | * rev is assumed to be existing. If not, an exception is set. |
|
571 | 569 | */ |
|
572 | 570 | static const char *index_node_existing(indexObject *self, Py_ssize_t pos) |
|
573 | 571 | { |
|
574 | 572 | const char *node = index_node(self, pos); |
|
575 | 573 | if (node == NULL) { |
|
576 | 574 | PyErr_Format(PyExc_IndexError, "could not access rev %d", |
|
577 | 575 | (int)pos); |
|
578 | 576 | } |
|
579 | 577 | return node; |
|
580 | 578 | } |
|
581 | 579 | |
|
582 | 580 | static int nt_insert(nodetree *self, const char *node, int rev); |
|
583 | 581 | |
|
584 | 582 | static int node_check(Py_ssize_t nodelen, PyObject *obj, char **node) |
|
585 | 583 | { |
|
586 | 584 | Py_ssize_t thisnodelen; |
|
587 | 585 | if (PyBytes_AsStringAndSize(obj, node, &thisnodelen) == -1) |
|
588 | 586 | return -1; |
|
589 | 587 | if (nodelen == thisnodelen) |
|
590 | 588 | return 0; |
|
591 | 589 | PyErr_Format(PyExc_ValueError, "node len %zd != expected node len %zd", |
|
592 | 590 | thisnodelen, nodelen); |
|
593 | 591 | return -1; |
|
594 | 592 | } |
|
595 | 593 | |
|
596 | 594 | static PyObject *index_append(indexObject *self, PyObject *obj) |
|
597 | 595 | { |
|
598 | 596 | uint64_t offset_flags, sidedata_offset; |
|
599 | 597 | int rev, comp_len, uncomp_len, base_rev, link_rev, parent_1, parent_2, |
|
600 | 598 | sidedata_comp_len, rank; |
|
601 | 599 | char data_comp_mode, sidedata_comp_mode; |
|
602 | 600 | Py_ssize_t c_node_id_len; |
|
603 | 601 | const char *c_node_id; |
|
604 | 602 | char comp_field; |
|
605 | 603 | char *data; |
|
606 | 604 | |
|
607 | 605 | if (!PyArg_ParseTuple(obj, tuple_format, &offset_flags, &comp_len, |
|
608 | 606 | &uncomp_len, &base_rev, &link_rev, &parent_1, |
|
609 | 607 | &parent_2, &c_node_id, &c_node_id_len, |
|
610 | 608 | &sidedata_offset, &sidedata_comp_len, |
|
611 | 609 | &data_comp_mode, &sidedata_comp_mode, &rank)) { |
|
612 | 610 | PyErr_SetString(PyExc_TypeError, "12-tuple required"); |
|
613 | 611 | return NULL; |
|
614 | 612 | } |
|
615 | 613 | |
|
616 | 614 | if (c_node_id_len != self->nodelen) { |
|
617 | 615 | PyErr_SetString(PyExc_TypeError, "invalid node"); |
|
618 | 616 | return NULL; |
|
619 | 617 | } |
|
620 | 618 | if (self->format_version == format_v1) { |
|
621 | 619 | |
|
622 | 620 | if (data_comp_mode != comp_mode_inline) { |
|
623 | 621 | PyErr_Format(PyExc_ValueError, |
|
624 | 622 | "invalid data compression mode: %i", |
|
625 | 623 | data_comp_mode); |
|
626 | 624 | return NULL; |
|
627 | 625 | } |
|
628 | 626 | if (sidedata_comp_mode != comp_mode_inline) { |
|
629 | 627 | PyErr_Format(PyExc_ValueError, |
|
630 | 628 | "invalid sidedata compression mode: %i", |
|
631 | 629 | sidedata_comp_mode); |
|
632 | 630 | return NULL; |
|
633 | 631 | } |
|
634 | 632 | } |
|
635 | 633 | |
|
636 | 634 | if (self->new_length == self->added_length) { |
|
637 | 635 | size_t new_added_length = |
|
638 | 636 | self->added_length ? self->added_length * 2 : 4096; |
|
639 | 637 | void *new_added = PyMem_Realloc( |
|
640 | 638 | self->added, new_added_length * self->entry_size); |
|
641 | 639 | if (!new_added) |
|
642 | 640 | return PyErr_NoMemory(); |
|
643 | 641 | self->added = new_added; |
|
644 | 642 | self->added_length = new_added_length; |
|
645 | 643 | } |
|
646 | 644 | rev = self->length + self->new_length; |
|
647 | 645 | data = self->added + self->entry_size * self->new_length++; |
|
648 | 646 | |
|
649 | 647 | memset(data, 0, self->entry_size); |
|
650 | 648 | |
|
651 | 649 | if (self->format_version == format_v1) { |
|
652 | 650 | putbe32(offset_flags >> 32, data + entry_v1_offset_high); |
|
653 | 651 | putbe32(offset_flags & 0xffffffffU, |
|
654 | 652 | data + entry_v1_offset_offset_flags); |
|
655 | 653 | putbe32(comp_len, data + entry_v1_offset_comp_len); |
|
656 | 654 | putbe32(uncomp_len, data + entry_v1_offset_uncomp_len); |
|
657 | 655 | putbe32(base_rev, data + entry_v1_offset_base_rev); |
|
658 | 656 | putbe32(link_rev, data + entry_v1_offset_link_rev); |
|
659 | 657 | putbe32(parent_1, data + entry_v1_offset_parent_1); |
|
660 | 658 | putbe32(parent_2, data + entry_v1_offset_parent_2); |
|
661 | 659 | memcpy(data + entry_v1_offset_node_id, c_node_id, |
|
662 | 660 | c_node_id_len); |
|
663 | 661 | } else if (self->format_version == format_v2) { |
|
664 | 662 | putbe32(offset_flags >> 32, data + entry_v2_offset_high); |
|
665 | 663 | putbe32(offset_flags & 0xffffffffU, |
|
666 | 664 | data + entry_v2_offset_offset_flags); |
|
667 | 665 | putbe32(comp_len, data + entry_v2_offset_comp_len); |
|
668 | 666 | putbe32(uncomp_len, data + entry_v2_offset_uncomp_len); |
|
669 | 667 | putbe32(base_rev, data + entry_v2_offset_base_rev); |
|
670 | 668 | putbe32(link_rev, data + entry_v2_offset_link_rev); |
|
671 | 669 | putbe32(parent_1, data + entry_v2_offset_parent_1); |
|
672 | 670 | putbe32(parent_2, data + entry_v2_offset_parent_2); |
|
673 | 671 | memcpy(data + entry_v2_offset_node_id, c_node_id, |
|
674 | 672 | c_node_id_len); |
|
675 | 673 | putbe64(sidedata_offset, |
|
676 | 674 | data + entry_v2_offset_sidedata_offset); |
|
677 | 675 | putbe32(sidedata_comp_len, |
|
678 | 676 | data + entry_v2_offset_sidedata_comp_len); |
|
679 | 677 | comp_field = data_comp_mode & 3; |
|
680 | 678 | comp_field = comp_field | (sidedata_comp_mode & 3) << 2; |
|
681 | 679 | data[entry_v2_offset_all_comp_mode] = comp_field; |
|
682 | 680 | } else if (self->format_version == format_cl2) { |
|
683 | 681 | putbe32(offset_flags >> 32, data + entry_cl2_offset_high); |
|
684 | 682 | putbe32(offset_flags & 0xffffffffU, |
|
685 | 683 | data + entry_cl2_offset_offset_flags); |
|
686 | 684 | putbe32(comp_len, data + entry_cl2_offset_comp_len); |
|
687 | 685 | putbe32(uncomp_len, data + entry_cl2_offset_uncomp_len); |
|
688 | 686 | putbe32(parent_1, data + entry_cl2_offset_parent_1); |
|
689 | 687 | putbe32(parent_2, data + entry_cl2_offset_parent_2); |
|
690 | 688 | memcpy(data + entry_cl2_offset_node_id, c_node_id, |
|
691 | 689 | c_node_id_len); |
|
692 | 690 | putbe64(sidedata_offset, |
|
693 | 691 | data + entry_cl2_offset_sidedata_offset); |
|
694 | 692 | putbe32(sidedata_comp_len, |
|
695 | 693 | data + entry_cl2_offset_sidedata_comp_len); |
|
696 | 694 | comp_field = data_comp_mode & 3; |
|
697 | 695 | comp_field = comp_field | (sidedata_comp_mode & 3) << 2; |
|
698 | 696 | data[entry_cl2_offset_all_comp_mode] = comp_field; |
|
699 | 697 | putbe32(rank, data + entry_cl2_offset_rank); |
|
700 | 698 | } else { |
|
701 | 699 | raise_revlog_error(); |
|
702 | 700 | return NULL; |
|
703 | 701 | } |
|
704 | 702 | |
|
705 | 703 | if (self->ntinitialized) |
|
706 | 704 | nt_insert(&self->nt, c_node_id, rev); |
|
707 | 705 | |
|
708 | 706 | Py_CLEAR(self->headrevs); |
|
709 | 707 | Py_RETURN_NONE; |
|
710 | 708 | } |
|
711 | 709 | |
|
712 | 710 | /* Replace an existing index entry's sidedata offset and length with new ones. |
|
713 | 711 | This cannot be used outside of the context of sidedata rewriting, |
|
714 | 712 | inside the transaction that creates the given revision. */ |
|
715 | 713 | static PyObject *index_replace_sidedata_info(indexObject *self, PyObject *args) |
|
716 | 714 | { |
|
717 | 715 | uint64_t offset_flags, sidedata_offset; |
|
718 | 716 | Py_ssize_t rev; |
|
719 | 717 | int sidedata_comp_len; |
|
720 | 718 | char comp_mode; |
|
721 | 719 | char *data; |
|
722 | 720 | #if LONG_MAX == 0x7fffffffL |
|
723 |
const char *const sidedata_format = |
|
|
721 | const char *const sidedata_format = "nKiKB"; | |
|
724 | 722 | #else |
|
725 |
const char *const sidedata_format = |
|
|
723 | const char *const sidedata_format = "nkikB"; | |
|
726 | 724 | #endif |
|
727 | 725 | |
|
728 | 726 | if (self->entry_size == v1_entry_size || self->inlined) { |
|
729 | 727 | /* |
|
730 | 728 | There is a bug in the transaction handling when going from an |
|
731 | 729 | inline revlog to a separate index and data file. Turn it off until |
|
732 | 730 | it's fixed, since v2 revlogs sometimes get rewritten on exchange. |
|
733 | 731 | See issue6485. |
|
734 | 732 | */ |
|
735 | 733 | raise_revlog_error(); |
|
736 | 734 | return NULL; |
|
737 | 735 | } |
|
738 | 736 | |
|
739 | 737 | if (!PyArg_ParseTuple(args, sidedata_format, &rev, &sidedata_offset, |
|
740 | 738 | &sidedata_comp_len, &offset_flags, &comp_mode)) |
|
741 | 739 | return NULL; |
|
742 | 740 | |
|
743 | 741 | if (rev < 0 || rev >= index_length(self)) { |
|
744 | 742 | PyErr_SetString(PyExc_IndexError, "revision outside index"); |
|
745 | 743 | return NULL; |
|
746 | 744 | } |
|
747 | 745 | if (rev < self->length) { |
|
748 | 746 | PyErr_SetString( |
|
749 | 747 | PyExc_IndexError, |
|
750 | 748 | "cannot rewrite entries outside of this transaction"); |
|
751 | 749 | return NULL; |
|
752 | 750 | } |
|
753 | 751 | |
|
754 | 752 | /* Find the newly added node, offset from the "already on-disk" length |
|
755 | 753 | */ |
|
756 | 754 | data = self->added + self->entry_size * (rev - self->length); |
|
757 | 755 | if (self->format_version == format_v2) { |
|
758 | 756 | putbe64(offset_flags, data + entry_v2_offset_high); |
|
759 | 757 | putbe64(sidedata_offset, |
|
760 | 758 | data + entry_v2_offset_sidedata_offset); |
|
761 | 759 | putbe32(sidedata_comp_len, |
|
762 | 760 | data + entry_v2_offset_sidedata_comp_len); |
|
763 | 761 | data[entry_v2_offset_all_comp_mode] = |
|
764 | 762 | (data[entry_v2_offset_all_comp_mode] & ~(3 << 2)) | |
|
765 | 763 | ((comp_mode & 3) << 2); |
|
766 | 764 | } else if (self->format_version == format_cl2) { |
|
767 | 765 | putbe64(offset_flags, data + entry_cl2_offset_high); |
|
768 | 766 | putbe64(sidedata_offset, |
|
769 | 767 | data + entry_cl2_offset_sidedata_offset); |
|
770 | 768 | putbe32(sidedata_comp_len, |
|
771 | 769 | data + entry_cl2_offset_sidedata_comp_len); |
|
772 | 770 | data[entry_cl2_offset_all_comp_mode] = |
|
773 | 771 | (data[entry_cl2_offset_all_comp_mode] & ~(3 << 2)) | |
|
774 | 772 | ((comp_mode & 3) << 2); |
|
775 | 773 | } else { |
|
776 | 774 | raise_revlog_error(); |
|
777 | 775 | return NULL; |
|
778 | 776 | } |
|
779 | 777 | |
|
780 | 778 | Py_RETURN_NONE; |
|
781 | 779 | } |
|
782 | 780 | |
|
783 | 781 | static PyObject *index_stats(indexObject *self) |
|
784 | 782 | { |
|
785 | 783 | PyObject *obj = PyDict_New(); |
|
786 | 784 | PyObject *s = NULL; |
|
787 | 785 | PyObject *t = NULL; |
|
788 | 786 | |
|
789 | 787 | if (obj == NULL) |
|
790 | 788 | return NULL; |
|
791 | 789 | |
|
792 | 790 | #define istat(__n, __d) \ |
|
793 | 791 | do { \ |
|
794 | 792 | s = PyBytes_FromString(__d); \ |
|
795 | 793 | t = PyLong_FromSsize_t(self->__n); \ |
|
796 | 794 | if (!s || !t) \ |
|
797 | 795 | goto bail; \ |
|
798 | 796 | if (PyDict_SetItem(obj, s, t) == -1) \ |
|
799 | 797 | goto bail; \ |
|
800 | 798 | Py_CLEAR(s); \ |
|
801 | 799 | Py_CLEAR(t); \ |
|
802 | 800 | } while (0) |
|
803 | 801 | |
|
804 | 802 | if (self->added_length) |
|
805 | 803 | istat(new_length, "index entries added"); |
|
806 | 804 | istat(length, "revs in memory"); |
|
807 | 805 | istat(ntlookups, "node trie lookups"); |
|
808 | 806 | istat(ntmisses, "node trie misses"); |
|
809 | 807 | istat(ntrev, "node trie last rev scanned"); |
|
810 | 808 | if (self->ntinitialized) { |
|
811 | 809 | istat(nt.capacity, "node trie capacity"); |
|
812 | 810 | istat(nt.depth, "node trie depth"); |
|
813 | 811 | istat(nt.length, "node trie count"); |
|
814 | 812 | istat(nt.splits, "node trie splits"); |
|
815 | 813 | } |
|
816 | 814 | |
|
817 | 815 | #undef istat |
|
818 | 816 | |
|
819 | 817 | return obj; |
|
820 | 818 | |
|
821 | 819 | bail: |
|
822 | 820 | Py_XDECREF(obj); |
|
823 | 821 | Py_XDECREF(s); |
|
824 | 822 | Py_XDECREF(t); |
|
825 | 823 | return NULL; |
|
826 | 824 | } |
|
827 | 825 | |
|
828 | 826 | /* |
|
829 | 827 | * When we cache a list, we want to be sure the caller can't mutate |
|
830 | 828 | * the cached copy. |
|
831 | 829 | */ |
|
832 | 830 | static PyObject *list_copy(PyObject *list) |
|
833 | 831 | { |
|
834 | 832 | Py_ssize_t len = PyList_GET_SIZE(list); |
|
835 | 833 | PyObject *newlist = PyList_New(len); |
|
836 | 834 | Py_ssize_t i; |
|
837 | 835 | |
|
838 | 836 | if (newlist == NULL) |
|
839 | 837 | return NULL; |
|
840 | 838 | |
|
841 | 839 | for (i = 0; i < len; i++) { |
|
842 | 840 | PyObject *obj = PyList_GET_ITEM(list, i); |
|
843 | 841 | Py_INCREF(obj); |
|
844 | 842 | PyList_SET_ITEM(newlist, i, obj); |
|
845 | 843 | } |
|
846 | 844 | |
|
847 | 845 | return newlist; |
|
848 | 846 | } |
|
849 | 847 | |
|
850 | 848 | static int check_filter(PyObject *filter, Py_ssize_t arg) |
|
851 | 849 | { |
|
852 | 850 | if (filter) { |
|
853 | 851 | PyObject *arglist, *result; |
|
854 | 852 | int isfiltered; |
|
855 | 853 | |
|
856 | 854 | arglist = Py_BuildValue("(n)", arg); |
|
857 | 855 | if (!arglist) { |
|
858 | 856 | return -1; |
|
859 | 857 | } |
|
860 | 858 | |
|
861 | 859 | result = PyObject_Call(filter, arglist, NULL); |
|
862 | 860 | Py_DECREF(arglist); |
|
863 | 861 | if (!result) { |
|
864 | 862 | return -1; |
|
865 | 863 | } |
|
866 | 864 | |
|
867 | 865 | /* PyObject_IsTrue returns 1 if true, 0 if false, -1 if error, |
|
868 | 866 | * same as this function, so we can just return it directly.*/ |
|
869 | 867 | isfiltered = PyObject_IsTrue(result); |
|
870 | 868 | Py_DECREF(result); |
|
871 | 869 | return isfiltered; |
|
872 | 870 | } else { |
|
873 | 871 | return 0; |
|
874 | 872 | } |
|
875 | 873 | } |
|
876 | 874 | |
|
877 | 875 | static inline void set_phase_from_parents(char *phases, int parent_1, |
|
878 | 876 | int parent_2, Py_ssize_t i) |
|
879 | 877 | { |
|
880 | 878 | if (parent_1 >= 0 && phases[parent_1] > phases[i]) |
|
881 | 879 | phases[i] = phases[parent_1]; |
|
882 | 880 | if (parent_2 >= 0 && phases[parent_2] > phases[i]) |
|
883 | 881 | phases[i] = phases[parent_2]; |
|
884 | 882 | } |
|
885 | 883 | |
|
886 | 884 | static PyObject *reachableroots2(indexObject *self, PyObject *args) |
|
887 | 885 | { |
|
888 | 886 | |
|
889 | 887 | /* Input */ |
|
890 | 888 | long minroot; |
|
891 | 889 | PyObject *includepatharg = NULL; |
|
892 | 890 | int includepath = 0; |
|
893 | 891 | /* heads and roots are lists */ |
|
894 | 892 | PyObject *heads = NULL; |
|
895 | 893 | PyObject *roots = NULL; |
|
896 | 894 | PyObject *reachable = NULL; |
|
897 | 895 | |
|
898 | 896 | PyObject *val; |
|
899 | 897 | Py_ssize_t len = index_length(self); |
|
900 | 898 | long revnum; |
|
901 | 899 | Py_ssize_t k; |
|
902 | 900 | Py_ssize_t i; |
|
903 | 901 | Py_ssize_t l; |
|
904 | 902 | int r; |
|
905 | 903 | int parents[2]; |
|
906 | 904 | |
|
907 | 905 | /* Internal data structure: |
|
908 | 906 | * tovisit: array of length len+1 (all revs + nullrev), filled upto |
|
909 | 907 | * lentovisit |
|
910 | 908 | * |
|
911 | 909 | * revstates: array of length len+1 (all revs + nullrev) */ |
|
912 | 910 | int *tovisit = NULL; |
|
913 | 911 | long lentovisit = 0; |
|
914 | 912 | enum { RS_SEEN = 1, RS_ROOT = 2, RS_REACHABLE = 4 }; |
|
915 | 913 | char *revstates = NULL; |
|
916 | 914 | |
|
917 | 915 | /* Get arguments */ |
|
918 | 916 | if (!PyArg_ParseTuple(args, "lO!O!O!", &minroot, &PyList_Type, &heads, |
|
919 | 917 | &PyList_Type, &roots, &PyBool_Type, |
|
920 | 918 | &includepatharg)) |
|
921 | 919 | goto bail; |
|
922 | 920 | |
|
923 | 921 | if (includepatharg == Py_True) |
|
924 | 922 | includepath = 1; |
|
925 | 923 | |
|
926 | 924 | /* Initialize return set */ |
|
927 | 925 | reachable = PyList_New(0); |
|
928 | 926 | if (reachable == NULL) |
|
929 | 927 | goto bail; |
|
930 | 928 | |
|
931 | 929 | /* Initialize internal datastructures */ |
|
932 | 930 | tovisit = (int *)malloc((len + 1) * sizeof(int)); |
|
933 | 931 | if (tovisit == NULL) { |
|
934 | 932 | PyErr_NoMemory(); |
|
935 | 933 | goto bail; |
|
936 | 934 | } |
|
937 | 935 | |
|
938 | 936 | revstates = (char *)calloc(len + 1, 1); |
|
939 | 937 | if (revstates == NULL) { |
|
940 | 938 | PyErr_NoMemory(); |
|
941 | 939 | goto bail; |
|
942 | 940 | } |
|
943 | 941 | |
|
944 | 942 | l = PyList_GET_SIZE(roots); |
|
945 | 943 | for (i = 0; i < l; i++) { |
|
946 | 944 | revnum = PyLong_AsLong(PyList_GET_ITEM(roots, i)); |
|
947 | 945 | if (revnum == -1 && PyErr_Occurred()) |
|
948 | 946 | goto bail; |
|
949 | 947 | /* If root is out of range, e.g. wdir(), it must be unreachable |
|
950 | 948 | * from heads. So we can just ignore it. */ |
|
951 | 949 | if (revnum + 1 < 0 || revnum + 1 >= len + 1) |
|
952 | 950 | continue; |
|
953 | 951 | revstates[revnum + 1] |= RS_ROOT; |
|
954 | 952 | } |
|
955 | 953 | |
|
956 | 954 | /* Populate tovisit with all the heads */ |
|
957 | 955 | l = PyList_GET_SIZE(heads); |
|
958 | 956 | for (i = 0; i < l; i++) { |
|
959 | 957 | revnum = PyLong_AsLong(PyList_GET_ITEM(heads, i)); |
|
960 | 958 | if (revnum == -1 && PyErr_Occurred()) |
|
961 | 959 | goto bail; |
|
962 | 960 | if (revnum + 1 < 0 || revnum + 1 >= len + 1) { |
|
963 | 961 | PyErr_SetString(PyExc_IndexError, "head out of range"); |
|
964 | 962 | goto bail; |
|
965 | 963 | } |
|
966 | 964 | if (!(revstates[revnum + 1] & RS_SEEN)) { |
|
967 | 965 | tovisit[lentovisit++] = (int)revnum; |
|
968 | 966 | revstates[revnum + 1] |= RS_SEEN; |
|
969 | 967 | } |
|
970 | 968 | } |
|
971 | 969 | |
|
972 | 970 | /* Visit the tovisit list and find the reachable roots */ |
|
973 | 971 | k = 0; |
|
974 | 972 | while (k < lentovisit) { |
|
975 | 973 | /* Add the node to reachable if it is a root*/ |
|
976 | 974 | revnum = tovisit[k++]; |
|
977 | 975 | if (revstates[revnum + 1] & RS_ROOT) { |
|
978 | 976 | revstates[revnum + 1] |= RS_REACHABLE; |
|
979 | 977 | val = PyLong_FromLong(revnum); |
|
980 | 978 | if (val == NULL) |
|
981 | 979 | goto bail; |
|
982 | 980 | r = PyList_Append(reachable, val); |
|
983 | 981 | Py_DECREF(val); |
|
984 | 982 | if (r < 0) |
|
985 | 983 | goto bail; |
|
986 | 984 | if (includepath == 0) |
|
987 | 985 | continue; |
|
988 | 986 | } |
|
989 | 987 | |
|
990 | 988 | /* Add its parents to the list of nodes to visit */ |
|
991 | 989 | if (revnum == nullrev) |
|
992 | 990 | continue; |
|
993 | 991 | r = index_get_parents(self, revnum, parents, (int)len - 1); |
|
994 | 992 | if (r < 0) |
|
995 | 993 | goto bail; |
|
996 | 994 | for (i = 0; i < 2; i++) { |
|
997 | 995 | if (!(revstates[parents[i] + 1] & RS_SEEN) && |
|
998 | 996 | parents[i] >= minroot) { |
|
999 | 997 | tovisit[lentovisit++] = parents[i]; |
|
1000 | 998 | revstates[parents[i] + 1] |= RS_SEEN; |
|
1001 | 999 | } |
|
1002 | 1000 | } |
|
1003 | 1001 | } |
|
1004 | 1002 | |
|
1005 | 1003 | /* Find all the nodes in between the roots we found and the heads |
|
1006 | 1004 | * and add them to the reachable set */ |
|
1007 | 1005 | if (includepath == 1) { |
|
1008 | 1006 | long minidx = minroot; |
|
1009 | 1007 | if (minidx < 0) |
|
1010 | 1008 | minidx = 0; |
|
1011 | 1009 | for (i = minidx; i < len; i++) { |
|
1012 | 1010 | if (!(revstates[i + 1] & RS_SEEN)) |
|
1013 | 1011 | continue; |
|
1014 | 1012 | r = index_get_parents(self, i, parents, (int)len - 1); |
|
1015 | 1013 | /* Corrupted index file, error is set from |
|
1016 | 1014 | * index_get_parents */ |
|
1017 | 1015 | if (r < 0) |
|
1018 | 1016 | goto bail; |
|
1019 | 1017 | if (((revstates[parents[0] + 1] | |
|
1020 | 1018 | revstates[parents[1] + 1]) & |
|
1021 | 1019 | RS_REACHABLE) && |
|
1022 | 1020 | !(revstates[i + 1] & RS_REACHABLE)) { |
|
1023 | 1021 | revstates[i + 1] |= RS_REACHABLE; |
|
1024 | 1022 | val = PyLong_FromSsize_t(i); |
|
1025 | 1023 | if (val == NULL) |
|
1026 | 1024 | goto bail; |
|
1027 | 1025 | r = PyList_Append(reachable, val); |
|
1028 | 1026 | Py_DECREF(val); |
|
1029 | 1027 | if (r < 0) |
|
1030 | 1028 | goto bail; |
|
1031 | 1029 | } |
|
1032 | 1030 | } |
|
1033 | 1031 | } |
|
1034 | 1032 | |
|
1035 | 1033 | free(revstates); |
|
1036 | 1034 | free(tovisit); |
|
1037 | 1035 | return reachable; |
|
1038 | 1036 | bail: |
|
1039 | 1037 | Py_XDECREF(reachable); |
|
1040 | 1038 | free(revstates); |
|
1041 | 1039 | free(tovisit); |
|
1042 | 1040 | return NULL; |
|
1043 | 1041 | } |
|
1044 | 1042 | |
|
1045 | 1043 | static int add_roots_get_min(indexObject *self, PyObject *roots, char *phases, |
|
1046 | 1044 | char phase) |
|
1047 | 1045 | { |
|
1048 | 1046 | Py_ssize_t len = index_length(self); |
|
1049 | 1047 | PyObject *item; |
|
1050 | 1048 | PyObject *iterator; |
|
1051 | 1049 | int rev, minrev = -1; |
|
1052 | 1050 | char *node; |
|
1053 | 1051 | |
|
1054 | 1052 | if (!PySet_Check(roots)) { |
|
1055 | 1053 | PyErr_SetString(PyExc_TypeError, |
|
1056 | 1054 | "roots must be a set of nodes"); |
|
1057 | 1055 | return -2; |
|
1058 | 1056 | } |
|
1059 | 1057 | iterator = PyObject_GetIter(roots); |
|
1060 | 1058 | if (iterator == NULL) |
|
1061 | 1059 | return -2; |
|
1062 | 1060 | while ((item = PyIter_Next(iterator))) { |
|
1063 | 1061 | if (node_check(self->nodelen, item, &node) == -1) |
|
1064 | 1062 | goto failed; |
|
1065 | 1063 | rev = index_find_node(self, node); |
|
1066 | 1064 | /* null is implicitly public, so negative is invalid */ |
|
1067 | 1065 | if (rev < 0 || rev >= len) |
|
1068 | 1066 | goto failed; |
|
1069 | 1067 | phases[rev] = phase; |
|
1070 | 1068 | if (minrev == -1 || minrev > rev) |
|
1071 | 1069 | minrev = rev; |
|
1072 | 1070 | Py_DECREF(item); |
|
1073 | 1071 | } |
|
1074 | 1072 | Py_DECREF(iterator); |
|
1075 | 1073 | return minrev; |
|
1076 | 1074 | failed: |
|
1077 | 1075 | Py_DECREF(iterator); |
|
1078 | 1076 | Py_DECREF(item); |
|
1079 | 1077 | return -2; |
|
1080 | 1078 | } |
|
1081 | 1079 | |
|
1082 | 1080 | static PyObject *compute_phases_map_sets(indexObject *self, PyObject *args) |
|
1083 | 1081 | { |
|
1084 | 1082 | /* 0: public (untracked), 1: draft, 2: secret, 32: archive, |
|
1085 | 1083 | 96: internal */ |
|
1086 | 1084 | static const char trackedphases[] = {1, 2, 32, 96}; |
|
1087 | 1085 | PyObject *roots = Py_None; |
|
1088 | 1086 | PyObject *phasesetsdict = NULL; |
|
1089 | 1087 | PyObject *phasesets[4] = {NULL, NULL, NULL, NULL}; |
|
1090 | 1088 | Py_ssize_t len = index_length(self); |
|
1091 | 1089 | char *phases = NULL; |
|
1092 | 1090 | int minphaserev = -1, rev, i; |
|
1093 | 1091 | const int numphases = (int)(sizeof(phasesets) / sizeof(phasesets[0])); |
|
1094 | 1092 | |
|
1095 | 1093 | if (!PyArg_ParseTuple(args, "O", &roots)) |
|
1096 | 1094 | return NULL; |
|
1097 | 1095 | if (roots == NULL || !PyDict_Check(roots)) { |
|
1098 | 1096 | PyErr_SetString(PyExc_TypeError, "roots must be a dictionary"); |
|
1099 | 1097 | return NULL; |
|
1100 | 1098 | } |
|
1101 | 1099 | |
|
1102 | 1100 | phases = calloc(len, 1); |
|
1103 | 1101 | if (phases == NULL) { |
|
1104 | 1102 | PyErr_NoMemory(); |
|
1105 | 1103 | return NULL; |
|
1106 | 1104 | } |
|
1107 | 1105 | |
|
1108 | 1106 | for (i = 0; i < numphases; ++i) { |
|
1109 | 1107 | PyObject *pyphase = PyLong_FromLong(trackedphases[i]); |
|
1110 | 1108 | PyObject *phaseroots = NULL; |
|
1111 | 1109 | if (pyphase == NULL) |
|
1112 | 1110 | goto release; |
|
1113 | 1111 | phaseroots = PyDict_GetItem(roots, pyphase); |
|
1114 | 1112 | Py_DECREF(pyphase); |
|
1115 | 1113 | if (phaseroots == NULL) |
|
1116 | 1114 | continue; |
|
1117 | 1115 | rev = add_roots_get_min(self, phaseroots, phases, |
|
1118 | 1116 | trackedphases[i]); |
|
1119 | 1117 | if (rev == -2) |
|
1120 | 1118 | goto release; |
|
1121 | 1119 | if (rev != -1 && (minphaserev == -1 || rev < minphaserev)) |
|
1122 | 1120 | minphaserev = rev; |
|
1123 | 1121 | } |
|
1124 | 1122 | |
|
1125 | 1123 | for (i = 0; i < numphases; ++i) { |
|
1126 | 1124 | phasesets[i] = PySet_New(NULL); |
|
1127 | 1125 | if (phasesets[i] == NULL) |
|
1128 | 1126 | goto release; |
|
1129 | 1127 | } |
|
1130 | 1128 | |
|
1131 | 1129 | if (minphaserev == -1) |
|
1132 | 1130 | minphaserev = len; |
|
1133 | 1131 | for (rev = minphaserev; rev < len; ++rev) { |
|
1134 | 1132 | PyObject *pyphase = NULL; |
|
1135 | 1133 | PyObject *pyrev = NULL; |
|
1136 | 1134 | int parents[2]; |
|
1137 | 1135 | /* |
|
1138 | 1136 | * The parent lookup could be skipped for phaseroots, but |
|
1139 | 1137 | * phase --force would historically not recompute them |
|
1140 | 1138 | * correctly, leaving descendents with a lower phase around. |
|
1141 | 1139 | * As such, unconditionally recompute the phase. |
|
1142 | 1140 | */ |
|
1143 | 1141 | if (index_get_parents(self, rev, parents, (int)len - 1) < 0) |
|
1144 | 1142 | goto release; |
|
1145 | 1143 | set_phase_from_parents(phases, parents[0], parents[1], rev); |
|
1146 | 1144 | switch (phases[rev]) { |
|
1147 | 1145 | case 0: |
|
1148 | 1146 | continue; |
|
1149 | 1147 | case 1: |
|
1150 | 1148 | pyphase = phasesets[0]; |
|
1151 | 1149 | break; |
|
1152 | 1150 | case 2: |
|
1153 | 1151 | pyphase = phasesets[1]; |
|
1154 | 1152 | break; |
|
1155 | 1153 | case 32: |
|
1156 | 1154 | pyphase = phasesets[2]; |
|
1157 | 1155 | break; |
|
1158 | 1156 | case 96: |
|
1159 | 1157 | pyphase = phasesets[3]; |
|
1160 | 1158 | break; |
|
1161 | 1159 | default: |
|
1162 | 1160 | /* this should never happen since the phase number is |
|
1163 | 1161 | * specified by this function. */ |
|
1164 | 1162 | PyErr_SetString(PyExc_SystemError, |
|
1165 | 1163 | "bad phase number in internal list"); |
|
1166 | 1164 | goto release; |
|
1167 | 1165 | } |
|
1168 | 1166 | pyrev = PyLong_FromLong(rev); |
|
1169 | 1167 | if (pyrev == NULL) |
|
1170 | 1168 | goto release; |
|
1171 | 1169 | if (PySet_Add(pyphase, pyrev) == -1) { |
|
1172 | 1170 | Py_DECREF(pyrev); |
|
1173 | 1171 | goto release; |
|
1174 | 1172 | } |
|
1175 | 1173 | Py_DECREF(pyrev); |
|
1176 | 1174 | } |
|
1177 | 1175 | |
|
1178 | 1176 | phasesetsdict = _dict_new_presized(numphases); |
|
1179 | 1177 | if (phasesetsdict == NULL) |
|
1180 | 1178 | goto release; |
|
1181 | 1179 | for (i = 0; i < numphases; ++i) { |
|
1182 | 1180 | PyObject *pyphase = PyLong_FromLong(trackedphases[i]); |
|
1183 | 1181 | if (pyphase == NULL) |
|
1184 | 1182 | goto release; |
|
1185 | 1183 | if (PyDict_SetItem(phasesetsdict, pyphase, phasesets[i]) == |
|
1186 | 1184 | -1) { |
|
1187 | 1185 | Py_DECREF(pyphase); |
|
1188 | 1186 | goto release; |
|
1189 | 1187 | } |
|
1190 | 1188 | Py_DECREF(phasesets[i]); |
|
1191 | 1189 | phasesets[i] = NULL; |
|
1192 | 1190 | } |
|
1193 | 1191 | |
|
1194 | 1192 | free(phases); |
|
1195 | 1193 | return Py_BuildValue("nN", len, phasesetsdict); |
|
1196 | 1194 | |
|
1197 | 1195 | release: |
|
1198 | 1196 | for (i = 0; i < numphases; ++i) |
|
1199 | 1197 | Py_XDECREF(phasesets[i]); |
|
1200 | 1198 | Py_XDECREF(phasesetsdict); |
|
1201 | 1199 | |
|
1202 | 1200 | free(phases); |
|
1203 | 1201 | return NULL; |
|
1204 | 1202 | } |
|
1205 | 1203 | |
|
1206 | 1204 | static PyObject *index_headrevs(indexObject *self, PyObject *args) |
|
1207 | 1205 | { |
|
1208 | 1206 | Py_ssize_t i, j, len; |
|
1209 | 1207 | char *nothead = NULL; |
|
1210 | 1208 | PyObject *heads = NULL; |
|
1211 | 1209 | PyObject *filter = NULL; |
|
1212 | 1210 | PyObject *filteredrevs = Py_None; |
|
1213 | 1211 | |
|
1214 | 1212 | if (!PyArg_ParseTuple(args, "|O", &filteredrevs)) { |
|
1215 | 1213 | return NULL; |
|
1216 | 1214 | } |
|
1217 | 1215 | |
|
1218 | 1216 | if (self->headrevs && filteredrevs == self->filteredrevs) |
|
1219 | 1217 | return list_copy(self->headrevs); |
|
1220 | 1218 | |
|
1221 | 1219 | Py_DECREF(self->filteredrevs); |
|
1222 | 1220 | self->filteredrevs = filteredrevs; |
|
1223 | 1221 | Py_INCREF(filteredrevs); |
|
1224 | 1222 | |
|
1225 | 1223 | if (filteredrevs != Py_None) { |
|
1226 | 1224 | filter = PyObject_GetAttrString(filteredrevs, "__contains__"); |
|
1227 | 1225 | if (!filter) { |
|
1228 | 1226 | PyErr_SetString( |
|
1229 | 1227 | PyExc_TypeError, |
|
1230 | 1228 | "filteredrevs has no attribute __contains__"); |
|
1231 | 1229 | goto bail; |
|
1232 | 1230 | } |
|
1233 | 1231 | } |
|
1234 | 1232 | |
|
1235 | 1233 | len = index_length(self); |
|
1236 | 1234 | heads = PyList_New(0); |
|
1237 | 1235 | if (heads == NULL) |
|
1238 | 1236 | goto bail; |
|
1239 | 1237 | if (len == 0) { |
|
1240 | 1238 | PyObject *nullid = PyLong_FromLong(-1); |
|
1241 | 1239 | if (nullid == NULL || PyList_Append(heads, nullid) == -1) { |
|
1242 | 1240 | Py_XDECREF(nullid); |
|
1243 | 1241 | goto bail; |
|
1244 | 1242 | } |
|
1245 | 1243 | goto done; |
|
1246 | 1244 | } |
|
1247 | 1245 | |
|
1248 | 1246 | nothead = calloc(len, 1); |
|
1249 | 1247 | if (nothead == NULL) { |
|
1250 | 1248 | PyErr_NoMemory(); |
|
1251 | 1249 | goto bail; |
|
1252 | 1250 | } |
|
1253 | 1251 | |
|
1254 | 1252 | for (i = len - 1; i >= 0; i--) { |
|
1255 | 1253 | int isfiltered; |
|
1256 | 1254 | int parents[2]; |
|
1257 | 1255 | |
|
1258 | 1256 | /* If nothead[i] == 1, it means we've seen an unfiltered child |
|
1259 | 1257 | * of this node already, and therefore this node is not |
|
1260 | 1258 | * filtered. So we can skip the expensive check_filter step. |
|
1261 | 1259 | */ |
|
1262 | 1260 | if (nothead[i] != 1) { |
|
1263 | 1261 | isfiltered = check_filter(filter, i); |
|
1264 | 1262 | if (isfiltered == -1) { |
|
1265 | 1263 | PyErr_SetString(PyExc_TypeError, |
|
1266 | 1264 | "unable to check filter"); |
|
1267 | 1265 | goto bail; |
|
1268 | 1266 | } |
|
1269 | 1267 | |
|
1270 | 1268 | if (isfiltered) { |
|
1271 | 1269 | nothead[i] = 1; |
|
1272 | 1270 | continue; |
|
1273 | 1271 | } |
|
1274 | 1272 | } |
|
1275 | 1273 | |
|
1276 | 1274 | if (index_get_parents(self, i, parents, (int)len - 1) < 0) |
|
1277 | 1275 | goto bail; |
|
1278 | 1276 | for (j = 0; j < 2; j++) { |
|
1279 | 1277 | if (parents[j] >= 0) |
|
1280 | 1278 | nothead[parents[j]] = 1; |
|
1281 | 1279 | } |
|
1282 | 1280 | } |
|
1283 | 1281 | |
|
1284 | 1282 | for (i = 0; i < len; i++) { |
|
1285 | 1283 | PyObject *head; |
|
1286 | 1284 | |
|
1287 | 1285 | if (nothead[i]) |
|
1288 | 1286 | continue; |
|
1289 | 1287 | head = PyLong_FromSsize_t(i); |
|
1290 | 1288 | if (head == NULL || PyList_Append(heads, head) == -1) { |
|
1291 | 1289 | Py_XDECREF(head); |
|
1292 | 1290 | goto bail; |
|
1293 | 1291 | } |
|
1294 | 1292 | } |
|
1295 | 1293 | |
|
1296 | 1294 | done: |
|
1297 | 1295 | self->headrevs = heads; |
|
1298 | 1296 | Py_XDECREF(filter); |
|
1299 | 1297 | free(nothead); |
|
1300 | 1298 | return list_copy(self->headrevs); |
|
1301 | 1299 | bail: |
|
1302 | 1300 | Py_XDECREF(filter); |
|
1303 | 1301 | Py_XDECREF(heads); |
|
1304 | 1302 | free(nothead); |
|
1305 | 1303 | return NULL; |
|
1306 | 1304 | } |
|
1307 | 1305 | |
|
1308 | 1306 | /** |
|
1309 | 1307 | * Obtain the base revision index entry. |
|
1310 | 1308 | * |
|
1311 | 1309 | * Callers must ensure that rev >= 0 or illegal memory access may occur. |
|
1312 | 1310 | */ |
|
1313 | 1311 | static inline int index_baserev(indexObject *self, int rev) |
|
1314 | 1312 | { |
|
1315 | 1313 | const char *data; |
|
1316 | 1314 | int result; |
|
1317 | 1315 | |
|
1318 | 1316 | data = index_deref(self, rev); |
|
1319 | 1317 | if (data == NULL) |
|
1320 | 1318 | return -2; |
|
1321 | 1319 | |
|
1322 | 1320 | if (self->format_version == format_v1) { |
|
1323 | 1321 | result = getbe32(data + entry_v1_offset_base_rev); |
|
1324 | 1322 | } else if (self->format_version == format_v2) { |
|
1325 | 1323 | result = getbe32(data + entry_v2_offset_base_rev); |
|
1326 | 1324 | } else if (self->format_version == format_cl2) { |
|
1327 | 1325 | return rev; |
|
1328 | 1326 | } else { |
|
1329 | 1327 | raise_revlog_error(); |
|
1330 | 1328 | return -1; |
|
1331 | 1329 | } |
|
1332 | 1330 | |
|
1333 | 1331 | if (result > rev) { |
|
1334 | 1332 | PyErr_Format( |
|
1335 | 1333 | PyExc_ValueError, |
|
1336 | 1334 | "corrupted revlog, revision base above revision: %d, %d", |
|
1337 | 1335 | rev, result); |
|
1338 | 1336 | return -2; |
|
1339 | 1337 | } |
|
1340 | 1338 | if (result < -1) { |
|
1341 | 1339 | PyErr_Format( |
|
1342 | 1340 | PyExc_ValueError, |
|
1343 | 1341 | "corrupted revlog, revision base out of range: %d, %d", rev, |
|
1344 | 1342 | result); |
|
1345 | 1343 | return -2; |
|
1346 | 1344 | } |
|
1347 | 1345 | return result; |
|
1348 | 1346 | } |
|
1349 | 1347 | |
|
1350 | 1348 | /** |
|
1351 | 1349 | * Find if a revision is a snapshot or not |
|
1352 | 1350 | * |
|
1353 | 1351 | * Only relevant for sparse-revlog case. |
|
1354 | 1352 | * Callers must ensure that rev is in a valid range. |
|
1355 | 1353 | */ |
|
1356 | 1354 | static int index_issnapshotrev(indexObject *self, Py_ssize_t rev) |
|
1357 | 1355 | { |
|
1358 | 1356 | int ps[2]; |
|
1359 | 1357 | Py_ssize_t base; |
|
1360 | 1358 | while (rev >= 0) { |
|
1361 | 1359 | base = (Py_ssize_t)index_baserev(self, rev); |
|
1362 | 1360 | if (base == rev) { |
|
1363 | 1361 | base = -1; |
|
1364 | 1362 | } |
|
1365 | 1363 | if (base == -2) { |
|
1366 | 1364 | assert(PyErr_Occurred()); |
|
1367 | 1365 | return -1; |
|
1368 | 1366 | } |
|
1369 | 1367 | if (base == -1) { |
|
1370 | 1368 | return 1; |
|
1371 | 1369 | } |
|
1372 | 1370 | if (index_get_parents(self, rev, ps, (int)rev) < 0) { |
|
1373 | 1371 | assert(PyErr_Occurred()); |
|
1374 | 1372 | return -1; |
|
1375 | 1373 | }; |
|
1376 | 1374 | if (base == ps[0] || base == ps[1]) { |
|
1377 | 1375 | return 0; |
|
1378 | 1376 | } |
|
1379 | 1377 | rev = base; |
|
1380 | 1378 | } |
|
1381 | 1379 | return rev == -1; |
|
1382 | 1380 | } |
|
1383 | 1381 | |
|
1384 | 1382 | static PyObject *index_issnapshot(indexObject *self, PyObject *value) |
|
1385 | 1383 | { |
|
1386 | 1384 | long rev; |
|
1387 | 1385 | int issnap; |
|
1388 | 1386 | Py_ssize_t length = index_length(self); |
|
1389 | 1387 | |
|
1390 | 1388 | if (!pylong_to_long(value, &rev)) { |
|
1391 | 1389 | return NULL; |
|
1392 | 1390 | } |
|
1393 | 1391 | if (rev < -1 || rev >= length) { |
|
1394 | 1392 | PyErr_Format(PyExc_ValueError, "revlog index out of range: %ld", |
|
1395 | 1393 | rev); |
|
1396 | 1394 | return NULL; |
|
1397 | 1395 | }; |
|
1398 | 1396 | issnap = index_issnapshotrev(self, (Py_ssize_t)rev); |
|
1399 | 1397 | if (issnap < 0) { |
|
1400 | 1398 | return NULL; |
|
1401 | 1399 | }; |
|
1402 | 1400 | return PyBool_FromLong((long)issnap); |
|
1403 | 1401 | } |
|
1404 | 1402 | |
|
1405 | 1403 | static PyObject *index_findsnapshots(indexObject *self, PyObject *args) |
|
1406 | 1404 | { |
|
1407 | 1405 | Py_ssize_t start_rev; |
|
1408 | 1406 | PyObject *cache; |
|
1409 | 1407 | Py_ssize_t base; |
|
1410 | 1408 | Py_ssize_t rev; |
|
1411 | 1409 | PyObject *key = NULL; |
|
1412 | 1410 | PyObject *value = NULL; |
|
1413 | 1411 | const Py_ssize_t length = index_length(self); |
|
1414 | 1412 | if (!PyArg_ParseTuple(args, "O!n", &PyDict_Type, &cache, &start_rev)) { |
|
1415 | 1413 | return NULL; |
|
1416 | 1414 | } |
|
1417 | 1415 | for (rev = start_rev; rev < length; rev++) { |
|
1418 | 1416 | int issnap; |
|
1419 | 1417 | PyObject *allvalues = NULL; |
|
1420 | 1418 | issnap = index_issnapshotrev(self, rev); |
|
1421 | 1419 | if (issnap < 0) { |
|
1422 | 1420 | goto bail; |
|
1423 | 1421 | } |
|
1424 | 1422 | if (issnap == 0) { |
|
1425 | 1423 | continue; |
|
1426 | 1424 | } |
|
1427 | 1425 | base = (Py_ssize_t)index_baserev(self, rev); |
|
1428 | 1426 | if (base == rev) { |
|
1429 | 1427 | base = -1; |
|
1430 | 1428 | } |
|
1431 | 1429 | if (base == -2) { |
|
1432 | 1430 | assert(PyErr_Occurred()); |
|
1433 | 1431 | goto bail; |
|
1434 | 1432 | } |
|
1435 | 1433 | key = PyLong_FromSsize_t(base); |
|
1436 | 1434 | allvalues = PyDict_GetItem(cache, key); |
|
1437 | 1435 | if (allvalues == NULL && PyErr_Occurred()) { |
|
1438 | 1436 | goto bail; |
|
1439 | 1437 | } |
|
1440 | 1438 | if (allvalues == NULL) { |
|
1441 | 1439 | int r; |
|
1442 | 1440 | allvalues = PyList_New(0); |
|
1443 | 1441 | if (!allvalues) { |
|
1444 | 1442 | goto bail; |
|
1445 | 1443 | } |
|
1446 | 1444 | r = PyDict_SetItem(cache, key, allvalues); |
|
1447 | 1445 | Py_DECREF(allvalues); |
|
1448 | 1446 | if (r < 0) { |
|
1449 | 1447 | goto bail; |
|
1450 | 1448 | } |
|
1451 | 1449 | } |
|
1452 | 1450 | value = PyLong_FromSsize_t(rev); |
|
1453 | 1451 | if (PyList_Append(allvalues, value)) { |
|
1454 | 1452 | goto bail; |
|
1455 | 1453 | } |
|
1456 | 1454 | Py_CLEAR(key); |
|
1457 | 1455 | Py_CLEAR(value); |
|
1458 | 1456 | } |
|
1459 | 1457 | Py_RETURN_NONE; |
|
1460 | 1458 | bail: |
|
1461 | 1459 | Py_XDECREF(key); |
|
1462 | 1460 | Py_XDECREF(value); |
|
1463 | 1461 | return NULL; |
|
1464 | 1462 | } |
|
1465 | 1463 | |
|
1466 | 1464 | static PyObject *index_deltachain(indexObject *self, PyObject *args) |
|
1467 | 1465 | { |
|
1468 | 1466 | int rev, generaldelta; |
|
1469 | 1467 | PyObject *stoparg; |
|
1470 | 1468 | int stoprev, iterrev, baserev = -1; |
|
1471 | 1469 | int stopped; |
|
1472 | 1470 | PyObject *chain = NULL, *result = NULL; |
|
1473 | 1471 | const Py_ssize_t length = index_length(self); |
|
1474 | 1472 | |
|
1475 | 1473 | if (!PyArg_ParseTuple(args, "iOi", &rev, &stoparg, &generaldelta)) { |
|
1476 | 1474 | return NULL; |
|
1477 | 1475 | } |
|
1478 | 1476 | |
|
1479 | 1477 | if (PyLong_Check(stoparg)) { |
|
1480 | 1478 | stoprev = (int)PyLong_AsLong(stoparg); |
|
1481 | 1479 | if (stoprev == -1 && PyErr_Occurred()) { |
|
1482 | 1480 | return NULL; |
|
1483 | 1481 | } |
|
1484 | 1482 | } else if (stoparg == Py_None) { |
|
1485 | 1483 | stoprev = -2; |
|
1486 | 1484 | } else { |
|
1487 | 1485 | PyErr_SetString(PyExc_ValueError, |
|
1488 | 1486 | "stoprev must be integer or None"); |
|
1489 | 1487 | return NULL; |
|
1490 | 1488 | } |
|
1491 | 1489 | |
|
1492 | 1490 | if (rev < 0 || rev >= length) { |
|
1493 | 1491 | PyErr_SetString(PyExc_ValueError, "revlog index out of range"); |
|
1494 | 1492 | return NULL; |
|
1495 | 1493 | } |
|
1496 | 1494 | |
|
1497 | 1495 | chain = PyList_New(0); |
|
1498 | 1496 | if (chain == NULL) { |
|
1499 | 1497 | return NULL; |
|
1500 | 1498 | } |
|
1501 | 1499 | |
|
1502 | 1500 | baserev = index_baserev(self, rev); |
|
1503 | 1501 | |
|
1504 | 1502 | /* This should never happen. */ |
|
1505 | 1503 | if (baserev <= -2) { |
|
1506 | 1504 | /* Error should be set by index_deref() */ |
|
1507 | 1505 | assert(PyErr_Occurred()); |
|
1508 | 1506 | goto bail; |
|
1509 | 1507 | } |
|
1510 | 1508 | |
|
1511 | 1509 | iterrev = rev; |
|
1512 | 1510 | |
|
1513 | 1511 | while (iterrev != baserev && iterrev != stoprev) { |
|
1514 | 1512 | PyObject *value = PyLong_FromLong(iterrev); |
|
1515 | 1513 | if (value == NULL) { |
|
1516 | 1514 | goto bail; |
|
1517 | 1515 | } |
|
1518 | 1516 | if (PyList_Append(chain, value)) { |
|
1519 | 1517 | Py_DECREF(value); |
|
1520 | 1518 | goto bail; |
|
1521 | 1519 | } |
|
1522 | 1520 | Py_DECREF(value); |
|
1523 | 1521 | |
|
1524 | 1522 | if (generaldelta) { |
|
1525 | 1523 | iterrev = baserev; |
|
1526 | 1524 | } else { |
|
1527 | 1525 | iterrev--; |
|
1528 | 1526 | } |
|
1529 | 1527 | |
|
1530 | 1528 | if (iterrev < 0) { |
|
1531 | 1529 | break; |
|
1532 | 1530 | } |
|
1533 | 1531 | |
|
1534 | 1532 | if (iterrev >= length) { |
|
1535 | 1533 | PyErr_SetString(PyExc_IndexError, |
|
1536 | 1534 | "revision outside index"); |
|
1537 | 1535 | return NULL; |
|
1538 | 1536 | } |
|
1539 | 1537 | |
|
1540 | 1538 | baserev = index_baserev(self, iterrev); |
|
1541 | 1539 | |
|
1542 | 1540 | /* This should never happen. */ |
|
1543 | 1541 | if (baserev <= -2) { |
|
1544 | 1542 | /* Error should be set by index_deref() */ |
|
1545 | 1543 | assert(PyErr_Occurred()); |
|
1546 | 1544 | goto bail; |
|
1547 | 1545 | } |
|
1548 | 1546 | } |
|
1549 | 1547 | |
|
1550 | 1548 | if (iterrev == stoprev) { |
|
1551 | 1549 | stopped = 1; |
|
1552 | 1550 | } else { |
|
1553 | 1551 | PyObject *value = PyLong_FromLong(iterrev); |
|
1554 | 1552 | if (value == NULL) { |
|
1555 | 1553 | goto bail; |
|
1556 | 1554 | } |
|
1557 | 1555 | if (PyList_Append(chain, value)) { |
|
1558 | 1556 | Py_DECREF(value); |
|
1559 | 1557 | goto bail; |
|
1560 | 1558 | } |
|
1561 | 1559 | Py_DECREF(value); |
|
1562 | 1560 | |
|
1563 | 1561 | stopped = 0; |
|
1564 | 1562 | } |
|
1565 | 1563 | |
|
1566 | 1564 | if (PyList_Reverse(chain)) { |
|
1567 | 1565 | goto bail; |
|
1568 | 1566 | } |
|
1569 | 1567 | |
|
1570 | 1568 | result = Py_BuildValue("OO", chain, stopped ? Py_True : Py_False); |
|
1571 | 1569 | Py_DECREF(chain); |
|
1572 | 1570 | return result; |
|
1573 | 1571 | |
|
1574 | 1572 | bail: |
|
1575 | 1573 | Py_DECREF(chain); |
|
1576 | 1574 | return NULL; |
|
1577 | 1575 | } |
|
1578 | 1576 | |
|
1579 | 1577 | static inline int64_t |
|
1580 | 1578 | index_segment_span(indexObject *self, Py_ssize_t start_rev, Py_ssize_t end_rev) |
|
1581 | 1579 | { |
|
1582 | 1580 | int64_t start_offset; |
|
1583 | 1581 | int64_t end_offset; |
|
1584 | 1582 | int end_size; |
|
1585 | 1583 | start_offset = index_get_start(self, start_rev); |
|
1586 | 1584 | if (start_offset < 0) { |
|
1587 | 1585 | return -1; |
|
1588 | 1586 | } |
|
1589 | 1587 | end_offset = index_get_start(self, end_rev); |
|
1590 | 1588 | if (end_offset < 0) { |
|
1591 | 1589 | return -1; |
|
1592 | 1590 | } |
|
1593 | 1591 | end_size = index_get_length(self, end_rev); |
|
1594 | 1592 | if (end_size < 0) { |
|
1595 | 1593 | return -1; |
|
1596 | 1594 | } |
|
1597 | 1595 | if (end_offset < start_offset) { |
|
1598 | 1596 | PyErr_Format(PyExc_ValueError, |
|
1599 | 1597 | "corrupted revlog index: inconsistent offset " |
|
1600 | 1598 | "between revisions (%zd) and (%zd)", |
|
1601 | 1599 | start_rev, end_rev); |
|
1602 | 1600 | return -1; |
|
1603 | 1601 | } |
|
1604 | 1602 | return (end_offset - start_offset) + (int64_t)end_size; |
|
1605 | 1603 | } |
|
1606 | 1604 | |
|
1607 | 1605 | /* returns endidx so that revs[startidx:endidx] has no empty trailing revs */ |
|
1608 | 1606 | static Py_ssize_t trim_endidx(indexObject *self, const Py_ssize_t *revs, |
|
1609 | 1607 | Py_ssize_t startidx, Py_ssize_t endidx) |
|
1610 | 1608 | { |
|
1611 | 1609 | int length; |
|
1612 | 1610 | while (endidx > 1 && endidx > startidx) { |
|
1613 | 1611 | length = index_get_length(self, revs[endidx - 1]); |
|
1614 | 1612 | if (length < 0) { |
|
1615 | 1613 | return -1; |
|
1616 | 1614 | } |
|
1617 | 1615 | if (length != 0) { |
|
1618 | 1616 | break; |
|
1619 | 1617 | } |
|
1620 | 1618 | endidx -= 1; |
|
1621 | 1619 | } |
|
1622 | 1620 | return endidx; |
|
1623 | 1621 | } |
|
1624 | 1622 | |
|
1625 | 1623 | struct Gap { |
|
1626 | 1624 | int64_t size; |
|
1627 | 1625 | Py_ssize_t idx; |
|
1628 | 1626 | }; |
|
1629 | 1627 | |
|
1630 | 1628 | static int gap_compare(const void *left, const void *right) |
|
1631 | 1629 | { |
|
1632 | 1630 | const struct Gap *l_left = ((const struct Gap *)left); |
|
1633 | 1631 | const struct Gap *l_right = ((const struct Gap *)right); |
|
1634 | 1632 | if (l_left->size < l_right->size) { |
|
1635 | 1633 | return -1; |
|
1636 | 1634 | } else if (l_left->size > l_right->size) { |
|
1637 | 1635 | return 1; |
|
1638 | 1636 | } |
|
1639 | 1637 | return 0; |
|
1640 | 1638 | } |
|
1641 | 1639 | static int Py_ssize_t_compare(const void *left, const void *right) |
|
1642 | 1640 | { |
|
1643 | 1641 | const Py_ssize_t l_left = *(const Py_ssize_t *)left; |
|
1644 | 1642 | const Py_ssize_t l_right = *(const Py_ssize_t *)right; |
|
1645 | 1643 | if (l_left < l_right) { |
|
1646 | 1644 | return -1; |
|
1647 | 1645 | } else if (l_left > l_right) { |
|
1648 | 1646 | return 1; |
|
1649 | 1647 | } |
|
1650 | 1648 | return 0; |
|
1651 | 1649 | } |
|
1652 | 1650 | |
|
1653 | 1651 | static PyObject *index_slicechunktodensity(indexObject *self, PyObject *args) |
|
1654 | 1652 | { |
|
1655 | 1653 | /* method arguments */ |
|
1656 | 1654 | PyObject *list_revs = NULL; /* revisions in the chain */ |
|
1657 | 1655 | double targetdensity = 0; /* min density to achieve */ |
|
1658 | 1656 | Py_ssize_t mingapsize = 0; /* threshold to ignore gaps */ |
|
1659 | 1657 | |
|
1660 | 1658 | /* other core variables */ |
|
1661 | 1659 | Py_ssize_t idxlen = index_length(self); |
|
1662 | 1660 | Py_ssize_t i; /* used for various iteration */ |
|
1663 | 1661 | PyObject *result = NULL; /* the final return of the function */ |
|
1664 | 1662 | |
|
1665 | 1663 | /* generic information about the delta chain being slice */ |
|
1666 | 1664 | Py_ssize_t num_revs = 0; /* size of the full delta chain */ |
|
1667 | 1665 | Py_ssize_t *revs = NULL; /* native array of revision in the chain */ |
|
1668 | 1666 | int64_t chainpayload = 0; /* sum of all delta in the chain */ |
|
1669 | 1667 | int64_t deltachainspan = 0; /* distance from first byte to last byte */ |
|
1670 | 1668 | |
|
1671 | 1669 | /* variable used for slicing the delta chain */ |
|
1672 | 1670 | int64_t readdata = 0; /* amount of data currently planned to be read */ |
|
1673 | 1671 | double density = 0; /* ration of payload data compared to read ones */ |
|
1674 | 1672 | int64_t previous_end; |
|
1675 | 1673 | struct Gap *gaps = NULL; /* array of notable gap in the chain */ |
|
1676 | 1674 | Py_ssize_t num_gaps = |
|
1677 | 1675 | 0; /* total number of notable gap recorded so far */ |
|
1678 | 1676 | Py_ssize_t *selected_indices = NULL; /* indices of gap skipped over */ |
|
1679 | 1677 | Py_ssize_t num_selected = 0; /* number of gaps skipped */ |
|
1680 | 1678 | PyObject *chunk = NULL; /* individual slice */ |
|
1681 | 1679 | PyObject *allchunks = NULL; /* all slices */ |
|
1682 | 1680 | Py_ssize_t previdx; |
|
1683 | 1681 | |
|
1684 | 1682 | /* parsing argument */ |
|
1685 | 1683 | if (!PyArg_ParseTuple(args, "O!dn", &PyList_Type, &list_revs, |
|
1686 | 1684 | &targetdensity, &mingapsize)) { |
|
1687 | 1685 | goto bail; |
|
1688 | 1686 | } |
|
1689 | 1687 | |
|
1690 | 1688 | /* If the delta chain contains a single element, we do not need slicing |
|
1691 | 1689 | */ |
|
1692 | 1690 | num_revs = PyList_GET_SIZE(list_revs); |
|
1693 | 1691 | if (num_revs <= 1) { |
|
1694 | 1692 | result = PyTuple_Pack(1, list_revs); |
|
1695 | 1693 | goto done; |
|
1696 | 1694 | } |
|
1697 | 1695 | |
|
1698 | 1696 | /* Turn the python list into a native integer array (for efficiency) */ |
|
1699 | 1697 | revs = (Py_ssize_t *)calloc(num_revs, sizeof(Py_ssize_t)); |
|
1700 | 1698 | if (revs == NULL) { |
|
1701 | 1699 | PyErr_NoMemory(); |
|
1702 | 1700 | goto bail; |
|
1703 | 1701 | } |
|
1704 | 1702 | for (i = 0; i < num_revs; i++) { |
|
1705 | 1703 | Py_ssize_t revnum = |
|
1706 | 1704 | PyLong_AsLong(PyList_GET_ITEM(list_revs, i)); |
|
1707 | 1705 | if (revnum == -1 && PyErr_Occurred()) { |
|
1708 | 1706 | goto bail; |
|
1709 | 1707 | } |
|
1710 | 1708 | if (revnum < nullrev || revnum >= idxlen) { |
|
1711 | 1709 | PyErr_Format(PyExc_IndexError, |
|
1712 | 1710 | "index out of range: %zd", revnum); |
|
1713 | 1711 | goto bail; |
|
1714 | 1712 | } |
|
1715 | 1713 | revs[i] = revnum; |
|
1716 | 1714 | } |
|
1717 | 1715 | |
|
1718 | 1716 | /* Compute and check various property of the unsliced delta chain */ |
|
1719 | 1717 | deltachainspan = index_segment_span(self, revs[0], revs[num_revs - 1]); |
|
1720 | 1718 | if (deltachainspan < 0) { |
|
1721 | 1719 | goto bail; |
|
1722 | 1720 | } |
|
1723 | 1721 | |
|
1724 | 1722 | if (deltachainspan <= mingapsize) { |
|
1725 | 1723 | result = PyTuple_Pack(1, list_revs); |
|
1726 | 1724 | goto done; |
|
1727 | 1725 | } |
|
1728 | 1726 | chainpayload = 0; |
|
1729 | 1727 | for (i = 0; i < num_revs; i++) { |
|
1730 | 1728 | int tmp = index_get_length(self, revs[i]); |
|
1731 | 1729 | if (tmp < 0) { |
|
1732 | 1730 | goto bail; |
|
1733 | 1731 | } |
|
1734 | 1732 | chainpayload += tmp; |
|
1735 | 1733 | } |
|
1736 | 1734 | |
|
1737 | 1735 | readdata = deltachainspan; |
|
1738 | 1736 | density = 1.0; |
|
1739 | 1737 | |
|
1740 | 1738 | if (0 < deltachainspan) { |
|
1741 | 1739 | density = (double)chainpayload / (double)deltachainspan; |
|
1742 | 1740 | } |
|
1743 | 1741 | |
|
1744 | 1742 | if (density >= targetdensity) { |
|
1745 | 1743 | result = PyTuple_Pack(1, list_revs); |
|
1746 | 1744 | goto done; |
|
1747 | 1745 | } |
|
1748 | 1746 | |
|
1749 | 1747 | /* if chain is too sparse, look for relevant gaps */ |
|
1750 | 1748 | gaps = (struct Gap *)calloc(num_revs, sizeof(struct Gap)); |
|
1751 | 1749 | if (gaps == NULL) { |
|
1752 | 1750 | PyErr_NoMemory(); |
|
1753 | 1751 | goto bail; |
|
1754 | 1752 | } |
|
1755 | 1753 | |
|
1756 | 1754 | previous_end = -1; |
|
1757 | 1755 | for (i = 0; i < num_revs; i++) { |
|
1758 | 1756 | int64_t revstart; |
|
1759 | 1757 | int revsize; |
|
1760 | 1758 | revstart = index_get_start(self, revs[i]); |
|
1761 | 1759 | if (revstart < 0) { |
|
1762 | 1760 | goto bail; |
|
1763 | 1761 | }; |
|
1764 | 1762 | revsize = index_get_length(self, revs[i]); |
|
1765 | 1763 | if (revsize < 0) { |
|
1766 | 1764 | goto bail; |
|
1767 | 1765 | }; |
|
1768 | 1766 | if (revsize == 0) { |
|
1769 | 1767 | continue; |
|
1770 | 1768 | } |
|
1771 | 1769 | if (previous_end >= 0) { |
|
1772 | 1770 | int64_t gapsize = revstart - previous_end; |
|
1773 | 1771 | if (gapsize > mingapsize) { |
|
1774 | 1772 | gaps[num_gaps].size = gapsize; |
|
1775 | 1773 | gaps[num_gaps].idx = i; |
|
1776 | 1774 | num_gaps += 1; |
|
1777 | 1775 | } |
|
1778 | 1776 | } |
|
1779 | 1777 | previous_end = revstart + revsize; |
|
1780 | 1778 | } |
|
1781 | 1779 | if (num_gaps == 0) { |
|
1782 | 1780 | result = PyTuple_Pack(1, list_revs); |
|
1783 | 1781 | goto done; |
|
1784 | 1782 | } |
|
1785 | 1783 | qsort(gaps, num_gaps, sizeof(struct Gap), &gap_compare); |
|
1786 | 1784 | |
|
1787 | 1785 | /* Slice the largest gap first, they improve the density the most */ |
|
1788 | 1786 | selected_indices = |
|
1789 | 1787 | (Py_ssize_t *)malloc((num_gaps + 1) * sizeof(Py_ssize_t)); |
|
1790 | 1788 | if (selected_indices == NULL) { |
|
1791 | 1789 | PyErr_NoMemory(); |
|
1792 | 1790 | goto bail; |
|
1793 | 1791 | } |
|
1794 | 1792 | |
|
1795 | 1793 | for (i = num_gaps - 1; i >= 0; i--) { |
|
1796 | 1794 | selected_indices[num_selected] = gaps[i].idx; |
|
1797 | 1795 | readdata -= gaps[i].size; |
|
1798 | 1796 | num_selected += 1; |
|
1799 | 1797 | if (readdata <= 0) { |
|
1800 | 1798 | density = 1.0; |
|
1801 | 1799 | } else { |
|
1802 | 1800 | density = (double)chainpayload / (double)readdata; |
|
1803 | 1801 | } |
|
1804 | 1802 | if (density >= targetdensity) { |
|
1805 | 1803 | break; |
|
1806 | 1804 | } |
|
1807 | 1805 | } |
|
1808 | 1806 | qsort(selected_indices, num_selected, sizeof(Py_ssize_t), |
|
1809 | 1807 | &Py_ssize_t_compare); |
|
1810 | 1808 | |
|
1811 | 1809 | /* create the resulting slice */ |
|
1812 | 1810 | allchunks = PyList_New(0); |
|
1813 | 1811 | if (allchunks == NULL) { |
|
1814 | 1812 | goto bail; |
|
1815 | 1813 | } |
|
1816 | 1814 | previdx = 0; |
|
1817 | 1815 | selected_indices[num_selected] = num_revs; |
|
1818 | 1816 | for (i = 0; i <= num_selected; i++) { |
|
1819 | 1817 | Py_ssize_t idx = selected_indices[i]; |
|
1820 | 1818 | Py_ssize_t endidx = trim_endidx(self, revs, previdx, idx); |
|
1821 | 1819 | if (endidx < 0) { |
|
1822 | 1820 | goto bail; |
|
1823 | 1821 | } |
|
1824 | 1822 | if (previdx < endidx) { |
|
1825 | 1823 | chunk = PyList_GetSlice(list_revs, previdx, endidx); |
|
1826 | 1824 | if (chunk == NULL) { |
|
1827 | 1825 | goto bail; |
|
1828 | 1826 | } |
|
1829 | 1827 | if (PyList_Append(allchunks, chunk) == -1) { |
|
1830 | 1828 | goto bail; |
|
1831 | 1829 | } |
|
1832 | 1830 | Py_DECREF(chunk); |
|
1833 | 1831 | chunk = NULL; |
|
1834 | 1832 | } |
|
1835 | 1833 | previdx = idx; |
|
1836 | 1834 | } |
|
1837 | 1835 | result = allchunks; |
|
1838 | 1836 | goto done; |
|
1839 | 1837 | |
|
1840 | 1838 | bail: |
|
1841 | 1839 | Py_XDECREF(allchunks); |
|
1842 | 1840 | Py_XDECREF(chunk); |
|
1843 | 1841 | done: |
|
1844 | 1842 | free(revs); |
|
1845 | 1843 | free(gaps); |
|
1846 | 1844 | free(selected_indices); |
|
1847 | 1845 | return result; |
|
1848 | 1846 | } |
|
1849 | 1847 | |
|
1850 | 1848 | static inline int nt_level(const char *node, Py_ssize_t level) |
|
1851 | 1849 | { |
|
1852 | 1850 | int v = node[level >> 1]; |
|
1853 | 1851 | if (!(level & 1)) |
|
1854 | 1852 | v >>= 4; |
|
1855 | 1853 | return v & 0xf; |
|
1856 | 1854 | } |
|
1857 | 1855 | |
|
1858 | 1856 | /* |
|
1859 | 1857 | * Return values: |
|
1860 | 1858 | * |
|
1861 | 1859 | * -4: match is ambiguous (multiple candidates) |
|
1862 | 1860 | * -2: not found |
|
1863 | 1861 | * rest: valid rev |
|
1864 | 1862 | */ |
|
1865 | 1863 | static int nt_find(nodetree *self, const char *node, Py_ssize_t nodelen, |
|
1866 | 1864 | int hex) |
|
1867 | 1865 | { |
|
1868 | 1866 | int (*getnybble)(const char *, Py_ssize_t) = hex ? hexdigit : nt_level; |
|
1869 | 1867 | int level, maxlevel, off; |
|
1870 | 1868 | |
|
1871 | 1869 | /* If the input is binary, do a fast check for the nullid first. */ |
|
1872 | 1870 | if (!hex && nodelen == self->nodelen && node[0] == '\0' && |
|
1873 | 1871 | node[1] == '\0' && memcmp(node, nullid, self->nodelen) == 0) |
|
1874 | 1872 | return -1; |
|
1875 | 1873 | |
|
1876 | 1874 | if (hex) |
|
1877 | 1875 | maxlevel = nodelen; |
|
1878 | 1876 | else |
|
1879 | 1877 | maxlevel = 2 * nodelen; |
|
1880 | 1878 | if (maxlevel > 2 * self->nodelen) |
|
1881 | 1879 | maxlevel = 2 * self->nodelen; |
|
1882 | 1880 | |
|
1883 | 1881 | for (level = off = 0; level < maxlevel; level++) { |
|
1884 | 1882 | int k = getnybble(node, level); |
|
1885 | 1883 | nodetreenode *n = &self->nodes[off]; |
|
1886 | 1884 | int v = n->children[k]; |
|
1887 | 1885 | |
|
1888 | 1886 | if (v < 0) { |
|
1889 | 1887 | const char *n; |
|
1890 | 1888 | Py_ssize_t i; |
|
1891 | 1889 | |
|
1892 | 1890 | v = -(v + 2); |
|
1893 | 1891 | n = index_node(self->index, v); |
|
1894 | 1892 | if (n == NULL) |
|
1895 | 1893 | return -2; |
|
1896 | 1894 | for (i = level; i < maxlevel; i++) |
|
1897 | 1895 | if (getnybble(node, i) != nt_level(n, i)) |
|
1898 | 1896 | return -2; |
|
1899 | 1897 | return v; |
|
1900 | 1898 | } |
|
1901 | 1899 | if (v == 0) |
|
1902 | 1900 | return -2; |
|
1903 | 1901 | off = v; |
|
1904 | 1902 | } |
|
1905 | 1903 | /* multiple matches against an ambiguous prefix */ |
|
1906 | 1904 | return -4; |
|
1907 | 1905 | } |
|
1908 | 1906 | |
|
1909 | 1907 | static int nt_new(nodetree *self) |
|
1910 | 1908 | { |
|
1911 | 1909 | if (self->length == self->capacity) { |
|
1912 | 1910 | size_t newcapacity; |
|
1913 | 1911 | nodetreenode *newnodes; |
|
1914 | 1912 | newcapacity = self->capacity * 2; |
|
1915 | 1913 | if (newcapacity >= SIZE_MAX / sizeof(nodetreenode)) { |
|
1916 | 1914 | PyErr_SetString(PyExc_MemoryError, |
|
1917 | 1915 | "overflow in nt_new"); |
|
1918 | 1916 | return -1; |
|
1919 | 1917 | } |
|
1920 | 1918 | newnodes = |
|
1921 | 1919 | realloc(self->nodes, newcapacity * sizeof(nodetreenode)); |
|
1922 | 1920 | if (newnodes == NULL) { |
|
1923 | 1921 | PyErr_SetString(PyExc_MemoryError, "out of memory"); |
|
1924 | 1922 | return -1; |
|
1925 | 1923 | } |
|
1926 | 1924 | self->capacity = newcapacity; |
|
1927 | 1925 | self->nodes = newnodes; |
|
1928 | 1926 | memset(&self->nodes[self->length], 0, |
|
1929 | 1927 | sizeof(nodetreenode) * (self->capacity - self->length)); |
|
1930 | 1928 | } |
|
1931 | 1929 | return self->length++; |
|
1932 | 1930 | } |
|
1933 | 1931 | |
|
1934 | 1932 | static int nt_insert(nodetree *self, const char *node, int rev) |
|
1935 | 1933 | { |
|
1936 | 1934 | int level = 0; |
|
1937 | 1935 | int off = 0; |
|
1938 | 1936 | |
|
1939 | 1937 | while (level < 2 * self->nodelen) { |
|
1940 | 1938 | int k = nt_level(node, level); |
|
1941 | 1939 | nodetreenode *n; |
|
1942 | 1940 | int v; |
|
1943 | 1941 | |
|
1944 | 1942 | n = &self->nodes[off]; |
|
1945 | 1943 | v = n->children[k]; |
|
1946 | 1944 | |
|
1947 | 1945 | if (v == 0) { |
|
1948 | 1946 | n->children[k] = -rev - 2; |
|
1949 | 1947 | return 0; |
|
1950 | 1948 | } |
|
1951 | 1949 | if (v < 0) { |
|
1952 | 1950 | const char *oldnode = |
|
1953 | 1951 | index_node_existing(self->index, -(v + 2)); |
|
1954 | 1952 | int noff; |
|
1955 | 1953 | |
|
1956 | 1954 | if (oldnode == NULL) |
|
1957 | 1955 | return -1; |
|
1958 | 1956 | if (!memcmp(oldnode, node, self->nodelen)) { |
|
1959 | 1957 | n->children[k] = -rev - 2; |
|
1960 | 1958 | return 0; |
|
1961 | 1959 | } |
|
1962 | 1960 | noff = nt_new(self); |
|
1963 | 1961 | if (noff == -1) |
|
1964 | 1962 | return -1; |
|
1965 | 1963 | /* self->nodes may have been changed by realloc */ |
|
1966 | 1964 | self->nodes[off].children[k] = noff; |
|
1967 | 1965 | off = noff; |
|
1968 | 1966 | n = &self->nodes[off]; |
|
1969 | 1967 | n->children[nt_level(oldnode, ++level)] = v; |
|
1970 | 1968 | if (level > self->depth) |
|
1971 | 1969 | self->depth = level; |
|
1972 | 1970 | self->splits += 1; |
|
1973 | 1971 | } else { |
|
1974 | 1972 | level += 1; |
|
1975 | 1973 | off = v; |
|
1976 | 1974 | } |
|
1977 | 1975 | } |
|
1978 | 1976 | |
|
1979 | 1977 | return -1; |
|
1980 | 1978 | } |
|
1981 | 1979 | |
|
1982 | 1980 | static PyObject *ntobj_insert(nodetreeObject *self, PyObject *args) |
|
1983 | 1981 | { |
|
1984 | 1982 | Py_ssize_t rev; |
|
1985 | 1983 | const char *node; |
|
1986 | 1984 | Py_ssize_t length; |
|
1987 | 1985 | if (!PyArg_ParseTuple(args, "n", &rev)) |
|
1988 | 1986 | return NULL; |
|
1989 | 1987 | length = index_length(self->nt.index); |
|
1990 | 1988 | if (rev < 0 || rev >= length) { |
|
1991 | 1989 | PyErr_SetString(PyExc_ValueError, "revlog index out of range"); |
|
1992 | 1990 | return NULL; |
|
1993 | 1991 | } |
|
1994 | 1992 | node = index_node_existing(self->nt.index, rev); |
|
1995 | 1993 | if (nt_insert(&self->nt, node, (int)rev) == -1) |
|
1996 | 1994 | return NULL; |
|
1997 | 1995 | Py_RETURN_NONE; |
|
1998 | 1996 | } |
|
1999 | 1997 | |
|
2000 | 1998 | static int nt_delete_node(nodetree *self, const char *node) |
|
2001 | 1999 | { |
|
2002 | 2000 | /* rev==-2 happens to get encoded as 0, which is interpreted as not set |
|
2003 | 2001 | */ |
|
2004 | 2002 | return nt_insert(self, node, -2); |
|
2005 | 2003 | } |
|
2006 | 2004 | |
|
2007 | 2005 | static int nt_init(nodetree *self, indexObject *index, unsigned capacity) |
|
2008 | 2006 | { |
|
2009 | 2007 | /* Initialize before overflow-checking to avoid nt_dealloc() crash. */ |
|
2010 | 2008 | self->nodes = NULL; |
|
2011 | 2009 | |
|
2012 | 2010 | self->index = index; |
|
2013 | 2011 | /* The input capacity is in terms of revisions, while the field is in |
|
2014 | 2012 | * terms of nodetree nodes. */ |
|
2015 | 2013 | self->capacity = (capacity < 4 ? 4 : capacity / 2); |
|
2016 | 2014 | self->nodelen = index->nodelen; |
|
2017 | 2015 | self->depth = 0; |
|
2018 | 2016 | self->splits = 0; |
|
2019 | 2017 | if (self->capacity > SIZE_MAX / sizeof(nodetreenode)) { |
|
2020 | 2018 | PyErr_SetString(PyExc_ValueError, "overflow in init_nt"); |
|
2021 | 2019 | return -1; |
|
2022 | 2020 | } |
|
2023 | 2021 | self->nodes = calloc(self->capacity, sizeof(nodetreenode)); |
|
2024 | 2022 | if (self->nodes == NULL) { |
|
2025 | 2023 | PyErr_NoMemory(); |
|
2026 | 2024 | return -1; |
|
2027 | 2025 | } |
|
2028 | 2026 | self->length = 1; |
|
2029 | 2027 | return 0; |
|
2030 | 2028 | } |
|
2031 | 2029 | |
|
2032 | 2030 | static int ntobj_init(nodetreeObject *self, PyObject *args) |
|
2033 | 2031 | { |
|
2034 | 2032 | PyObject *index; |
|
2035 | 2033 | unsigned capacity; |
|
2036 | 2034 | if (!PyArg_ParseTuple(args, "O!I", &HgRevlogIndex_Type, &index, |
|
2037 | 2035 | &capacity)) |
|
2038 | 2036 | return -1; |
|
2039 | 2037 | Py_INCREF(index); |
|
2040 | 2038 | return nt_init(&self->nt, (indexObject *)index, capacity); |
|
2041 | 2039 | } |
|
2042 | 2040 | |
|
2043 | 2041 | static int nt_partialmatch(nodetree *self, const char *node, Py_ssize_t nodelen) |
|
2044 | 2042 | { |
|
2045 | 2043 | return nt_find(self, node, nodelen, 1); |
|
2046 | 2044 | } |
|
2047 | 2045 | |
|
2048 | 2046 | /* |
|
2049 | 2047 | * Find the length of the shortest unique prefix of node. |
|
2050 | 2048 | * |
|
2051 | 2049 | * Return values: |
|
2052 | 2050 | * |
|
2053 | 2051 | * -3: error (exception set) |
|
2054 | 2052 | * -2: not found (no exception set) |
|
2055 | 2053 | * rest: length of shortest prefix |
|
2056 | 2054 | */ |
|
2057 | 2055 | static int nt_shortest(nodetree *self, const char *node) |
|
2058 | 2056 | { |
|
2059 | 2057 | int level, off; |
|
2060 | 2058 | |
|
2061 | 2059 | for (level = off = 0; level < 2 * self->nodelen; level++) { |
|
2062 | 2060 | int k, v; |
|
2063 | 2061 | nodetreenode *n = &self->nodes[off]; |
|
2064 | 2062 | k = nt_level(node, level); |
|
2065 | 2063 | v = n->children[k]; |
|
2066 | 2064 | if (v < 0) { |
|
2067 | 2065 | const char *n; |
|
2068 | 2066 | v = -(v + 2); |
|
2069 | 2067 | n = index_node_existing(self->index, v); |
|
2070 | 2068 | if (n == NULL) |
|
2071 | 2069 | return -3; |
|
2072 | 2070 | if (memcmp(node, n, self->nodelen) != 0) |
|
2073 | 2071 | /* |
|
2074 | 2072 | * Found a unique prefix, but it wasn't for the |
|
2075 | 2073 | * requested node (i.e the requested node does |
|
2076 | 2074 | * not exist). |
|
2077 | 2075 | */ |
|
2078 | 2076 | return -2; |
|
2079 | 2077 | return level + 1; |
|
2080 | 2078 | } |
|
2081 | 2079 | if (v == 0) |
|
2082 | 2080 | return -2; |
|
2083 | 2081 | off = v; |
|
2084 | 2082 | } |
|
2085 | 2083 | /* |
|
2086 | 2084 | * The node was still not unique after 40 hex digits, so this won't |
|
2087 | 2085 | * happen. Also, if we get here, then there's a programming error in |
|
2088 | 2086 | * this file that made us insert a node longer than 40 hex digits. |
|
2089 | 2087 | */ |
|
2090 | 2088 | PyErr_SetString(PyExc_Exception, "broken node tree"); |
|
2091 | 2089 | return -3; |
|
2092 | 2090 | } |
|
2093 | 2091 | |
|
2094 | 2092 | static PyObject *ntobj_shortest(nodetreeObject *self, PyObject *args) |
|
2095 | 2093 | { |
|
2096 | 2094 | PyObject *val; |
|
2097 | 2095 | char *node; |
|
2098 | 2096 | int length; |
|
2099 | 2097 | |
|
2100 | 2098 | if (!PyArg_ParseTuple(args, "O", &val)) |
|
2101 | 2099 | return NULL; |
|
2102 | 2100 | if (node_check(self->nt.nodelen, val, &node) == -1) |
|
2103 | 2101 | return NULL; |
|
2104 | 2102 | |
|
2105 | 2103 | length = nt_shortest(&self->nt, node); |
|
2106 | 2104 | if (length == -3) |
|
2107 | 2105 | return NULL; |
|
2108 | 2106 | if (length == -2) { |
|
2109 | 2107 | raise_revlog_error(); |
|
2110 | 2108 | return NULL; |
|
2111 | 2109 | } |
|
2112 | 2110 | return PyLong_FromLong(length); |
|
2113 | 2111 | } |
|
2114 | 2112 | |
|
2115 | 2113 | static void nt_dealloc(nodetree *self) |
|
2116 | 2114 | { |
|
2117 | 2115 | free(self->nodes); |
|
2118 | 2116 | self->nodes = NULL; |
|
2119 | 2117 | } |
|
2120 | 2118 | |
|
2121 | 2119 | static void ntobj_dealloc(nodetreeObject *self) |
|
2122 | 2120 | { |
|
2123 | 2121 | Py_XDECREF(self->nt.index); |
|
2124 | 2122 | nt_dealloc(&self->nt); |
|
2125 | 2123 | PyObject_Del(self); |
|
2126 | 2124 | } |
|
2127 | 2125 | |
|
2128 | 2126 | static PyMethodDef ntobj_methods[] = { |
|
2129 | 2127 | {"insert", (PyCFunction)ntobj_insert, METH_VARARGS, |
|
2130 | 2128 | "insert an index entry"}, |
|
2131 | 2129 | {"shortest", (PyCFunction)ntobj_shortest, METH_VARARGS, |
|
2132 | 2130 | "find length of shortest hex nodeid of a binary ID"}, |
|
2133 | 2131 | {NULL} /* Sentinel */ |
|
2134 | 2132 | }; |
|
2135 | 2133 | |
|
2136 | 2134 | static PyTypeObject nodetreeType = { |
|
2137 | 2135 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ |
|
2138 | 2136 | "parsers.nodetree", /* tp_name */ |
|
2139 | 2137 | sizeof(nodetreeObject), /* tp_basicsize */ |
|
2140 | 2138 | 0, /* tp_itemsize */ |
|
2141 | 2139 | (destructor)ntobj_dealloc, /* tp_dealloc */ |
|
2142 | 2140 | 0, /* tp_print */ |
|
2143 | 2141 | 0, /* tp_getattr */ |
|
2144 | 2142 | 0, /* tp_setattr */ |
|
2145 | 2143 | 0, /* tp_compare */ |
|
2146 | 2144 | 0, /* tp_repr */ |
|
2147 | 2145 | 0, /* tp_as_number */ |
|
2148 | 2146 | 0, /* tp_as_sequence */ |
|
2149 | 2147 | 0, /* tp_as_mapping */ |
|
2150 | 2148 | 0, /* tp_hash */ |
|
2151 | 2149 | 0, /* tp_call */ |
|
2152 | 2150 | 0, /* tp_str */ |
|
2153 | 2151 | 0, /* tp_getattro */ |
|
2154 | 2152 | 0, /* tp_setattro */ |
|
2155 | 2153 | 0, /* tp_as_buffer */ |
|
2156 | 2154 | Py_TPFLAGS_DEFAULT, /* tp_flags */ |
|
2157 | 2155 | "nodetree", /* tp_doc */ |
|
2158 | 2156 | 0, /* tp_traverse */ |
|
2159 | 2157 | 0, /* tp_clear */ |
|
2160 | 2158 | 0, /* tp_richcompare */ |
|
2161 | 2159 | 0, /* tp_weaklistoffset */ |
|
2162 | 2160 | 0, /* tp_iter */ |
|
2163 | 2161 | 0, /* tp_iternext */ |
|
2164 | 2162 | ntobj_methods, /* tp_methods */ |
|
2165 | 2163 | 0, /* tp_members */ |
|
2166 | 2164 | 0, /* tp_getset */ |
|
2167 | 2165 | 0, /* tp_base */ |
|
2168 | 2166 | 0, /* tp_dict */ |
|
2169 | 2167 | 0, /* tp_descr_get */ |
|
2170 | 2168 | 0, /* tp_descr_set */ |
|
2171 | 2169 | 0, /* tp_dictoffset */ |
|
2172 | 2170 | (initproc)ntobj_init, /* tp_init */ |
|
2173 | 2171 | 0, /* tp_alloc */ |
|
2174 | 2172 | }; |
|
2175 | 2173 | |
|
2176 | 2174 | static int index_init_nt(indexObject *self) |
|
2177 | 2175 | { |
|
2178 | 2176 | if (!self->ntinitialized) { |
|
2179 | 2177 | if (nt_init(&self->nt, self, (int)self->length) == -1) { |
|
2180 | 2178 | nt_dealloc(&self->nt); |
|
2181 | 2179 | return -1; |
|
2182 | 2180 | } |
|
2183 | 2181 | if (nt_insert(&self->nt, nullid, -1) == -1) { |
|
2184 | 2182 | nt_dealloc(&self->nt); |
|
2185 | 2183 | return -1; |
|
2186 | 2184 | } |
|
2187 | 2185 | self->ntinitialized = 1; |
|
2188 | 2186 | self->ntrev = (int)index_length(self); |
|
2189 | 2187 | self->ntlookups = 1; |
|
2190 | 2188 | self->ntmisses = 0; |
|
2191 | 2189 | } |
|
2192 | 2190 | return 0; |
|
2193 | 2191 | } |
|
2194 | 2192 | |
|
2195 | 2193 | /* |
|
2196 | 2194 | * Return values: |
|
2197 | 2195 | * |
|
2198 | 2196 | * -3: error (exception set) |
|
2199 | 2197 | * -2: not found (no exception set) |
|
2200 | 2198 | * rest: valid rev |
|
2201 | 2199 | */ |
|
2202 | 2200 | static int index_find_node(indexObject *self, const char *node) |
|
2203 | 2201 | { |
|
2204 | 2202 | int rev; |
|
2205 | 2203 | |
|
2206 | 2204 | if (index_init_nt(self) == -1) |
|
2207 | 2205 | return -3; |
|
2208 | 2206 | |
|
2209 | 2207 | self->ntlookups++; |
|
2210 | 2208 | rev = nt_find(&self->nt, node, self->nodelen, 0); |
|
2211 | 2209 | if (rev >= -1) |
|
2212 | 2210 | return rev; |
|
2213 | 2211 | |
|
2214 | 2212 | /* |
|
2215 | 2213 | * For the first handful of lookups, we scan the entire index, |
|
2216 | 2214 | * and cache only the matching nodes. This optimizes for cases |
|
2217 | 2215 | * like "hg tip", where only a few nodes are accessed. |
|
2218 | 2216 | * |
|
2219 | 2217 | * After that, we cache every node we visit, using a single |
|
2220 | 2218 | * scan amortized over multiple lookups. This gives the best |
|
2221 | 2219 | * bulk performance, e.g. for "hg log". |
|
2222 | 2220 | */ |
|
2223 | 2221 | if (self->ntmisses++ < 4) { |
|
2224 | 2222 | for (rev = self->ntrev - 1; rev >= 0; rev--) { |
|
2225 | 2223 | const char *n = index_node_existing(self, rev); |
|
2226 | 2224 | if (n == NULL) |
|
2227 | 2225 | return -3; |
|
2228 | 2226 | if (memcmp(node, n, self->nodelen) == 0) { |
|
2229 | 2227 | if (nt_insert(&self->nt, n, rev) == -1) |
|
2230 | 2228 | return -3; |
|
2231 | 2229 | break; |
|
2232 | 2230 | } |
|
2233 | 2231 | } |
|
2234 | 2232 | } else { |
|
2235 | 2233 | for (rev = self->ntrev - 1; rev >= 0; rev--) { |
|
2236 | 2234 | const char *n = index_node_existing(self, rev); |
|
2237 | 2235 | if (n == NULL) |
|
2238 | 2236 | return -3; |
|
2239 | 2237 | if (nt_insert(&self->nt, n, rev) == -1) { |
|
2240 | 2238 | self->ntrev = rev + 1; |
|
2241 | 2239 | return -3; |
|
2242 | 2240 | } |
|
2243 | 2241 | if (memcmp(node, n, self->nodelen) == 0) { |
|
2244 | 2242 | break; |
|
2245 | 2243 | } |
|
2246 | 2244 | } |
|
2247 | 2245 | self->ntrev = rev; |
|
2248 | 2246 | } |
|
2249 | 2247 | |
|
2250 | 2248 | if (rev >= 0) |
|
2251 | 2249 | return rev; |
|
2252 | 2250 | return -2; |
|
2253 | 2251 | } |
|
2254 | 2252 | |
|
2255 | 2253 | static PyObject *index_getitem(indexObject *self, PyObject *value) |
|
2256 | 2254 | { |
|
2257 | 2255 | char *node; |
|
2258 | 2256 | int rev; |
|
2259 | 2257 | |
|
2260 | 2258 | if (PyLong_Check(value)) { |
|
2261 | 2259 | long idx; |
|
2262 | 2260 | if (!pylong_to_long(value, &idx)) { |
|
2263 | 2261 | return NULL; |
|
2264 | 2262 | } |
|
2265 | 2263 | return index_get(self, idx); |
|
2266 | 2264 | } |
|
2267 | 2265 | |
|
2268 | 2266 | if (node_check(self->nodelen, value, &node) == -1) |
|
2269 | 2267 | return NULL; |
|
2270 | 2268 | rev = index_find_node(self, node); |
|
2271 | 2269 | if (rev >= -1) |
|
2272 | 2270 | return PyLong_FromLong(rev); |
|
2273 | 2271 | if (rev == -2) |
|
2274 | 2272 | raise_revlog_error(); |
|
2275 | 2273 | return NULL; |
|
2276 | 2274 | } |
|
2277 | 2275 | |
|
2278 | 2276 | /* |
|
2279 | 2277 | * Fully populate the radix tree. |
|
2280 | 2278 | */ |
|
2281 | 2279 | static int index_populate_nt(indexObject *self) |
|
2282 | 2280 | { |
|
2283 | 2281 | int rev; |
|
2284 | 2282 | if (self->ntrev > 0) { |
|
2285 | 2283 | for (rev = self->ntrev - 1; rev >= 0; rev--) { |
|
2286 | 2284 | const char *n = index_node_existing(self, rev); |
|
2287 | 2285 | if (n == NULL) |
|
2288 | 2286 | return -1; |
|
2289 | 2287 | if (nt_insert(&self->nt, n, rev) == -1) |
|
2290 | 2288 | return -1; |
|
2291 | 2289 | } |
|
2292 | 2290 | self->ntrev = -1; |
|
2293 | 2291 | } |
|
2294 | 2292 | return 0; |
|
2295 | 2293 | } |
|
2296 | 2294 | |
|
2297 | 2295 | static PyObject *index_partialmatch(indexObject *self, PyObject *args) |
|
2298 | 2296 | { |
|
2299 | 2297 | const char *fullnode; |
|
2300 | 2298 | Py_ssize_t nodelen; |
|
2301 | 2299 | char *node; |
|
2302 | 2300 | int rev, i; |
|
2303 | 2301 | |
|
2304 |
if (!PyArg_ParseTuple(args, |
|
|
2302 | if (!PyArg_ParseTuple(args, "y#", &node, &nodelen)) | |
|
2305 | 2303 | return NULL; |
|
2306 | 2304 | |
|
2307 | 2305 | if (nodelen < 1) { |
|
2308 | 2306 | PyErr_SetString(PyExc_ValueError, "key too short"); |
|
2309 | 2307 | return NULL; |
|
2310 | 2308 | } |
|
2311 | 2309 | |
|
2312 | 2310 | if (nodelen > 2 * self->nodelen) { |
|
2313 | 2311 | PyErr_SetString(PyExc_ValueError, "key too long"); |
|
2314 | 2312 | return NULL; |
|
2315 | 2313 | } |
|
2316 | 2314 | |
|
2317 | 2315 | for (i = 0; i < nodelen; i++) |
|
2318 | 2316 | hexdigit(node, i); |
|
2319 | 2317 | if (PyErr_Occurred()) { |
|
2320 | 2318 | /* input contains non-hex characters */ |
|
2321 | 2319 | PyErr_Clear(); |
|
2322 | 2320 | Py_RETURN_NONE; |
|
2323 | 2321 | } |
|
2324 | 2322 | |
|
2325 | 2323 | if (index_init_nt(self) == -1) |
|
2326 | 2324 | return NULL; |
|
2327 | 2325 | if (index_populate_nt(self) == -1) |
|
2328 | 2326 | return NULL; |
|
2329 | 2327 | rev = nt_partialmatch(&self->nt, node, nodelen); |
|
2330 | 2328 | |
|
2331 | 2329 | switch (rev) { |
|
2332 | 2330 | case -4: |
|
2333 | 2331 | raise_revlog_error(); |
|
2334 | 2332 | return NULL; |
|
2335 | 2333 | case -2: |
|
2336 | 2334 | Py_RETURN_NONE; |
|
2337 | 2335 | case -1: |
|
2338 | 2336 | return PyBytes_FromStringAndSize(nullid, self->nodelen); |
|
2339 | 2337 | } |
|
2340 | 2338 | |
|
2341 | 2339 | fullnode = index_node_existing(self, rev); |
|
2342 | 2340 | if (fullnode == NULL) { |
|
2343 | 2341 | return NULL; |
|
2344 | 2342 | } |
|
2345 | 2343 | return PyBytes_FromStringAndSize(fullnode, self->nodelen); |
|
2346 | 2344 | } |
|
2347 | 2345 | |
|
2348 | 2346 | static PyObject *index_shortest(indexObject *self, PyObject *args) |
|
2349 | 2347 | { |
|
2350 | 2348 | PyObject *val; |
|
2351 | 2349 | char *node; |
|
2352 | 2350 | int length; |
|
2353 | 2351 | |
|
2354 | 2352 | if (!PyArg_ParseTuple(args, "O", &val)) |
|
2355 | 2353 | return NULL; |
|
2356 | 2354 | if (node_check(self->nodelen, val, &node) == -1) |
|
2357 | 2355 | return NULL; |
|
2358 | 2356 | |
|
2359 | 2357 | self->ntlookups++; |
|
2360 | 2358 | if (index_init_nt(self) == -1) |
|
2361 | 2359 | return NULL; |
|
2362 | 2360 | if (index_populate_nt(self) == -1) |
|
2363 | 2361 | return NULL; |
|
2364 | 2362 | length = nt_shortest(&self->nt, node); |
|
2365 | 2363 | if (length == -3) |
|
2366 | 2364 | return NULL; |
|
2367 | 2365 | if (length == -2) { |
|
2368 | 2366 | raise_revlog_error(); |
|
2369 | 2367 | return NULL; |
|
2370 | 2368 | } |
|
2371 | 2369 | return PyLong_FromLong(length); |
|
2372 | 2370 | } |
|
2373 | 2371 | |
|
2374 | 2372 | static PyObject *index_m_get(indexObject *self, PyObject *args) |
|
2375 | 2373 | { |
|
2376 | 2374 | PyObject *val; |
|
2377 | 2375 | char *node; |
|
2378 | 2376 | int rev; |
|
2379 | 2377 | |
|
2380 | 2378 | if (!PyArg_ParseTuple(args, "O", &val)) |
|
2381 | 2379 | return NULL; |
|
2382 | 2380 | if (node_check(self->nodelen, val, &node) == -1) |
|
2383 | 2381 | return NULL; |
|
2384 | 2382 | rev = index_find_node(self, node); |
|
2385 | 2383 | if (rev == -3) |
|
2386 | 2384 | return NULL; |
|
2387 | 2385 | if (rev == -2) |
|
2388 | 2386 | Py_RETURN_NONE; |
|
2389 | 2387 | return PyLong_FromLong(rev); |
|
2390 | 2388 | } |
|
2391 | 2389 | |
|
2392 | 2390 | static int index_contains(indexObject *self, PyObject *value) |
|
2393 | 2391 | { |
|
2394 | 2392 | char *node; |
|
2395 | 2393 | |
|
2396 | 2394 | if (PyLong_Check(value)) { |
|
2397 | 2395 | long rev; |
|
2398 | 2396 | if (!pylong_to_long(value, &rev)) { |
|
2399 | 2397 | return -1; |
|
2400 | 2398 | } |
|
2401 | 2399 | return rev >= -1 && rev < index_length(self); |
|
2402 | 2400 | } |
|
2403 | 2401 | |
|
2404 | 2402 | if (node_check(self->nodelen, value, &node) == -1) |
|
2405 | 2403 | return -1; |
|
2406 | 2404 | |
|
2407 | 2405 | switch (index_find_node(self, node)) { |
|
2408 | 2406 | case -3: |
|
2409 | 2407 | return -1; |
|
2410 | 2408 | case -2: |
|
2411 | 2409 | return 0; |
|
2412 | 2410 | default: |
|
2413 | 2411 | return 1; |
|
2414 | 2412 | } |
|
2415 | 2413 | } |
|
2416 | 2414 | |
|
2417 | 2415 | static PyObject *index_m_has_node(indexObject *self, PyObject *args) |
|
2418 | 2416 | { |
|
2419 | 2417 | int ret = index_contains(self, args); |
|
2420 | 2418 | if (ret < 0) |
|
2421 | 2419 | return NULL; |
|
2422 | 2420 | return PyBool_FromLong((long)ret); |
|
2423 | 2421 | } |
|
2424 | 2422 | |
|
2425 | 2423 | static PyObject *index_m_rev(indexObject *self, PyObject *val) |
|
2426 | 2424 | { |
|
2427 | 2425 | char *node; |
|
2428 | 2426 | int rev; |
|
2429 | 2427 | |
|
2430 | 2428 | if (node_check(self->nodelen, val, &node) == -1) |
|
2431 | 2429 | return NULL; |
|
2432 | 2430 | rev = index_find_node(self, node); |
|
2433 | 2431 | if (rev >= -1) |
|
2434 | 2432 | return PyLong_FromLong(rev); |
|
2435 | 2433 | if (rev == -2) |
|
2436 | 2434 | raise_revlog_error(); |
|
2437 | 2435 | return NULL; |
|
2438 | 2436 | } |
|
2439 | 2437 | |
|
2440 | 2438 | typedef uint64_t bitmask; |
|
2441 | 2439 | |
|
2442 | 2440 | /* |
|
2443 | 2441 | * Given a disjoint set of revs, return all candidates for the |
|
2444 | 2442 | * greatest common ancestor. In revset notation, this is the set |
|
2445 | 2443 | * "heads(::a and ::b and ...)" |
|
2446 | 2444 | */ |
|
2447 | 2445 | static PyObject *find_gca_candidates(indexObject *self, const int *revs, |
|
2448 | 2446 | int revcount) |
|
2449 | 2447 | { |
|
2450 | 2448 | const bitmask allseen = (1ull << revcount) - 1; |
|
2451 | 2449 | const bitmask poison = 1ull << revcount; |
|
2452 | 2450 | PyObject *gca = PyList_New(0); |
|
2453 | 2451 | int i, v, interesting; |
|
2454 | 2452 | int maxrev = -1; |
|
2455 | 2453 | bitmask sp; |
|
2456 | 2454 | bitmask *seen; |
|
2457 | 2455 | |
|
2458 | 2456 | if (gca == NULL) |
|
2459 | 2457 | return PyErr_NoMemory(); |
|
2460 | 2458 | |
|
2461 | 2459 | for (i = 0; i < revcount; i++) { |
|
2462 | 2460 | if (revs[i] > maxrev) |
|
2463 | 2461 | maxrev = revs[i]; |
|
2464 | 2462 | } |
|
2465 | 2463 | |
|
2466 | 2464 | seen = calloc(sizeof(*seen), maxrev + 1); |
|
2467 | 2465 | if (seen == NULL) { |
|
2468 | 2466 | Py_DECREF(gca); |
|
2469 | 2467 | return PyErr_NoMemory(); |
|
2470 | 2468 | } |
|
2471 | 2469 | |
|
2472 | 2470 | for (i = 0; i < revcount; i++) |
|
2473 | 2471 | seen[revs[i]] = 1ull << i; |
|
2474 | 2472 | |
|
2475 | 2473 | interesting = revcount; |
|
2476 | 2474 | |
|
2477 | 2475 | for (v = maxrev; v >= 0 && interesting; v--) { |
|
2478 | 2476 | bitmask sv = seen[v]; |
|
2479 | 2477 | int parents[2]; |
|
2480 | 2478 | |
|
2481 | 2479 | if (!sv) |
|
2482 | 2480 | continue; |
|
2483 | 2481 | |
|
2484 | 2482 | if (sv < poison) { |
|
2485 | 2483 | interesting -= 1; |
|
2486 | 2484 | if (sv == allseen) { |
|
2487 | 2485 | PyObject *obj = PyLong_FromLong(v); |
|
2488 | 2486 | if (obj == NULL) |
|
2489 | 2487 | goto bail; |
|
2490 | 2488 | if (PyList_Append(gca, obj) == -1) { |
|
2491 | 2489 | Py_DECREF(obj); |
|
2492 | 2490 | goto bail; |
|
2493 | 2491 | } |
|
2494 | 2492 | sv |= poison; |
|
2495 | 2493 | for (i = 0; i < revcount; i++) { |
|
2496 | 2494 | if (revs[i] == v) |
|
2497 | 2495 | goto done; |
|
2498 | 2496 | } |
|
2499 | 2497 | } |
|
2500 | 2498 | } |
|
2501 | 2499 | if (index_get_parents(self, v, parents, maxrev) < 0) |
|
2502 | 2500 | goto bail; |
|
2503 | 2501 | |
|
2504 | 2502 | for (i = 0; i < 2; i++) { |
|
2505 | 2503 | int p = parents[i]; |
|
2506 | 2504 | if (p == -1) |
|
2507 | 2505 | continue; |
|
2508 | 2506 | sp = seen[p]; |
|
2509 | 2507 | if (sv < poison) { |
|
2510 | 2508 | if (sp == 0) { |
|
2511 | 2509 | seen[p] = sv; |
|
2512 | 2510 | interesting++; |
|
2513 | 2511 | } else if (sp != sv) |
|
2514 | 2512 | seen[p] |= sv; |
|
2515 | 2513 | } else { |
|
2516 | 2514 | if (sp && sp < poison) |
|
2517 | 2515 | interesting--; |
|
2518 | 2516 | seen[p] = sv; |
|
2519 | 2517 | } |
|
2520 | 2518 | } |
|
2521 | 2519 | } |
|
2522 | 2520 | |
|
2523 | 2521 | done: |
|
2524 | 2522 | free(seen); |
|
2525 | 2523 | return gca; |
|
2526 | 2524 | bail: |
|
2527 | 2525 | free(seen); |
|
2528 | 2526 | Py_XDECREF(gca); |
|
2529 | 2527 | return NULL; |
|
2530 | 2528 | } |
|
2531 | 2529 | |
|
2532 | 2530 | /* |
|
2533 | 2531 | * Given a disjoint set of revs, return the subset with the longest |
|
2534 | 2532 | * path to the root. |
|
2535 | 2533 | */ |
|
2536 | 2534 | static PyObject *find_deepest(indexObject *self, PyObject *revs) |
|
2537 | 2535 | { |
|
2538 | 2536 | const Py_ssize_t revcount = PyList_GET_SIZE(revs); |
|
2539 | 2537 | static const Py_ssize_t capacity = 24; |
|
2540 | 2538 | int *depth, *interesting = NULL; |
|
2541 | 2539 | int i, j, v, ninteresting; |
|
2542 | 2540 | PyObject *dict = NULL, *keys = NULL; |
|
2543 | 2541 | long *seen = NULL; |
|
2544 | 2542 | int maxrev = -1; |
|
2545 | 2543 | long final; |
|
2546 | 2544 | |
|
2547 | 2545 | if (revcount > capacity) { |
|
2548 | 2546 | PyErr_Format(PyExc_OverflowError, |
|
2549 | 2547 | "bitset size (%ld) > capacity (%ld)", |
|
2550 | 2548 | (long)revcount, (long)capacity); |
|
2551 | 2549 | return NULL; |
|
2552 | 2550 | } |
|
2553 | 2551 | |
|
2554 | 2552 | for (i = 0; i < revcount; i++) { |
|
2555 | 2553 | int n = (int)PyLong_AsLong(PyList_GET_ITEM(revs, i)); |
|
2556 | 2554 | if (n > maxrev) |
|
2557 | 2555 | maxrev = n; |
|
2558 | 2556 | } |
|
2559 | 2557 | |
|
2560 | 2558 | depth = calloc(sizeof(*depth), maxrev + 1); |
|
2561 | 2559 | if (depth == NULL) |
|
2562 | 2560 | return PyErr_NoMemory(); |
|
2563 | 2561 | |
|
2564 | 2562 | seen = calloc(sizeof(*seen), maxrev + 1); |
|
2565 | 2563 | if (seen == NULL) { |
|
2566 | 2564 | PyErr_NoMemory(); |
|
2567 | 2565 | goto bail; |
|
2568 | 2566 | } |
|
2569 | 2567 | |
|
2570 | 2568 | interesting = calloc(sizeof(*interesting), ((size_t)1) << revcount); |
|
2571 | 2569 | if (interesting == NULL) { |
|
2572 | 2570 | PyErr_NoMemory(); |
|
2573 | 2571 | goto bail; |
|
2574 | 2572 | } |
|
2575 | 2573 | |
|
2576 | 2574 | if (PyList_Sort(revs) == -1) |
|
2577 | 2575 | goto bail; |
|
2578 | 2576 | |
|
2579 | 2577 | for (i = 0; i < revcount; i++) { |
|
2580 | 2578 | int n = (int)PyLong_AsLong(PyList_GET_ITEM(revs, i)); |
|
2581 | 2579 | long b = 1l << i; |
|
2582 | 2580 | depth[n] = 1; |
|
2583 | 2581 | seen[n] = b; |
|
2584 | 2582 | interesting[b] = 1; |
|
2585 | 2583 | } |
|
2586 | 2584 | |
|
2587 | 2585 | /* invariant: ninteresting is the number of non-zero entries in |
|
2588 | 2586 | * interesting. */ |
|
2589 | 2587 | ninteresting = (int)revcount; |
|
2590 | 2588 | |
|
2591 | 2589 | for (v = maxrev; v >= 0 && ninteresting > 1; v--) { |
|
2592 | 2590 | int dv = depth[v]; |
|
2593 | 2591 | int parents[2]; |
|
2594 | 2592 | long sv; |
|
2595 | 2593 | |
|
2596 | 2594 | if (dv == 0) |
|
2597 | 2595 | continue; |
|
2598 | 2596 | |
|
2599 | 2597 | sv = seen[v]; |
|
2600 | 2598 | if (index_get_parents(self, v, parents, maxrev) < 0) |
|
2601 | 2599 | goto bail; |
|
2602 | 2600 | |
|
2603 | 2601 | for (i = 0; i < 2; i++) { |
|
2604 | 2602 | int p = parents[i]; |
|
2605 | 2603 | long sp; |
|
2606 | 2604 | int dp; |
|
2607 | 2605 | |
|
2608 | 2606 | if (p == -1) |
|
2609 | 2607 | continue; |
|
2610 | 2608 | |
|
2611 | 2609 | dp = depth[p]; |
|
2612 | 2610 | sp = seen[p]; |
|
2613 | 2611 | if (dp <= dv) { |
|
2614 | 2612 | depth[p] = dv + 1; |
|
2615 | 2613 | if (sp != sv) { |
|
2616 | 2614 | interesting[sv] += 1; |
|
2617 | 2615 | seen[p] = sv; |
|
2618 | 2616 | if (sp) { |
|
2619 | 2617 | interesting[sp] -= 1; |
|
2620 | 2618 | if (interesting[sp] == 0) |
|
2621 | 2619 | ninteresting -= 1; |
|
2622 | 2620 | } |
|
2623 | 2621 | } |
|
2624 | 2622 | } else if (dv == dp - 1) { |
|
2625 | 2623 | long nsp = sp | sv; |
|
2626 | 2624 | if (nsp == sp) |
|
2627 | 2625 | continue; |
|
2628 | 2626 | seen[p] = nsp; |
|
2629 | 2627 | interesting[sp] -= 1; |
|
2630 | 2628 | if (interesting[sp] == 0) |
|
2631 | 2629 | ninteresting -= 1; |
|
2632 | 2630 | if (interesting[nsp] == 0) |
|
2633 | 2631 | ninteresting += 1; |
|
2634 | 2632 | interesting[nsp] += 1; |
|
2635 | 2633 | } |
|
2636 | 2634 | } |
|
2637 | 2635 | interesting[sv] -= 1; |
|
2638 | 2636 | if (interesting[sv] == 0) |
|
2639 | 2637 | ninteresting -= 1; |
|
2640 | 2638 | } |
|
2641 | 2639 | |
|
2642 | 2640 | final = 0; |
|
2643 | 2641 | j = ninteresting; |
|
2644 | 2642 | for (i = 0; i < (int)(2 << revcount) && j > 0; i++) { |
|
2645 | 2643 | if (interesting[i] == 0) |
|
2646 | 2644 | continue; |
|
2647 | 2645 | final |= i; |
|
2648 | 2646 | j -= 1; |
|
2649 | 2647 | } |
|
2650 | 2648 | if (final == 0) { |
|
2651 | 2649 | keys = PyList_New(0); |
|
2652 | 2650 | goto bail; |
|
2653 | 2651 | } |
|
2654 | 2652 | |
|
2655 | 2653 | dict = PyDict_New(); |
|
2656 | 2654 | if (dict == NULL) |
|
2657 | 2655 | goto bail; |
|
2658 | 2656 | |
|
2659 | 2657 | for (i = 0; i < revcount; i++) { |
|
2660 | 2658 | PyObject *key; |
|
2661 | 2659 | |
|
2662 | 2660 | if ((final & (1 << i)) == 0) |
|
2663 | 2661 | continue; |
|
2664 | 2662 | |
|
2665 | 2663 | key = PyList_GET_ITEM(revs, i); |
|
2666 | 2664 | Py_INCREF(key); |
|
2667 | 2665 | Py_INCREF(Py_None); |
|
2668 | 2666 | if (PyDict_SetItem(dict, key, Py_None) == -1) { |
|
2669 | 2667 | Py_DECREF(key); |
|
2670 | 2668 | Py_DECREF(Py_None); |
|
2671 | 2669 | goto bail; |
|
2672 | 2670 | } |
|
2673 | 2671 | } |
|
2674 | 2672 | |
|
2675 | 2673 | keys = PyDict_Keys(dict); |
|
2676 | 2674 | |
|
2677 | 2675 | bail: |
|
2678 | 2676 | free(depth); |
|
2679 | 2677 | free(seen); |
|
2680 | 2678 | free(interesting); |
|
2681 | 2679 | Py_XDECREF(dict); |
|
2682 | 2680 | |
|
2683 | 2681 | return keys; |
|
2684 | 2682 | } |
|
2685 | 2683 | |
|
2686 | 2684 | /* |
|
2687 | 2685 | * Given a (possibly overlapping) set of revs, return all the |
|
2688 | 2686 | * common ancestors heads: heads(::args[0] and ::a[1] and ...) |
|
2689 | 2687 | */ |
|
2690 | 2688 | static PyObject *index_commonancestorsheads(indexObject *self, PyObject *args) |
|
2691 | 2689 | { |
|
2692 | 2690 | PyObject *ret = NULL; |
|
2693 | 2691 | Py_ssize_t argcount, i, len; |
|
2694 | 2692 | bitmask repeat = 0; |
|
2695 | 2693 | int revcount = 0; |
|
2696 | 2694 | int *revs; |
|
2697 | 2695 | |
|
2698 | 2696 | argcount = PySequence_Length(args); |
|
2699 | 2697 | revs = PyMem_Malloc(argcount * sizeof(*revs)); |
|
2700 | 2698 | if (argcount > 0 && revs == NULL) |
|
2701 | 2699 | return PyErr_NoMemory(); |
|
2702 | 2700 | len = index_length(self); |
|
2703 | 2701 | |
|
2704 | 2702 | for (i = 0; i < argcount; i++) { |
|
2705 | 2703 | static const int capacity = 24; |
|
2706 | 2704 | PyObject *obj = PySequence_GetItem(args, i); |
|
2707 | 2705 | bitmask x; |
|
2708 | 2706 | long val; |
|
2709 | 2707 | |
|
2710 | 2708 | if (!PyLong_Check(obj)) { |
|
2711 | 2709 | PyErr_SetString(PyExc_TypeError, |
|
2712 | 2710 | "arguments must all be ints"); |
|
2713 | 2711 | Py_DECREF(obj); |
|
2714 | 2712 | goto bail; |
|
2715 | 2713 | } |
|
2716 | 2714 | val = PyLong_AsLong(obj); |
|
2717 | 2715 | Py_DECREF(obj); |
|
2718 | 2716 | if (val == -1) { |
|
2719 | 2717 | ret = PyList_New(0); |
|
2720 | 2718 | goto done; |
|
2721 | 2719 | } |
|
2722 | 2720 | if (val < 0 || val >= len) { |
|
2723 | 2721 | PyErr_SetString(PyExc_IndexError, "index out of range"); |
|
2724 | 2722 | goto bail; |
|
2725 | 2723 | } |
|
2726 | 2724 | /* this cheesy bloom filter lets us avoid some more |
|
2727 | 2725 | * expensive duplicate checks in the common set-is-disjoint |
|
2728 | 2726 | * case */ |
|
2729 | 2727 | x = 1ull << (val & 0x3f); |
|
2730 | 2728 | if (repeat & x) { |
|
2731 | 2729 | int k; |
|
2732 | 2730 | for (k = 0; k < revcount; k++) { |
|
2733 | 2731 | if (val == revs[k]) |
|
2734 | 2732 | goto duplicate; |
|
2735 | 2733 | } |
|
2736 | 2734 | } else |
|
2737 | 2735 | repeat |= x; |
|
2738 | 2736 | if (revcount >= capacity) { |
|
2739 | 2737 | PyErr_Format(PyExc_OverflowError, |
|
2740 | 2738 | "bitset size (%d) > capacity (%d)", |
|
2741 | 2739 | revcount, capacity); |
|
2742 | 2740 | goto bail; |
|
2743 | 2741 | } |
|
2744 | 2742 | revs[revcount++] = (int)val; |
|
2745 | 2743 | duplicate:; |
|
2746 | 2744 | } |
|
2747 | 2745 | |
|
2748 | 2746 | if (revcount == 0) { |
|
2749 | 2747 | ret = PyList_New(0); |
|
2750 | 2748 | goto done; |
|
2751 | 2749 | } |
|
2752 | 2750 | if (revcount == 1) { |
|
2753 | 2751 | PyObject *obj; |
|
2754 | 2752 | ret = PyList_New(1); |
|
2755 | 2753 | if (ret == NULL) |
|
2756 | 2754 | goto bail; |
|
2757 | 2755 | obj = PyLong_FromLong(revs[0]); |
|
2758 | 2756 | if (obj == NULL) |
|
2759 | 2757 | goto bail; |
|
2760 | 2758 | PyList_SET_ITEM(ret, 0, obj); |
|
2761 | 2759 | goto done; |
|
2762 | 2760 | } |
|
2763 | 2761 | |
|
2764 | 2762 | ret = find_gca_candidates(self, revs, revcount); |
|
2765 | 2763 | if (ret == NULL) |
|
2766 | 2764 | goto bail; |
|
2767 | 2765 | |
|
2768 | 2766 | done: |
|
2769 | 2767 | PyMem_Free(revs); |
|
2770 | 2768 | return ret; |
|
2771 | 2769 | |
|
2772 | 2770 | bail: |
|
2773 | 2771 | PyMem_Free(revs); |
|
2774 | 2772 | Py_XDECREF(ret); |
|
2775 | 2773 | return NULL; |
|
2776 | 2774 | } |
|
2777 | 2775 | |
|
2778 | 2776 | /* |
|
2779 | 2777 | * Given a (possibly overlapping) set of revs, return the greatest |
|
2780 | 2778 | * common ancestors: those with the longest path to the root. |
|
2781 | 2779 | */ |
|
2782 | 2780 | static PyObject *index_ancestors(indexObject *self, PyObject *args) |
|
2783 | 2781 | { |
|
2784 | 2782 | PyObject *ret; |
|
2785 | 2783 | PyObject *gca = index_commonancestorsheads(self, args); |
|
2786 | 2784 | if (gca == NULL) |
|
2787 | 2785 | return NULL; |
|
2788 | 2786 | |
|
2789 | 2787 | if (PyList_GET_SIZE(gca) <= 1) { |
|
2790 | 2788 | return gca; |
|
2791 | 2789 | } |
|
2792 | 2790 | |
|
2793 | 2791 | ret = find_deepest(self, gca); |
|
2794 | 2792 | Py_DECREF(gca); |
|
2795 | 2793 | return ret; |
|
2796 | 2794 | } |
|
2797 | 2795 | |
|
2798 | 2796 | /* |
|
2799 | 2797 | * Invalidate any trie entries introduced by added revs. |
|
2800 | 2798 | */ |
|
2801 | 2799 | static void index_invalidate_added(indexObject *self, Py_ssize_t start) |
|
2802 | 2800 | { |
|
2803 | 2801 | Py_ssize_t i, len; |
|
2804 | 2802 | |
|
2805 | 2803 | len = self->length + self->new_length; |
|
2806 | 2804 | i = start - self->length; |
|
2807 | 2805 | if (i < 0) |
|
2808 | 2806 | return; |
|
2809 | 2807 | |
|
2810 | 2808 | for (i = start; i < len; i++) { |
|
2811 | 2809 | const char *node = index_node(self, i); |
|
2812 | 2810 | nt_delete_node(&self->nt, node); |
|
2813 | 2811 | } |
|
2814 | 2812 | |
|
2815 | 2813 | self->new_length = start - self->length; |
|
2816 | 2814 | } |
|
2817 | 2815 | |
|
2818 | 2816 | /* |
|
2819 | 2817 | * Delete a numeric range of revs, which must be at the end of the |
|
2820 | 2818 | * range. |
|
2821 | 2819 | */ |
|
2822 | 2820 | static int index_slice_del(indexObject *self, PyObject *item) |
|
2823 | 2821 | { |
|
2824 | 2822 | Py_ssize_t start, stop, step, slicelength; |
|
2825 | 2823 | Py_ssize_t length = index_length(self) + 1; |
|
2826 | 2824 | int ret = 0; |
|
2827 | 2825 | |
|
2828 | 2826 | if (PySlice_GetIndicesEx(item, length, &start, &stop, &step, |
|
2829 | 2827 | &slicelength) < 0) |
|
2830 | 2828 | return -1; |
|
2831 | 2829 | |
|
2832 | 2830 | if (slicelength <= 0) |
|
2833 | 2831 | return 0; |
|
2834 | 2832 | |
|
2835 | 2833 | if ((step < 0 && start < stop) || (step > 0 && start > stop)) |
|
2836 | 2834 | stop = start; |
|
2837 | 2835 | |
|
2838 | 2836 | if (step < 0) { |
|
2839 | 2837 | stop = start + 1; |
|
2840 | 2838 | start = stop + step * (slicelength - 1) - 1; |
|
2841 | 2839 | step = -step; |
|
2842 | 2840 | } |
|
2843 | 2841 | |
|
2844 | 2842 | if (step != 1) { |
|
2845 | 2843 | PyErr_SetString(PyExc_ValueError, |
|
2846 | 2844 | "revlog index delete requires step size of 1"); |
|
2847 | 2845 | return -1; |
|
2848 | 2846 | } |
|
2849 | 2847 | |
|
2850 | 2848 | if (stop != length - 1) { |
|
2851 | 2849 | PyErr_SetString(PyExc_IndexError, |
|
2852 | 2850 | "revlog index deletion indices are invalid"); |
|
2853 | 2851 | return -1; |
|
2854 | 2852 | } |
|
2855 | 2853 | |
|
2856 | 2854 | if (start < self->length) { |
|
2857 | 2855 | if (self->ntinitialized) { |
|
2858 | 2856 | Py_ssize_t i; |
|
2859 | 2857 | |
|
2860 | 2858 | for (i = start; i < self->length; i++) { |
|
2861 | 2859 | const char *node = index_node_existing(self, i); |
|
2862 | 2860 | if (node == NULL) |
|
2863 | 2861 | return -1; |
|
2864 | 2862 | |
|
2865 | 2863 | nt_delete_node(&self->nt, node); |
|
2866 | 2864 | } |
|
2867 | 2865 | if (self->new_length) |
|
2868 | 2866 | index_invalidate_added(self, self->length); |
|
2869 | 2867 | if (self->ntrev > start) |
|
2870 | 2868 | self->ntrev = (int)start; |
|
2871 | 2869 | } else if (self->new_length) { |
|
2872 | 2870 | self->new_length = 0; |
|
2873 | 2871 | } |
|
2874 | 2872 | |
|
2875 | 2873 | self->length = start; |
|
2876 | 2874 | goto done; |
|
2877 | 2875 | } |
|
2878 | 2876 | |
|
2879 | 2877 | if (self->ntinitialized) { |
|
2880 | 2878 | index_invalidate_added(self, start); |
|
2881 | 2879 | if (self->ntrev > start) |
|
2882 | 2880 | self->ntrev = (int)start; |
|
2883 | 2881 | } else { |
|
2884 | 2882 | self->new_length = start - self->length; |
|
2885 | 2883 | } |
|
2886 | 2884 | done: |
|
2887 | 2885 | Py_CLEAR(self->headrevs); |
|
2888 | 2886 | return ret; |
|
2889 | 2887 | } |
|
2890 | 2888 | |
|
2891 | 2889 | /* |
|
2892 | 2890 | * Supported ops: |
|
2893 | 2891 | * |
|
2894 | 2892 | * slice deletion |
|
2895 | 2893 | * string assignment (extend node->rev mapping) |
|
2896 | 2894 | * string deletion (shrink node->rev mapping) |
|
2897 | 2895 | */ |
|
2898 | 2896 | static int index_assign_subscript(indexObject *self, PyObject *item, |
|
2899 | 2897 | PyObject *value) |
|
2900 | 2898 | { |
|
2901 | 2899 | char *node; |
|
2902 | 2900 | long rev; |
|
2903 | 2901 | |
|
2904 | 2902 | if (PySlice_Check(item) && value == NULL) |
|
2905 | 2903 | return index_slice_del(self, item); |
|
2906 | 2904 | |
|
2907 | 2905 | if (node_check(self->nodelen, item, &node) == -1) |
|
2908 | 2906 | return -1; |
|
2909 | 2907 | |
|
2910 | 2908 | if (value == NULL) |
|
2911 | 2909 | return self->ntinitialized ? nt_delete_node(&self->nt, node) |
|
2912 | 2910 | : 0; |
|
2913 | 2911 | rev = PyLong_AsLong(value); |
|
2914 | 2912 | if (rev > INT_MAX || rev < 0) { |
|
2915 | 2913 | if (!PyErr_Occurred()) |
|
2916 | 2914 | PyErr_SetString(PyExc_ValueError, "rev out of range"); |
|
2917 | 2915 | return -1; |
|
2918 | 2916 | } |
|
2919 | 2917 | |
|
2920 | 2918 | if (index_init_nt(self) == -1) |
|
2921 | 2919 | return -1; |
|
2922 | 2920 | return nt_insert(&self->nt, node, (int)rev); |
|
2923 | 2921 | } |
|
2924 | 2922 | |
|
2925 | 2923 | /* |
|
2926 | 2924 | * Find all RevlogNG entries in an index that has inline data. Update |
|
2927 | 2925 | * the optional "offsets" table with those entries. |
|
2928 | 2926 | */ |
|
2929 | 2927 | static Py_ssize_t inline_scan(indexObject *self, const char **offsets) |
|
2930 | 2928 | { |
|
2931 | 2929 | const char *data = (const char *)self->buf.buf; |
|
2932 | 2930 | Py_ssize_t pos = 0; |
|
2933 | 2931 | Py_ssize_t end = self->buf.len; |
|
2934 | 2932 | long incr = self->entry_size; |
|
2935 | 2933 | Py_ssize_t len = 0; |
|
2936 | 2934 | |
|
2937 | 2935 | while (pos + self->entry_size <= end && pos >= 0) { |
|
2938 | 2936 | uint32_t comp_len, sidedata_comp_len = 0; |
|
2939 | 2937 | /* 3rd element of header is length of compressed inline data */ |
|
2940 | 2938 | if (self->format_version == format_v1) { |
|
2941 | 2939 | comp_len = |
|
2942 | 2940 | getbe32(data + pos + entry_v1_offset_comp_len); |
|
2943 | 2941 | sidedata_comp_len = 0; |
|
2944 | 2942 | } else if (self->format_version == format_v2) { |
|
2945 | 2943 | comp_len = |
|
2946 | 2944 | getbe32(data + pos + entry_v2_offset_comp_len); |
|
2947 | 2945 | sidedata_comp_len = getbe32( |
|
2948 | 2946 | data + pos + entry_v2_offset_sidedata_comp_len); |
|
2949 | 2947 | } else { |
|
2950 | 2948 | raise_revlog_error(); |
|
2951 | 2949 | return -1; |
|
2952 | 2950 | } |
|
2953 | 2951 | incr = self->entry_size + comp_len + sidedata_comp_len; |
|
2954 | 2952 | if (offsets) |
|
2955 | 2953 | offsets[len] = data + pos; |
|
2956 | 2954 | len++; |
|
2957 | 2955 | pos += incr; |
|
2958 | 2956 | } |
|
2959 | 2957 | |
|
2960 | 2958 | if (pos != end) { |
|
2961 | 2959 | if (!PyErr_Occurred()) |
|
2962 | 2960 | PyErr_SetString(PyExc_ValueError, "corrupt index file"); |
|
2963 | 2961 | return -1; |
|
2964 | 2962 | } |
|
2965 | 2963 | |
|
2966 | 2964 | return len; |
|
2967 | 2965 | } |
|
2968 | 2966 | |
|
2969 | 2967 | static int index_init(indexObject *self, PyObject *args, PyObject *kwargs) |
|
2970 | 2968 | { |
|
2971 | 2969 | PyObject *data_obj, *inlined_obj; |
|
2972 | 2970 | Py_ssize_t size; |
|
2973 | 2971 | |
|
2974 | 2972 | static char *kwlist[] = {"data", "inlined", "format", NULL}; |
|
2975 | 2973 | |
|
2976 | 2974 | /* Initialize before argument-checking to avoid index_dealloc() crash. |
|
2977 | 2975 | */ |
|
2978 | 2976 | self->added = NULL; |
|
2979 | 2977 | self->new_length = 0; |
|
2980 | 2978 | self->added_length = 0; |
|
2981 | 2979 | self->data = NULL; |
|
2982 | 2980 | memset(&self->buf, 0, sizeof(self->buf)); |
|
2983 | 2981 | self->headrevs = NULL; |
|
2984 | 2982 | self->filteredrevs = Py_None; |
|
2985 | 2983 | Py_INCREF(Py_None); |
|
2986 | 2984 | self->ntinitialized = 0; |
|
2987 | 2985 | self->offsets = NULL; |
|
2988 | 2986 | self->nodelen = 20; |
|
2989 | 2987 | self->nullentry = NULL; |
|
2990 | 2988 | self->rust_ext_compat = 1; |
|
2991 | 2989 | self->format_version = format_v1; |
|
2992 | 2990 | |
|
2993 | 2991 | if (!PyArg_ParseTupleAndKeywords(args, kwargs, "OO|l", kwlist, |
|
2994 | 2992 | &data_obj, &inlined_obj, |
|
2995 | 2993 | &(self->format_version))) |
|
2996 | 2994 | return -1; |
|
2997 | 2995 | if (!PyObject_CheckBuffer(data_obj)) { |
|
2998 | 2996 | PyErr_SetString(PyExc_TypeError, |
|
2999 | 2997 | "data does not support buffer interface"); |
|
3000 | 2998 | return -1; |
|
3001 | 2999 | } |
|
3002 | 3000 | if (self->nodelen < 20 || self->nodelen > (Py_ssize_t)sizeof(nullid)) { |
|
3003 | 3001 | PyErr_SetString(PyExc_RuntimeError, "unsupported node size"); |
|
3004 | 3002 | return -1; |
|
3005 | 3003 | } |
|
3006 | 3004 | |
|
3007 | 3005 | if (self->format_version == format_v1) { |
|
3008 | 3006 | self->entry_size = v1_entry_size; |
|
3009 | 3007 | } else if (self->format_version == format_v2) { |
|
3010 | 3008 | self->entry_size = v2_entry_size; |
|
3011 | 3009 | } else if (self->format_version == format_cl2) { |
|
3012 | 3010 | self->entry_size = cl2_entry_size; |
|
3013 | 3011 | } |
|
3014 | 3012 | |
|
3015 | self->nullentry = | |
|
3016 | Py_BuildValue(PY23("iiiiiiis#iiBBi", "iiiiiiiy#iiBBi"), 0, 0, 0, -1, | |
|
3017 | -1, -1, -1, nullid, self->nodelen, 0, 0, | |
|
3018 | comp_mode_inline, comp_mode_inline, rank_unknown); | |
|
3013 | self->nullentry = Py_BuildValue( | |
|
3014 | "iiiiiiiy#iiBBi", 0, 0, 0, -1, -1, -1, -1, nullid, self->nodelen, 0, | |
|
3015 | 0, comp_mode_inline, comp_mode_inline, rank_unknown); | |
|
3019 | 3016 | |
|
3020 | 3017 | if (!self->nullentry) |
|
3021 | 3018 | return -1; |
|
3022 | 3019 | PyObject_GC_UnTrack(self->nullentry); |
|
3023 | 3020 | |
|
3024 | 3021 | if (PyObject_GetBuffer(data_obj, &self->buf, PyBUF_SIMPLE) == -1) |
|
3025 | 3022 | return -1; |
|
3026 | 3023 | size = self->buf.len; |
|
3027 | 3024 | |
|
3028 | 3025 | self->inlined = inlined_obj && PyObject_IsTrue(inlined_obj); |
|
3029 | 3026 | self->data = data_obj; |
|
3030 | 3027 | |
|
3031 | 3028 | self->ntlookups = self->ntmisses = 0; |
|
3032 | 3029 | self->ntrev = -1; |
|
3033 | 3030 | Py_INCREF(self->data); |
|
3034 | 3031 | |
|
3035 | 3032 | if (self->inlined) { |
|
3036 | 3033 | Py_ssize_t len = inline_scan(self, NULL); |
|
3037 | 3034 | if (len == -1) |
|
3038 | 3035 | goto bail; |
|
3039 | 3036 | self->length = len; |
|
3040 | 3037 | } else { |
|
3041 | 3038 | if (size % self->entry_size) { |
|
3042 | 3039 | PyErr_SetString(PyExc_ValueError, "corrupt index file"); |
|
3043 | 3040 | goto bail; |
|
3044 | 3041 | } |
|
3045 | 3042 | self->length = size / self->entry_size; |
|
3046 | 3043 | } |
|
3047 | 3044 | |
|
3048 | 3045 | return 0; |
|
3049 | 3046 | bail: |
|
3050 | 3047 | return -1; |
|
3051 | 3048 | } |
|
3052 | 3049 | |
|
3053 | 3050 | static PyObject *index_nodemap(indexObject *self) |
|
3054 | 3051 | { |
|
3055 | 3052 | Py_INCREF(self); |
|
3056 | 3053 | return (PyObject *)self; |
|
3057 | 3054 | } |
|
3058 | 3055 | |
|
3059 | 3056 | static void _index_clearcaches(indexObject *self) |
|
3060 | 3057 | { |
|
3061 | 3058 | if (self->offsets) { |
|
3062 | 3059 | PyMem_Free((void *)self->offsets); |
|
3063 | 3060 | self->offsets = NULL; |
|
3064 | 3061 | } |
|
3065 | 3062 | if (self->ntinitialized) { |
|
3066 | 3063 | nt_dealloc(&self->nt); |
|
3067 | 3064 | } |
|
3068 | 3065 | self->ntinitialized = 0; |
|
3069 | 3066 | Py_CLEAR(self->headrevs); |
|
3070 | 3067 | } |
|
3071 | 3068 | |
|
3072 | 3069 | static PyObject *index_clearcaches(indexObject *self) |
|
3073 | 3070 | { |
|
3074 | 3071 | _index_clearcaches(self); |
|
3075 | 3072 | self->ntrev = -1; |
|
3076 | 3073 | self->ntlookups = self->ntmisses = 0; |
|
3077 | 3074 | Py_RETURN_NONE; |
|
3078 | 3075 | } |
|
3079 | 3076 | |
|
3080 | 3077 | static void index_dealloc(indexObject *self) |
|
3081 | 3078 | { |
|
3082 | 3079 | _index_clearcaches(self); |
|
3083 | 3080 | Py_XDECREF(self->filteredrevs); |
|
3084 | 3081 | if (self->buf.buf) { |
|
3085 | 3082 | PyBuffer_Release(&self->buf); |
|
3086 | 3083 | memset(&self->buf, 0, sizeof(self->buf)); |
|
3087 | 3084 | } |
|
3088 | 3085 | Py_XDECREF(self->data); |
|
3089 | 3086 | PyMem_Free(self->added); |
|
3090 | 3087 | Py_XDECREF(self->nullentry); |
|
3091 | 3088 | PyObject_Del(self); |
|
3092 | 3089 | } |
|
3093 | 3090 | |
|
3094 | 3091 | static PySequenceMethods index_sequence_methods = { |
|
3095 | 3092 | (lenfunc)index_length, /* sq_length */ |
|
3096 | 3093 | 0, /* sq_concat */ |
|
3097 | 3094 | 0, /* sq_repeat */ |
|
3098 | 3095 | (ssizeargfunc)index_get, /* sq_item */ |
|
3099 | 3096 | 0, /* sq_slice */ |
|
3100 | 3097 | 0, /* sq_ass_item */ |
|
3101 | 3098 | 0, /* sq_ass_slice */ |
|
3102 | 3099 | (objobjproc)index_contains, /* sq_contains */ |
|
3103 | 3100 | }; |
|
3104 | 3101 | |
|
3105 | 3102 | static PyMappingMethods index_mapping_methods = { |
|
3106 | 3103 | (lenfunc)index_length, /* mp_length */ |
|
3107 | 3104 | (binaryfunc)index_getitem, /* mp_subscript */ |
|
3108 | 3105 | (objobjargproc)index_assign_subscript, /* mp_ass_subscript */ |
|
3109 | 3106 | }; |
|
3110 | 3107 | |
|
3111 | 3108 | static PyMethodDef index_methods[] = { |
|
3112 | 3109 | {"ancestors", (PyCFunction)index_ancestors, METH_VARARGS, |
|
3113 | 3110 | "return the gca set of the given revs"}, |
|
3114 | 3111 | {"commonancestorsheads", (PyCFunction)index_commonancestorsheads, |
|
3115 | 3112 | METH_VARARGS, |
|
3116 | 3113 | "return the heads of the common ancestors of the given revs"}, |
|
3117 | 3114 | {"clearcaches", (PyCFunction)index_clearcaches, METH_NOARGS, |
|
3118 | 3115 | "clear the index caches"}, |
|
3119 | 3116 | {"get", (PyCFunction)index_m_get, METH_VARARGS, "get an index entry"}, |
|
3120 | 3117 | {"get_rev", (PyCFunction)index_m_get, METH_VARARGS, |
|
3121 | 3118 | "return `rev` associated with a node or None"}, |
|
3122 | 3119 | {"has_node", (PyCFunction)index_m_has_node, METH_O, |
|
3123 | 3120 | "return True if the node exist in the index"}, |
|
3124 | 3121 | {"rev", (PyCFunction)index_m_rev, METH_O, |
|
3125 | 3122 | "return `rev` associated with a node or raise RevlogError"}, |
|
3126 | 3123 | {"computephasesmapsets", (PyCFunction)compute_phases_map_sets, METH_VARARGS, |
|
3127 | 3124 | "compute phases"}, |
|
3128 | 3125 | {"reachableroots2", (PyCFunction)reachableroots2, METH_VARARGS, |
|
3129 | 3126 | "reachableroots"}, |
|
3130 | 3127 | {"replace_sidedata_info", (PyCFunction)index_replace_sidedata_info, |
|
3131 | 3128 | METH_VARARGS, "replace an existing index entry with a new value"}, |
|
3132 | 3129 | {"headrevs", (PyCFunction)index_headrevs, METH_VARARGS, |
|
3133 | 3130 | "get head revisions"}, /* Can do filtering since 3.2 */ |
|
3134 | 3131 | {"headrevsfiltered", (PyCFunction)index_headrevs, METH_VARARGS, |
|
3135 | 3132 | "get filtered head revisions"}, /* Can always do filtering */ |
|
3136 | 3133 | {"issnapshot", (PyCFunction)index_issnapshot, METH_O, |
|
3137 | 3134 | "True if the object is a snapshot"}, |
|
3138 | 3135 | {"findsnapshots", (PyCFunction)index_findsnapshots, METH_VARARGS, |
|
3139 | 3136 | "Gather snapshot data in a cache dict"}, |
|
3140 | 3137 | {"deltachain", (PyCFunction)index_deltachain, METH_VARARGS, |
|
3141 | 3138 | "determine revisions with deltas to reconstruct fulltext"}, |
|
3142 | 3139 | {"slicechunktodensity", (PyCFunction)index_slicechunktodensity, |
|
3143 | 3140 | METH_VARARGS, "determine revisions with deltas to reconstruct fulltext"}, |
|
3144 | 3141 | {"append", (PyCFunction)index_append, METH_O, "append an index entry"}, |
|
3145 | 3142 | {"partialmatch", (PyCFunction)index_partialmatch, METH_VARARGS, |
|
3146 | 3143 | "match a potentially ambiguous node ID"}, |
|
3147 | 3144 | {"shortest", (PyCFunction)index_shortest, METH_VARARGS, |
|
3148 | 3145 | "find length of shortest hex nodeid of a binary ID"}, |
|
3149 | 3146 | {"stats", (PyCFunction)index_stats, METH_NOARGS, "stats for the index"}, |
|
3150 | 3147 | {"entry_binary", (PyCFunction)index_entry_binary, METH_O, |
|
3151 | 3148 | "return an entry in binary form"}, |
|
3152 | 3149 | {"pack_header", (PyCFunction)index_pack_header, METH_VARARGS, |
|
3153 | 3150 | "pack the revlog header information into binary"}, |
|
3154 | 3151 | {NULL} /* Sentinel */ |
|
3155 | 3152 | }; |
|
3156 | 3153 | |
|
3157 | 3154 | static PyGetSetDef index_getset[] = { |
|
3158 | 3155 | {"nodemap", (getter)index_nodemap, NULL, "nodemap", NULL}, |
|
3159 | 3156 | {NULL} /* Sentinel */ |
|
3160 | 3157 | }; |
|
3161 | 3158 | |
|
3162 | 3159 | static PyMemberDef index_members[] = { |
|
3163 | 3160 | {"entry_size", T_LONG, offsetof(indexObject, entry_size), 0, |
|
3164 | 3161 | "size of an index entry"}, |
|
3165 | 3162 | {"rust_ext_compat", T_LONG, offsetof(indexObject, rust_ext_compat), 0, |
|
3166 | 3163 | "size of an index entry"}, |
|
3167 | 3164 | {NULL} /* Sentinel */ |
|
3168 | 3165 | }; |
|
3169 | 3166 | |
|
3170 | 3167 | PyTypeObject HgRevlogIndex_Type = { |
|
3171 | 3168 | PyVarObject_HEAD_INIT(NULL, 0) /* header */ |
|
3172 | 3169 | "parsers.index", /* tp_name */ |
|
3173 | 3170 | sizeof(indexObject), /* tp_basicsize */ |
|
3174 | 3171 | 0, /* tp_itemsize */ |
|
3175 | 3172 | (destructor)index_dealloc, /* tp_dealloc */ |
|
3176 | 3173 | 0, /* tp_print */ |
|
3177 | 3174 | 0, /* tp_getattr */ |
|
3178 | 3175 | 0, /* tp_setattr */ |
|
3179 | 3176 | 0, /* tp_compare */ |
|
3180 | 3177 | 0, /* tp_repr */ |
|
3181 | 3178 | 0, /* tp_as_number */ |
|
3182 | 3179 | &index_sequence_methods, /* tp_as_sequence */ |
|
3183 | 3180 | &index_mapping_methods, /* tp_as_mapping */ |
|
3184 | 3181 | 0, /* tp_hash */ |
|
3185 | 3182 | 0, /* tp_call */ |
|
3186 | 3183 | 0, /* tp_str */ |
|
3187 | 3184 | 0, /* tp_getattro */ |
|
3188 | 3185 | 0, /* tp_setattro */ |
|
3189 | 3186 | 0, /* tp_as_buffer */ |
|
3190 | 3187 | Py_TPFLAGS_DEFAULT, /* tp_flags */ |
|
3191 | 3188 | "revlog index", /* tp_doc */ |
|
3192 | 3189 | 0, /* tp_traverse */ |
|
3193 | 3190 | 0, /* tp_clear */ |
|
3194 | 3191 | 0, /* tp_richcompare */ |
|
3195 | 3192 | 0, /* tp_weaklistoffset */ |
|
3196 | 3193 | 0, /* tp_iter */ |
|
3197 | 3194 | 0, /* tp_iternext */ |
|
3198 | 3195 | index_methods, /* tp_methods */ |
|
3199 | 3196 | index_members, /* tp_members */ |
|
3200 | 3197 | index_getset, /* tp_getset */ |
|
3201 | 3198 | 0, /* tp_base */ |
|
3202 | 3199 | 0, /* tp_dict */ |
|
3203 | 3200 | 0, /* tp_descr_get */ |
|
3204 | 3201 | 0, /* tp_descr_set */ |
|
3205 | 3202 | 0, /* tp_dictoffset */ |
|
3206 | 3203 | (initproc)index_init, /* tp_init */ |
|
3207 | 3204 | 0, /* tp_alloc */ |
|
3208 | 3205 | }; |
|
3209 | 3206 | |
|
3210 | 3207 | /* |
|
3211 | 3208 | * returns a tuple of the form (index, cache) with elements as |
|
3212 | 3209 | * follows: |
|
3213 | 3210 | * |
|
3214 | 3211 | * index: an index object that lazily parses Revlog (v1 or v2) records |
|
3215 | 3212 | * cache: if data is inlined, a tuple (0, index_file_content), else None |
|
3216 | 3213 | * index_file_content could be a string, or a buffer |
|
3217 | 3214 | * |
|
3218 | 3215 | * added complications are for backwards compatibility |
|
3219 | 3216 | */ |
|
3220 | 3217 | PyObject *parse_index2(PyObject *self, PyObject *args, PyObject *kwargs) |
|
3221 | 3218 | { |
|
3222 | 3219 | PyObject *cache = NULL; |
|
3223 | 3220 | indexObject *idx; |
|
3224 | 3221 | int ret; |
|
3225 | 3222 | |
|
3226 | 3223 | idx = PyObject_New(indexObject, &HgRevlogIndex_Type); |
|
3227 | 3224 | if (idx == NULL) |
|
3228 | 3225 | goto bail; |
|
3229 | 3226 | |
|
3230 | 3227 | ret = index_init(idx, args, kwargs); |
|
3231 | 3228 | if (ret == -1) |
|
3232 | 3229 | goto bail; |
|
3233 | 3230 | |
|
3234 | 3231 | if (idx->inlined) { |
|
3235 | 3232 | cache = Py_BuildValue("iO", 0, idx->data); |
|
3236 | 3233 | if (cache == NULL) |
|
3237 | 3234 | goto bail; |
|
3238 | 3235 | } else { |
|
3239 | 3236 | cache = Py_None; |
|
3240 | 3237 | Py_INCREF(cache); |
|
3241 | 3238 | } |
|
3242 | 3239 | |
|
3243 | 3240 | return Py_BuildValue("NN", idx, cache); |
|
3244 | 3241 | |
|
3245 | 3242 | bail: |
|
3246 | 3243 | Py_XDECREF(idx); |
|
3247 | 3244 | Py_XDECREF(cache); |
|
3248 | 3245 | return NULL; |
|
3249 | 3246 | } |
|
3250 | 3247 | |
|
3251 | 3248 | static Revlog_CAPI CAPI = { |
|
3252 | 3249 | /* increment the abi_version field upon each change in the Revlog_CAPI |
|
3253 | 3250 | struct or in the ABI of the listed functions */ |
|
3254 | 3251 | 2, |
|
3255 | 3252 | index_length, |
|
3256 | 3253 | index_node, |
|
3257 | 3254 | HgRevlogIndex_GetParents, |
|
3258 | 3255 | }; |
|
3259 | 3256 | |
|
3260 | 3257 | void revlog_module_init(PyObject *mod) |
|
3261 | 3258 | { |
|
3262 | 3259 | PyObject *caps = NULL; |
|
3263 | 3260 | HgRevlogIndex_Type.tp_new = PyType_GenericNew; |
|
3264 | 3261 | if (PyType_Ready(&HgRevlogIndex_Type) < 0) |
|
3265 | 3262 | return; |
|
3266 | 3263 | Py_INCREF(&HgRevlogIndex_Type); |
|
3267 | 3264 | PyModule_AddObject(mod, "index", (PyObject *)&HgRevlogIndex_Type); |
|
3268 | 3265 | |
|
3269 | 3266 | nodetreeType.tp_new = PyType_GenericNew; |
|
3270 | 3267 | if (PyType_Ready(&nodetreeType) < 0) |
|
3271 | 3268 | return; |
|
3272 | 3269 | Py_INCREF(&nodetreeType); |
|
3273 | 3270 | PyModule_AddObject(mod, "nodetree", (PyObject *)&nodetreeType); |
|
3274 | 3271 | |
|
3275 | 3272 | caps = PyCapsule_New(&CAPI, "mercurial.cext.parsers.revlog_CAPI", NULL); |
|
3276 | 3273 | if (caps != NULL) |
|
3277 | 3274 | PyModule_AddObject(mod, "revlog_CAPI", caps); |
|
3278 | 3275 | } |
@@ -1,92 +1,85 b'' | |||
|
1 | 1 | /* |
|
2 | 2 | util.h - utility functions for interfacing with the various python APIs. |
|
3 | 3 | |
|
4 | 4 | This software may be used and distributed according to the terms of |
|
5 | 5 | the GNU General Public License, incorporated herein by reference. |
|
6 | 6 | */ |
|
7 | 7 | |
|
8 | 8 | #ifndef _HG_UTIL_H_ |
|
9 | 9 | #define _HG_UTIL_H_ |
|
10 | 10 | |
|
11 | 11 | #include "compat.h" |
|
12 | 12 | |
|
13 | 13 | #if PY_MAJOR_VERSION >= 3 |
|
14 | 14 | #define IS_PY3K |
|
15 | 15 | #endif |
|
16 | 16 | |
|
17 | /* helper to switch things like string literal depending on Python version */ | |
|
18 | #ifdef IS_PY3K | |
|
19 | #define PY23(py2, py3) py3 | |
|
20 | #else | |
|
21 | #define PY23(py2, py3) py2 | |
|
22 | #endif | |
|
23 | ||
|
24 | 17 | /* clang-format off */ |
|
25 | 18 | typedef struct { |
|
26 | 19 | PyObject_HEAD |
|
27 | 20 | int flags; |
|
28 | 21 | int mode; |
|
29 | 22 | int size; |
|
30 | 23 | int mtime_s; |
|
31 | 24 | int mtime_ns; |
|
32 | 25 | } dirstateItemObject; |
|
33 | 26 | /* clang-format on */ |
|
34 | 27 | |
|
35 | 28 | static const int dirstate_flag_wc_tracked = 1 << 0; |
|
36 | 29 | static const int dirstate_flag_p1_tracked = 1 << 1; |
|
37 | 30 | static const int dirstate_flag_p2_info = 1 << 2; |
|
38 | 31 | static const int dirstate_flag_mode_exec_perm = 1 << 3; |
|
39 | 32 | static const int dirstate_flag_mode_is_symlink = 1 << 4; |
|
40 | 33 | static const int dirstate_flag_has_fallback_exec = 1 << 5; |
|
41 | 34 | static const int dirstate_flag_fallback_exec = 1 << 6; |
|
42 | 35 | static const int dirstate_flag_has_fallback_symlink = 1 << 7; |
|
43 | 36 | static const int dirstate_flag_fallback_symlink = 1 << 8; |
|
44 | 37 | static const int dirstate_flag_expected_state_is_modified = 1 << 9; |
|
45 | 38 | static const int dirstate_flag_has_meaningful_data = 1 << 10; |
|
46 | 39 | static const int dirstate_flag_has_mtime = 1 << 11; |
|
47 | 40 | static const int dirstate_flag_mtime_second_ambiguous = 1 << 12; |
|
48 | 41 | static const int dirstate_flag_directory = 1 << 13; |
|
49 | 42 | static const int dirstate_flag_all_unknown_recorded = 1 << 14; |
|
50 | 43 | static const int dirstate_flag_all_ignored_recorded = 1 << 15; |
|
51 | 44 | |
|
52 | 45 | extern PyTypeObject dirstateItemType; |
|
53 | 46 | #define dirstate_tuple_check(op) (Py_TYPE(op) == &dirstateItemType) |
|
54 | 47 | |
|
55 | 48 | #ifndef MIN |
|
56 | 49 | #define MIN(a, b) (((a) < (b)) ? (a) : (b)) |
|
57 | 50 | #endif |
|
58 | 51 | /* VC9 doesn't include bool and lacks stdbool.h based on my searching */ |
|
59 | 52 | #if defined(_MSC_VER) || __STDC_VERSION__ < 199901L |
|
60 | 53 | #define true 1 |
|
61 | 54 | #define false 0 |
|
62 | 55 | typedef unsigned char bool; |
|
63 | 56 | #else |
|
64 | 57 | #include <stdbool.h> |
|
65 | 58 | #endif |
|
66 | 59 | |
|
67 | 60 | static inline PyObject *_dict_new_presized(Py_ssize_t expected_size) |
|
68 | 61 | { |
|
69 | 62 | /* _PyDict_NewPresized expects a minused parameter, but it actually |
|
70 | 63 | creates a dictionary that's the nearest power of two bigger than the |
|
71 | 64 | parameter. For example, with the initial minused = 1000, the |
|
72 | 65 | dictionary created has size 1024. Of course in a lot of cases that |
|
73 | 66 | can be greater than the maximum load factor Python's dict object |
|
74 | 67 | expects (= 2/3), so as soon as we cross the threshold we'll resize |
|
75 | 68 | anyway. So create a dictionary that's at least 3/2 the size. */ |
|
76 | 69 | return _PyDict_NewPresized(((1 + expected_size) / 2) * 3); |
|
77 | 70 | } |
|
78 | 71 | |
|
79 | 72 | /* Convert a PyInt or PyLong to a long. Returns false if there is an |
|
80 | 73 | error, in which case an exception will already have been set. */ |
|
81 | 74 | static inline bool pylong_to_long(PyObject *pylong, long *out) |
|
82 | 75 | { |
|
83 | 76 | *out = PyLong_AsLong(pylong); |
|
84 | 77 | /* Fast path to avoid hitting PyErr_Occurred if the value was obviously |
|
85 | 78 | * not an error. */ |
|
86 | 79 | if (*out != -1) { |
|
87 | 80 | return true; |
|
88 | 81 | } |
|
89 | 82 | return PyErr_Occurred() == NULL; |
|
90 | 83 | } |
|
91 | 84 | |
|
92 | 85 | #endif /* _HG_UTIL_H_ */ |
General Comments 0
You need to be logged in to leave comments.
Login now