##// END OF EJS Templates
cext: remove unused variables...
Joerg Sonnenberger -
r45701:9719e118 default
parent child Browse files
Show More
@@ -1,1014 +1,1011 b''
1 /*
1 /*
2 * manifest.c - manifest type that does on-demand parsing.
2 * manifest.c - manifest type that does on-demand parsing.
3 *
3 *
4 * Copyright 2015, Google Inc.
4 * Copyright 2015, Google Inc.
5 *
5 *
6 * This software may be used and distributed according to the terms of
6 * This software may be used and distributed according to the terms of
7 * the GNU General Public License, incorporated herein by reference.
7 * the GNU General Public License, incorporated herein by reference.
8 */
8 */
9 #include <Python.h>
9 #include <Python.h>
10
10
11 #include <assert.h>
11 #include <assert.h>
12 #include <stdlib.h>
12 #include <stdlib.h>
13 #include <string.h>
13 #include <string.h>
14
14
15 #include "charencode.h"
15 #include "charencode.h"
16 #include "util.h"
16 #include "util.h"
17
17
18 #define DEFAULT_LINES 100000
18 #define DEFAULT_LINES 100000
19
19
20 typedef struct {
20 typedef struct {
21 char *start;
21 char *start;
22 Py_ssize_t len; /* length of line including terminal newline */
22 Py_ssize_t len; /* length of line including terminal newline */
23 char hash_suffix;
23 char hash_suffix;
24 bool from_malloc;
24 bool from_malloc;
25 bool deleted;
25 bool deleted;
26 } line;
26 } line;
27
27
28 typedef struct {
28 typedef struct {
29 PyObject_HEAD
29 PyObject_HEAD
30 PyObject *pydata;
30 PyObject *pydata;
31 line *lines;
31 line *lines;
32 int numlines; /* number of line entries */
32 int numlines; /* number of line entries */
33 int livelines; /* number of non-deleted lines */
33 int livelines; /* number of non-deleted lines */
34 int maxlines; /* allocated number of lines */
34 int maxlines; /* allocated number of lines */
35 bool dirty;
35 bool dirty;
36 } lazymanifest;
36 } lazymanifest;
37
37
38 #define MANIFEST_OOM -1
38 #define MANIFEST_OOM -1
39 #define MANIFEST_NOT_SORTED -2
39 #define MANIFEST_NOT_SORTED -2
40 #define MANIFEST_MALFORMED -3
40 #define MANIFEST_MALFORMED -3
41 #define MANIFEST_BOGUS_FILENAME -4
41 #define MANIFEST_BOGUS_FILENAME -4
42 #define MANIFEST_TOO_SHORT_LINE -5
42 #define MANIFEST_TOO_SHORT_LINE -5
43
43
44 /* get the length of the path for a line */
44 /* get the length of the path for a line */
45 static Py_ssize_t pathlen(line *l)
45 static Py_ssize_t pathlen(line *l)
46 {
46 {
47 const char *end = memchr(l->start, '\0', l->len);
47 const char *end = memchr(l->start, '\0', l->len);
48 return (end) ? (Py_ssize_t)(end - l->start) : l->len;
48 return (end) ? (Py_ssize_t)(end - l->start) : l->len;
49 }
49 }
50
50
51 /* get the node value of a single line */
51 /* get the node value of a single line */
52 static PyObject *nodeof(line *l, char *flag)
52 static PyObject *nodeof(line *l, char *flag)
53 {
53 {
54 char *s = l->start;
54 char *s = l->start;
55 Py_ssize_t llen = pathlen(l);
55 Py_ssize_t llen = pathlen(l);
56 Py_ssize_t hlen = l->len - llen - 2;
56 Py_ssize_t hlen = l->len - llen - 2;
57 Py_ssize_t hlen_raw;
57 Py_ssize_t hlen_raw;
58 PyObject *hash;
58 PyObject *hash;
59 if (llen + 1 + 40 + 1 > l->len) { /* path '\0' hash '\n' */
59 if (llen + 1 + 40 + 1 > l->len) { /* path '\0' hash '\n' */
60 PyErr_SetString(PyExc_ValueError, "manifest line too short");
60 PyErr_SetString(PyExc_ValueError, "manifest line too short");
61 return NULL;
61 return NULL;
62 }
62 }
63 /* Detect flags after the hash first. */
63 /* Detect flags after the hash first. */
64 switch (s[llen + hlen]) {
64 switch (s[llen + hlen]) {
65 case 'l':
65 case 'l':
66 case 't':
66 case 't':
67 case 'x':
67 case 'x':
68 *flag = s[llen + hlen];
68 *flag = s[llen + hlen];
69 --hlen;
69 --hlen;
70 break;
70 break;
71 default:
71 default:
72 *flag = '\0';
72 *flag = '\0';
73 break;
73 break;
74 }
74 }
75
75
76 switch (hlen) {
76 switch (hlen) {
77 case 40: /* sha1 */
77 case 40: /* sha1 */
78 hlen_raw = 20;
78 hlen_raw = 20;
79 break;
79 break;
80 case 64: /* new hash */
80 case 64: /* new hash */
81 hlen_raw = 32;
81 hlen_raw = 32;
82 break;
82 break;
83 default:
83 default:
84 PyErr_SetString(PyExc_ValueError, "invalid node length in manifest");
84 PyErr_SetString(PyExc_ValueError, "invalid node length in manifest");
85 return NULL;
85 return NULL;
86 }
86 }
87 hash = unhexlify(s + llen + 1, hlen_raw * 2);
87 hash = unhexlify(s + llen + 1, hlen_raw * 2);
88 if (!hash) {
88 if (!hash) {
89 return NULL;
89 return NULL;
90 }
90 }
91 if (l->hash_suffix != '\0') {
91 if (l->hash_suffix != '\0') {
92 char newhash[33];
92 char newhash[33];
93 memcpy(newhash, PyBytes_AsString(hash), hlen_raw);
93 memcpy(newhash, PyBytes_AsString(hash), hlen_raw);
94 Py_DECREF(hash);
94 Py_DECREF(hash);
95 newhash[hlen_raw] = l->hash_suffix;
95 newhash[hlen_raw] = l->hash_suffix;
96 hash = PyBytes_FromStringAndSize(newhash, hlen_raw+1);
96 hash = PyBytes_FromStringAndSize(newhash, hlen_raw+1);
97 }
97 }
98 return hash;
98 return hash;
99 }
99 }
100
100
101 /* get the node hash and flags of a line as a tuple */
101 /* get the node hash and flags of a line as a tuple */
102 static PyObject *hashflags(line *l)
102 static PyObject *hashflags(line *l)
103 {
103 {
104 char flag;
104 char flag;
105 PyObject *hash = nodeof(l, &flag);
105 PyObject *hash = nodeof(l, &flag);
106 ssize_t hlen;
107 Py_ssize_t hplen, flen;
108 PyObject *flags;
106 PyObject *flags;
109 PyObject *tup;
107 PyObject *tup;
110
108
111 if (!hash)
109 if (!hash)
112 return NULL;
110 return NULL;
113 flags = PyBytes_FromStringAndSize(&flag, flag ? 1 : 0);
111 flags = PyBytes_FromStringAndSize(&flag, flag ? 1 : 0);
114 if (!flags) {
112 if (!flags) {
115 Py_DECREF(hash);
113 Py_DECREF(hash);
116 return NULL;
114 return NULL;
117 }
115 }
118 tup = PyTuple_Pack(2, hash, flags);
116 tup = PyTuple_Pack(2, hash, flags);
119 Py_DECREF(flags);
117 Py_DECREF(flags);
120 Py_DECREF(hash);
118 Py_DECREF(hash);
121 return tup;
119 return tup;
122 }
120 }
123
121
124 /* if we're about to run out of space in the line index, add more */
122 /* if we're about to run out of space in the line index, add more */
125 static bool realloc_if_full(lazymanifest *self)
123 static bool realloc_if_full(lazymanifest *self)
126 {
124 {
127 if (self->numlines == self->maxlines) {
125 if (self->numlines == self->maxlines) {
128 self->maxlines *= 2;
126 self->maxlines *= 2;
129 self->lines = realloc(self->lines, self->maxlines * sizeof(line));
127 self->lines = realloc(self->lines, self->maxlines * sizeof(line));
130 }
128 }
131 return !!self->lines;
129 return !!self->lines;
132 }
130 }
133
131
134 /*
132 /*
135 * Find the line boundaries in the manifest that 'data' points to and store
133 * Find the line boundaries in the manifest that 'data' points to and store
136 * information about each line in 'self'.
134 * information about each line in 'self'.
137 */
135 */
138 static int find_lines(lazymanifest *self, char *data, Py_ssize_t len)
136 static int find_lines(lazymanifest *self, char *data, Py_ssize_t len)
139 {
137 {
140 char *prev = NULL;
138 char *prev = NULL;
141 while (len > 0) {
139 while (len > 0) {
142 line *l;
140 line *l;
143 char *next;
141 char *next;
144 if (*data == '\0') {
142 if (*data == '\0') {
145 /* It's implausible there's no filename, don't
143 /* It's implausible there's no filename, don't
146 * even bother looking for the newline. */
144 * even bother looking for the newline. */
147 return MANIFEST_BOGUS_FILENAME;
145 return MANIFEST_BOGUS_FILENAME;
148 }
146 }
149 next = memchr(data, '\n', len);
147 next = memchr(data, '\n', len);
150 if (!next) {
148 if (!next) {
151 return MANIFEST_MALFORMED;
149 return MANIFEST_MALFORMED;
152 }
150 }
153 if ((next - data) < 42) {
151 if ((next - data) < 42) {
154 /* We should have at least 42 bytes in a line:
152 /* We should have at least 42 bytes in a line:
155 1 byte filename
153 1 byte filename
156 1 NUL
154 1 NUL
157 40 bytes of hash
155 40 bytes of hash
158 so we can give up here.
156 so we can give up here.
159 */
157 */
160 return MANIFEST_TOO_SHORT_LINE;
158 return MANIFEST_TOO_SHORT_LINE;
161 }
159 }
162 next++; /* advance past newline */
160 next++; /* advance past newline */
163 if (prev && strcmp(prev, data) > -1) {
161 if (prev && strcmp(prev, data) > -1) {
164 /* This data isn't sorted, so we have to abort. */
162 /* This data isn't sorted, so we have to abort. */
165 return MANIFEST_NOT_SORTED;
163 return MANIFEST_NOT_SORTED;
166 }
164 }
167 if (!realloc_if_full(self)) {
165 if (!realloc_if_full(self)) {
168 return MANIFEST_OOM; /* no memory */
166 return MANIFEST_OOM; /* no memory */
169 }
167 }
170 l = self->lines + ((self->numlines)++);
168 l = self->lines + ((self->numlines)++);
171 l->start = data;
169 l->start = data;
172 l->len = next - data;
170 l->len = next - data;
173 l->hash_suffix = '\0';
171 l->hash_suffix = '\0';
174 l->from_malloc = false;
172 l->from_malloc = false;
175 l->deleted = false;
173 l->deleted = false;
176 len = len - l->len;
174 len = len - l->len;
177 prev = data;
175 prev = data;
178 data = next;
176 data = next;
179 }
177 }
180 self->livelines = self->numlines;
178 self->livelines = self->numlines;
181 return 0;
179 return 0;
182 }
180 }
183
181
184 static void lazymanifest_init_early(lazymanifest *self)
182 static void lazymanifest_init_early(lazymanifest *self)
185 {
183 {
186 self->pydata = NULL;
184 self->pydata = NULL;
187 self->lines = NULL;
185 self->lines = NULL;
188 self->numlines = 0;
186 self->numlines = 0;
189 self->maxlines = 0;
187 self->maxlines = 0;
190 }
188 }
191
189
192 static int lazymanifest_init(lazymanifest *self, PyObject *args)
190 static int lazymanifest_init(lazymanifest *self, PyObject *args)
193 {
191 {
194 char *data;
192 char *data;
195 Py_ssize_t len;
193 Py_ssize_t len;
196 int err, ret;
194 int err, ret;
197 PyObject *pydata;
195 PyObject *pydata;
198
196
199 lazymanifest_init_early(self);
197 lazymanifest_init_early(self);
200 if (!PyArg_ParseTuple(args, "S", &pydata)) {
198 if (!PyArg_ParseTuple(args, "S", &pydata)) {
201 return -1;
199 return -1;
202 }
200 }
203 err = PyBytes_AsStringAndSize(pydata, &data, &len);
201 err = PyBytes_AsStringAndSize(pydata, &data, &len);
204
202
205 self->dirty = false;
203 self->dirty = false;
206 if (err == -1)
204 if (err == -1)
207 return -1;
205 return -1;
208 self->pydata = pydata;
206 self->pydata = pydata;
209 Py_INCREF(self->pydata);
207 Py_INCREF(self->pydata);
210 Py_BEGIN_ALLOW_THREADS
208 Py_BEGIN_ALLOW_THREADS
211 self->lines = malloc(DEFAULT_LINES * sizeof(line));
209 self->lines = malloc(DEFAULT_LINES * sizeof(line));
212 self->maxlines = DEFAULT_LINES;
210 self->maxlines = DEFAULT_LINES;
213 self->numlines = 0;
211 self->numlines = 0;
214 if (!self->lines)
212 if (!self->lines)
215 ret = MANIFEST_OOM;
213 ret = MANIFEST_OOM;
216 else
214 else
217 ret = find_lines(self, data, len);
215 ret = find_lines(self, data, len);
218 Py_END_ALLOW_THREADS
216 Py_END_ALLOW_THREADS
219 switch (ret) {
217 switch (ret) {
220 case 0:
218 case 0:
221 break;
219 break;
222 case MANIFEST_OOM:
220 case MANIFEST_OOM:
223 PyErr_NoMemory();
221 PyErr_NoMemory();
224 break;
222 break;
225 case MANIFEST_NOT_SORTED:
223 case MANIFEST_NOT_SORTED:
226 PyErr_Format(PyExc_ValueError,
224 PyErr_Format(PyExc_ValueError,
227 "Manifest lines not in sorted order.");
225 "Manifest lines not in sorted order.");
228 break;
226 break;
229 case MANIFEST_MALFORMED:
227 case MANIFEST_MALFORMED:
230 PyErr_Format(PyExc_ValueError,
228 PyErr_Format(PyExc_ValueError,
231 "Manifest did not end in a newline.");
229 "Manifest did not end in a newline.");
232 break;
230 break;
233 case MANIFEST_BOGUS_FILENAME:
231 case MANIFEST_BOGUS_FILENAME:
234 PyErr_Format(
232 PyErr_Format(
235 PyExc_ValueError,
233 PyExc_ValueError,
236 "Manifest had an entry with a zero-length filename.");
234 "Manifest had an entry with a zero-length filename.");
237 break;
235 break;
238 case MANIFEST_TOO_SHORT_LINE:
236 case MANIFEST_TOO_SHORT_LINE:
239 PyErr_Format(
237 PyErr_Format(
240 PyExc_ValueError,
238 PyExc_ValueError,
241 "Manifest had implausibly-short line.");
239 "Manifest had implausibly-short line.");
242 break;
240 break;
243 default:
241 default:
244 PyErr_Format(PyExc_ValueError,
242 PyErr_Format(PyExc_ValueError,
245 "Unknown problem parsing manifest.");
243 "Unknown problem parsing manifest.");
246 }
244 }
247 return ret == 0 ? 0 : -1;
245 return ret == 0 ? 0 : -1;
248 }
246 }
249
247
250 static void lazymanifest_dealloc(lazymanifest *self)
248 static void lazymanifest_dealloc(lazymanifest *self)
251 {
249 {
252 /* free any extra lines we had to allocate */
250 /* free any extra lines we had to allocate */
253 int i;
251 int i;
254 for (i = 0; self->lines && (i < self->numlines); i++) {
252 for (i = 0; self->lines && (i < self->numlines); i++) {
255 if (self->lines[i].from_malloc) {
253 if (self->lines[i].from_malloc) {
256 free(self->lines[i].start);
254 free(self->lines[i].start);
257 }
255 }
258 }
256 }
259 free(self->lines);
257 free(self->lines);
260 self->lines = NULL;
258 self->lines = NULL;
261 if (self->pydata) {
259 if (self->pydata) {
262 Py_DECREF(self->pydata);
260 Py_DECREF(self->pydata);
263 self->pydata = NULL;
261 self->pydata = NULL;
264 }
262 }
265 PyObject_Del(self);
263 PyObject_Del(self);
266 }
264 }
267
265
268 /* iteration support */
266 /* iteration support */
269
267
270 typedef struct {
268 typedef struct {
271 PyObject_HEAD lazymanifest *m;
269 PyObject_HEAD lazymanifest *m;
272 Py_ssize_t pos;
270 Py_ssize_t pos;
273 } lmIter;
271 } lmIter;
274
272
275 static void lmiter_dealloc(PyObject *o)
273 static void lmiter_dealloc(PyObject *o)
276 {
274 {
277 lmIter *self = (lmIter *)o;
275 lmIter *self = (lmIter *)o;
278 Py_DECREF(self->m);
276 Py_DECREF(self->m);
279 PyObject_Del(self);
277 PyObject_Del(self);
280 }
278 }
281
279
282 static line *lmiter_nextline(lmIter *self)
280 static line *lmiter_nextline(lmIter *self)
283 {
281 {
284 do {
282 do {
285 self->pos++;
283 self->pos++;
286 if (self->pos >= self->m->numlines) {
284 if (self->pos >= self->m->numlines) {
287 return NULL;
285 return NULL;
288 }
286 }
289 /* skip over deleted manifest entries */
287 /* skip over deleted manifest entries */
290 } while (self->m->lines[self->pos].deleted);
288 } while (self->m->lines[self->pos].deleted);
291 return self->m->lines + self->pos;
289 return self->m->lines + self->pos;
292 }
290 }
293
291
294 static PyObject *lmiter_iterentriesnext(PyObject *o)
292 static PyObject *lmiter_iterentriesnext(PyObject *o)
295 {
293 {
296 Py_ssize_t pl;
294 Py_ssize_t pl;
297 line *l;
295 line *l;
298 char flag;
296 char flag;
299 Py_ssize_t consumed;
300 PyObject *ret = NULL, *path = NULL, *hash = NULL, *flags = NULL;
297 PyObject *ret = NULL, *path = NULL, *hash = NULL, *flags = NULL;
301 l = lmiter_nextline((lmIter *)o);
298 l = lmiter_nextline((lmIter *)o);
302 if (!l) {
299 if (!l) {
303 goto done;
300 goto done;
304 }
301 }
305 pl = pathlen(l);
302 pl = pathlen(l);
306 path = PyBytes_FromStringAndSize(l->start, pl);
303 path = PyBytes_FromStringAndSize(l->start, pl);
307 hash = nodeof(l, &flag);
304 hash = nodeof(l, &flag);
308 if (!path || !hash) {
305 if (!path || !hash) {
309 goto done;
306 goto done;
310 }
307 }
311 flags = PyBytes_FromStringAndSize(&flag, flag ? 1 : 0);
308 flags = PyBytes_FromStringAndSize(&flag, flag ? 1 : 0);
312 if (!flags) {
309 if (!flags) {
313 goto done;
310 goto done;
314 }
311 }
315 ret = PyTuple_Pack(3, path, hash, flags);
312 ret = PyTuple_Pack(3, path, hash, flags);
316 done:
313 done:
317 Py_XDECREF(path);
314 Py_XDECREF(path);
318 Py_XDECREF(hash);
315 Py_XDECREF(hash);
319 Py_XDECREF(flags);
316 Py_XDECREF(flags);
320 return ret;
317 return ret;
321 }
318 }
322
319
323 #ifdef IS_PY3K
320 #ifdef IS_PY3K
324 #define LAZYMANIFESTENTRIESITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT
321 #define LAZYMANIFESTENTRIESITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT
325 #else
322 #else
326 #define LAZYMANIFESTENTRIESITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT \
323 #define LAZYMANIFESTENTRIESITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT \
327 | Py_TPFLAGS_HAVE_ITER
324 | Py_TPFLAGS_HAVE_ITER
328 #endif
325 #endif
329
326
330 static PyTypeObject lazymanifestEntriesIterator = {
327 static PyTypeObject lazymanifestEntriesIterator = {
331 PyVarObject_HEAD_INIT(NULL, 0) /* header */
328 PyVarObject_HEAD_INIT(NULL, 0) /* header */
332 "parsers.lazymanifest.entriesiterator", /*tp_name */
329 "parsers.lazymanifest.entriesiterator", /*tp_name */
333 sizeof(lmIter), /*tp_basicsize */
330 sizeof(lmIter), /*tp_basicsize */
334 0, /*tp_itemsize */
331 0, /*tp_itemsize */
335 lmiter_dealloc, /*tp_dealloc */
332 lmiter_dealloc, /*tp_dealloc */
336 0, /*tp_print */
333 0, /*tp_print */
337 0, /*tp_getattr */
334 0, /*tp_getattr */
338 0, /*tp_setattr */
335 0, /*tp_setattr */
339 0, /*tp_compare */
336 0, /*tp_compare */
340 0, /*tp_repr */
337 0, /*tp_repr */
341 0, /*tp_as_number */
338 0, /*tp_as_number */
342 0, /*tp_as_sequence */
339 0, /*tp_as_sequence */
343 0, /*tp_as_mapping */
340 0, /*tp_as_mapping */
344 0, /*tp_hash */
341 0, /*tp_hash */
345 0, /*tp_call */
342 0, /*tp_call */
346 0, /*tp_str */
343 0, /*tp_str */
347 0, /*tp_getattro */
344 0, /*tp_getattro */
348 0, /*tp_setattro */
345 0, /*tp_setattro */
349 0, /*tp_as_buffer */
346 0, /*tp_as_buffer */
350 LAZYMANIFESTENTRIESITERATOR_TPFLAGS, /* tp_flags */
347 LAZYMANIFESTENTRIESITERATOR_TPFLAGS, /* tp_flags */
351 "Iterator for 3-tuples in a lazymanifest.", /* tp_doc */
348 "Iterator for 3-tuples in a lazymanifest.", /* tp_doc */
352 0, /* tp_traverse */
349 0, /* tp_traverse */
353 0, /* tp_clear */
350 0, /* tp_clear */
354 0, /* tp_richcompare */
351 0, /* tp_richcompare */
355 0, /* tp_weaklistoffset */
352 0, /* tp_weaklistoffset */
356 PyObject_SelfIter, /* tp_iter: __iter__() method */
353 PyObject_SelfIter, /* tp_iter: __iter__() method */
357 lmiter_iterentriesnext, /* tp_iternext: next() method */
354 lmiter_iterentriesnext, /* tp_iternext: next() method */
358 };
355 };
359
356
360 static PyObject *lmiter_iterkeysnext(PyObject *o)
357 static PyObject *lmiter_iterkeysnext(PyObject *o)
361 {
358 {
362 Py_ssize_t pl;
359 Py_ssize_t pl;
363 line *l = lmiter_nextline((lmIter *)o);
360 line *l = lmiter_nextline((lmIter *)o);
364 if (!l) {
361 if (!l) {
365 return NULL;
362 return NULL;
366 }
363 }
367 pl = pathlen(l);
364 pl = pathlen(l);
368 return PyBytes_FromStringAndSize(l->start, pl);
365 return PyBytes_FromStringAndSize(l->start, pl);
369 }
366 }
370
367
371 #ifdef IS_PY3K
368 #ifdef IS_PY3K
372 #define LAZYMANIFESTKEYSITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT
369 #define LAZYMANIFESTKEYSITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT
373 #else
370 #else
374 #define LAZYMANIFESTKEYSITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT \
371 #define LAZYMANIFESTKEYSITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT \
375 | Py_TPFLAGS_HAVE_ITER
372 | Py_TPFLAGS_HAVE_ITER
376 #endif
373 #endif
377
374
378 static PyTypeObject lazymanifestKeysIterator = {
375 static PyTypeObject lazymanifestKeysIterator = {
379 PyVarObject_HEAD_INIT(NULL, 0) /* header */
376 PyVarObject_HEAD_INIT(NULL, 0) /* header */
380 "parsers.lazymanifest.keysiterator", /*tp_name */
377 "parsers.lazymanifest.keysiterator", /*tp_name */
381 sizeof(lmIter), /*tp_basicsize */
378 sizeof(lmIter), /*tp_basicsize */
382 0, /*tp_itemsize */
379 0, /*tp_itemsize */
383 lmiter_dealloc, /*tp_dealloc */
380 lmiter_dealloc, /*tp_dealloc */
384 0, /*tp_print */
381 0, /*tp_print */
385 0, /*tp_getattr */
382 0, /*tp_getattr */
386 0, /*tp_setattr */
383 0, /*tp_setattr */
387 0, /*tp_compare */
384 0, /*tp_compare */
388 0, /*tp_repr */
385 0, /*tp_repr */
389 0, /*tp_as_number */
386 0, /*tp_as_number */
390 0, /*tp_as_sequence */
387 0, /*tp_as_sequence */
391 0, /*tp_as_mapping */
388 0, /*tp_as_mapping */
392 0, /*tp_hash */
389 0, /*tp_hash */
393 0, /*tp_call */
390 0, /*tp_call */
394 0, /*tp_str */
391 0, /*tp_str */
395 0, /*tp_getattro */
392 0, /*tp_getattro */
396 0, /*tp_setattro */
393 0, /*tp_setattro */
397 0, /*tp_as_buffer */
394 0, /*tp_as_buffer */
398 LAZYMANIFESTKEYSITERATOR_TPFLAGS, /* tp_flags */
395 LAZYMANIFESTKEYSITERATOR_TPFLAGS, /* tp_flags */
399 "Keys iterator for a lazymanifest.", /* tp_doc */
396 "Keys iterator for a lazymanifest.", /* tp_doc */
400 0, /* tp_traverse */
397 0, /* tp_traverse */
401 0, /* tp_clear */
398 0, /* tp_clear */
402 0, /* tp_richcompare */
399 0, /* tp_richcompare */
403 0, /* tp_weaklistoffset */
400 0, /* tp_weaklistoffset */
404 PyObject_SelfIter, /* tp_iter: __iter__() method */
401 PyObject_SelfIter, /* tp_iter: __iter__() method */
405 lmiter_iterkeysnext, /* tp_iternext: next() method */
402 lmiter_iterkeysnext, /* tp_iternext: next() method */
406 };
403 };
407
404
408 static lazymanifest *lazymanifest_copy(lazymanifest *self);
405 static lazymanifest *lazymanifest_copy(lazymanifest *self);
409
406
410 static PyObject *lazymanifest_getentriesiter(lazymanifest *self)
407 static PyObject *lazymanifest_getentriesiter(lazymanifest *self)
411 {
408 {
412 lmIter *i = NULL;
409 lmIter *i = NULL;
413 lazymanifest *t = lazymanifest_copy(self);
410 lazymanifest *t = lazymanifest_copy(self);
414 if (!t) {
411 if (!t) {
415 PyErr_NoMemory();
412 PyErr_NoMemory();
416 return NULL;
413 return NULL;
417 }
414 }
418 i = PyObject_New(lmIter, &lazymanifestEntriesIterator);
415 i = PyObject_New(lmIter, &lazymanifestEntriesIterator);
419 if (i) {
416 if (i) {
420 i->m = t;
417 i->m = t;
421 i->pos = -1;
418 i->pos = -1;
422 } else {
419 } else {
423 Py_DECREF(t);
420 Py_DECREF(t);
424 PyErr_NoMemory();
421 PyErr_NoMemory();
425 }
422 }
426 return (PyObject *)i;
423 return (PyObject *)i;
427 }
424 }
428
425
429 static PyObject *lazymanifest_getkeysiter(lazymanifest *self)
426 static PyObject *lazymanifest_getkeysiter(lazymanifest *self)
430 {
427 {
431 lmIter *i = NULL;
428 lmIter *i = NULL;
432 lazymanifest *t = lazymanifest_copy(self);
429 lazymanifest *t = lazymanifest_copy(self);
433 if (!t) {
430 if (!t) {
434 PyErr_NoMemory();
431 PyErr_NoMemory();
435 return NULL;
432 return NULL;
436 }
433 }
437 i = PyObject_New(lmIter, &lazymanifestKeysIterator);
434 i = PyObject_New(lmIter, &lazymanifestKeysIterator);
438 if (i) {
435 if (i) {
439 i->m = t;
436 i->m = t;
440 i->pos = -1;
437 i->pos = -1;
441 } else {
438 } else {
442 Py_DECREF(t);
439 Py_DECREF(t);
443 PyErr_NoMemory();
440 PyErr_NoMemory();
444 }
441 }
445 return (PyObject *)i;
442 return (PyObject *)i;
446 }
443 }
447
444
448 /* __getitem__ and __setitem__ support */
445 /* __getitem__ and __setitem__ support */
449
446
450 static Py_ssize_t lazymanifest_size(lazymanifest *self)
447 static Py_ssize_t lazymanifest_size(lazymanifest *self)
451 {
448 {
452 return self->livelines;
449 return self->livelines;
453 }
450 }
454
451
455 static int linecmp(const void *left, const void *right)
452 static int linecmp(const void *left, const void *right)
456 {
453 {
457 return strcmp(((const line *)left)->start,
454 return strcmp(((const line *)left)->start,
458 ((const line *)right)->start);
455 ((const line *)right)->start);
459 }
456 }
460
457
461 static PyObject *lazymanifest_getitem(lazymanifest *self, PyObject *key)
458 static PyObject *lazymanifest_getitem(lazymanifest *self, PyObject *key)
462 {
459 {
463 line needle;
460 line needle;
464 line *hit;
461 line *hit;
465 if (!PyBytes_Check(key)) {
462 if (!PyBytes_Check(key)) {
466 PyErr_Format(PyExc_TypeError,
463 PyErr_Format(PyExc_TypeError,
467 "getitem: manifest keys must be a string.");
464 "getitem: manifest keys must be a string.");
468 return NULL;
465 return NULL;
469 }
466 }
470 needle.start = PyBytes_AsString(key);
467 needle.start = PyBytes_AsString(key);
471 hit = bsearch(&needle, self->lines, self->numlines, sizeof(line),
468 hit = bsearch(&needle, self->lines, self->numlines, sizeof(line),
472 &linecmp);
469 &linecmp);
473 if (!hit || hit->deleted) {
470 if (!hit || hit->deleted) {
474 PyErr_Format(PyExc_KeyError, "No such manifest entry.");
471 PyErr_Format(PyExc_KeyError, "No such manifest entry.");
475 return NULL;
472 return NULL;
476 }
473 }
477 return hashflags(hit);
474 return hashflags(hit);
478 }
475 }
479
476
480 static int lazymanifest_delitem(lazymanifest *self, PyObject *key)
477 static int lazymanifest_delitem(lazymanifest *self, PyObject *key)
481 {
478 {
482 line needle;
479 line needle;
483 line *hit;
480 line *hit;
484 if (!PyBytes_Check(key)) {
481 if (!PyBytes_Check(key)) {
485 PyErr_Format(PyExc_TypeError,
482 PyErr_Format(PyExc_TypeError,
486 "delitem: manifest keys must be a string.");
483 "delitem: manifest keys must be a string.");
487 return -1;
484 return -1;
488 }
485 }
489 needle.start = PyBytes_AsString(key);
486 needle.start = PyBytes_AsString(key);
490 hit = bsearch(&needle, self->lines, self->numlines, sizeof(line),
487 hit = bsearch(&needle, self->lines, self->numlines, sizeof(line),
491 &linecmp);
488 &linecmp);
492 if (!hit || hit->deleted) {
489 if (!hit || hit->deleted) {
493 PyErr_Format(PyExc_KeyError,
490 PyErr_Format(PyExc_KeyError,
494 "Tried to delete nonexistent manifest entry.");
491 "Tried to delete nonexistent manifest entry.");
495 return -1;
492 return -1;
496 }
493 }
497 self->dirty = true;
494 self->dirty = true;
498 hit->deleted = true;
495 hit->deleted = true;
499 self->livelines--;
496 self->livelines--;
500 return 0;
497 return 0;
501 }
498 }
502
499
503 /* Do a binary search for the insertion point for new, creating the
500 /* Do a binary search for the insertion point for new, creating the
504 * new entry if needed. */
501 * new entry if needed. */
505 static int internalsetitem(lazymanifest *self, line *new)
502 static int internalsetitem(lazymanifest *self, line *new)
506 {
503 {
507 int start = 0, end = self->numlines;
504 int start = 0, end = self->numlines;
508 while (start < end) {
505 while (start < end) {
509 int pos = start + (end - start) / 2;
506 int pos = start + (end - start) / 2;
510 int c = linecmp(new, self->lines + pos);
507 int c = linecmp(new, self->lines + pos);
511 if (c < 0)
508 if (c < 0)
512 end = pos;
509 end = pos;
513 else if (c > 0)
510 else if (c > 0)
514 start = pos + 1;
511 start = pos + 1;
515 else {
512 else {
516 if (self->lines[pos].deleted)
513 if (self->lines[pos].deleted)
517 self->livelines++;
514 self->livelines++;
518 if (self->lines[pos].from_malloc)
515 if (self->lines[pos].from_malloc)
519 free(self->lines[pos].start);
516 free(self->lines[pos].start);
520 start = pos;
517 start = pos;
521 goto finish;
518 goto finish;
522 }
519 }
523 }
520 }
524 /* being here means we need to do an insert */
521 /* being here means we need to do an insert */
525 if (!realloc_if_full(self)) {
522 if (!realloc_if_full(self)) {
526 PyErr_NoMemory();
523 PyErr_NoMemory();
527 return -1;
524 return -1;
528 }
525 }
529 memmove(self->lines + start + 1, self->lines + start,
526 memmove(self->lines + start + 1, self->lines + start,
530 (self->numlines - start) * sizeof(line));
527 (self->numlines - start) * sizeof(line));
531 self->numlines++;
528 self->numlines++;
532 self->livelines++;
529 self->livelines++;
533 finish:
530 finish:
534 self->lines[start] = *new;
531 self->lines[start] = *new;
535 self->dirty = true;
532 self->dirty = true;
536 return 0;
533 return 0;
537 }
534 }
538
535
539 static int lazymanifest_setitem(
536 static int lazymanifest_setitem(
540 lazymanifest *self, PyObject *key, PyObject *value)
537 lazymanifest *self, PyObject *key, PyObject *value)
541 {
538 {
542 char *path;
539 char *path;
543 Py_ssize_t plen;
540 Py_ssize_t plen;
544 PyObject *pyhash;
541 PyObject *pyhash;
545 Py_ssize_t hlen;
542 Py_ssize_t hlen;
546 char *hash;
543 char *hash;
547 PyObject *pyflags;
544 PyObject *pyflags;
548 char *flags;
545 char *flags;
549 Py_ssize_t flen;
546 Py_ssize_t flen;
550 Py_ssize_t dlen;
547 Py_ssize_t dlen;
551 char *dest;
548 char *dest;
552 int i;
549 int i;
553 line new;
550 line new;
554 if (!PyBytes_Check(key)) {
551 if (!PyBytes_Check(key)) {
555 PyErr_Format(PyExc_TypeError,
552 PyErr_Format(PyExc_TypeError,
556 "setitem: manifest keys must be a string.");
553 "setitem: manifest keys must be a string.");
557 return -1;
554 return -1;
558 }
555 }
559 if (!value) {
556 if (!value) {
560 return lazymanifest_delitem(self, key);
557 return lazymanifest_delitem(self, key);
561 }
558 }
562 if (!PyTuple_Check(value) || PyTuple_Size(value) != 2) {
559 if (!PyTuple_Check(value) || PyTuple_Size(value) != 2) {
563 PyErr_Format(PyExc_TypeError,
560 PyErr_Format(PyExc_TypeError,
564 "Manifest values must be a tuple of (node, flags).");
561 "Manifest values must be a tuple of (node, flags).");
565 return -1;
562 return -1;
566 }
563 }
567 if (PyBytes_AsStringAndSize(key, &path, &plen) == -1) {
564 if (PyBytes_AsStringAndSize(key, &path, &plen) == -1) {
568 return -1;
565 return -1;
569 }
566 }
570
567
571 pyhash = PyTuple_GetItem(value, 0);
568 pyhash = PyTuple_GetItem(value, 0);
572 if (!PyBytes_Check(pyhash)) {
569 if (!PyBytes_Check(pyhash)) {
573 PyErr_Format(PyExc_TypeError,
570 PyErr_Format(PyExc_TypeError,
574 "node must be a 20 or 32 bytes string");
571 "node must be a 20 or 32 bytes string");
575 return -1;
572 return -1;
576 }
573 }
577 hlen = PyBytes_Size(pyhash);
574 hlen = PyBytes_Size(pyhash);
578 if (hlen != 20 && hlen != 32) {
575 if (hlen != 20 && hlen != 32) {
579 PyErr_Format(PyExc_TypeError,
576 PyErr_Format(PyExc_TypeError,
580 "node must be a 20 or 32 bytes string");
577 "node must be a 20 or 32 bytes string");
581 return -1;
578 return -1;
582 }
579 }
583 hash = PyBytes_AsString(pyhash);
580 hash = PyBytes_AsString(pyhash);
584
581
585 pyflags = PyTuple_GetItem(value, 1);
582 pyflags = PyTuple_GetItem(value, 1);
586 if (!PyBytes_Check(pyflags) || PyBytes_Size(pyflags) > 1) {
583 if (!PyBytes_Check(pyflags) || PyBytes_Size(pyflags) > 1) {
587 PyErr_Format(PyExc_TypeError,
584 PyErr_Format(PyExc_TypeError,
588 "flags must a 0 or 1 bytes string");
585 "flags must a 0 or 1 bytes string");
589 return -1;
586 return -1;
590 }
587 }
591 if (PyBytes_AsStringAndSize(pyflags, &flags, &flen) == -1) {
588 if (PyBytes_AsStringAndSize(pyflags, &flags, &flen) == -1) {
592 return -1;
589 return -1;
593 }
590 }
594 if (flen == 1) {
591 if (flen == 1) {
595 switch (*flags) {
592 switch (*flags) {
596 case 'l':
593 case 'l':
597 case 't':
594 case 't':
598 case 'x':
595 case 'x':
599 break;
596 break;
600 default:
597 default:
601 PyErr_Format(PyExc_TypeError, "invalid manifest flag");
598 PyErr_Format(PyExc_TypeError, "invalid manifest flag");
602 return -1;
599 return -1;
603 }
600 }
604 }
601 }
605 /* one null byte and one newline */
602 /* one null byte and one newline */
606 dlen = plen + hlen * 2 + 1 + flen + 1;
603 dlen = plen + hlen * 2 + 1 + flen + 1;
607 dest = malloc(dlen);
604 dest = malloc(dlen);
608 if (!dest) {
605 if (!dest) {
609 PyErr_NoMemory();
606 PyErr_NoMemory();
610 return -1;
607 return -1;
611 }
608 }
612 memcpy(dest, path, plen + 1);
609 memcpy(dest, path, plen + 1);
613 for (i = 0; i < hlen; i++) {
610 for (i = 0; i < hlen; i++) {
614 /* Cast to unsigned, so it will not get sign-extended when promoted
611 /* Cast to unsigned, so it will not get sign-extended when promoted
615 * to int (as is done when passing to a variadic function)
612 * to int (as is done when passing to a variadic function)
616 */
613 */
617 sprintf(dest + plen + 1 + (i * 2), "%02x", (unsigned char)hash[i]);
614 sprintf(dest + plen + 1 + (i * 2), "%02x", (unsigned char)hash[i]);
618 }
615 }
619 memcpy(dest + plen + 2 * hlen + 1, flags, flen);
616 memcpy(dest + plen + 2 * hlen + 1, flags, flen);
620 dest[plen + 2 * hlen + 1 + flen] = '\n';
617 dest[plen + 2 * hlen + 1 + flen] = '\n';
621 new.start = dest;
618 new.start = dest;
622 new.len = dlen;
619 new.len = dlen;
623 new.hash_suffix = '\0';
620 new.hash_suffix = '\0';
624 if (hlen > 20) {
621 if (hlen > 20) {
625 new.hash_suffix = hash[20];
622 new.hash_suffix = hash[20];
626 }
623 }
627 new.from_malloc = true; /* is `start` a pointer we allocated? */
624 new.from_malloc = true; /* is `start` a pointer we allocated? */
628 new.deleted = false; /* is this entry deleted? */
625 new.deleted = false; /* is this entry deleted? */
629 if (internalsetitem(self, &new)) {
626 if (internalsetitem(self, &new)) {
630 return -1;
627 return -1;
631 }
628 }
632 return 0;
629 return 0;
633 }
630 }
634
631
635 static PyMappingMethods lazymanifest_mapping_methods = {
632 static PyMappingMethods lazymanifest_mapping_methods = {
636 (lenfunc)lazymanifest_size, /* mp_length */
633 (lenfunc)lazymanifest_size, /* mp_length */
637 (binaryfunc)lazymanifest_getitem, /* mp_subscript */
634 (binaryfunc)lazymanifest_getitem, /* mp_subscript */
638 (objobjargproc)lazymanifest_setitem, /* mp_ass_subscript */
635 (objobjargproc)lazymanifest_setitem, /* mp_ass_subscript */
639 };
636 };
640
637
641 /* sequence methods (important or __contains__ builds an iterator) */
638 /* sequence methods (important or __contains__ builds an iterator) */
642
639
643 static int lazymanifest_contains(lazymanifest *self, PyObject *key)
640 static int lazymanifest_contains(lazymanifest *self, PyObject *key)
644 {
641 {
645 line needle;
642 line needle;
646 line *hit;
643 line *hit;
647 if (!PyBytes_Check(key)) {
644 if (!PyBytes_Check(key)) {
648 /* Our keys are always strings, so if the contains
645 /* Our keys are always strings, so if the contains
649 * check is for a non-string, just return false. */
646 * check is for a non-string, just return false. */
650 return 0;
647 return 0;
651 }
648 }
652 needle.start = PyBytes_AsString(key);
649 needle.start = PyBytes_AsString(key);
653 hit = bsearch(&needle, self->lines, self->numlines, sizeof(line),
650 hit = bsearch(&needle, self->lines, self->numlines, sizeof(line),
654 &linecmp);
651 &linecmp);
655 if (!hit || hit->deleted) {
652 if (!hit || hit->deleted) {
656 return 0;
653 return 0;
657 }
654 }
658 return 1;
655 return 1;
659 }
656 }
660
657
661 static PySequenceMethods lazymanifest_seq_meths = {
658 static PySequenceMethods lazymanifest_seq_meths = {
662 (lenfunc)lazymanifest_size, /* sq_length */
659 (lenfunc)lazymanifest_size, /* sq_length */
663 0, /* sq_concat */
660 0, /* sq_concat */
664 0, /* sq_repeat */
661 0, /* sq_repeat */
665 0, /* sq_item */
662 0, /* sq_item */
666 0, /* sq_slice */
663 0, /* sq_slice */
667 0, /* sq_ass_item */
664 0, /* sq_ass_item */
668 0, /* sq_ass_slice */
665 0, /* sq_ass_slice */
669 (objobjproc)lazymanifest_contains, /* sq_contains */
666 (objobjproc)lazymanifest_contains, /* sq_contains */
670 0, /* sq_inplace_concat */
667 0, /* sq_inplace_concat */
671 0, /* sq_inplace_repeat */
668 0, /* sq_inplace_repeat */
672 };
669 };
673
670
674
671
675 /* Other methods (copy, diff, etc) */
672 /* Other methods (copy, diff, etc) */
676 static PyTypeObject lazymanifestType;
673 static PyTypeObject lazymanifestType;
677
674
678 /* If the manifest has changes, build the new manifest text and reindex it. */
675 /* If the manifest has changes, build the new manifest text and reindex it. */
679 static int compact(lazymanifest *self)
676 static int compact(lazymanifest *self)
680 {
677 {
681 int i;
678 int i;
682 ssize_t need = 0;
679 ssize_t need = 0;
683 char *data;
680 char *data;
684 line *src, *dst;
681 line *src, *dst;
685 PyObject *pydata;
682 PyObject *pydata;
686 if (!self->dirty)
683 if (!self->dirty)
687 return 0;
684 return 0;
688 for (i = 0; i < self->numlines; i++) {
685 for (i = 0; i < self->numlines; i++) {
689 if (!self->lines[i].deleted) {
686 if (!self->lines[i].deleted) {
690 need += self->lines[i].len;
687 need += self->lines[i].len;
691 }
688 }
692 }
689 }
693 pydata = PyBytes_FromStringAndSize(NULL, need);
690 pydata = PyBytes_FromStringAndSize(NULL, need);
694 if (!pydata)
691 if (!pydata)
695 return -1;
692 return -1;
696 data = PyBytes_AsString(pydata);
693 data = PyBytes_AsString(pydata);
697 if (!data) {
694 if (!data) {
698 return -1;
695 return -1;
699 }
696 }
700 src = self->lines;
697 src = self->lines;
701 dst = self->lines;
698 dst = self->lines;
702 for (i = 0; i < self->numlines; i++, src++) {
699 for (i = 0; i < self->numlines; i++, src++) {
703 char *tofree = NULL;
700 char *tofree = NULL;
704 if (src->from_malloc) {
701 if (src->from_malloc) {
705 tofree = src->start;
702 tofree = src->start;
706 }
703 }
707 if (!src->deleted) {
704 if (!src->deleted) {
708 memcpy(data, src->start, src->len);
705 memcpy(data, src->start, src->len);
709 *dst = *src;
706 *dst = *src;
710 dst->start = data;
707 dst->start = data;
711 dst->from_malloc = false;
708 dst->from_malloc = false;
712 data += dst->len;
709 data += dst->len;
713 dst++;
710 dst++;
714 }
711 }
715 free(tofree);
712 free(tofree);
716 }
713 }
717 Py_DECREF(self->pydata);
714 Py_DECREF(self->pydata);
718 self->pydata = pydata;
715 self->pydata = pydata;
719 self->numlines = self->livelines;
716 self->numlines = self->livelines;
720 self->dirty = false;
717 self->dirty = false;
721 return 0;
718 return 0;
722 }
719 }
723
720
724 static PyObject *lazymanifest_text(lazymanifest *self)
721 static PyObject *lazymanifest_text(lazymanifest *self)
725 {
722 {
726 if (compact(self) != 0) {
723 if (compact(self) != 0) {
727 PyErr_NoMemory();
724 PyErr_NoMemory();
728 return NULL;
725 return NULL;
729 }
726 }
730 Py_INCREF(self->pydata);
727 Py_INCREF(self->pydata);
731 return self->pydata;
728 return self->pydata;
732 }
729 }
733
730
734 static lazymanifest *lazymanifest_copy(lazymanifest *self)
731 static lazymanifest *lazymanifest_copy(lazymanifest *self)
735 {
732 {
736 lazymanifest *copy = NULL;
733 lazymanifest *copy = NULL;
737 if (compact(self) != 0) {
734 if (compact(self) != 0) {
738 goto nomem;
735 goto nomem;
739 }
736 }
740 copy = PyObject_New(lazymanifest, &lazymanifestType);
737 copy = PyObject_New(lazymanifest, &lazymanifestType);
741 if (!copy) {
738 if (!copy) {
742 goto nomem;
739 goto nomem;
743 }
740 }
744 lazymanifest_init_early(copy);
741 lazymanifest_init_early(copy);
745 copy->numlines = self->numlines;
742 copy->numlines = self->numlines;
746 copy->livelines = self->livelines;
743 copy->livelines = self->livelines;
747 copy->dirty = false;
744 copy->dirty = false;
748 copy->lines = malloc(self->maxlines *sizeof(line));
745 copy->lines = malloc(self->maxlines *sizeof(line));
749 if (!copy->lines) {
746 if (!copy->lines) {
750 goto nomem;
747 goto nomem;
751 }
748 }
752 memcpy(copy->lines, self->lines, self->numlines * sizeof(line));
749 memcpy(copy->lines, self->lines, self->numlines * sizeof(line));
753 copy->maxlines = self->maxlines;
750 copy->maxlines = self->maxlines;
754 copy->pydata = self->pydata;
751 copy->pydata = self->pydata;
755 Py_INCREF(copy->pydata);
752 Py_INCREF(copy->pydata);
756 return copy;
753 return copy;
757 nomem:
754 nomem:
758 PyErr_NoMemory();
755 PyErr_NoMemory();
759 Py_XDECREF(copy);
756 Py_XDECREF(copy);
760 return NULL;
757 return NULL;
761 }
758 }
762
759
763 static lazymanifest *lazymanifest_filtercopy(
760 static lazymanifest *lazymanifest_filtercopy(
764 lazymanifest *self, PyObject *matchfn)
761 lazymanifest *self, PyObject *matchfn)
765 {
762 {
766 lazymanifest *copy = NULL;
763 lazymanifest *copy = NULL;
767 int i;
764 int i;
768 if (!PyCallable_Check(matchfn)) {
765 if (!PyCallable_Check(matchfn)) {
769 PyErr_SetString(PyExc_TypeError, "matchfn must be callable");
766 PyErr_SetString(PyExc_TypeError, "matchfn must be callable");
770 return NULL;
767 return NULL;
771 }
768 }
772 /* compact ourselves first to avoid double-frees later when we
769 /* compact ourselves first to avoid double-frees later when we
773 * compact tmp so that it doesn't have random pointers to our
770 * compact tmp so that it doesn't have random pointers to our
774 * underlying from_malloc-data (self->pydata is safe) */
771 * underlying from_malloc-data (self->pydata is safe) */
775 if (compact(self) != 0) {
772 if (compact(self) != 0) {
776 goto nomem;
773 goto nomem;
777 }
774 }
778 copy = PyObject_New(lazymanifest, &lazymanifestType);
775 copy = PyObject_New(lazymanifest, &lazymanifestType);
779 if (!copy) {
776 if (!copy) {
780 goto nomem;
777 goto nomem;
781 }
778 }
782 lazymanifest_init_early(copy);
779 lazymanifest_init_early(copy);
783 copy->dirty = true;
780 copy->dirty = true;
784 copy->lines = malloc(self->maxlines * sizeof(line));
781 copy->lines = malloc(self->maxlines * sizeof(line));
785 if (!copy->lines) {
782 if (!copy->lines) {
786 goto nomem;
783 goto nomem;
787 }
784 }
788 copy->maxlines = self->maxlines;
785 copy->maxlines = self->maxlines;
789 copy->numlines = 0;
786 copy->numlines = 0;
790 copy->pydata = self->pydata;
787 copy->pydata = self->pydata;
791 Py_INCREF(copy->pydata);
788 Py_INCREF(copy->pydata);
792 for (i = 0; i < self->numlines; i++) {
789 for (i = 0; i < self->numlines; i++) {
793 PyObject *arglist = NULL, *result = NULL;
790 PyObject *arglist = NULL, *result = NULL;
794 arglist = Py_BuildValue(PY23("(s)", "(y)"),
791 arglist = Py_BuildValue(PY23("(s)", "(y)"),
795 self->lines[i].start);
792 self->lines[i].start);
796 if (!arglist) {
793 if (!arglist) {
797 goto bail;
794 goto bail;
798 }
795 }
799 result = PyObject_CallObject(matchfn, arglist);
796 result = PyObject_CallObject(matchfn, arglist);
800 Py_DECREF(arglist);
797 Py_DECREF(arglist);
801 /* if the callback raised an exception, just let it
798 /* if the callback raised an exception, just let it
802 * through and give up */
799 * through and give up */
803 if (!result) {
800 if (!result) {
804 goto bail;
801 goto bail;
805 }
802 }
806 if (PyObject_IsTrue(result)) {
803 if (PyObject_IsTrue(result)) {
807 assert(!(self->lines[i].from_malloc));
804 assert(!(self->lines[i].from_malloc));
808 copy->lines[copy->numlines++] = self->lines[i];
805 copy->lines[copy->numlines++] = self->lines[i];
809 }
806 }
810 Py_DECREF(result);
807 Py_DECREF(result);
811 }
808 }
812 copy->livelines = copy->numlines;
809 copy->livelines = copy->numlines;
813 return copy;
810 return copy;
814 nomem:
811 nomem:
815 PyErr_NoMemory();
812 PyErr_NoMemory();
816 bail:
813 bail:
817 Py_XDECREF(copy);
814 Py_XDECREF(copy);
818 return NULL;
815 return NULL;
819 }
816 }
820
817
821 static PyObject *lazymanifest_diff(lazymanifest *self, PyObject *args)
818 static PyObject *lazymanifest_diff(lazymanifest *self, PyObject *args)
822 {
819 {
823 lazymanifest *other;
820 lazymanifest *other;
824 PyObject *pyclean = NULL;
821 PyObject *pyclean = NULL;
825 bool listclean;
822 bool listclean;
826 PyObject *emptyTup = NULL, *ret = NULL;
823 PyObject *emptyTup = NULL, *ret = NULL;
827 PyObject *es;
824 PyObject *es;
828 int sneedle = 0, oneedle = 0;
825 int sneedle = 0, oneedle = 0;
829 if (!PyArg_ParseTuple(args, "O!|O", &lazymanifestType, &other, &pyclean)) {
826 if (!PyArg_ParseTuple(args, "O!|O", &lazymanifestType, &other, &pyclean)) {
830 return NULL;
827 return NULL;
831 }
828 }
832 listclean = (!pyclean) ? false : PyObject_IsTrue(pyclean);
829 listclean = (!pyclean) ? false : PyObject_IsTrue(pyclean);
833 es = PyBytes_FromString("");
830 es = PyBytes_FromString("");
834 if (!es) {
831 if (!es) {
835 goto nomem;
832 goto nomem;
836 }
833 }
837 emptyTup = PyTuple_Pack(2, Py_None, es);
834 emptyTup = PyTuple_Pack(2, Py_None, es);
838 Py_DECREF(es);
835 Py_DECREF(es);
839 if (!emptyTup) {
836 if (!emptyTup) {
840 goto nomem;
837 goto nomem;
841 }
838 }
842 ret = PyDict_New();
839 ret = PyDict_New();
843 if (!ret) {
840 if (!ret) {
844 goto nomem;
841 goto nomem;
845 }
842 }
846 while (sneedle != self->numlines || oneedle != other->numlines) {
843 while (sneedle != self->numlines || oneedle != other->numlines) {
847 line *left = self->lines + sneedle;
844 line *left = self->lines + sneedle;
848 line *right = other->lines + oneedle;
845 line *right = other->lines + oneedle;
849 int result;
846 int result;
850 PyObject *key;
847 PyObject *key;
851 PyObject *outer;
848 PyObject *outer;
852 /* If we're looking at a deleted entry and it's not
849 /* If we're looking at a deleted entry and it's not
853 * the end of the manifest, just skip it. */
850 * the end of the manifest, just skip it. */
854 if (sneedle < self->numlines && left->deleted) {
851 if (sneedle < self->numlines && left->deleted) {
855 sneedle++;
852 sneedle++;
856 continue;
853 continue;
857 }
854 }
858 if (oneedle < other->numlines && right->deleted) {
855 if (oneedle < other->numlines && right->deleted) {
859 oneedle++;
856 oneedle++;
860 continue;
857 continue;
861 }
858 }
862 /* if we're at the end of either manifest, then we
859 /* if we're at the end of either manifest, then we
863 * know the remaining items are adds so we can skip
860 * know the remaining items are adds so we can skip
864 * the strcmp. */
861 * the strcmp. */
865 if (sneedle == self->numlines) {
862 if (sneedle == self->numlines) {
866 result = 1;
863 result = 1;
867 } else if (oneedle == other->numlines) {
864 } else if (oneedle == other->numlines) {
868 result = -1;
865 result = -1;
869 } else {
866 } else {
870 result = linecmp(left, right);
867 result = linecmp(left, right);
871 }
868 }
872 key = result <= 0 ?
869 key = result <= 0 ?
873 PyBytes_FromString(left->start) :
870 PyBytes_FromString(left->start) :
874 PyBytes_FromString(right->start);
871 PyBytes_FromString(right->start);
875 if (!key)
872 if (!key)
876 goto nomem;
873 goto nomem;
877 if (result < 0) {
874 if (result < 0) {
878 PyObject *l = hashflags(left);
875 PyObject *l = hashflags(left);
879 if (!l) {
876 if (!l) {
880 goto nomem;
877 goto nomem;
881 }
878 }
882 outer = PyTuple_Pack(2, l, emptyTup);
879 outer = PyTuple_Pack(2, l, emptyTup);
883 Py_DECREF(l);
880 Py_DECREF(l);
884 if (!outer) {
881 if (!outer) {
885 goto nomem;
882 goto nomem;
886 }
883 }
887 PyDict_SetItem(ret, key, outer);
884 PyDict_SetItem(ret, key, outer);
888 Py_DECREF(outer);
885 Py_DECREF(outer);
889 sneedle++;
886 sneedle++;
890 } else if (result > 0) {
887 } else if (result > 0) {
891 PyObject *r = hashflags(right);
888 PyObject *r = hashflags(right);
892 if (!r) {
889 if (!r) {
893 goto nomem;
890 goto nomem;
894 }
891 }
895 outer = PyTuple_Pack(2, emptyTup, r);
892 outer = PyTuple_Pack(2, emptyTup, r);
896 Py_DECREF(r);
893 Py_DECREF(r);
897 if (!outer) {
894 if (!outer) {
898 goto nomem;
895 goto nomem;
899 }
896 }
900 PyDict_SetItem(ret, key, outer);
897 PyDict_SetItem(ret, key, outer);
901 Py_DECREF(outer);
898 Py_DECREF(outer);
902 oneedle++;
899 oneedle++;
903 } else {
900 } else {
904 /* file exists in both manifests */
901 /* file exists in both manifests */
905 if (left->len != right->len
902 if (left->len != right->len
906 || memcmp(left->start, right->start, left->len)
903 || memcmp(left->start, right->start, left->len)
907 || left->hash_suffix != right->hash_suffix) {
904 || left->hash_suffix != right->hash_suffix) {
908 PyObject *l = hashflags(left);
905 PyObject *l = hashflags(left);
909 PyObject *r;
906 PyObject *r;
910 if (!l) {
907 if (!l) {
911 goto nomem;
908 goto nomem;
912 }
909 }
913 r = hashflags(right);
910 r = hashflags(right);
914 if (!r) {
911 if (!r) {
915 Py_DECREF(l);
912 Py_DECREF(l);
916 goto nomem;
913 goto nomem;
917 }
914 }
918 outer = PyTuple_Pack(2, l, r);
915 outer = PyTuple_Pack(2, l, r);
919 Py_DECREF(l);
916 Py_DECREF(l);
920 Py_DECREF(r);
917 Py_DECREF(r);
921 if (!outer) {
918 if (!outer) {
922 goto nomem;
919 goto nomem;
923 }
920 }
924 PyDict_SetItem(ret, key, outer);
921 PyDict_SetItem(ret, key, outer);
925 Py_DECREF(outer);
922 Py_DECREF(outer);
926 } else if (listclean) {
923 } else if (listclean) {
927 PyDict_SetItem(ret, key, Py_None);
924 PyDict_SetItem(ret, key, Py_None);
928 }
925 }
929 sneedle++;
926 sneedle++;
930 oneedle++;
927 oneedle++;
931 }
928 }
932 Py_DECREF(key);
929 Py_DECREF(key);
933 }
930 }
934 Py_DECREF(emptyTup);
931 Py_DECREF(emptyTup);
935 return ret;
932 return ret;
936 nomem:
933 nomem:
937 PyErr_NoMemory();
934 PyErr_NoMemory();
938 Py_XDECREF(ret);
935 Py_XDECREF(ret);
939 Py_XDECREF(emptyTup);
936 Py_XDECREF(emptyTup);
940 return NULL;
937 return NULL;
941 }
938 }
942
939
943 static PyMethodDef lazymanifest_methods[] = {
940 static PyMethodDef lazymanifest_methods[] = {
944 {"iterkeys", (PyCFunction)lazymanifest_getkeysiter, METH_NOARGS,
941 {"iterkeys", (PyCFunction)lazymanifest_getkeysiter, METH_NOARGS,
945 "Iterate over file names in this lazymanifest."},
942 "Iterate over file names in this lazymanifest."},
946 {"iterentries", (PyCFunction)lazymanifest_getentriesiter, METH_NOARGS,
943 {"iterentries", (PyCFunction)lazymanifest_getentriesiter, METH_NOARGS,
947 "Iterate over (path, nodeid, flags) tuples in this lazymanifest."},
944 "Iterate over (path, nodeid, flags) tuples in this lazymanifest."},
948 {"copy", (PyCFunction)lazymanifest_copy, METH_NOARGS,
945 {"copy", (PyCFunction)lazymanifest_copy, METH_NOARGS,
949 "Make a copy of this lazymanifest."},
946 "Make a copy of this lazymanifest."},
950 {"filtercopy", (PyCFunction)lazymanifest_filtercopy, METH_O,
947 {"filtercopy", (PyCFunction)lazymanifest_filtercopy, METH_O,
951 "Make a copy of this manifest filtered by matchfn."},
948 "Make a copy of this manifest filtered by matchfn."},
952 {"diff", (PyCFunction)lazymanifest_diff, METH_VARARGS,
949 {"diff", (PyCFunction)lazymanifest_diff, METH_VARARGS,
953 "Compare this lazymanifest to another one."},
950 "Compare this lazymanifest to another one."},
954 {"text", (PyCFunction)lazymanifest_text, METH_NOARGS,
951 {"text", (PyCFunction)lazymanifest_text, METH_NOARGS,
955 "Encode this manifest to text."},
952 "Encode this manifest to text."},
956 {NULL},
953 {NULL},
957 };
954 };
958
955
959 #ifdef IS_PY3K
956 #ifdef IS_PY3K
960 #define LAZYMANIFEST_TPFLAGS Py_TPFLAGS_DEFAULT
957 #define LAZYMANIFEST_TPFLAGS Py_TPFLAGS_DEFAULT
961 #else
958 #else
962 #define LAZYMANIFEST_TPFLAGS Py_TPFLAGS_DEFAULT | Py_TPFLAGS_HAVE_SEQUENCE_IN
959 #define LAZYMANIFEST_TPFLAGS Py_TPFLAGS_DEFAULT | Py_TPFLAGS_HAVE_SEQUENCE_IN
963 #endif
960 #endif
964
961
965 static PyTypeObject lazymanifestType = {
962 static PyTypeObject lazymanifestType = {
966 PyVarObject_HEAD_INIT(NULL, 0) /* header */
963 PyVarObject_HEAD_INIT(NULL, 0) /* header */
967 "parsers.lazymanifest", /* tp_name */
964 "parsers.lazymanifest", /* tp_name */
968 sizeof(lazymanifest), /* tp_basicsize */
965 sizeof(lazymanifest), /* tp_basicsize */
969 0, /* tp_itemsize */
966 0, /* tp_itemsize */
970 (destructor)lazymanifest_dealloc, /* tp_dealloc */
967 (destructor)lazymanifest_dealloc, /* tp_dealloc */
971 0, /* tp_print */
968 0, /* tp_print */
972 0, /* tp_getattr */
969 0, /* tp_getattr */
973 0, /* tp_setattr */
970 0, /* tp_setattr */
974 0, /* tp_compare */
971 0, /* tp_compare */
975 0, /* tp_repr */
972 0, /* tp_repr */
976 0, /* tp_as_number */
973 0, /* tp_as_number */
977 &lazymanifest_seq_meths, /* tp_as_sequence */
974 &lazymanifest_seq_meths, /* tp_as_sequence */
978 &lazymanifest_mapping_methods, /* tp_as_mapping */
975 &lazymanifest_mapping_methods, /* tp_as_mapping */
979 0, /* tp_hash */
976 0, /* tp_hash */
980 0, /* tp_call */
977 0, /* tp_call */
981 0, /* tp_str */
978 0, /* tp_str */
982 0, /* tp_getattro */
979 0, /* tp_getattro */
983 0, /* tp_setattro */
980 0, /* tp_setattro */
984 0, /* tp_as_buffer */
981 0, /* tp_as_buffer */
985 LAZYMANIFEST_TPFLAGS, /* tp_flags */
982 LAZYMANIFEST_TPFLAGS, /* tp_flags */
986 "TODO(augie)", /* tp_doc */
983 "TODO(augie)", /* tp_doc */
987 0, /* tp_traverse */
984 0, /* tp_traverse */
988 0, /* tp_clear */
985 0, /* tp_clear */
989 0, /* tp_richcompare */
986 0, /* tp_richcompare */
990 0, /* tp_weaklistoffset */
987 0, /* tp_weaklistoffset */
991 (getiterfunc)lazymanifest_getkeysiter, /* tp_iter */
988 (getiterfunc)lazymanifest_getkeysiter, /* tp_iter */
992 0, /* tp_iternext */
989 0, /* tp_iternext */
993 lazymanifest_methods, /* tp_methods */
990 lazymanifest_methods, /* tp_methods */
994 0, /* tp_members */
991 0, /* tp_members */
995 0, /* tp_getset */
992 0, /* tp_getset */
996 0, /* tp_base */
993 0, /* tp_base */
997 0, /* tp_dict */
994 0, /* tp_dict */
998 0, /* tp_descr_get */
995 0, /* tp_descr_get */
999 0, /* tp_descr_set */
996 0, /* tp_descr_set */
1000 0, /* tp_dictoffset */
997 0, /* tp_dictoffset */
1001 (initproc)lazymanifest_init, /* tp_init */
998 (initproc)lazymanifest_init, /* tp_init */
1002 0, /* tp_alloc */
999 0, /* tp_alloc */
1003 };
1000 };
1004
1001
1005 void manifest_module_init(PyObject * mod)
1002 void manifest_module_init(PyObject * mod)
1006 {
1003 {
1007 lazymanifestType.tp_new = PyType_GenericNew;
1004 lazymanifestType.tp_new = PyType_GenericNew;
1008 if (PyType_Ready(&lazymanifestType) < 0)
1005 if (PyType_Ready(&lazymanifestType) < 0)
1009 return;
1006 return;
1010 Py_INCREF(&lazymanifestType);
1007 Py_INCREF(&lazymanifestType);
1011
1008
1012 PyModule_AddObject(mod, "lazymanifest",
1009 PyModule_AddObject(mod, "lazymanifest",
1013 (PyObject *)&lazymanifestType);
1010 (PyObject *)&lazymanifestType);
1014 }
1011 }
@@ -1,2959 +1,2956 b''
1 /*
1 /*
2 parsers.c - efficient content parsing
2 parsers.c - efficient content parsing
3
3
4 Copyright 2008 Matt Mackall <mpm@selenic.com> and others
4 Copyright 2008 Matt Mackall <mpm@selenic.com> and others
5
5
6 This software may be used and distributed according to the terms of
6 This software may be used and distributed according to the terms of
7 the GNU General Public License, incorporated herein by reference.
7 the GNU General Public License, incorporated herein by reference.
8 */
8 */
9
9
10 #define PY_SSIZE_T_CLEAN
10 #define PY_SSIZE_T_CLEAN
11 #include <Python.h>
11 #include <Python.h>
12 #include <assert.h>
12 #include <assert.h>
13 #include <ctype.h>
13 #include <ctype.h>
14 #include <limits.h>
14 #include <limits.h>
15 #include <stddef.h>
15 #include <stddef.h>
16 #include <stdlib.h>
16 #include <stdlib.h>
17 #include <string.h>
17 #include <string.h>
18
18
19 #include "bitmanipulation.h"
19 #include "bitmanipulation.h"
20 #include "charencode.h"
20 #include "charencode.h"
21 #include "revlog.h"
21 #include "revlog.h"
22 #include "util.h"
22 #include "util.h"
23
23
24 #ifdef IS_PY3K
24 #ifdef IS_PY3K
25 /* The mapping of Python types is meant to be temporary to get Python
25 /* The mapping of Python types is meant to be temporary to get Python
26 * 3 to compile. We should remove this once Python 3 support is fully
26 * 3 to compile. We should remove this once Python 3 support is fully
27 * supported and proper types are used in the extensions themselves. */
27 * supported and proper types are used in the extensions themselves. */
28 #define PyInt_Check PyLong_Check
28 #define PyInt_Check PyLong_Check
29 #define PyInt_FromLong PyLong_FromLong
29 #define PyInt_FromLong PyLong_FromLong
30 #define PyInt_FromSsize_t PyLong_FromSsize_t
30 #define PyInt_FromSsize_t PyLong_FromSsize_t
31 #define PyInt_AsLong PyLong_AsLong
31 #define PyInt_AsLong PyLong_AsLong
32 #endif
32 #endif
33
33
34 typedef struct indexObjectStruct indexObject;
34 typedef struct indexObjectStruct indexObject;
35
35
36 typedef struct {
36 typedef struct {
37 int children[16];
37 int children[16];
38 } nodetreenode;
38 } nodetreenode;
39
39
40 typedef struct {
40 typedef struct {
41 int abi_version;
41 int abi_version;
42 Py_ssize_t (*index_length)(const indexObject *);
42 Py_ssize_t (*index_length)(const indexObject *);
43 const char *(*index_node)(indexObject *, Py_ssize_t);
43 const char *(*index_node)(indexObject *, Py_ssize_t);
44 int (*index_parents)(PyObject *, int, int *);
44 int (*index_parents)(PyObject *, int, int *);
45 } Revlog_CAPI;
45 } Revlog_CAPI;
46
46
47 /*
47 /*
48 * A base-16 trie for fast node->rev mapping.
48 * A base-16 trie for fast node->rev mapping.
49 *
49 *
50 * Positive value is index of the next node in the trie
50 * Positive value is index of the next node in the trie
51 * Negative value is a leaf: -(rev + 2)
51 * Negative value is a leaf: -(rev + 2)
52 * Zero is empty
52 * Zero is empty
53 */
53 */
54 typedef struct {
54 typedef struct {
55 indexObject *index;
55 indexObject *index;
56 nodetreenode *nodes;
56 nodetreenode *nodes;
57 unsigned length; /* # nodes in use */
57 unsigned length; /* # nodes in use */
58 unsigned capacity; /* # nodes allocated */
58 unsigned capacity; /* # nodes allocated */
59 int depth; /* maximum depth of tree */
59 int depth; /* maximum depth of tree */
60 int splits; /* # splits performed */
60 int splits; /* # splits performed */
61 } nodetree;
61 } nodetree;
62
62
63 typedef struct {
63 typedef struct {
64 PyObject_HEAD /* ; */
64 PyObject_HEAD /* ; */
65 nodetree nt;
65 nodetree nt;
66 } nodetreeObject;
66 } nodetreeObject;
67
67
68 /*
68 /*
69 * This class has two behaviors.
69 * This class has two behaviors.
70 *
70 *
71 * When used in a list-like way (with integer keys), we decode an
71 * When used in a list-like way (with integer keys), we decode an
72 * entry in a RevlogNG index file on demand. We have limited support for
72 * entry in a RevlogNG index file on demand. We have limited support for
73 * integer-keyed insert and delete, only at elements right before the
73 * integer-keyed insert and delete, only at elements right before the
74 * end.
74 * end.
75 *
75 *
76 * With string keys, we lazily perform a reverse mapping from node to
76 * With string keys, we lazily perform a reverse mapping from node to
77 * rev, using a base-16 trie.
77 * rev, using a base-16 trie.
78 */
78 */
79 struct indexObjectStruct {
79 struct indexObjectStruct {
80 PyObject_HEAD
80 PyObject_HEAD
81 /* Type-specific fields go here. */
81 /* Type-specific fields go here. */
82 PyObject *data; /* raw bytes of index */
82 PyObject *data; /* raw bytes of index */
83 Py_buffer buf; /* buffer of data */
83 Py_buffer buf; /* buffer of data */
84 PyObject **cache; /* cached tuples */
84 PyObject **cache; /* cached tuples */
85 const char **offsets; /* populated on demand */
85 const char **offsets; /* populated on demand */
86 Py_ssize_t raw_length; /* original number of elements */
86 Py_ssize_t raw_length; /* original number of elements */
87 Py_ssize_t length; /* current number of elements */
87 Py_ssize_t length; /* current number of elements */
88 PyObject *added; /* populated on demand */
88 PyObject *added; /* populated on demand */
89 PyObject *headrevs; /* cache, invalidated on changes */
89 PyObject *headrevs; /* cache, invalidated on changes */
90 PyObject *filteredrevs; /* filtered revs set */
90 PyObject *filteredrevs; /* filtered revs set */
91 nodetree nt; /* base-16 trie */
91 nodetree nt; /* base-16 trie */
92 int ntinitialized; /* 0 or 1 */
92 int ntinitialized; /* 0 or 1 */
93 int ntrev; /* last rev scanned */
93 int ntrev; /* last rev scanned */
94 int ntlookups; /* # lookups */
94 int ntlookups; /* # lookups */
95 int ntmisses; /* # lookups that miss the cache */
95 int ntmisses; /* # lookups that miss the cache */
96 int inlined;
96 int inlined;
97 };
97 };
98
98
99 static Py_ssize_t index_length(const indexObject *self)
99 static Py_ssize_t index_length(const indexObject *self)
100 {
100 {
101 if (self->added == NULL)
101 if (self->added == NULL)
102 return self->length;
102 return self->length;
103 return self->length + PyList_GET_SIZE(self->added);
103 return self->length + PyList_GET_SIZE(self->added);
104 }
104 }
105
105
106 static PyObject *nullentry = NULL;
106 static PyObject *nullentry = NULL;
107 static const char nullid[20] = {0};
107 static const char nullid[20] = {0};
108 static const Py_ssize_t nullrev = -1;
108 static const Py_ssize_t nullrev = -1;
109
109
110 static Py_ssize_t inline_scan(indexObject *self, const char **offsets);
110 static Py_ssize_t inline_scan(indexObject *self, const char **offsets);
111
111
112 static int index_find_node(indexObject *self, const char *node,
112 static int index_find_node(indexObject *self, const char *node,
113 Py_ssize_t nodelen);
113 Py_ssize_t nodelen);
114
114
115 #if LONG_MAX == 0x7fffffffL
115 #if LONG_MAX == 0x7fffffffL
116 static const char *const tuple_format = PY23("Kiiiiiis#", "Kiiiiiiy#");
116 static const char *const tuple_format = PY23("Kiiiiiis#", "Kiiiiiiy#");
117 #else
117 #else
118 static const char *const tuple_format = PY23("kiiiiiis#", "kiiiiiiy#");
118 static const char *const tuple_format = PY23("kiiiiiis#", "kiiiiiiy#");
119 #endif
119 #endif
120
120
121 /* A RevlogNG v1 index entry is 64 bytes long. */
121 /* A RevlogNG v1 index entry is 64 bytes long. */
122 static const long v1_hdrsize = 64;
122 static const long v1_hdrsize = 64;
123
123
124 static void raise_revlog_error(void)
124 static void raise_revlog_error(void)
125 {
125 {
126 PyObject *mod = NULL, *dict = NULL, *errclass = NULL;
126 PyObject *mod = NULL, *dict = NULL, *errclass = NULL;
127
127
128 mod = PyImport_ImportModule("mercurial.error");
128 mod = PyImport_ImportModule("mercurial.error");
129 if (mod == NULL) {
129 if (mod == NULL) {
130 goto cleanup;
130 goto cleanup;
131 }
131 }
132
132
133 dict = PyModule_GetDict(mod);
133 dict = PyModule_GetDict(mod);
134 if (dict == NULL) {
134 if (dict == NULL) {
135 goto cleanup;
135 goto cleanup;
136 }
136 }
137 Py_INCREF(dict);
137 Py_INCREF(dict);
138
138
139 errclass = PyDict_GetItemString(dict, "RevlogError");
139 errclass = PyDict_GetItemString(dict, "RevlogError");
140 if (errclass == NULL) {
140 if (errclass == NULL) {
141 PyErr_SetString(PyExc_SystemError,
141 PyErr_SetString(PyExc_SystemError,
142 "could not find RevlogError");
142 "could not find RevlogError");
143 goto cleanup;
143 goto cleanup;
144 }
144 }
145
145
146 /* value of exception is ignored by callers */
146 /* value of exception is ignored by callers */
147 PyErr_SetString(errclass, "RevlogError");
147 PyErr_SetString(errclass, "RevlogError");
148
148
149 cleanup:
149 cleanup:
150 Py_XDECREF(dict);
150 Py_XDECREF(dict);
151 Py_XDECREF(mod);
151 Py_XDECREF(mod);
152 }
152 }
153
153
154 /*
154 /*
155 * Return a pointer to the beginning of a RevlogNG record.
155 * Return a pointer to the beginning of a RevlogNG record.
156 */
156 */
157 static const char *index_deref(indexObject *self, Py_ssize_t pos)
157 static const char *index_deref(indexObject *self, Py_ssize_t pos)
158 {
158 {
159 if (self->inlined && pos > 0) {
159 if (self->inlined && pos > 0) {
160 if (self->offsets == NULL) {
160 if (self->offsets == NULL) {
161 Py_ssize_t ret;
161 Py_ssize_t ret;
162 self->offsets = PyMem_Malloc(self->raw_length *
162 self->offsets = PyMem_Malloc(self->raw_length *
163 sizeof(*self->offsets));
163 sizeof(*self->offsets));
164 if (self->offsets == NULL)
164 if (self->offsets == NULL)
165 return (const char *)PyErr_NoMemory();
165 return (const char *)PyErr_NoMemory();
166 ret = inline_scan(self, self->offsets);
166 ret = inline_scan(self, self->offsets);
167 if (ret == -1) {
167 if (ret == -1) {
168 return NULL;
168 return NULL;
169 };
169 };
170 }
170 }
171 return self->offsets[pos];
171 return self->offsets[pos];
172 }
172 }
173
173
174 return (const char *)(self->buf.buf) + pos * v1_hdrsize;
174 return (const char *)(self->buf.buf) + pos * v1_hdrsize;
175 }
175 }
176
176
177 /*
177 /*
178 * Get parents of the given rev.
178 * Get parents of the given rev.
179 *
179 *
180 * The specified rev must be valid and must not be nullrev. A returned
180 * The specified rev must be valid and must not be nullrev. A returned
181 * parent revision may be nullrev, but is guaranteed to be in valid range.
181 * parent revision may be nullrev, but is guaranteed to be in valid range.
182 */
182 */
183 static inline int index_get_parents(indexObject *self, Py_ssize_t rev, int *ps,
183 static inline int index_get_parents(indexObject *self, Py_ssize_t rev, int *ps,
184 int maxrev)
184 int maxrev)
185 {
185 {
186 if (rev >= self->length) {
186 if (rev >= self->length) {
187 long tmp;
187 long tmp;
188 PyObject *tuple =
188 PyObject *tuple =
189 PyList_GET_ITEM(self->added, rev - self->length);
189 PyList_GET_ITEM(self->added, rev - self->length);
190 if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 5), &tmp)) {
190 if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 5), &tmp)) {
191 return -1;
191 return -1;
192 }
192 }
193 ps[0] = (int)tmp;
193 ps[0] = (int)tmp;
194 if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 6), &tmp)) {
194 if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 6), &tmp)) {
195 return -1;
195 return -1;
196 }
196 }
197 ps[1] = (int)tmp;
197 ps[1] = (int)tmp;
198 } else {
198 } else {
199 const char *data = index_deref(self, rev);
199 const char *data = index_deref(self, rev);
200 ps[0] = getbe32(data + 24);
200 ps[0] = getbe32(data + 24);
201 ps[1] = getbe32(data + 28);
201 ps[1] = getbe32(data + 28);
202 }
202 }
203 /* If index file is corrupted, ps[] may point to invalid revisions. So
203 /* If index file is corrupted, ps[] may point to invalid revisions. So
204 * there is a risk of buffer overflow to trust them unconditionally. */
204 * there is a risk of buffer overflow to trust them unconditionally. */
205 if (ps[0] < -1 || ps[0] > maxrev || ps[1] < -1 || ps[1] > maxrev) {
205 if (ps[0] < -1 || ps[0] > maxrev || ps[1] < -1 || ps[1] > maxrev) {
206 PyErr_SetString(PyExc_ValueError, "parent out of range");
206 PyErr_SetString(PyExc_ValueError, "parent out of range");
207 return -1;
207 return -1;
208 }
208 }
209 return 0;
209 return 0;
210 }
210 }
211
211
212 /*
212 /*
213 * Get parents of the given rev.
213 * Get parents of the given rev.
214 *
214 *
215 * If the specified rev is out of range, IndexError will be raised. If the
215 * If the specified rev is out of range, IndexError will be raised. If the
216 * revlog entry is corrupted, ValueError may be raised.
216 * revlog entry is corrupted, ValueError may be raised.
217 *
217 *
218 * Returns 0 on success or -1 on failure.
218 * Returns 0 on success or -1 on failure.
219 */
219 */
220 static int HgRevlogIndex_GetParents(PyObject *op, int rev, int *ps)
220 static int HgRevlogIndex_GetParents(PyObject *op, int rev, int *ps)
221 {
221 {
222 int tiprev;
222 int tiprev;
223 if (!op || !HgRevlogIndex_Check(op) || !ps) {
223 if (!op || !HgRevlogIndex_Check(op) || !ps) {
224 PyErr_BadInternalCall();
224 PyErr_BadInternalCall();
225 return -1;
225 return -1;
226 }
226 }
227 tiprev = (int)index_length((indexObject *)op) - 1;
227 tiprev = (int)index_length((indexObject *)op) - 1;
228 if (rev < -1 || rev > tiprev) {
228 if (rev < -1 || rev > tiprev) {
229 PyErr_Format(PyExc_IndexError, "rev out of range: %d", rev);
229 PyErr_Format(PyExc_IndexError, "rev out of range: %d", rev);
230 return -1;
230 return -1;
231 } else if (rev == -1) {
231 } else if (rev == -1) {
232 ps[0] = ps[1] = -1;
232 ps[0] = ps[1] = -1;
233 return 0;
233 return 0;
234 } else {
234 } else {
235 return index_get_parents((indexObject *)op, rev, ps, tiprev);
235 return index_get_parents((indexObject *)op, rev, ps, tiprev);
236 }
236 }
237 }
237 }
238
238
239 static inline int64_t index_get_start(indexObject *self, Py_ssize_t rev)
239 static inline int64_t index_get_start(indexObject *self, Py_ssize_t rev)
240 {
240 {
241 uint64_t offset;
241 uint64_t offset;
242 if (rev == nullrev) {
242 if (rev == nullrev) {
243 return 0;
243 return 0;
244 }
244 }
245 if (rev >= self->length) {
245 if (rev >= self->length) {
246 PyObject *tuple;
246 PyObject *tuple;
247 PyObject *pylong;
247 PyObject *pylong;
248 PY_LONG_LONG tmp;
248 PY_LONG_LONG tmp;
249 tuple = PyList_GET_ITEM(self->added, rev - self->length);
249 tuple = PyList_GET_ITEM(self->added, rev - self->length);
250 pylong = PyTuple_GET_ITEM(tuple, 0);
250 pylong = PyTuple_GET_ITEM(tuple, 0);
251 tmp = PyLong_AsLongLong(pylong);
251 tmp = PyLong_AsLongLong(pylong);
252 if (tmp == -1 && PyErr_Occurred()) {
252 if (tmp == -1 && PyErr_Occurred()) {
253 return -1;
253 return -1;
254 }
254 }
255 if (tmp < 0) {
255 if (tmp < 0) {
256 PyErr_Format(PyExc_OverflowError,
256 PyErr_Format(PyExc_OverflowError,
257 "revlog entry size out of bound (%lld)",
257 "revlog entry size out of bound (%lld)",
258 (long long)tmp);
258 (long long)tmp);
259 return -1;
259 return -1;
260 }
260 }
261 offset = (uint64_t)tmp;
261 offset = (uint64_t)tmp;
262 } else {
262 } else {
263 const char *data = index_deref(self, rev);
263 const char *data = index_deref(self, rev);
264 offset = getbe32(data + 4);
264 offset = getbe32(data + 4);
265 if (rev == 0) {
265 if (rev == 0) {
266 /* mask out version number for the first entry */
266 /* mask out version number for the first entry */
267 offset &= 0xFFFF;
267 offset &= 0xFFFF;
268 } else {
268 } else {
269 uint32_t offset_high = getbe32(data);
269 uint32_t offset_high = getbe32(data);
270 offset |= ((uint64_t)offset_high) << 32;
270 offset |= ((uint64_t)offset_high) << 32;
271 }
271 }
272 }
272 }
273 return (int64_t)(offset >> 16);
273 return (int64_t)(offset >> 16);
274 }
274 }
275
275
276 static inline int index_get_length(indexObject *self, Py_ssize_t rev)
276 static inline int index_get_length(indexObject *self, Py_ssize_t rev)
277 {
277 {
278 if (rev == nullrev) {
278 if (rev == nullrev) {
279 return 0;
279 return 0;
280 }
280 }
281 if (rev >= self->length) {
281 if (rev >= self->length) {
282 PyObject *tuple;
282 PyObject *tuple;
283 PyObject *pylong;
283 PyObject *pylong;
284 long ret;
284 long ret;
285 tuple = PyList_GET_ITEM(self->added, rev - self->length);
285 tuple = PyList_GET_ITEM(self->added, rev - self->length);
286 pylong = PyTuple_GET_ITEM(tuple, 1);
286 pylong = PyTuple_GET_ITEM(tuple, 1);
287 ret = PyInt_AsLong(pylong);
287 ret = PyInt_AsLong(pylong);
288 if (ret == -1 && PyErr_Occurred()) {
288 if (ret == -1 && PyErr_Occurred()) {
289 return -1;
289 return -1;
290 }
290 }
291 if (ret < 0 || ret > (long)INT_MAX) {
291 if (ret < 0 || ret > (long)INT_MAX) {
292 PyErr_Format(PyExc_OverflowError,
292 PyErr_Format(PyExc_OverflowError,
293 "revlog entry size out of bound (%ld)",
293 "revlog entry size out of bound (%ld)",
294 ret);
294 ret);
295 return -1;
295 return -1;
296 }
296 }
297 return (int)ret;
297 return (int)ret;
298 } else {
298 } else {
299 const char *data = index_deref(self, rev);
299 const char *data = index_deref(self, rev);
300 int tmp = (int)getbe32(data + 8);
300 int tmp = (int)getbe32(data + 8);
301 if (tmp < 0) {
301 if (tmp < 0) {
302 PyErr_Format(PyExc_OverflowError,
302 PyErr_Format(PyExc_OverflowError,
303 "revlog entry size out of bound (%d)",
303 "revlog entry size out of bound (%d)",
304 tmp);
304 tmp);
305 return -1;
305 return -1;
306 }
306 }
307 return tmp;
307 return tmp;
308 }
308 }
309 }
309 }
310
310
311 /*
311 /*
312 * RevlogNG format (all in big endian, data may be inlined):
312 * RevlogNG format (all in big endian, data may be inlined):
313 * 6 bytes: offset
313 * 6 bytes: offset
314 * 2 bytes: flags
314 * 2 bytes: flags
315 * 4 bytes: compressed length
315 * 4 bytes: compressed length
316 * 4 bytes: uncompressed length
316 * 4 bytes: uncompressed length
317 * 4 bytes: base revision
317 * 4 bytes: base revision
318 * 4 bytes: link revision
318 * 4 bytes: link revision
319 * 4 bytes: parent 1 revision
319 * 4 bytes: parent 1 revision
320 * 4 bytes: parent 2 revision
320 * 4 bytes: parent 2 revision
321 * 32 bytes: nodeid (only 20 bytes used)
321 * 32 bytes: nodeid (only 20 bytes used)
322 */
322 */
323 static PyObject *index_get(indexObject *self, Py_ssize_t pos)
323 static PyObject *index_get(indexObject *self, Py_ssize_t pos)
324 {
324 {
325 uint64_t offset_flags;
325 uint64_t offset_flags;
326 int comp_len, uncomp_len, base_rev, link_rev, parent_1, parent_2;
326 int comp_len, uncomp_len, base_rev, link_rev, parent_1, parent_2;
327 const char *c_node_id;
327 const char *c_node_id;
328 const char *data;
328 const char *data;
329 Py_ssize_t length = index_length(self);
329 Py_ssize_t length = index_length(self);
330 PyObject *entry;
330 PyObject *entry;
331
331
332 if (pos == nullrev) {
332 if (pos == nullrev) {
333 Py_INCREF(nullentry);
333 Py_INCREF(nullentry);
334 return nullentry;
334 return nullentry;
335 }
335 }
336
336
337 if (pos < 0 || pos >= length) {
337 if (pos < 0 || pos >= length) {
338 PyErr_SetString(PyExc_IndexError, "revlog index out of range");
338 PyErr_SetString(PyExc_IndexError, "revlog index out of range");
339 return NULL;
339 return NULL;
340 }
340 }
341
341
342 if (pos >= self->length) {
342 if (pos >= self->length) {
343 PyObject *obj;
343 PyObject *obj;
344 obj = PyList_GET_ITEM(self->added, pos - self->length);
344 obj = PyList_GET_ITEM(self->added, pos - self->length);
345 Py_INCREF(obj);
345 Py_INCREF(obj);
346 return obj;
346 return obj;
347 }
347 }
348
348
349 if (self->cache) {
349 if (self->cache) {
350 if (self->cache[pos]) {
350 if (self->cache[pos]) {
351 Py_INCREF(self->cache[pos]);
351 Py_INCREF(self->cache[pos]);
352 return self->cache[pos];
352 return self->cache[pos];
353 }
353 }
354 } else {
354 } else {
355 self->cache = calloc(self->raw_length, sizeof(PyObject *));
355 self->cache = calloc(self->raw_length, sizeof(PyObject *));
356 if (self->cache == NULL)
356 if (self->cache == NULL)
357 return PyErr_NoMemory();
357 return PyErr_NoMemory();
358 }
358 }
359
359
360 data = index_deref(self, pos);
360 data = index_deref(self, pos);
361 if (data == NULL)
361 if (data == NULL)
362 return NULL;
362 return NULL;
363
363
364 offset_flags = getbe32(data + 4);
364 offset_flags = getbe32(data + 4);
365 if (pos == 0) /* mask out version number for the first entry */
365 if (pos == 0) /* mask out version number for the first entry */
366 offset_flags &= 0xFFFF;
366 offset_flags &= 0xFFFF;
367 else {
367 else {
368 uint32_t offset_high = getbe32(data);
368 uint32_t offset_high = getbe32(data);
369 offset_flags |= ((uint64_t)offset_high) << 32;
369 offset_flags |= ((uint64_t)offset_high) << 32;
370 }
370 }
371
371
372 comp_len = getbe32(data + 8);
372 comp_len = getbe32(data + 8);
373 uncomp_len = getbe32(data + 12);
373 uncomp_len = getbe32(data + 12);
374 base_rev = getbe32(data + 16);
374 base_rev = getbe32(data + 16);
375 link_rev = getbe32(data + 20);
375 link_rev = getbe32(data + 20);
376 parent_1 = getbe32(data + 24);
376 parent_1 = getbe32(data + 24);
377 parent_2 = getbe32(data + 28);
377 parent_2 = getbe32(data + 28);
378 c_node_id = data + 32;
378 c_node_id = data + 32;
379
379
380 entry = Py_BuildValue(tuple_format, offset_flags, comp_len, uncomp_len,
380 entry = Py_BuildValue(tuple_format, offset_flags, comp_len, uncomp_len,
381 base_rev, link_rev, parent_1, parent_2, c_node_id,
381 base_rev, link_rev, parent_1, parent_2, c_node_id,
382 (Py_ssize_t)20);
382 (Py_ssize_t)20);
383
383
384 if (entry) {
384 if (entry) {
385 PyObject_GC_UnTrack(entry);
385 PyObject_GC_UnTrack(entry);
386 Py_INCREF(entry);
386 Py_INCREF(entry);
387 }
387 }
388
388
389 self->cache[pos] = entry;
389 self->cache[pos] = entry;
390
390
391 return entry;
391 return entry;
392 }
392 }
393
393
394 /*
394 /*
395 * Return the 20-byte SHA of the node corresponding to the given rev.
395 * Return the 20-byte SHA of the node corresponding to the given rev.
396 */
396 */
397 static const char *index_node(indexObject *self, Py_ssize_t pos)
397 static const char *index_node(indexObject *self, Py_ssize_t pos)
398 {
398 {
399 Py_ssize_t length = index_length(self);
399 Py_ssize_t length = index_length(self);
400 const char *data;
400 const char *data;
401
401
402 if (pos == nullrev)
402 if (pos == nullrev)
403 return nullid;
403 return nullid;
404
404
405 if (pos >= length)
405 if (pos >= length)
406 return NULL;
406 return NULL;
407
407
408 if (pos >= self->length) {
408 if (pos >= self->length) {
409 PyObject *tuple, *str;
409 PyObject *tuple, *str;
410 tuple = PyList_GET_ITEM(self->added, pos - self->length);
410 tuple = PyList_GET_ITEM(self->added, pos - self->length);
411 str = PyTuple_GetItem(tuple, 7);
411 str = PyTuple_GetItem(tuple, 7);
412 return str ? PyBytes_AS_STRING(str) : NULL;
412 return str ? PyBytes_AS_STRING(str) : NULL;
413 }
413 }
414
414
415 data = index_deref(self, pos);
415 data = index_deref(self, pos);
416 return data ? data + 32 : NULL;
416 return data ? data + 32 : NULL;
417 }
417 }
418
418
419 /*
419 /*
420 * Return the 20-byte SHA of the node corresponding to the given rev. The
420 * Return the 20-byte SHA of the node corresponding to the given rev. The
421 * rev is assumed to be existing. If not, an exception is set.
421 * rev is assumed to be existing. If not, an exception is set.
422 */
422 */
423 static const char *index_node_existing(indexObject *self, Py_ssize_t pos)
423 static const char *index_node_existing(indexObject *self, Py_ssize_t pos)
424 {
424 {
425 const char *node = index_node(self, pos);
425 const char *node = index_node(self, pos);
426 if (node == NULL) {
426 if (node == NULL) {
427 PyErr_Format(PyExc_IndexError, "could not access rev %d",
427 PyErr_Format(PyExc_IndexError, "could not access rev %d",
428 (int)pos);
428 (int)pos);
429 }
429 }
430 return node;
430 return node;
431 }
431 }
432
432
433 static int nt_insert(nodetree *self, const char *node, int rev);
433 static int nt_insert(nodetree *self, const char *node, int rev);
434
434
435 static int node_check(PyObject *obj, char **node)
435 static int node_check(PyObject *obj, char **node)
436 {
436 {
437 Py_ssize_t nodelen;
437 Py_ssize_t nodelen;
438 if (PyBytes_AsStringAndSize(obj, node, &nodelen) == -1)
438 if (PyBytes_AsStringAndSize(obj, node, &nodelen) == -1)
439 return -1;
439 return -1;
440 if (nodelen == 20)
440 if (nodelen == 20)
441 return 0;
441 return 0;
442 PyErr_SetString(PyExc_ValueError, "20-byte hash required");
442 PyErr_SetString(PyExc_ValueError, "20-byte hash required");
443 return -1;
443 return -1;
444 }
444 }
445
445
446 static PyObject *index_append(indexObject *self, PyObject *obj)
446 static PyObject *index_append(indexObject *self, PyObject *obj)
447 {
447 {
448 char *node;
448 char *node;
449 Py_ssize_t len;
449 Py_ssize_t len;
450
450
451 if (!PyTuple_Check(obj) || PyTuple_GET_SIZE(obj) != 8) {
451 if (!PyTuple_Check(obj) || PyTuple_GET_SIZE(obj) != 8) {
452 PyErr_SetString(PyExc_TypeError, "8-tuple required");
452 PyErr_SetString(PyExc_TypeError, "8-tuple required");
453 return NULL;
453 return NULL;
454 }
454 }
455
455
456 if (node_check(PyTuple_GET_ITEM(obj, 7), &node) == -1)
456 if (node_check(PyTuple_GET_ITEM(obj, 7), &node) == -1)
457 return NULL;
457 return NULL;
458
458
459 len = index_length(self);
459 len = index_length(self);
460
460
461 if (self->added == NULL) {
461 if (self->added == NULL) {
462 self->added = PyList_New(0);
462 self->added = PyList_New(0);
463 if (self->added == NULL)
463 if (self->added == NULL)
464 return NULL;
464 return NULL;
465 }
465 }
466
466
467 if (PyList_Append(self->added, obj) == -1)
467 if (PyList_Append(self->added, obj) == -1)
468 return NULL;
468 return NULL;
469
469
470 if (self->ntinitialized)
470 if (self->ntinitialized)
471 nt_insert(&self->nt, node, (int)len);
471 nt_insert(&self->nt, node, (int)len);
472
472
473 Py_CLEAR(self->headrevs);
473 Py_CLEAR(self->headrevs);
474 Py_RETURN_NONE;
474 Py_RETURN_NONE;
475 }
475 }
476
476
477 static PyObject *index_stats(indexObject *self)
477 static PyObject *index_stats(indexObject *self)
478 {
478 {
479 PyObject *obj = PyDict_New();
479 PyObject *obj = PyDict_New();
480 PyObject *s = NULL;
480 PyObject *s = NULL;
481 PyObject *t = NULL;
481 PyObject *t = NULL;
482
482
483 if (obj == NULL)
483 if (obj == NULL)
484 return NULL;
484 return NULL;
485
485
486 #define istat(__n, __d) \
486 #define istat(__n, __d) \
487 do { \
487 do { \
488 s = PyBytes_FromString(__d); \
488 s = PyBytes_FromString(__d); \
489 t = PyInt_FromSsize_t(self->__n); \
489 t = PyInt_FromSsize_t(self->__n); \
490 if (!s || !t) \
490 if (!s || !t) \
491 goto bail; \
491 goto bail; \
492 if (PyDict_SetItem(obj, s, t) == -1) \
492 if (PyDict_SetItem(obj, s, t) == -1) \
493 goto bail; \
493 goto bail; \
494 Py_CLEAR(s); \
494 Py_CLEAR(s); \
495 Py_CLEAR(t); \
495 Py_CLEAR(t); \
496 } while (0)
496 } while (0)
497
497
498 if (self->added) {
498 if (self->added) {
499 Py_ssize_t len = PyList_GET_SIZE(self->added);
499 Py_ssize_t len = PyList_GET_SIZE(self->added);
500 s = PyBytes_FromString("index entries added");
500 s = PyBytes_FromString("index entries added");
501 t = PyInt_FromSsize_t(len);
501 t = PyInt_FromSsize_t(len);
502 if (!s || !t)
502 if (!s || !t)
503 goto bail;
503 goto bail;
504 if (PyDict_SetItem(obj, s, t) == -1)
504 if (PyDict_SetItem(obj, s, t) == -1)
505 goto bail;
505 goto bail;
506 Py_CLEAR(s);
506 Py_CLEAR(s);
507 Py_CLEAR(t);
507 Py_CLEAR(t);
508 }
508 }
509
509
510 if (self->raw_length != self->length)
510 if (self->raw_length != self->length)
511 istat(raw_length, "revs on disk");
511 istat(raw_length, "revs on disk");
512 istat(length, "revs in memory");
512 istat(length, "revs in memory");
513 istat(ntlookups, "node trie lookups");
513 istat(ntlookups, "node trie lookups");
514 istat(ntmisses, "node trie misses");
514 istat(ntmisses, "node trie misses");
515 istat(ntrev, "node trie last rev scanned");
515 istat(ntrev, "node trie last rev scanned");
516 if (self->ntinitialized) {
516 if (self->ntinitialized) {
517 istat(nt.capacity, "node trie capacity");
517 istat(nt.capacity, "node trie capacity");
518 istat(nt.depth, "node trie depth");
518 istat(nt.depth, "node trie depth");
519 istat(nt.length, "node trie count");
519 istat(nt.length, "node trie count");
520 istat(nt.splits, "node trie splits");
520 istat(nt.splits, "node trie splits");
521 }
521 }
522
522
523 #undef istat
523 #undef istat
524
524
525 return obj;
525 return obj;
526
526
527 bail:
527 bail:
528 Py_XDECREF(obj);
528 Py_XDECREF(obj);
529 Py_XDECREF(s);
529 Py_XDECREF(s);
530 Py_XDECREF(t);
530 Py_XDECREF(t);
531 return NULL;
531 return NULL;
532 }
532 }
533
533
534 /*
534 /*
535 * When we cache a list, we want to be sure the caller can't mutate
535 * When we cache a list, we want to be sure the caller can't mutate
536 * the cached copy.
536 * the cached copy.
537 */
537 */
538 static PyObject *list_copy(PyObject *list)
538 static PyObject *list_copy(PyObject *list)
539 {
539 {
540 Py_ssize_t len = PyList_GET_SIZE(list);
540 Py_ssize_t len = PyList_GET_SIZE(list);
541 PyObject *newlist = PyList_New(len);
541 PyObject *newlist = PyList_New(len);
542 Py_ssize_t i;
542 Py_ssize_t i;
543
543
544 if (newlist == NULL)
544 if (newlist == NULL)
545 return NULL;
545 return NULL;
546
546
547 for (i = 0; i < len; i++) {
547 for (i = 0; i < len; i++) {
548 PyObject *obj = PyList_GET_ITEM(list, i);
548 PyObject *obj = PyList_GET_ITEM(list, i);
549 Py_INCREF(obj);
549 Py_INCREF(obj);
550 PyList_SET_ITEM(newlist, i, obj);
550 PyList_SET_ITEM(newlist, i, obj);
551 }
551 }
552
552
553 return newlist;
553 return newlist;
554 }
554 }
555
555
556 static int check_filter(PyObject *filter, Py_ssize_t arg)
556 static int check_filter(PyObject *filter, Py_ssize_t arg)
557 {
557 {
558 if (filter) {
558 if (filter) {
559 PyObject *arglist, *result;
559 PyObject *arglist, *result;
560 int isfiltered;
560 int isfiltered;
561
561
562 arglist = Py_BuildValue("(n)", arg);
562 arglist = Py_BuildValue("(n)", arg);
563 if (!arglist) {
563 if (!arglist) {
564 return -1;
564 return -1;
565 }
565 }
566
566
567 result = PyEval_CallObject(filter, arglist);
567 result = PyEval_CallObject(filter, arglist);
568 Py_DECREF(arglist);
568 Py_DECREF(arglist);
569 if (!result) {
569 if (!result) {
570 return -1;
570 return -1;
571 }
571 }
572
572
573 /* PyObject_IsTrue returns 1 if true, 0 if false, -1 if error,
573 /* PyObject_IsTrue returns 1 if true, 0 if false, -1 if error,
574 * same as this function, so we can just return it directly.*/
574 * same as this function, so we can just return it directly.*/
575 isfiltered = PyObject_IsTrue(result);
575 isfiltered = PyObject_IsTrue(result);
576 Py_DECREF(result);
576 Py_DECREF(result);
577 return isfiltered;
577 return isfiltered;
578 } else {
578 } else {
579 return 0;
579 return 0;
580 }
580 }
581 }
581 }
582
582
583 static inline void set_phase_from_parents(char *phases, int parent_1,
583 static inline void set_phase_from_parents(char *phases, int parent_1,
584 int parent_2, Py_ssize_t i)
584 int parent_2, Py_ssize_t i)
585 {
585 {
586 if (parent_1 >= 0 && phases[parent_1] > phases[i])
586 if (parent_1 >= 0 && phases[parent_1] > phases[i])
587 phases[i] = phases[parent_1];
587 phases[i] = phases[parent_1];
588 if (parent_2 >= 0 && phases[parent_2] > phases[i])
588 if (parent_2 >= 0 && phases[parent_2] > phases[i])
589 phases[i] = phases[parent_2];
589 phases[i] = phases[parent_2];
590 }
590 }
591
591
592 static PyObject *reachableroots2(indexObject *self, PyObject *args)
592 static PyObject *reachableroots2(indexObject *self, PyObject *args)
593 {
593 {
594
594
595 /* Input */
595 /* Input */
596 long minroot;
596 long minroot;
597 PyObject *includepatharg = NULL;
597 PyObject *includepatharg = NULL;
598 int includepath = 0;
598 int includepath = 0;
599 /* heads and roots are lists */
599 /* heads and roots are lists */
600 PyObject *heads = NULL;
600 PyObject *heads = NULL;
601 PyObject *roots = NULL;
601 PyObject *roots = NULL;
602 PyObject *reachable = NULL;
602 PyObject *reachable = NULL;
603
603
604 PyObject *val;
604 PyObject *val;
605 Py_ssize_t len = index_length(self);
605 Py_ssize_t len = index_length(self);
606 long revnum;
606 long revnum;
607 Py_ssize_t k;
607 Py_ssize_t k;
608 Py_ssize_t i;
608 Py_ssize_t i;
609 Py_ssize_t l;
609 Py_ssize_t l;
610 int r;
610 int r;
611 int parents[2];
611 int parents[2];
612
612
613 /* Internal data structure:
613 /* Internal data structure:
614 * tovisit: array of length len+1 (all revs + nullrev), filled upto
614 * tovisit: array of length len+1 (all revs + nullrev), filled upto
615 * lentovisit
615 * lentovisit
616 *
616 *
617 * revstates: array of length len+1 (all revs + nullrev) */
617 * revstates: array of length len+1 (all revs + nullrev) */
618 int *tovisit = NULL;
618 int *tovisit = NULL;
619 long lentovisit = 0;
619 long lentovisit = 0;
620 enum { RS_SEEN = 1, RS_ROOT = 2, RS_REACHABLE = 4 };
620 enum { RS_SEEN = 1, RS_ROOT = 2, RS_REACHABLE = 4 };
621 char *revstates = NULL;
621 char *revstates = NULL;
622
622
623 /* Get arguments */
623 /* Get arguments */
624 if (!PyArg_ParseTuple(args, "lO!O!O!", &minroot, &PyList_Type, &heads,
624 if (!PyArg_ParseTuple(args, "lO!O!O!", &minroot, &PyList_Type, &heads,
625 &PyList_Type, &roots, &PyBool_Type,
625 &PyList_Type, &roots, &PyBool_Type,
626 &includepatharg))
626 &includepatharg))
627 goto bail;
627 goto bail;
628
628
629 if (includepatharg == Py_True)
629 if (includepatharg == Py_True)
630 includepath = 1;
630 includepath = 1;
631
631
632 /* Initialize return set */
632 /* Initialize return set */
633 reachable = PyList_New(0);
633 reachable = PyList_New(0);
634 if (reachable == NULL)
634 if (reachable == NULL)
635 goto bail;
635 goto bail;
636
636
637 /* Initialize internal datastructures */
637 /* Initialize internal datastructures */
638 tovisit = (int *)malloc((len + 1) * sizeof(int));
638 tovisit = (int *)malloc((len + 1) * sizeof(int));
639 if (tovisit == NULL) {
639 if (tovisit == NULL) {
640 PyErr_NoMemory();
640 PyErr_NoMemory();
641 goto bail;
641 goto bail;
642 }
642 }
643
643
644 revstates = (char *)calloc(len + 1, 1);
644 revstates = (char *)calloc(len + 1, 1);
645 if (revstates == NULL) {
645 if (revstates == NULL) {
646 PyErr_NoMemory();
646 PyErr_NoMemory();
647 goto bail;
647 goto bail;
648 }
648 }
649
649
650 l = PyList_GET_SIZE(roots);
650 l = PyList_GET_SIZE(roots);
651 for (i = 0; i < l; i++) {
651 for (i = 0; i < l; i++) {
652 revnum = PyInt_AsLong(PyList_GET_ITEM(roots, i));
652 revnum = PyInt_AsLong(PyList_GET_ITEM(roots, i));
653 if (revnum == -1 && PyErr_Occurred())
653 if (revnum == -1 && PyErr_Occurred())
654 goto bail;
654 goto bail;
655 /* If root is out of range, e.g. wdir(), it must be unreachable
655 /* If root is out of range, e.g. wdir(), it must be unreachable
656 * from heads. So we can just ignore it. */
656 * from heads. So we can just ignore it. */
657 if (revnum + 1 < 0 || revnum + 1 >= len + 1)
657 if (revnum + 1 < 0 || revnum + 1 >= len + 1)
658 continue;
658 continue;
659 revstates[revnum + 1] |= RS_ROOT;
659 revstates[revnum + 1] |= RS_ROOT;
660 }
660 }
661
661
662 /* Populate tovisit with all the heads */
662 /* Populate tovisit with all the heads */
663 l = PyList_GET_SIZE(heads);
663 l = PyList_GET_SIZE(heads);
664 for (i = 0; i < l; i++) {
664 for (i = 0; i < l; i++) {
665 revnum = PyInt_AsLong(PyList_GET_ITEM(heads, i));
665 revnum = PyInt_AsLong(PyList_GET_ITEM(heads, i));
666 if (revnum == -1 && PyErr_Occurred())
666 if (revnum == -1 && PyErr_Occurred())
667 goto bail;
667 goto bail;
668 if (revnum + 1 < 0 || revnum + 1 >= len + 1) {
668 if (revnum + 1 < 0 || revnum + 1 >= len + 1) {
669 PyErr_SetString(PyExc_IndexError, "head out of range");
669 PyErr_SetString(PyExc_IndexError, "head out of range");
670 goto bail;
670 goto bail;
671 }
671 }
672 if (!(revstates[revnum + 1] & RS_SEEN)) {
672 if (!(revstates[revnum + 1] & RS_SEEN)) {
673 tovisit[lentovisit++] = (int)revnum;
673 tovisit[lentovisit++] = (int)revnum;
674 revstates[revnum + 1] |= RS_SEEN;
674 revstates[revnum + 1] |= RS_SEEN;
675 }
675 }
676 }
676 }
677
677
678 /* Visit the tovisit list and find the reachable roots */
678 /* Visit the tovisit list and find the reachable roots */
679 k = 0;
679 k = 0;
680 while (k < lentovisit) {
680 while (k < lentovisit) {
681 /* Add the node to reachable if it is a root*/
681 /* Add the node to reachable if it is a root*/
682 revnum = tovisit[k++];
682 revnum = tovisit[k++];
683 if (revstates[revnum + 1] & RS_ROOT) {
683 if (revstates[revnum + 1] & RS_ROOT) {
684 revstates[revnum + 1] |= RS_REACHABLE;
684 revstates[revnum + 1] |= RS_REACHABLE;
685 val = PyInt_FromLong(revnum);
685 val = PyInt_FromLong(revnum);
686 if (val == NULL)
686 if (val == NULL)
687 goto bail;
687 goto bail;
688 r = PyList_Append(reachable, val);
688 r = PyList_Append(reachable, val);
689 Py_DECREF(val);
689 Py_DECREF(val);
690 if (r < 0)
690 if (r < 0)
691 goto bail;
691 goto bail;
692 if (includepath == 0)
692 if (includepath == 0)
693 continue;
693 continue;
694 }
694 }
695
695
696 /* Add its parents to the list of nodes to visit */
696 /* Add its parents to the list of nodes to visit */
697 if (revnum == nullrev)
697 if (revnum == nullrev)
698 continue;
698 continue;
699 r = index_get_parents(self, revnum, parents, (int)len - 1);
699 r = index_get_parents(self, revnum, parents, (int)len - 1);
700 if (r < 0)
700 if (r < 0)
701 goto bail;
701 goto bail;
702 for (i = 0; i < 2; i++) {
702 for (i = 0; i < 2; i++) {
703 if (!(revstates[parents[i] + 1] & RS_SEEN) &&
703 if (!(revstates[parents[i] + 1] & RS_SEEN) &&
704 parents[i] >= minroot) {
704 parents[i] >= minroot) {
705 tovisit[lentovisit++] = parents[i];
705 tovisit[lentovisit++] = parents[i];
706 revstates[parents[i] + 1] |= RS_SEEN;
706 revstates[parents[i] + 1] |= RS_SEEN;
707 }
707 }
708 }
708 }
709 }
709 }
710
710
711 /* Find all the nodes in between the roots we found and the heads
711 /* Find all the nodes in between the roots we found and the heads
712 * and add them to the reachable set */
712 * and add them to the reachable set */
713 if (includepath == 1) {
713 if (includepath == 1) {
714 long minidx = minroot;
714 long minidx = minroot;
715 if (minidx < 0)
715 if (minidx < 0)
716 minidx = 0;
716 minidx = 0;
717 for (i = minidx; i < len; i++) {
717 for (i = minidx; i < len; i++) {
718 if (!(revstates[i + 1] & RS_SEEN))
718 if (!(revstates[i + 1] & RS_SEEN))
719 continue;
719 continue;
720 r = index_get_parents(self, i, parents, (int)len - 1);
720 r = index_get_parents(self, i, parents, (int)len - 1);
721 /* Corrupted index file, error is set from
721 /* Corrupted index file, error is set from
722 * index_get_parents */
722 * index_get_parents */
723 if (r < 0)
723 if (r < 0)
724 goto bail;
724 goto bail;
725 if (((revstates[parents[0] + 1] |
725 if (((revstates[parents[0] + 1] |
726 revstates[parents[1] + 1]) &
726 revstates[parents[1] + 1]) &
727 RS_REACHABLE) &&
727 RS_REACHABLE) &&
728 !(revstates[i + 1] & RS_REACHABLE)) {
728 !(revstates[i + 1] & RS_REACHABLE)) {
729 revstates[i + 1] |= RS_REACHABLE;
729 revstates[i + 1] |= RS_REACHABLE;
730 val = PyInt_FromSsize_t(i);
730 val = PyInt_FromSsize_t(i);
731 if (val == NULL)
731 if (val == NULL)
732 goto bail;
732 goto bail;
733 r = PyList_Append(reachable, val);
733 r = PyList_Append(reachable, val);
734 Py_DECREF(val);
734 Py_DECREF(val);
735 if (r < 0)
735 if (r < 0)
736 goto bail;
736 goto bail;
737 }
737 }
738 }
738 }
739 }
739 }
740
740
741 free(revstates);
741 free(revstates);
742 free(tovisit);
742 free(tovisit);
743 return reachable;
743 return reachable;
744 bail:
744 bail:
745 Py_XDECREF(reachable);
745 Py_XDECREF(reachable);
746 free(revstates);
746 free(revstates);
747 free(tovisit);
747 free(tovisit);
748 return NULL;
748 return NULL;
749 }
749 }
750
750
751 static int add_roots_get_min(indexObject *self, PyObject *roots, char *phases,
751 static int add_roots_get_min(indexObject *self, PyObject *roots, char *phases,
752 char phase)
752 char phase)
753 {
753 {
754 Py_ssize_t len = index_length(self);
754 Py_ssize_t len = index_length(self);
755 PyObject *iter;
756 PyObject *item;
755 PyObject *item;
757 PyObject *iterator;
756 PyObject *iterator;
758 int rev, minrev = -1;
757 int rev, minrev = -1;
759 char *node;
758 char *node;
760
759
761 if (!PySet_Check(roots))
760 if (!PySet_Check(roots))
762 return -2;
761 return -2;
763 iterator = PyObject_GetIter(roots);
762 iterator = PyObject_GetIter(roots);
764 if (iterator == NULL)
763 if (iterator == NULL)
765 return -2;
764 return -2;
766 while ((item = PyIter_Next(iterator))) {
765 while ((item = PyIter_Next(iterator))) {
767 if (node_check(item, &node) == -1)
766 if (node_check(item, &node) == -1)
768 goto failed;
767 goto failed;
769 rev = index_find_node(self, node, 20);
768 rev = index_find_node(self, node, 20);
770 /* null is implicitly public, so negative is invalid */
769 /* null is implicitly public, so negative is invalid */
771 if (rev < 0 || rev >= len)
770 if (rev < 0 || rev >= len)
772 goto failed;
771 goto failed;
773 phases[rev] = phase;
772 phases[rev] = phase;
774 if (minrev == -1 || minrev > rev)
773 if (minrev == -1 || minrev > rev)
775 minrev = rev;
774 minrev = rev;
776 Py_DECREF(item);
775 Py_DECREF(item);
777 }
776 }
778 Py_DECREF(iterator);
777 Py_DECREF(iterator);
779 return minrev;
778 return minrev;
780 failed:
779 failed:
781 Py_DECREF(iterator);
780 Py_DECREF(iterator);
782 Py_DECREF(item);
781 Py_DECREF(item);
783 return -2;
782 return -2;
784 }
783 }
785
784
786 static PyObject *compute_phases_map_sets(indexObject *self, PyObject *args)
785 static PyObject *compute_phases_map_sets(indexObject *self, PyObject *args)
787 {
786 {
788 /* 0: public (untracked), 1: draft, 2: secret, 32: archive,
787 /* 0: public (untracked), 1: draft, 2: secret, 32: archive,
789 96: internal */
788 96: internal */
790 static const char trackedphases[] = {1, 2, 32, 96};
789 static const char trackedphases[] = {1, 2, 32, 96};
791 PyObject *ret = NULL;
790 PyObject *ret = NULL;
792 PyObject *roots = Py_None;
791 PyObject *roots = Py_None;
793 PyObject *idx = NULL;
794 PyObject *pyphase = NULL;
792 PyObject *pyphase = NULL;
795 PyObject *pyrev = NULL;
793 PyObject *pyrev = NULL;
796 PyObject *phaseroots = NULL;
794 PyObject *phaseroots = NULL;
797 PyObject *phasessize = NULL;
795 PyObject *phasessize = NULL;
798 PyObject *phasesets[4] = {NULL, NULL, NULL, NULL};
796 PyObject *phasesets[4] = {NULL, NULL, NULL, NULL};
799 Py_ssize_t len = index_length(self);
797 Py_ssize_t len = index_length(self);
800 const char *currentphase;
801 char *phases = NULL;
798 char *phases = NULL;
802 int minphaserev = -1, rev, i;
799 int minphaserev = -1, rev, i;
803 const int numphases = (int)(sizeof(phasesets) / sizeof(phasesets[0]));
800 const int numphases = (int)(sizeof(phasesets) / sizeof(phasesets[0]));
804
801
805 if (!PyArg_ParseTuple(args, "O", &roots))
802 if (!PyArg_ParseTuple(args, "O", &roots))
806 return NULL;
803 return NULL;
807 if (roots == NULL || !PyDict_Check(roots)) {
804 if (roots == NULL || !PyDict_Check(roots)) {
808 PyErr_SetString(PyExc_TypeError, "roots must be a dictionary");
805 PyErr_SetString(PyExc_TypeError, "roots must be a dictionary");
809 return NULL;
806 return NULL;
810 }
807 }
811
808
812 phases = calloc(len, 1);
809 phases = calloc(len, 1);
813 if (phases == NULL) {
810 if (phases == NULL) {
814 PyErr_NoMemory();
811 PyErr_NoMemory();
815 return NULL;
812 return NULL;
816 }
813 }
817
814
818 for (i = 0; i < numphases; ++i) {
815 for (i = 0; i < numphases; ++i) {
819 pyphase = PyInt_FromLong(trackedphases[i]);
816 pyphase = PyInt_FromLong(trackedphases[i]);
820 if (pyphase == NULL)
817 if (pyphase == NULL)
821 goto release;
818 goto release;
822 phaseroots = PyDict_GetItem(roots, pyphase);
819 phaseroots = PyDict_GetItem(roots, pyphase);
823 Py_DECREF(pyphase);
820 Py_DECREF(pyphase);
824 if (phaseroots == NULL)
821 if (phaseroots == NULL)
825 continue;
822 continue;
826 rev = add_roots_get_min(self, phaseroots, phases, trackedphases[i]);
823 rev = add_roots_get_min(self, phaseroots, phases, trackedphases[i]);
827 phaseroots = NULL;
824 phaseroots = NULL;
828 if (rev == -2)
825 if (rev == -2)
829 goto release;
826 goto release;
830 if (rev != -1 && (minphaserev == -1 || rev < minphaserev))
827 if (rev != -1 && (minphaserev == -1 || rev < minphaserev))
831 minphaserev = rev;
828 minphaserev = rev;
832 }
829 }
833
830
834 for (i = 0; i < numphases; ++i) {
831 for (i = 0; i < numphases; ++i) {
835 phasesets[i] = PySet_New(NULL);
832 phasesets[i] = PySet_New(NULL);
836 if (phasesets[i] == NULL)
833 if (phasesets[i] == NULL)
837 goto release;
834 goto release;
838 }
835 }
839
836
840 if (minphaserev == -1)
837 if (minphaserev == -1)
841 minphaserev = len;
838 minphaserev = len;
842 for (rev = minphaserev; rev < len; ++rev) {
839 for (rev = minphaserev; rev < len; ++rev) {
843 int parents[2];
840 int parents[2];
844 /*
841 /*
845 * The parent lookup could be skipped for phaseroots, but
842 * The parent lookup could be skipped for phaseroots, but
846 * phase --force would historically not recompute them
843 * phase --force would historically not recompute them
847 * correctly, leaving descendents with a lower phase around.
844 * correctly, leaving descendents with a lower phase around.
848 * As such, unconditionally recompute the phase.
845 * As such, unconditionally recompute the phase.
849 */
846 */
850 if (index_get_parents(self, rev, parents, (int)len - 1) < 0)
847 if (index_get_parents(self, rev, parents, (int)len - 1) < 0)
851 goto release;
848 goto release;
852 set_phase_from_parents(phases, parents[0], parents[1], rev);
849 set_phase_from_parents(phases, parents[0], parents[1], rev);
853 switch (phases[rev]) {
850 switch (phases[rev]) {
854 case 0:
851 case 0:
855 continue;
852 continue;
856 case 1:
853 case 1:
857 pyphase = phasesets[0];
854 pyphase = phasesets[0];
858 break;
855 break;
859 case 2:
856 case 2:
860 pyphase = phasesets[1];
857 pyphase = phasesets[1];
861 break;
858 break;
862 case 32:
859 case 32:
863 pyphase = phasesets[2];
860 pyphase = phasesets[2];
864 break;
861 break;
865 case 96:
862 case 96:
866 pyphase = phasesets[3];
863 pyphase = phasesets[3];
867 break;
864 break;
868 default:
865 default:
869 goto release;
866 goto release;
870 }
867 }
871 pyrev = PyInt_FromLong(rev);
868 pyrev = PyInt_FromLong(rev);
872 if (pyrev == NULL)
869 if (pyrev == NULL)
873 goto release;
870 goto release;
874 if (PySet_Add(pyphase, pyrev) == -1) {
871 if (PySet_Add(pyphase, pyrev) == -1) {
875 Py_DECREF(pyrev);
872 Py_DECREF(pyrev);
876 goto release;
873 goto release;
877 }
874 }
878 Py_DECREF(pyrev);
875 Py_DECREF(pyrev);
879 }
876 }
880 phaseroots = _dict_new_presized(numphases);
877 phaseroots = _dict_new_presized(numphases);
881 if (phaseroots == NULL)
878 if (phaseroots == NULL)
882 goto release;
879 goto release;
883 for (int i = 0; i < numphases; ++i) {
880 for (i = 0; i < numphases; ++i) {
884 pyphase = PyInt_FromLong(trackedphases[i]);
881 pyphase = PyInt_FromLong(trackedphases[i]);
885 if (pyphase == NULL)
882 if (pyphase == NULL)
886 goto release;
883 goto release;
887 if (PyDict_SetItem(phaseroots, pyphase, phasesets[i]) == -1) {
884 if (PyDict_SetItem(phaseroots, pyphase, phasesets[i]) == -1) {
888 Py_DECREF(pyphase);
885 Py_DECREF(pyphase);
889 goto release;
886 goto release;
890 }
887 }
891 Py_DECREF(phasesets[i]);
888 Py_DECREF(phasesets[i]);
892 phasesets[i] = NULL;
889 phasesets[i] = NULL;
893 }
890 }
894 phasessize = PyInt_FromSsize_t(len);
891 phasessize = PyInt_FromSsize_t(len);
895 if (phasessize == NULL)
892 if (phasessize == NULL)
896 goto release;
893 goto release;
897
894
898 ret = PyTuple_Pack(2, phasessize, phaseroots);
895 ret = PyTuple_Pack(2, phasessize, phaseroots);
899 Py_DECREF(phasessize);
896 Py_DECREF(phasessize);
900 Py_DECREF(phaseroots);
897 Py_DECREF(phaseroots);
901 return ret;
898 return ret;
902
899
903 release:
900 release:
904 for (i = 0; i < numphases; ++i)
901 for (i = 0; i < numphases; ++i)
905 Py_XDECREF(phasesets[i]);
902 Py_XDECREF(phasesets[i]);
906 Py_XDECREF(phaseroots);
903 Py_XDECREF(phaseroots);
907
904
908 free(phases);
905 free(phases);
909 return NULL;
906 return NULL;
910 }
907 }
911
908
912 static PyObject *index_headrevs(indexObject *self, PyObject *args)
909 static PyObject *index_headrevs(indexObject *self, PyObject *args)
913 {
910 {
914 Py_ssize_t i, j, len;
911 Py_ssize_t i, j, len;
915 char *nothead = NULL;
912 char *nothead = NULL;
916 PyObject *heads = NULL;
913 PyObject *heads = NULL;
917 PyObject *filter = NULL;
914 PyObject *filter = NULL;
918 PyObject *filteredrevs = Py_None;
915 PyObject *filteredrevs = Py_None;
919
916
920 if (!PyArg_ParseTuple(args, "|O", &filteredrevs)) {
917 if (!PyArg_ParseTuple(args, "|O", &filteredrevs)) {
921 return NULL;
918 return NULL;
922 }
919 }
923
920
924 if (self->headrevs && filteredrevs == self->filteredrevs)
921 if (self->headrevs && filteredrevs == self->filteredrevs)
925 return list_copy(self->headrevs);
922 return list_copy(self->headrevs);
926
923
927 Py_DECREF(self->filteredrevs);
924 Py_DECREF(self->filteredrevs);
928 self->filteredrevs = filteredrevs;
925 self->filteredrevs = filteredrevs;
929 Py_INCREF(filteredrevs);
926 Py_INCREF(filteredrevs);
930
927
931 if (filteredrevs != Py_None) {
928 if (filteredrevs != Py_None) {
932 filter = PyObject_GetAttrString(filteredrevs, "__contains__");
929 filter = PyObject_GetAttrString(filteredrevs, "__contains__");
933 if (!filter) {
930 if (!filter) {
934 PyErr_SetString(
931 PyErr_SetString(
935 PyExc_TypeError,
932 PyExc_TypeError,
936 "filteredrevs has no attribute __contains__");
933 "filteredrevs has no attribute __contains__");
937 goto bail;
934 goto bail;
938 }
935 }
939 }
936 }
940
937
941 len = index_length(self);
938 len = index_length(self);
942 heads = PyList_New(0);
939 heads = PyList_New(0);
943 if (heads == NULL)
940 if (heads == NULL)
944 goto bail;
941 goto bail;
945 if (len == 0) {
942 if (len == 0) {
946 PyObject *nullid = PyInt_FromLong(-1);
943 PyObject *nullid = PyInt_FromLong(-1);
947 if (nullid == NULL || PyList_Append(heads, nullid) == -1) {
944 if (nullid == NULL || PyList_Append(heads, nullid) == -1) {
948 Py_XDECREF(nullid);
945 Py_XDECREF(nullid);
949 goto bail;
946 goto bail;
950 }
947 }
951 goto done;
948 goto done;
952 }
949 }
953
950
954 nothead = calloc(len, 1);
951 nothead = calloc(len, 1);
955 if (nothead == NULL) {
952 if (nothead == NULL) {
956 PyErr_NoMemory();
953 PyErr_NoMemory();
957 goto bail;
954 goto bail;
958 }
955 }
959
956
960 for (i = len - 1; i >= 0; i--) {
957 for (i = len - 1; i >= 0; i--) {
961 int isfiltered;
958 int isfiltered;
962 int parents[2];
959 int parents[2];
963
960
964 /* If nothead[i] == 1, it means we've seen an unfiltered child
961 /* If nothead[i] == 1, it means we've seen an unfiltered child
965 * of this node already, and therefore this node is not
962 * of this node already, and therefore this node is not
966 * filtered. So we can skip the expensive check_filter step.
963 * filtered. So we can skip the expensive check_filter step.
967 */
964 */
968 if (nothead[i] != 1) {
965 if (nothead[i] != 1) {
969 isfiltered = check_filter(filter, i);
966 isfiltered = check_filter(filter, i);
970 if (isfiltered == -1) {
967 if (isfiltered == -1) {
971 PyErr_SetString(PyExc_TypeError,
968 PyErr_SetString(PyExc_TypeError,
972 "unable to check filter");
969 "unable to check filter");
973 goto bail;
970 goto bail;
974 }
971 }
975
972
976 if (isfiltered) {
973 if (isfiltered) {
977 nothead[i] = 1;
974 nothead[i] = 1;
978 continue;
975 continue;
979 }
976 }
980 }
977 }
981
978
982 if (index_get_parents(self, i, parents, (int)len - 1) < 0)
979 if (index_get_parents(self, i, parents, (int)len - 1) < 0)
983 goto bail;
980 goto bail;
984 for (j = 0; j < 2; j++) {
981 for (j = 0; j < 2; j++) {
985 if (parents[j] >= 0)
982 if (parents[j] >= 0)
986 nothead[parents[j]] = 1;
983 nothead[parents[j]] = 1;
987 }
984 }
988 }
985 }
989
986
990 for (i = 0; i < len; i++) {
987 for (i = 0; i < len; i++) {
991 PyObject *head;
988 PyObject *head;
992
989
993 if (nothead[i])
990 if (nothead[i])
994 continue;
991 continue;
995 head = PyInt_FromSsize_t(i);
992 head = PyInt_FromSsize_t(i);
996 if (head == NULL || PyList_Append(heads, head) == -1) {
993 if (head == NULL || PyList_Append(heads, head) == -1) {
997 Py_XDECREF(head);
994 Py_XDECREF(head);
998 goto bail;
995 goto bail;
999 }
996 }
1000 }
997 }
1001
998
1002 done:
999 done:
1003 self->headrevs = heads;
1000 self->headrevs = heads;
1004 Py_XDECREF(filter);
1001 Py_XDECREF(filter);
1005 free(nothead);
1002 free(nothead);
1006 return list_copy(self->headrevs);
1003 return list_copy(self->headrevs);
1007 bail:
1004 bail:
1008 Py_XDECREF(filter);
1005 Py_XDECREF(filter);
1009 Py_XDECREF(heads);
1006 Py_XDECREF(heads);
1010 free(nothead);
1007 free(nothead);
1011 return NULL;
1008 return NULL;
1012 }
1009 }
1013
1010
1014 /**
1011 /**
1015 * Obtain the base revision index entry.
1012 * Obtain the base revision index entry.
1016 *
1013 *
1017 * Callers must ensure that rev >= 0 or illegal memory access may occur.
1014 * Callers must ensure that rev >= 0 or illegal memory access may occur.
1018 */
1015 */
1019 static inline int index_baserev(indexObject *self, int rev)
1016 static inline int index_baserev(indexObject *self, int rev)
1020 {
1017 {
1021 const char *data;
1018 const char *data;
1022 int result;
1019 int result;
1023
1020
1024 if (rev >= self->length) {
1021 if (rev >= self->length) {
1025 PyObject *tuple =
1022 PyObject *tuple =
1026 PyList_GET_ITEM(self->added, rev - self->length);
1023 PyList_GET_ITEM(self->added, rev - self->length);
1027 long ret;
1024 long ret;
1028 if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 3), &ret)) {
1025 if (!pylong_to_long(PyTuple_GET_ITEM(tuple, 3), &ret)) {
1029 return -2;
1026 return -2;
1030 }
1027 }
1031 result = (int)ret;
1028 result = (int)ret;
1032 } else {
1029 } else {
1033 data = index_deref(self, rev);
1030 data = index_deref(self, rev);
1034 if (data == NULL) {
1031 if (data == NULL) {
1035 return -2;
1032 return -2;
1036 }
1033 }
1037
1034
1038 result = getbe32(data + 16);
1035 result = getbe32(data + 16);
1039 }
1036 }
1040 if (result > rev) {
1037 if (result > rev) {
1041 PyErr_Format(
1038 PyErr_Format(
1042 PyExc_ValueError,
1039 PyExc_ValueError,
1043 "corrupted revlog, revision base above revision: %d, %d",
1040 "corrupted revlog, revision base above revision: %d, %d",
1044 rev, result);
1041 rev, result);
1045 return -2;
1042 return -2;
1046 }
1043 }
1047 if (result < -1) {
1044 if (result < -1) {
1048 PyErr_Format(
1045 PyErr_Format(
1049 PyExc_ValueError,
1046 PyExc_ValueError,
1050 "corrupted revlog, revision base out of range: %d, %d", rev,
1047 "corrupted revlog, revision base out of range: %d, %d", rev,
1051 result);
1048 result);
1052 return -2;
1049 return -2;
1053 }
1050 }
1054 return result;
1051 return result;
1055 }
1052 }
1056
1053
1057 /**
1054 /**
1058 * Find if a revision is a snapshot or not
1055 * Find if a revision is a snapshot or not
1059 *
1056 *
1060 * Only relevant for sparse-revlog case.
1057 * Only relevant for sparse-revlog case.
1061 * Callers must ensure that rev is in a valid range.
1058 * Callers must ensure that rev is in a valid range.
1062 */
1059 */
1063 static int index_issnapshotrev(indexObject *self, Py_ssize_t rev)
1060 static int index_issnapshotrev(indexObject *self, Py_ssize_t rev)
1064 {
1061 {
1065 int ps[2];
1062 int ps[2];
1066 Py_ssize_t base;
1063 Py_ssize_t base;
1067 while (rev >= 0) {
1064 while (rev >= 0) {
1068 base = (Py_ssize_t)index_baserev(self, rev);
1065 base = (Py_ssize_t)index_baserev(self, rev);
1069 if (base == rev) {
1066 if (base == rev) {
1070 base = -1;
1067 base = -1;
1071 }
1068 }
1072 if (base == -2) {
1069 if (base == -2) {
1073 assert(PyErr_Occurred());
1070 assert(PyErr_Occurred());
1074 return -1;
1071 return -1;
1075 }
1072 }
1076 if (base == -1) {
1073 if (base == -1) {
1077 return 1;
1074 return 1;
1078 }
1075 }
1079 if (index_get_parents(self, rev, ps, (int)rev) < 0) {
1076 if (index_get_parents(self, rev, ps, (int)rev) < 0) {
1080 assert(PyErr_Occurred());
1077 assert(PyErr_Occurred());
1081 return -1;
1078 return -1;
1082 };
1079 };
1083 if (base == ps[0] || base == ps[1]) {
1080 if (base == ps[0] || base == ps[1]) {
1084 return 0;
1081 return 0;
1085 }
1082 }
1086 rev = base;
1083 rev = base;
1087 }
1084 }
1088 return rev == -1;
1085 return rev == -1;
1089 }
1086 }
1090
1087
1091 static PyObject *index_issnapshot(indexObject *self, PyObject *value)
1088 static PyObject *index_issnapshot(indexObject *self, PyObject *value)
1092 {
1089 {
1093 long rev;
1090 long rev;
1094 int issnap;
1091 int issnap;
1095 Py_ssize_t length = index_length(self);
1092 Py_ssize_t length = index_length(self);
1096
1093
1097 if (!pylong_to_long(value, &rev)) {
1094 if (!pylong_to_long(value, &rev)) {
1098 return NULL;
1095 return NULL;
1099 }
1096 }
1100 if (rev < -1 || rev >= length) {
1097 if (rev < -1 || rev >= length) {
1101 PyErr_Format(PyExc_ValueError, "revlog index out of range: %ld",
1098 PyErr_Format(PyExc_ValueError, "revlog index out of range: %ld",
1102 rev);
1099 rev);
1103 return NULL;
1100 return NULL;
1104 };
1101 };
1105 issnap = index_issnapshotrev(self, (Py_ssize_t)rev);
1102 issnap = index_issnapshotrev(self, (Py_ssize_t)rev);
1106 if (issnap < 0) {
1103 if (issnap < 0) {
1107 return NULL;
1104 return NULL;
1108 };
1105 };
1109 return PyBool_FromLong((long)issnap);
1106 return PyBool_FromLong((long)issnap);
1110 }
1107 }
1111
1108
1112 static PyObject *index_findsnapshots(indexObject *self, PyObject *args)
1109 static PyObject *index_findsnapshots(indexObject *self, PyObject *args)
1113 {
1110 {
1114 Py_ssize_t start_rev;
1111 Py_ssize_t start_rev;
1115 PyObject *cache;
1112 PyObject *cache;
1116 Py_ssize_t base;
1113 Py_ssize_t base;
1117 Py_ssize_t rev;
1114 Py_ssize_t rev;
1118 PyObject *key = NULL;
1115 PyObject *key = NULL;
1119 PyObject *value = NULL;
1116 PyObject *value = NULL;
1120 const Py_ssize_t length = index_length(self);
1117 const Py_ssize_t length = index_length(self);
1121 if (!PyArg_ParseTuple(args, "O!n", &PyDict_Type, &cache, &start_rev)) {
1118 if (!PyArg_ParseTuple(args, "O!n", &PyDict_Type, &cache, &start_rev)) {
1122 return NULL;
1119 return NULL;
1123 }
1120 }
1124 for (rev = start_rev; rev < length; rev++) {
1121 for (rev = start_rev; rev < length; rev++) {
1125 int issnap;
1122 int issnap;
1126 PyObject *allvalues = NULL;
1123 PyObject *allvalues = NULL;
1127 issnap = index_issnapshotrev(self, rev);
1124 issnap = index_issnapshotrev(self, rev);
1128 if (issnap < 0) {
1125 if (issnap < 0) {
1129 goto bail;
1126 goto bail;
1130 }
1127 }
1131 if (issnap == 0) {
1128 if (issnap == 0) {
1132 continue;
1129 continue;
1133 }
1130 }
1134 base = (Py_ssize_t)index_baserev(self, rev);
1131 base = (Py_ssize_t)index_baserev(self, rev);
1135 if (base == rev) {
1132 if (base == rev) {
1136 base = -1;
1133 base = -1;
1137 }
1134 }
1138 if (base == -2) {
1135 if (base == -2) {
1139 assert(PyErr_Occurred());
1136 assert(PyErr_Occurred());
1140 goto bail;
1137 goto bail;
1141 }
1138 }
1142 key = PyInt_FromSsize_t(base);
1139 key = PyInt_FromSsize_t(base);
1143 allvalues = PyDict_GetItem(cache, key);
1140 allvalues = PyDict_GetItem(cache, key);
1144 if (allvalues == NULL && PyErr_Occurred()) {
1141 if (allvalues == NULL && PyErr_Occurred()) {
1145 goto bail;
1142 goto bail;
1146 }
1143 }
1147 if (allvalues == NULL) {
1144 if (allvalues == NULL) {
1148 int r;
1145 int r;
1149 allvalues = PyList_New(0);
1146 allvalues = PyList_New(0);
1150 if (!allvalues) {
1147 if (!allvalues) {
1151 goto bail;
1148 goto bail;
1152 }
1149 }
1153 r = PyDict_SetItem(cache, key, allvalues);
1150 r = PyDict_SetItem(cache, key, allvalues);
1154 Py_DECREF(allvalues);
1151 Py_DECREF(allvalues);
1155 if (r < 0) {
1152 if (r < 0) {
1156 goto bail;
1153 goto bail;
1157 }
1154 }
1158 }
1155 }
1159 value = PyInt_FromSsize_t(rev);
1156 value = PyInt_FromSsize_t(rev);
1160 if (PyList_Append(allvalues, value)) {
1157 if (PyList_Append(allvalues, value)) {
1161 goto bail;
1158 goto bail;
1162 }
1159 }
1163 Py_CLEAR(key);
1160 Py_CLEAR(key);
1164 Py_CLEAR(value);
1161 Py_CLEAR(value);
1165 }
1162 }
1166 Py_RETURN_NONE;
1163 Py_RETURN_NONE;
1167 bail:
1164 bail:
1168 Py_XDECREF(key);
1165 Py_XDECREF(key);
1169 Py_XDECREF(value);
1166 Py_XDECREF(value);
1170 return NULL;
1167 return NULL;
1171 }
1168 }
1172
1169
1173 static PyObject *index_deltachain(indexObject *self, PyObject *args)
1170 static PyObject *index_deltachain(indexObject *self, PyObject *args)
1174 {
1171 {
1175 int rev, generaldelta;
1172 int rev, generaldelta;
1176 PyObject *stoparg;
1173 PyObject *stoparg;
1177 int stoprev, iterrev, baserev = -1;
1174 int stoprev, iterrev, baserev = -1;
1178 int stopped;
1175 int stopped;
1179 PyObject *chain = NULL, *result = NULL;
1176 PyObject *chain = NULL, *result = NULL;
1180 const Py_ssize_t length = index_length(self);
1177 const Py_ssize_t length = index_length(self);
1181
1178
1182 if (!PyArg_ParseTuple(args, "iOi", &rev, &stoparg, &generaldelta)) {
1179 if (!PyArg_ParseTuple(args, "iOi", &rev, &stoparg, &generaldelta)) {
1183 return NULL;
1180 return NULL;
1184 }
1181 }
1185
1182
1186 if (PyInt_Check(stoparg)) {
1183 if (PyInt_Check(stoparg)) {
1187 stoprev = (int)PyInt_AsLong(stoparg);
1184 stoprev = (int)PyInt_AsLong(stoparg);
1188 if (stoprev == -1 && PyErr_Occurred()) {
1185 if (stoprev == -1 && PyErr_Occurred()) {
1189 return NULL;
1186 return NULL;
1190 }
1187 }
1191 } else if (stoparg == Py_None) {
1188 } else if (stoparg == Py_None) {
1192 stoprev = -2;
1189 stoprev = -2;
1193 } else {
1190 } else {
1194 PyErr_SetString(PyExc_ValueError,
1191 PyErr_SetString(PyExc_ValueError,
1195 "stoprev must be integer or None");
1192 "stoprev must be integer or None");
1196 return NULL;
1193 return NULL;
1197 }
1194 }
1198
1195
1199 if (rev < 0 || rev >= length) {
1196 if (rev < 0 || rev >= length) {
1200 PyErr_SetString(PyExc_ValueError, "revlog index out of range");
1197 PyErr_SetString(PyExc_ValueError, "revlog index out of range");
1201 return NULL;
1198 return NULL;
1202 }
1199 }
1203
1200
1204 chain = PyList_New(0);
1201 chain = PyList_New(0);
1205 if (chain == NULL) {
1202 if (chain == NULL) {
1206 return NULL;
1203 return NULL;
1207 }
1204 }
1208
1205
1209 baserev = index_baserev(self, rev);
1206 baserev = index_baserev(self, rev);
1210
1207
1211 /* This should never happen. */
1208 /* This should never happen. */
1212 if (baserev <= -2) {
1209 if (baserev <= -2) {
1213 /* Error should be set by index_deref() */
1210 /* Error should be set by index_deref() */
1214 assert(PyErr_Occurred());
1211 assert(PyErr_Occurred());
1215 goto bail;
1212 goto bail;
1216 }
1213 }
1217
1214
1218 iterrev = rev;
1215 iterrev = rev;
1219
1216
1220 while (iterrev != baserev && iterrev != stoprev) {
1217 while (iterrev != baserev && iterrev != stoprev) {
1221 PyObject *value = PyInt_FromLong(iterrev);
1218 PyObject *value = PyInt_FromLong(iterrev);
1222 if (value == NULL) {
1219 if (value == NULL) {
1223 goto bail;
1220 goto bail;
1224 }
1221 }
1225 if (PyList_Append(chain, value)) {
1222 if (PyList_Append(chain, value)) {
1226 Py_DECREF(value);
1223 Py_DECREF(value);
1227 goto bail;
1224 goto bail;
1228 }
1225 }
1229 Py_DECREF(value);
1226 Py_DECREF(value);
1230
1227
1231 if (generaldelta) {
1228 if (generaldelta) {
1232 iterrev = baserev;
1229 iterrev = baserev;
1233 } else {
1230 } else {
1234 iterrev--;
1231 iterrev--;
1235 }
1232 }
1236
1233
1237 if (iterrev < 0) {
1234 if (iterrev < 0) {
1238 break;
1235 break;
1239 }
1236 }
1240
1237
1241 if (iterrev >= length) {
1238 if (iterrev >= length) {
1242 PyErr_SetString(PyExc_IndexError,
1239 PyErr_SetString(PyExc_IndexError,
1243 "revision outside index");
1240 "revision outside index");
1244 return NULL;
1241 return NULL;
1245 }
1242 }
1246
1243
1247 baserev = index_baserev(self, iterrev);
1244 baserev = index_baserev(self, iterrev);
1248
1245
1249 /* This should never happen. */
1246 /* This should never happen. */
1250 if (baserev <= -2) {
1247 if (baserev <= -2) {
1251 /* Error should be set by index_deref() */
1248 /* Error should be set by index_deref() */
1252 assert(PyErr_Occurred());
1249 assert(PyErr_Occurred());
1253 goto bail;
1250 goto bail;
1254 }
1251 }
1255 }
1252 }
1256
1253
1257 if (iterrev == stoprev) {
1254 if (iterrev == stoprev) {
1258 stopped = 1;
1255 stopped = 1;
1259 } else {
1256 } else {
1260 PyObject *value = PyInt_FromLong(iterrev);
1257 PyObject *value = PyInt_FromLong(iterrev);
1261 if (value == NULL) {
1258 if (value == NULL) {
1262 goto bail;
1259 goto bail;
1263 }
1260 }
1264 if (PyList_Append(chain, value)) {
1261 if (PyList_Append(chain, value)) {
1265 Py_DECREF(value);
1262 Py_DECREF(value);
1266 goto bail;
1263 goto bail;
1267 }
1264 }
1268 Py_DECREF(value);
1265 Py_DECREF(value);
1269
1266
1270 stopped = 0;
1267 stopped = 0;
1271 }
1268 }
1272
1269
1273 if (PyList_Reverse(chain)) {
1270 if (PyList_Reverse(chain)) {
1274 goto bail;
1271 goto bail;
1275 }
1272 }
1276
1273
1277 result = Py_BuildValue("OO", chain, stopped ? Py_True : Py_False);
1274 result = Py_BuildValue("OO", chain, stopped ? Py_True : Py_False);
1278 Py_DECREF(chain);
1275 Py_DECREF(chain);
1279 return result;
1276 return result;
1280
1277
1281 bail:
1278 bail:
1282 Py_DECREF(chain);
1279 Py_DECREF(chain);
1283 return NULL;
1280 return NULL;
1284 }
1281 }
1285
1282
1286 static inline int64_t
1283 static inline int64_t
1287 index_segment_span(indexObject *self, Py_ssize_t start_rev, Py_ssize_t end_rev)
1284 index_segment_span(indexObject *self, Py_ssize_t start_rev, Py_ssize_t end_rev)
1288 {
1285 {
1289 int64_t start_offset;
1286 int64_t start_offset;
1290 int64_t end_offset;
1287 int64_t end_offset;
1291 int end_size;
1288 int end_size;
1292 start_offset = index_get_start(self, start_rev);
1289 start_offset = index_get_start(self, start_rev);
1293 if (start_offset < 0) {
1290 if (start_offset < 0) {
1294 return -1;
1291 return -1;
1295 }
1292 }
1296 end_offset = index_get_start(self, end_rev);
1293 end_offset = index_get_start(self, end_rev);
1297 if (end_offset < 0) {
1294 if (end_offset < 0) {
1298 return -1;
1295 return -1;
1299 }
1296 }
1300 end_size = index_get_length(self, end_rev);
1297 end_size = index_get_length(self, end_rev);
1301 if (end_size < 0) {
1298 if (end_size < 0) {
1302 return -1;
1299 return -1;
1303 }
1300 }
1304 if (end_offset < start_offset) {
1301 if (end_offset < start_offset) {
1305 PyErr_Format(PyExc_ValueError,
1302 PyErr_Format(PyExc_ValueError,
1306 "corrupted revlog index: inconsistent offset "
1303 "corrupted revlog index: inconsistent offset "
1307 "between revisions (%zd) and (%zd)",
1304 "between revisions (%zd) and (%zd)",
1308 start_rev, end_rev);
1305 start_rev, end_rev);
1309 return -1;
1306 return -1;
1310 }
1307 }
1311 return (end_offset - start_offset) + (int64_t)end_size;
1308 return (end_offset - start_offset) + (int64_t)end_size;
1312 }
1309 }
1313
1310
1314 /* returns endidx so that revs[startidx:endidx] has no empty trailing revs */
1311 /* returns endidx so that revs[startidx:endidx] has no empty trailing revs */
1315 static Py_ssize_t trim_endidx(indexObject *self, const Py_ssize_t *revs,
1312 static Py_ssize_t trim_endidx(indexObject *self, const Py_ssize_t *revs,
1316 Py_ssize_t startidx, Py_ssize_t endidx)
1313 Py_ssize_t startidx, Py_ssize_t endidx)
1317 {
1314 {
1318 int length;
1315 int length;
1319 while (endidx > 1 && endidx > startidx) {
1316 while (endidx > 1 && endidx > startidx) {
1320 length = index_get_length(self, revs[endidx - 1]);
1317 length = index_get_length(self, revs[endidx - 1]);
1321 if (length < 0) {
1318 if (length < 0) {
1322 return -1;
1319 return -1;
1323 }
1320 }
1324 if (length != 0) {
1321 if (length != 0) {
1325 break;
1322 break;
1326 }
1323 }
1327 endidx -= 1;
1324 endidx -= 1;
1328 }
1325 }
1329 return endidx;
1326 return endidx;
1330 }
1327 }
1331
1328
1332 struct Gap {
1329 struct Gap {
1333 int64_t size;
1330 int64_t size;
1334 Py_ssize_t idx;
1331 Py_ssize_t idx;
1335 };
1332 };
1336
1333
1337 static int gap_compare(const void *left, const void *right)
1334 static int gap_compare(const void *left, const void *right)
1338 {
1335 {
1339 const struct Gap *l_left = ((const struct Gap *)left);
1336 const struct Gap *l_left = ((const struct Gap *)left);
1340 const struct Gap *l_right = ((const struct Gap *)right);
1337 const struct Gap *l_right = ((const struct Gap *)right);
1341 if (l_left->size < l_right->size) {
1338 if (l_left->size < l_right->size) {
1342 return -1;
1339 return -1;
1343 } else if (l_left->size > l_right->size) {
1340 } else if (l_left->size > l_right->size) {
1344 return 1;
1341 return 1;
1345 }
1342 }
1346 return 0;
1343 return 0;
1347 }
1344 }
1348 static int Py_ssize_t_compare(const void *left, const void *right)
1345 static int Py_ssize_t_compare(const void *left, const void *right)
1349 {
1346 {
1350 const Py_ssize_t l_left = *(const Py_ssize_t *)left;
1347 const Py_ssize_t l_left = *(const Py_ssize_t *)left;
1351 const Py_ssize_t l_right = *(const Py_ssize_t *)right;
1348 const Py_ssize_t l_right = *(const Py_ssize_t *)right;
1352 if (l_left < l_right) {
1349 if (l_left < l_right) {
1353 return -1;
1350 return -1;
1354 } else if (l_left > l_right) {
1351 } else if (l_left > l_right) {
1355 return 1;
1352 return 1;
1356 }
1353 }
1357 return 0;
1354 return 0;
1358 }
1355 }
1359
1356
1360 static PyObject *index_slicechunktodensity(indexObject *self, PyObject *args)
1357 static PyObject *index_slicechunktodensity(indexObject *self, PyObject *args)
1361 {
1358 {
1362 /* method arguments */
1359 /* method arguments */
1363 PyObject *list_revs = NULL; /* revisions in the chain */
1360 PyObject *list_revs = NULL; /* revisions in the chain */
1364 double targetdensity = 0; /* min density to achieve */
1361 double targetdensity = 0; /* min density to achieve */
1365 Py_ssize_t mingapsize = 0; /* threshold to ignore gaps */
1362 Py_ssize_t mingapsize = 0; /* threshold to ignore gaps */
1366
1363
1367 /* other core variables */
1364 /* other core variables */
1368 Py_ssize_t idxlen = index_length(self);
1365 Py_ssize_t idxlen = index_length(self);
1369 Py_ssize_t i; /* used for various iteration */
1366 Py_ssize_t i; /* used for various iteration */
1370 PyObject *result = NULL; /* the final return of the function */
1367 PyObject *result = NULL; /* the final return of the function */
1371
1368
1372 /* generic information about the delta chain being slice */
1369 /* generic information about the delta chain being slice */
1373 Py_ssize_t num_revs = 0; /* size of the full delta chain */
1370 Py_ssize_t num_revs = 0; /* size of the full delta chain */
1374 Py_ssize_t *revs = NULL; /* native array of revision in the chain */
1371 Py_ssize_t *revs = NULL; /* native array of revision in the chain */
1375 int64_t chainpayload = 0; /* sum of all delta in the chain */
1372 int64_t chainpayload = 0; /* sum of all delta in the chain */
1376 int64_t deltachainspan = 0; /* distance from first byte to last byte */
1373 int64_t deltachainspan = 0; /* distance from first byte to last byte */
1377
1374
1378 /* variable used for slicing the delta chain */
1375 /* variable used for slicing the delta chain */
1379 int64_t readdata = 0; /* amount of data currently planned to be read */
1376 int64_t readdata = 0; /* amount of data currently planned to be read */
1380 double density = 0; /* ration of payload data compared to read ones */
1377 double density = 0; /* ration of payload data compared to read ones */
1381 int64_t previous_end;
1378 int64_t previous_end;
1382 struct Gap *gaps = NULL; /* array of notable gap in the chain */
1379 struct Gap *gaps = NULL; /* array of notable gap in the chain */
1383 Py_ssize_t num_gaps =
1380 Py_ssize_t num_gaps =
1384 0; /* total number of notable gap recorded so far */
1381 0; /* total number of notable gap recorded so far */
1385 Py_ssize_t *selected_indices = NULL; /* indices of gap skipped over */
1382 Py_ssize_t *selected_indices = NULL; /* indices of gap skipped over */
1386 Py_ssize_t num_selected = 0; /* number of gaps skipped */
1383 Py_ssize_t num_selected = 0; /* number of gaps skipped */
1387 PyObject *chunk = NULL; /* individual slice */
1384 PyObject *chunk = NULL; /* individual slice */
1388 PyObject *allchunks = NULL; /* all slices */
1385 PyObject *allchunks = NULL; /* all slices */
1389 Py_ssize_t previdx;
1386 Py_ssize_t previdx;
1390
1387
1391 /* parsing argument */
1388 /* parsing argument */
1392 if (!PyArg_ParseTuple(args, "O!dn", &PyList_Type, &list_revs,
1389 if (!PyArg_ParseTuple(args, "O!dn", &PyList_Type, &list_revs,
1393 &targetdensity, &mingapsize)) {
1390 &targetdensity, &mingapsize)) {
1394 goto bail;
1391 goto bail;
1395 }
1392 }
1396
1393
1397 /* If the delta chain contains a single element, we do not need slicing
1394 /* If the delta chain contains a single element, we do not need slicing
1398 */
1395 */
1399 num_revs = PyList_GET_SIZE(list_revs);
1396 num_revs = PyList_GET_SIZE(list_revs);
1400 if (num_revs <= 1) {
1397 if (num_revs <= 1) {
1401 result = PyTuple_Pack(1, list_revs);
1398 result = PyTuple_Pack(1, list_revs);
1402 goto done;
1399 goto done;
1403 }
1400 }
1404
1401
1405 /* Turn the python list into a native integer array (for efficiency) */
1402 /* Turn the python list into a native integer array (for efficiency) */
1406 revs = (Py_ssize_t *)calloc(num_revs, sizeof(Py_ssize_t));
1403 revs = (Py_ssize_t *)calloc(num_revs, sizeof(Py_ssize_t));
1407 if (revs == NULL) {
1404 if (revs == NULL) {
1408 PyErr_NoMemory();
1405 PyErr_NoMemory();
1409 goto bail;
1406 goto bail;
1410 }
1407 }
1411 for (i = 0; i < num_revs; i++) {
1408 for (i = 0; i < num_revs; i++) {
1412 Py_ssize_t revnum = PyInt_AsLong(PyList_GET_ITEM(list_revs, i));
1409 Py_ssize_t revnum = PyInt_AsLong(PyList_GET_ITEM(list_revs, i));
1413 if (revnum == -1 && PyErr_Occurred()) {
1410 if (revnum == -1 && PyErr_Occurred()) {
1414 goto bail;
1411 goto bail;
1415 }
1412 }
1416 if (revnum < nullrev || revnum >= idxlen) {
1413 if (revnum < nullrev || revnum >= idxlen) {
1417 PyErr_Format(PyExc_IndexError,
1414 PyErr_Format(PyExc_IndexError,
1418 "index out of range: %zd", revnum);
1415 "index out of range: %zd", revnum);
1419 goto bail;
1416 goto bail;
1420 }
1417 }
1421 revs[i] = revnum;
1418 revs[i] = revnum;
1422 }
1419 }
1423
1420
1424 /* Compute and check various property of the unsliced delta chain */
1421 /* Compute and check various property of the unsliced delta chain */
1425 deltachainspan = index_segment_span(self, revs[0], revs[num_revs - 1]);
1422 deltachainspan = index_segment_span(self, revs[0], revs[num_revs - 1]);
1426 if (deltachainspan < 0) {
1423 if (deltachainspan < 0) {
1427 goto bail;
1424 goto bail;
1428 }
1425 }
1429
1426
1430 if (deltachainspan <= mingapsize) {
1427 if (deltachainspan <= mingapsize) {
1431 result = PyTuple_Pack(1, list_revs);
1428 result = PyTuple_Pack(1, list_revs);
1432 goto done;
1429 goto done;
1433 }
1430 }
1434 chainpayload = 0;
1431 chainpayload = 0;
1435 for (i = 0; i < num_revs; i++) {
1432 for (i = 0; i < num_revs; i++) {
1436 int tmp = index_get_length(self, revs[i]);
1433 int tmp = index_get_length(self, revs[i]);
1437 if (tmp < 0) {
1434 if (tmp < 0) {
1438 goto bail;
1435 goto bail;
1439 }
1436 }
1440 chainpayload += tmp;
1437 chainpayload += tmp;
1441 }
1438 }
1442
1439
1443 readdata = deltachainspan;
1440 readdata = deltachainspan;
1444 density = 1.0;
1441 density = 1.0;
1445
1442
1446 if (0 < deltachainspan) {
1443 if (0 < deltachainspan) {
1447 density = (double)chainpayload / (double)deltachainspan;
1444 density = (double)chainpayload / (double)deltachainspan;
1448 }
1445 }
1449
1446
1450 if (density >= targetdensity) {
1447 if (density >= targetdensity) {
1451 result = PyTuple_Pack(1, list_revs);
1448 result = PyTuple_Pack(1, list_revs);
1452 goto done;
1449 goto done;
1453 }
1450 }
1454
1451
1455 /* if chain is too sparse, look for relevant gaps */
1452 /* if chain is too sparse, look for relevant gaps */
1456 gaps = (struct Gap *)calloc(num_revs, sizeof(struct Gap));
1453 gaps = (struct Gap *)calloc(num_revs, sizeof(struct Gap));
1457 if (gaps == NULL) {
1454 if (gaps == NULL) {
1458 PyErr_NoMemory();
1455 PyErr_NoMemory();
1459 goto bail;
1456 goto bail;
1460 }
1457 }
1461
1458
1462 previous_end = -1;
1459 previous_end = -1;
1463 for (i = 0; i < num_revs; i++) {
1460 for (i = 0; i < num_revs; i++) {
1464 int64_t revstart;
1461 int64_t revstart;
1465 int revsize;
1462 int revsize;
1466 revstart = index_get_start(self, revs[i]);
1463 revstart = index_get_start(self, revs[i]);
1467 if (revstart < 0) {
1464 if (revstart < 0) {
1468 goto bail;
1465 goto bail;
1469 };
1466 };
1470 revsize = index_get_length(self, revs[i]);
1467 revsize = index_get_length(self, revs[i]);
1471 if (revsize < 0) {
1468 if (revsize < 0) {
1472 goto bail;
1469 goto bail;
1473 };
1470 };
1474 if (revsize == 0) {
1471 if (revsize == 0) {
1475 continue;
1472 continue;
1476 }
1473 }
1477 if (previous_end >= 0) {
1474 if (previous_end >= 0) {
1478 int64_t gapsize = revstart - previous_end;
1475 int64_t gapsize = revstart - previous_end;
1479 if (gapsize > mingapsize) {
1476 if (gapsize > mingapsize) {
1480 gaps[num_gaps].size = gapsize;
1477 gaps[num_gaps].size = gapsize;
1481 gaps[num_gaps].idx = i;
1478 gaps[num_gaps].idx = i;
1482 num_gaps += 1;
1479 num_gaps += 1;
1483 }
1480 }
1484 }
1481 }
1485 previous_end = revstart + revsize;
1482 previous_end = revstart + revsize;
1486 }
1483 }
1487 if (num_gaps == 0) {
1484 if (num_gaps == 0) {
1488 result = PyTuple_Pack(1, list_revs);
1485 result = PyTuple_Pack(1, list_revs);
1489 goto done;
1486 goto done;
1490 }
1487 }
1491 qsort(gaps, num_gaps, sizeof(struct Gap), &gap_compare);
1488 qsort(gaps, num_gaps, sizeof(struct Gap), &gap_compare);
1492
1489
1493 /* Slice the largest gap first, they improve the density the most */
1490 /* Slice the largest gap first, they improve the density the most */
1494 selected_indices =
1491 selected_indices =
1495 (Py_ssize_t *)malloc((num_gaps + 1) * sizeof(Py_ssize_t));
1492 (Py_ssize_t *)malloc((num_gaps + 1) * sizeof(Py_ssize_t));
1496 if (selected_indices == NULL) {
1493 if (selected_indices == NULL) {
1497 PyErr_NoMemory();
1494 PyErr_NoMemory();
1498 goto bail;
1495 goto bail;
1499 }
1496 }
1500
1497
1501 for (i = num_gaps - 1; i >= 0; i--) {
1498 for (i = num_gaps - 1; i >= 0; i--) {
1502 selected_indices[num_selected] = gaps[i].idx;
1499 selected_indices[num_selected] = gaps[i].idx;
1503 readdata -= gaps[i].size;
1500 readdata -= gaps[i].size;
1504 num_selected += 1;
1501 num_selected += 1;
1505 if (readdata <= 0) {
1502 if (readdata <= 0) {
1506 density = 1.0;
1503 density = 1.0;
1507 } else {
1504 } else {
1508 density = (double)chainpayload / (double)readdata;
1505 density = (double)chainpayload / (double)readdata;
1509 }
1506 }
1510 if (density >= targetdensity) {
1507 if (density >= targetdensity) {
1511 break;
1508 break;
1512 }
1509 }
1513 }
1510 }
1514 qsort(selected_indices, num_selected, sizeof(Py_ssize_t),
1511 qsort(selected_indices, num_selected, sizeof(Py_ssize_t),
1515 &Py_ssize_t_compare);
1512 &Py_ssize_t_compare);
1516
1513
1517 /* create the resulting slice */
1514 /* create the resulting slice */
1518 allchunks = PyList_New(0);
1515 allchunks = PyList_New(0);
1519 if (allchunks == NULL) {
1516 if (allchunks == NULL) {
1520 goto bail;
1517 goto bail;
1521 }
1518 }
1522 previdx = 0;
1519 previdx = 0;
1523 selected_indices[num_selected] = num_revs;
1520 selected_indices[num_selected] = num_revs;
1524 for (i = 0; i <= num_selected; i++) {
1521 for (i = 0; i <= num_selected; i++) {
1525 Py_ssize_t idx = selected_indices[i];
1522 Py_ssize_t idx = selected_indices[i];
1526 Py_ssize_t endidx = trim_endidx(self, revs, previdx, idx);
1523 Py_ssize_t endidx = trim_endidx(self, revs, previdx, idx);
1527 if (endidx < 0) {
1524 if (endidx < 0) {
1528 goto bail;
1525 goto bail;
1529 }
1526 }
1530 if (previdx < endidx) {
1527 if (previdx < endidx) {
1531 chunk = PyList_GetSlice(list_revs, previdx, endidx);
1528 chunk = PyList_GetSlice(list_revs, previdx, endidx);
1532 if (chunk == NULL) {
1529 if (chunk == NULL) {
1533 goto bail;
1530 goto bail;
1534 }
1531 }
1535 if (PyList_Append(allchunks, chunk) == -1) {
1532 if (PyList_Append(allchunks, chunk) == -1) {
1536 goto bail;
1533 goto bail;
1537 }
1534 }
1538 Py_DECREF(chunk);
1535 Py_DECREF(chunk);
1539 chunk = NULL;
1536 chunk = NULL;
1540 }
1537 }
1541 previdx = idx;
1538 previdx = idx;
1542 }
1539 }
1543 result = allchunks;
1540 result = allchunks;
1544 goto done;
1541 goto done;
1545
1542
1546 bail:
1543 bail:
1547 Py_XDECREF(allchunks);
1544 Py_XDECREF(allchunks);
1548 Py_XDECREF(chunk);
1545 Py_XDECREF(chunk);
1549 done:
1546 done:
1550 free(revs);
1547 free(revs);
1551 free(gaps);
1548 free(gaps);
1552 free(selected_indices);
1549 free(selected_indices);
1553 return result;
1550 return result;
1554 }
1551 }
1555
1552
1556 static inline int nt_level(const char *node, Py_ssize_t level)
1553 static inline int nt_level(const char *node, Py_ssize_t level)
1557 {
1554 {
1558 int v = node[level >> 1];
1555 int v = node[level >> 1];
1559 if (!(level & 1))
1556 if (!(level & 1))
1560 v >>= 4;
1557 v >>= 4;
1561 return v & 0xf;
1558 return v & 0xf;
1562 }
1559 }
1563
1560
1564 /*
1561 /*
1565 * Return values:
1562 * Return values:
1566 *
1563 *
1567 * -4: match is ambiguous (multiple candidates)
1564 * -4: match is ambiguous (multiple candidates)
1568 * -2: not found
1565 * -2: not found
1569 * rest: valid rev
1566 * rest: valid rev
1570 */
1567 */
1571 static int nt_find(nodetree *self, const char *node, Py_ssize_t nodelen,
1568 static int nt_find(nodetree *self, const char *node, Py_ssize_t nodelen,
1572 int hex)
1569 int hex)
1573 {
1570 {
1574 int (*getnybble)(const char *, Py_ssize_t) = hex ? hexdigit : nt_level;
1571 int (*getnybble)(const char *, Py_ssize_t) = hex ? hexdigit : nt_level;
1575 int level, maxlevel, off;
1572 int level, maxlevel, off;
1576
1573
1577 if (nodelen == 20 && node[0] == '\0' && memcmp(node, nullid, 20) == 0)
1574 if (nodelen == 20 && node[0] == '\0' && memcmp(node, nullid, 20) == 0)
1578 return -1;
1575 return -1;
1579
1576
1580 if (hex)
1577 if (hex)
1581 maxlevel = nodelen > 40 ? 40 : (int)nodelen;
1578 maxlevel = nodelen > 40 ? 40 : (int)nodelen;
1582 else
1579 else
1583 maxlevel = nodelen > 20 ? 40 : ((int)nodelen * 2);
1580 maxlevel = nodelen > 20 ? 40 : ((int)nodelen * 2);
1584
1581
1585 for (level = off = 0; level < maxlevel; level++) {
1582 for (level = off = 0; level < maxlevel; level++) {
1586 int k = getnybble(node, level);
1583 int k = getnybble(node, level);
1587 nodetreenode *n = &self->nodes[off];
1584 nodetreenode *n = &self->nodes[off];
1588 int v = n->children[k];
1585 int v = n->children[k];
1589
1586
1590 if (v < 0) {
1587 if (v < 0) {
1591 const char *n;
1588 const char *n;
1592 Py_ssize_t i;
1589 Py_ssize_t i;
1593
1590
1594 v = -(v + 2);
1591 v = -(v + 2);
1595 n = index_node(self->index, v);
1592 n = index_node(self->index, v);
1596 if (n == NULL)
1593 if (n == NULL)
1597 return -2;
1594 return -2;
1598 for (i = level; i < maxlevel; i++)
1595 for (i = level; i < maxlevel; i++)
1599 if (getnybble(node, i) != nt_level(n, i))
1596 if (getnybble(node, i) != nt_level(n, i))
1600 return -2;
1597 return -2;
1601 return v;
1598 return v;
1602 }
1599 }
1603 if (v == 0)
1600 if (v == 0)
1604 return -2;
1601 return -2;
1605 off = v;
1602 off = v;
1606 }
1603 }
1607 /* multiple matches against an ambiguous prefix */
1604 /* multiple matches against an ambiguous prefix */
1608 return -4;
1605 return -4;
1609 }
1606 }
1610
1607
1611 static int nt_new(nodetree *self)
1608 static int nt_new(nodetree *self)
1612 {
1609 {
1613 if (self->length == self->capacity) {
1610 if (self->length == self->capacity) {
1614 unsigned newcapacity;
1611 unsigned newcapacity;
1615 nodetreenode *newnodes;
1612 nodetreenode *newnodes;
1616 newcapacity = self->capacity * 2;
1613 newcapacity = self->capacity * 2;
1617 if (newcapacity >= INT_MAX / sizeof(nodetreenode)) {
1614 if (newcapacity >= INT_MAX / sizeof(nodetreenode)) {
1618 PyErr_SetString(PyExc_MemoryError,
1615 PyErr_SetString(PyExc_MemoryError,
1619 "overflow in nt_new");
1616 "overflow in nt_new");
1620 return -1;
1617 return -1;
1621 }
1618 }
1622 newnodes =
1619 newnodes =
1623 realloc(self->nodes, newcapacity * sizeof(nodetreenode));
1620 realloc(self->nodes, newcapacity * sizeof(nodetreenode));
1624 if (newnodes == NULL) {
1621 if (newnodes == NULL) {
1625 PyErr_SetString(PyExc_MemoryError, "out of memory");
1622 PyErr_SetString(PyExc_MemoryError, "out of memory");
1626 return -1;
1623 return -1;
1627 }
1624 }
1628 self->capacity = newcapacity;
1625 self->capacity = newcapacity;
1629 self->nodes = newnodes;
1626 self->nodes = newnodes;
1630 memset(&self->nodes[self->length], 0,
1627 memset(&self->nodes[self->length], 0,
1631 sizeof(nodetreenode) * (self->capacity - self->length));
1628 sizeof(nodetreenode) * (self->capacity - self->length));
1632 }
1629 }
1633 return self->length++;
1630 return self->length++;
1634 }
1631 }
1635
1632
1636 static int nt_insert(nodetree *self, const char *node, int rev)
1633 static int nt_insert(nodetree *self, const char *node, int rev)
1637 {
1634 {
1638 int level = 0;
1635 int level = 0;
1639 int off = 0;
1636 int off = 0;
1640
1637
1641 while (level < 40) {
1638 while (level < 40) {
1642 int k = nt_level(node, level);
1639 int k = nt_level(node, level);
1643 nodetreenode *n;
1640 nodetreenode *n;
1644 int v;
1641 int v;
1645
1642
1646 n = &self->nodes[off];
1643 n = &self->nodes[off];
1647 v = n->children[k];
1644 v = n->children[k];
1648
1645
1649 if (v == 0) {
1646 if (v == 0) {
1650 n->children[k] = -rev - 2;
1647 n->children[k] = -rev - 2;
1651 return 0;
1648 return 0;
1652 }
1649 }
1653 if (v < 0) {
1650 if (v < 0) {
1654 const char *oldnode =
1651 const char *oldnode =
1655 index_node_existing(self->index, -(v + 2));
1652 index_node_existing(self->index, -(v + 2));
1656 int noff;
1653 int noff;
1657
1654
1658 if (oldnode == NULL)
1655 if (oldnode == NULL)
1659 return -1;
1656 return -1;
1660 if (!memcmp(oldnode, node, 20)) {
1657 if (!memcmp(oldnode, node, 20)) {
1661 n->children[k] = -rev - 2;
1658 n->children[k] = -rev - 2;
1662 return 0;
1659 return 0;
1663 }
1660 }
1664 noff = nt_new(self);
1661 noff = nt_new(self);
1665 if (noff == -1)
1662 if (noff == -1)
1666 return -1;
1663 return -1;
1667 /* self->nodes may have been changed by realloc */
1664 /* self->nodes may have been changed by realloc */
1668 self->nodes[off].children[k] = noff;
1665 self->nodes[off].children[k] = noff;
1669 off = noff;
1666 off = noff;
1670 n = &self->nodes[off];
1667 n = &self->nodes[off];
1671 n->children[nt_level(oldnode, ++level)] = v;
1668 n->children[nt_level(oldnode, ++level)] = v;
1672 if (level > self->depth)
1669 if (level > self->depth)
1673 self->depth = level;
1670 self->depth = level;
1674 self->splits += 1;
1671 self->splits += 1;
1675 } else {
1672 } else {
1676 level += 1;
1673 level += 1;
1677 off = v;
1674 off = v;
1678 }
1675 }
1679 }
1676 }
1680
1677
1681 return -1;
1678 return -1;
1682 }
1679 }
1683
1680
1684 static PyObject *ntobj_insert(nodetreeObject *self, PyObject *args)
1681 static PyObject *ntobj_insert(nodetreeObject *self, PyObject *args)
1685 {
1682 {
1686 Py_ssize_t rev;
1683 Py_ssize_t rev;
1687 const char *node;
1684 const char *node;
1688 Py_ssize_t length;
1685 Py_ssize_t length;
1689 if (!PyArg_ParseTuple(args, "n", &rev))
1686 if (!PyArg_ParseTuple(args, "n", &rev))
1690 return NULL;
1687 return NULL;
1691 length = index_length(self->nt.index);
1688 length = index_length(self->nt.index);
1692 if (rev < 0 || rev >= length) {
1689 if (rev < 0 || rev >= length) {
1693 PyErr_SetString(PyExc_ValueError, "revlog index out of range");
1690 PyErr_SetString(PyExc_ValueError, "revlog index out of range");
1694 return NULL;
1691 return NULL;
1695 }
1692 }
1696 node = index_node_existing(self->nt.index, rev);
1693 node = index_node_existing(self->nt.index, rev);
1697 if (nt_insert(&self->nt, node, (int)rev) == -1)
1694 if (nt_insert(&self->nt, node, (int)rev) == -1)
1698 return NULL;
1695 return NULL;
1699 Py_RETURN_NONE;
1696 Py_RETURN_NONE;
1700 }
1697 }
1701
1698
1702 static int nt_delete_node(nodetree *self, const char *node)
1699 static int nt_delete_node(nodetree *self, const char *node)
1703 {
1700 {
1704 /* rev==-2 happens to get encoded as 0, which is interpreted as not set
1701 /* rev==-2 happens to get encoded as 0, which is interpreted as not set
1705 */
1702 */
1706 return nt_insert(self, node, -2);
1703 return nt_insert(self, node, -2);
1707 }
1704 }
1708
1705
1709 static int nt_init(nodetree *self, indexObject *index, unsigned capacity)
1706 static int nt_init(nodetree *self, indexObject *index, unsigned capacity)
1710 {
1707 {
1711 /* Initialize before overflow-checking to avoid nt_dealloc() crash. */
1708 /* Initialize before overflow-checking to avoid nt_dealloc() crash. */
1712 self->nodes = NULL;
1709 self->nodes = NULL;
1713
1710
1714 self->index = index;
1711 self->index = index;
1715 /* The input capacity is in terms of revisions, while the field is in
1712 /* The input capacity is in terms of revisions, while the field is in
1716 * terms of nodetree nodes. */
1713 * terms of nodetree nodes. */
1717 self->capacity = (capacity < 4 ? 4 : capacity / 2);
1714 self->capacity = (capacity < 4 ? 4 : capacity / 2);
1718 self->depth = 0;
1715 self->depth = 0;
1719 self->splits = 0;
1716 self->splits = 0;
1720 if ((size_t)self->capacity > INT_MAX / sizeof(nodetreenode)) {
1717 if ((size_t)self->capacity > INT_MAX / sizeof(nodetreenode)) {
1721 PyErr_SetString(PyExc_ValueError, "overflow in init_nt");
1718 PyErr_SetString(PyExc_ValueError, "overflow in init_nt");
1722 return -1;
1719 return -1;
1723 }
1720 }
1724 self->nodes = calloc(self->capacity, sizeof(nodetreenode));
1721 self->nodes = calloc(self->capacity, sizeof(nodetreenode));
1725 if (self->nodes == NULL) {
1722 if (self->nodes == NULL) {
1726 PyErr_NoMemory();
1723 PyErr_NoMemory();
1727 return -1;
1724 return -1;
1728 }
1725 }
1729 self->length = 1;
1726 self->length = 1;
1730 return 0;
1727 return 0;
1731 }
1728 }
1732
1729
1733 static int ntobj_init(nodetreeObject *self, PyObject *args)
1730 static int ntobj_init(nodetreeObject *self, PyObject *args)
1734 {
1731 {
1735 PyObject *index;
1732 PyObject *index;
1736 unsigned capacity;
1733 unsigned capacity;
1737 if (!PyArg_ParseTuple(args, "O!I", &HgRevlogIndex_Type, &index,
1734 if (!PyArg_ParseTuple(args, "O!I", &HgRevlogIndex_Type, &index,
1738 &capacity))
1735 &capacity))
1739 return -1;
1736 return -1;
1740 Py_INCREF(index);
1737 Py_INCREF(index);
1741 return nt_init(&self->nt, (indexObject *)index, capacity);
1738 return nt_init(&self->nt, (indexObject *)index, capacity);
1742 }
1739 }
1743
1740
1744 static int nt_partialmatch(nodetree *self, const char *node, Py_ssize_t nodelen)
1741 static int nt_partialmatch(nodetree *self, const char *node, Py_ssize_t nodelen)
1745 {
1742 {
1746 return nt_find(self, node, nodelen, 1);
1743 return nt_find(self, node, nodelen, 1);
1747 }
1744 }
1748
1745
1749 /*
1746 /*
1750 * Find the length of the shortest unique prefix of node.
1747 * Find the length of the shortest unique prefix of node.
1751 *
1748 *
1752 * Return values:
1749 * Return values:
1753 *
1750 *
1754 * -3: error (exception set)
1751 * -3: error (exception set)
1755 * -2: not found (no exception set)
1752 * -2: not found (no exception set)
1756 * rest: length of shortest prefix
1753 * rest: length of shortest prefix
1757 */
1754 */
1758 static int nt_shortest(nodetree *self, const char *node)
1755 static int nt_shortest(nodetree *self, const char *node)
1759 {
1756 {
1760 int level, off;
1757 int level, off;
1761
1758
1762 for (level = off = 0; level < 40; level++) {
1759 for (level = off = 0; level < 40; level++) {
1763 int k, v;
1760 int k, v;
1764 nodetreenode *n = &self->nodes[off];
1761 nodetreenode *n = &self->nodes[off];
1765 k = nt_level(node, level);
1762 k = nt_level(node, level);
1766 v = n->children[k];
1763 v = n->children[k];
1767 if (v < 0) {
1764 if (v < 0) {
1768 const char *n;
1765 const char *n;
1769 v = -(v + 2);
1766 v = -(v + 2);
1770 n = index_node_existing(self->index, v);
1767 n = index_node_existing(self->index, v);
1771 if (n == NULL)
1768 if (n == NULL)
1772 return -3;
1769 return -3;
1773 if (memcmp(node, n, 20) != 0)
1770 if (memcmp(node, n, 20) != 0)
1774 /*
1771 /*
1775 * Found a unique prefix, but it wasn't for the
1772 * Found a unique prefix, but it wasn't for the
1776 * requested node (i.e the requested node does
1773 * requested node (i.e the requested node does
1777 * not exist).
1774 * not exist).
1778 */
1775 */
1779 return -2;
1776 return -2;
1780 return level + 1;
1777 return level + 1;
1781 }
1778 }
1782 if (v == 0)
1779 if (v == 0)
1783 return -2;
1780 return -2;
1784 off = v;
1781 off = v;
1785 }
1782 }
1786 /*
1783 /*
1787 * The node was still not unique after 40 hex digits, so this won't
1784 * The node was still not unique after 40 hex digits, so this won't
1788 * happen. Also, if we get here, then there's a programming error in
1785 * happen. Also, if we get here, then there's a programming error in
1789 * this file that made us insert a node longer than 40 hex digits.
1786 * this file that made us insert a node longer than 40 hex digits.
1790 */
1787 */
1791 PyErr_SetString(PyExc_Exception, "broken node tree");
1788 PyErr_SetString(PyExc_Exception, "broken node tree");
1792 return -3;
1789 return -3;
1793 }
1790 }
1794
1791
1795 static PyObject *ntobj_shortest(nodetreeObject *self, PyObject *args)
1792 static PyObject *ntobj_shortest(nodetreeObject *self, PyObject *args)
1796 {
1793 {
1797 PyObject *val;
1794 PyObject *val;
1798 char *node;
1795 char *node;
1799 int length;
1796 int length;
1800
1797
1801 if (!PyArg_ParseTuple(args, "O", &val))
1798 if (!PyArg_ParseTuple(args, "O", &val))
1802 return NULL;
1799 return NULL;
1803 if (node_check(val, &node) == -1)
1800 if (node_check(val, &node) == -1)
1804 return NULL;
1801 return NULL;
1805
1802
1806 length = nt_shortest(&self->nt, node);
1803 length = nt_shortest(&self->nt, node);
1807 if (length == -3)
1804 if (length == -3)
1808 return NULL;
1805 return NULL;
1809 if (length == -2) {
1806 if (length == -2) {
1810 raise_revlog_error();
1807 raise_revlog_error();
1811 return NULL;
1808 return NULL;
1812 }
1809 }
1813 return PyInt_FromLong(length);
1810 return PyInt_FromLong(length);
1814 }
1811 }
1815
1812
1816 static void nt_dealloc(nodetree *self)
1813 static void nt_dealloc(nodetree *self)
1817 {
1814 {
1818 free(self->nodes);
1815 free(self->nodes);
1819 self->nodes = NULL;
1816 self->nodes = NULL;
1820 }
1817 }
1821
1818
1822 static void ntobj_dealloc(nodetreeObject *self)
1819 static void ntobj_dealloc(nodetreeObject *self)
1823 {
1820 {
1824 Py_XDECREF(self->nt.index);
1821 Py_XDECREF(self->nt.index);
1825 nt_dealloc(&self->nt);
1822 nt_dealloc(&self->nt);
1826 PyObject_Del(self);
1823 PyObject_Del(self);
1827 }
1824 }
1828
1825
1829 static PyMethodDef ntobj_methods[] = {
1826 static PyMethodDef ntobj_methods[] = {
1830 {"insert", (PyCFunction)ntobj_insert, METH_VARARGS,
1827 {"insert", (PyCFunction)ntobj_insert, METH_VARARGS,
1831 "insert an index entry"},
1828 "insert an index entry"},
1832 {"shortest", (PyCFunction)ntobj_shortest, METH_VARARGS,
1829 {"shortest", (PyCFunction)ntobj_shortest, METH_VARARGS,
1833 "find length of shortest hex nodeid of a binary ID"},
1830 "find length of shortest hex nodeid of a binary ID"},
1834 {NULL} /* Sentinel */
1831 {NULL} /* Sentinel */
1835 };
1832 };
1836
1833
1837 static PyTypeObject nodetreeType = {
1834 static PyTypeObject nodetreeType = {
1838 PyVarObject_HEAD_INIT(NULL, 0) /* header */
1835 PyVarObject_HEAD_INIT(NULL, 0) /* header */
1839 "parsers.nodetree", /* tp_name */
1836 "parsers.nodetree", /* tp_name */
1840 sizeof(nodetreeObject), /* tp_basicsize */
1837 sizeof(nodetreeObject), /* tp_basicsize */
1841 0, /* tp_itemsize */
1838 0, /* tp_itemsize */
1842 (destructor)ntobj_dealloc, /* tp_dealloc */
1839 (destructor)ntobj_dealloc, /* tp_dealloc */
1843 0, /* tp_print */
1840 0, /* tp_print */
1844 0, /* tp_getattr */
1841 0, /* tp_getattr */
1845 0, /* tp_setattr */
1842 0, /* tp_setattr */
1846 0, /* tp_compare */
1843 0, /* tp_compare */
1847 0, /* tp_repr */
1844 0, /* tp_repr */
1848 0, /* tp_as_number */
1845 0, /* tp_as_number */
1849 0, /* tp_as_sequence */
1846 0, /* tp_as_sequence */
1850 0, /* tp_as_mapping */
1847 0, /* tp_as_mapping */
1851 0, /* tp_hash */
1848 0, /* tp_hash */
1852 0, /* tp_call */
1849 0, /* tp_call */
1853 0, /* tp_str */
1850 0, /* tp_str */
1854 0, /* tp_getattro */
1851 0, /* tp_getattro */
1855 0, /* tp_setattro */
1852 0, /* tp_setattro */
1856 0, /* tp_as_buffer */
1853 0, /* tp_as_buffer */
1857 Py_TPFLAGS_DEFAULT, /* tp_flags */
1854 Py_TPFLAGS_DEFAULT, /* tp_flags */
1858 "nodetree", /* tp_doc */
1855 "nodetree", /* tp_doc */
1859 0, /* tp_traverse */
1856 0, /* tp_traverse */
1860 0, /* tp_clear */
1857 0, /* tp_clear */
1861 0, /* tp_richcompare */
1858 0, /* tp_richcompare */
1862 0, /* tp_weaklistoffset */
1859 0, /* tp_weaklistoffset */
1863 0, /* tp_iter */
1860 0, /* tp_iter */
1864 0, /* tp_iternext */
1861 0, /* tp_iternext */
1865 ntobj_methods, /* tp_methods */
1862 ntobj_methods, /* tp_methods */
1866 0, /* tp_members */
1863 0, /* tp_members */
1867 0, /* tp_getset */
1864 0, /* tp_getset */
1868 0, /* tp_base */
1865 0, /* tp_base */
1869 0, /* tp_dict */
1866 0, /* tp_dict */
1870 0, /* tp_descr_get */
1867 0, /* tp_descr_get */
1871 0, /* tp_descr_set */
1868 0, /* tp_descr_set */
1872 0, /* tp_dictoffset */
1869 0, /* tp_dictoffset */
1873 (initproc)ntobj_init, /* tp_init */
1870 (initproc)ntobj_init, /* tp_init */
1874 0, /* tp_alloc */
1871 0, /* tp_alloc */
1875 };
1872 };
1876
1873
1877 static int index_init_nt(indexObject *self)
1874 static int index_init_nt(indexObject *self)
1878 {
1875 {
1879 if (!self->ntinitialized) {
1876 if (!self->ntinitialized) {
1880 if (nt_init(&self->nt, self, (int)self->raw_length) == -1) {
1877 if (nt_init(&self->nt, self, (int)self->raw_length) == -1) {
1881 nt_dealloc(&self->nt);
1878 nt_dealloc(&self->nt);
1882 return -1;
1879 return -1;
1883 }
1880 }
1884 if (nt_insert(&self->nt, nullid, -1) == -1) {
1881 if (nt_insert(&self->nt, nullid, -1) == -1) {
1885 nt_dealloc(&self->nt);
1882 nt_dealloc(&self->nt);
1886 return -1;
1883 return -1;
1887 }
1884 }
1888 self->ntinitialized = 1;
1885 self->ntinitialized = 1;
1889 self->ntrev = (int)index_length(self);
1886 self->ntrev = (int)index_length(self);
1890 self->ntlookups = 1;
1887 self->ntlookups = 1;
1891 self->ntmisses = 0;
1888 self->ntmisses = 0;
1892 }
1889 }
1893 return 0;
1890 return 0;
1894 }
1891 }
1895
1892
1896 /*
1893 /*
1897 * Return values:
1894 * Return values:
1898 *
1895 *
1899 * -3: error (exception set)
1896 * -3: error (exception set)
1900 * -2: not found (no exception set)
1897 * -2: not found (no exception set)
1901 * rest: valid rev
1898 * rest: valid rev
1902 */
1899 */
1903 static int index_find_node(indexObject *self, const char *node,
1900 static int index_find_node(indexObject *self, const char *node,
1904 Py_ssize_t nodelen)
1901 Py_ssize_t nodelen)
1905 {
1902 {
1906 int rev;
1903 int rev;
1907
1904
1908 if (index_init_nt(self) == -1)
1905 if (index_init_nt(self) == -1)
1909 return -3;
1906 return -3;
1910
1907
1911 self->ntlookups++;
1908 self->ntlookups++;
1912 rev = nt_find(&self->nt, node, nodelen, 0);
1909 rev = nt_find(&self->nt, node, nodelen, 0);
1913 if (rev >= -1)
1910 if (rev >= -1)
1914 return rev;
1911 return rev;
1915
1912
1916 /*
1913 /*
1917 * For the first handful of lookups, we scan the entire index,
1914 * For the first handful of lookups, we scan the entire index,
1918 * and cache only the matching nodes. This optimizes for cases
1915 * and cache only the matching nodes. This optimizes for cases
1919 * like "hg tip", where only a few nodes are accessed.
1916 * like "hg tip", where only a few nodes are accessed.
1920 *
1917 *
1921 * After that, we cache every node we visit, using a single
1918 * After that, we cache every node we visit, using a single
1922 * scan amortized over multiple lookups. This gives the best
1919 * scan amortized over multiple lookups. This gives the best
1923 * bulk performance, e.g. for "hg log".
1920 * bulk performance, e.g. for "hg log".
1924 */
1921 */
1925 if (self->ntmisses++ < 4) {
1922 if (self->ntmisses++ < 4) {
1926 for (rev = self->ntrev - 1; rev >= 0; rev--) {
1923 for (rev = self->ntrev - 1; rev >= 0; rev--) {
1927 const char *n = index_node_existing(self, rev);
1924 const char *n = index_node_existing(self, rev);
1928 if (n == NULL)
1925 if (n == NULL)
1929 return -3;
1926 return -3;
1930 if (memcmp(node, n, nodelen > 20 ? 20 : nodelen) == 0) {
1927 if (memcmp(node, n, nodelen > 20 ? 20 : nodelen) == 0) {
1931 if (nt_insert(&self->nt, n, rev) == -1)
1928 if (nt_insert(&self->nt, n, rev) == -1)
1932 return -3;
1929 return -3;
1933 break;
1930 break;
1934 }
1931 }
1935 }
1932 }
1936 } else {
1933 } else {
1937 for (rev = self->ntrev - 1; rev >= 0; rev--) {
1934 for (rev = self->ntrev - 1; rev >= 0; rev--) {
1938 const char *n = index_node_existing(self, rev);
1935 const char *n = index_node_existing(self, rev);
1939 if (n == NULL)
1936 if (n == NULL)
1940 return -3;
1937 return -3;
1941 if (nt_insert(&self->nt, n, rev) == -1) {
1938 if (nt_insert(&self->nt, n, rev) == -1) {
1942 self->ntrev = rev + 1;
1939 self->ntrev = rev + 1;
1943 return -3;
1940 return -3;
1944 }
1941 }
1945 if (memcmp(node, n, nodelen > 20 ? 20 : nodelen) == 0) {
1942 if (memcmp(node, n, nodelen > 20 ? 20 : nodelen) == 0) {
1946 break;
1943 break;
1947 }
1944 }
1948 }
1945 }
1949 self->ntrev = rev;
1946 self->ntrev = rev;
1950 }
1947 }
1951
1948
1952 if (rev >= 0)
1949 if (rev >= 0)
1953 return rev;
1950 return rev;
1954 return -2;
1951 return -2;
1955 }
1952 }
1956
1953
1957 static PyObject *index_getitem(indexObject *self, PyObject *value)
1954 static PyObject *index_getitem(indexObject *self, PyObject *value)
1958 {
1955 {
1959 char *node;
1956 char *node;
1960 int rev;
1957 int rev;
1961
1958
1962 if (PyInt_Check(value)) {
1959 if (PyInt_Check(value)) {
1963 long idx;
1960 long idx;
1964 if (!pylong_to_long(value, &idx)) {
1961 if (!pylong_to_long(value, &idx)) {
1965 return NULL;
1962 return NULL;
1966 }
1963 }
1967 return index_get(self, idx);
1964 return index_get(self, idx);
1968 }
1965 }
1969
1966
1970 if (node_check(value, &node) == -1)
1967 if (node_check(value, &node) == -1)
1971 return NULL;
1968 return NULL;
1972 rev = index_find_node(self, node, 20);
1969 rev = index_find_node(self, node, 20);
1973 if (rev >= -1)
1970 if (rev >= -1)
1974 return PyInt_FromLong(rev);
1971 return PyInt_FromLong(rev);
1975 if (rev == -2)
1972 if (rev == -2)
1976 raise_revlog_error();
1973 raise_revlog_error();
1977 return NULL;
1974 return NULL;
1978 }
1975 }
1979
1976
1980 /*
1977 /*
1981 * Fully populate the radix tree.
1978 * Fully populate the radix tree.
1982 */
1979 */
1983 static int index_populate_nt(indexObject *self)
1980 static int index_populate_nt(indexObject *self)
1984 {
1981 {
1985 int rev;
1982 int rev;
1986 if (self->ntrev > 0) {
1983 if (self->ntrev > 0) {
1987 for (rev = self->ntrev - 1; rev >= 0; rev--) {
1984 for (rev = self->ntrev - 1; rev >= 0; rev--) {
1988 const char *n = index_node_existing(self, rev);
1985 const char *n = index_node_existing(self, rev);
1989 if (n == NULL)
1986 if (n == NULL)
1990 return -1;
1987 return -1;
1991 if (nt_insert(&self->nt, n, rev) == -1)
1988 if (nt_insert(&self->nt, n, rev) == -1)
1992 return -1;
1989 return -1;
1993 }
1990 }
1994 self->ntrev = -1;
1991 self->ntrev = -1;
1995 }
1992 }
1996 return 0;
1993 return 0;
1997 }
1994 }
1998
1995
1999 static PyObject *index_partialmatch(indexObject *self, PyObject *args)
1996 static PyObject *index_partialmatch(indexObject *self, PyObject *args)
2000 {
1997 {
2001 const char *fullnode;
1998 const char *fullnode;
2002 Py_ssize_t nodelen;
1999 Py_ssize_t nodelen;
2003 char *node;
2000 char *node;
2004 int rev, i;
2001 int rev, i;
2005
2002
2006 if (!PyArg_ParseTuple(args, PY23("s#", "y#"), &node, &nodelen))
2003 if (!PyArg_ParseTuple(args, PY23("s#", "y#"), &node, &nodelen))
2007 return NULL;
2004 return NULL;
2008
2005
2009 if (nodelen < 1) {
2006 if (nodelen < 1) {
2010 PyErr_SetString(PyExc_ValueError, "key too short");
2007 PyErr_SetString(PyExc_ValueError, "key too short");
2011 return NULL;
2008 return NULL;
2012 }
2009 }
2013
2010
2014 if (nodelen > 40) {
2011 if (nodelen > 40) {
2015 PyErr_SetString(PyExc_ValueError, "key too long");
2012 PyErr_SetString(PyExc_ValueError, "key too long");
2016 return NULL;
2013 return NULL;
2017 }
2014 }
2018
2015
2019 for (i = 0; i < nodelen; i++)
2016 for (i = 0; i < nodelen; i++)
2020 hexdigit(node, i);
2017 hexdigit(node, i);
2021 if (PyErr_Occurred()) {
2018 if (PyErr_Occurred()) {
2022 /* input contains non-hex characters */
2019 /* input contains non-hex characters */
2023 PyErr_Clear();
2020 PyErr_Clear();
2024 Py_RETURN_NONE;
2021 Py_RETURN_NONE;
2025 }
2022 }
2026
2023
2027 if (index_init_nt(self) == -1)
2024 if (index_init_nt(self) == -1)
2028 return NULL;
2025 return NULL;
2029 if (index_populate_nt(self) == -1)
2026 if (index_populate_nt(self) == -1)
2030 return NULL;
2027 return NULL;
2031 rev = nt_partialmatch(&self->nt, node, nodelen);
2028 rev = nt_partialmatch(&self->nt, node, nodelen);
2032
2029
2033 switch (rev) {
2030 switch (rev) {
2034 case -4:
2031 case -4:
2035 raise_revlog_error();
2032 raise_revlog_error();
2036 return NULL;
2033 return NULL;
2037 case -2:
2034 case -2:
2038 Py_RETURN_NONE;
2035 Py_RETURN_NONE;
2039 case -1:
2036 case -1:
2040 return PyBytes_FromStringAndSize(nullid, 20);
2037 return PyBytes_FromStringAndSize(nullid, 20);
2041 }
2038 }
2042
2039
2043 fullnode = index_node_existing(self, rev);
2040 fullnode = index_node_existing(self, rev);
2044 if (fullnode == NULL) {
2041 if (fullnode == NULL) {
2045 return NULL;
2042 return NULL;
2046 }
2043 }
2047 return PyBytes_FromStringAndSize(fullnode, 20);
2044 return PyBytes_FromStringAndSize(fullnode, 20);
2048 }
2045 }
2049
2046
2050 static PyObject *index_shortest(indexObject *self, PyObject *args)
2047 static PyObject *index_shortest(indexObject *self, PyObject *args)
2051 {
2048 {
2052 PyObject *val;
2049 PyObject *val;
2053 char *node;
2050 char *node;
2054 int length;
2051 int length;
2055
2052
2056 if (!PyArg_ParseTuple(args, "O", &val))
2053 if (!PyArg_ParseTuple(args, "O", &val))
2057 return NULL;
2054 return NULL;
2058 if (node_check(val, &node) == -1)
2055 if (node_check(val, &node) == -1)
2059 return NULL;
2056 return NULL;
2060
2057
2061 self->ntlookups++;
2058 self->ntlookups++;
2062 if (index_init_nt(self) == -1)
2059 if (index_init_nt(self) == -1)
2063 return NULL;
2060 return NULL;
2064 if (index_populate_nt(self) == -1)
2061 if (index_populate_nt(self) == -1)
2065 return NULL;
2062 return NULL;
2066 length = nt_shortest(&self->nt, node);
2063 length = nt_shortest(&self->nt, node);
2067 if (length == -3)
2064 if (length == -3)
2068 return NULL;
2065 return NULL;
2069 if (length == -2) {
2066 if (length == -2) {
2070 raise_revlog_error();
2067 raise_revlog_error();
2071 return NULL;
2068 return NULL;
2072 }
2069 }
2073 return PyInt_FromLong(length);
2070 return PyInt_FromLong(length);
2074 }
2071 }
2075
2072
2076 static PyObject *index_m_get(indexObject *self, PyObject *args)
2073 static PyObject *index_m_get(indexObject *self, PyObject *args)
2077 {
2074 {
2078 PyObject *val;
2075 PyObject *val;
2079 char *node;
2076 char *node;
2080 int rev;
2077 int rev;
2081
2078
2082 if (!PyArg_ParseTuple(args, "O", &val))
2079 if (!PyArg_ParseTuple(args, "O", &val))
2083 return NULL;
2080 return NULL;
2084 if (node_check(val, &node) == -1)
2081 if (node_check(val, &node) == -1)
2085 return NULL;
2082 return NULL;
2086 rev = index_find_node(self, node, 20);
2083 rev = index_find_node(self, node, 20);
2087 if (rev == -3)
2084 if (rev == -3)
2088 return NULL;
2085 return NULL;
2089 if (rev == -2)
2086 if (rev == -2)
2090 Py_RETURN_NONE;
2087 Py_RETURN_NONE;
2091 return PyInt_FromLong(rev);
2088 return PyInt_FromLong(rev);
2092 }
2089 }
2093
2090
2094 static int index_contains(indexObject *self, PyObject *value)
2091 static int index_contains(indexObject *self, PyObject *value)
2095 {
2092 {
2096 char *node;
2093 char *node;
2097
2094
2098 if (PyInt_Check(value)) {
2095 if (PyInt_Check(value)) {
2099 long rev;
2096 long rev;
2100 if (!pylong_to_long(value, &rev)) {
2097 if (!pylong_to_long(value, &rev)) {
2101 return -1;
2098 return -1;
2102 }
2099 }
2103 return rev >= -1 && rev < index_length(self);
2100 return rev >= -1 && rev < index_length(self);
2104 }
2101 }
2105
2102
2106 if (node_check(value, &node) == -1)
2103 if (node_check(value, &node) == -1)
2107 return -1;
2104 return -1;
2108
2105
2109 switch (index_find_node(self, node, 20)) {
2106 switch (index_find_node(self, node, 20)) {
2110 case -3:
2107 case -3:
2111 return -1;
2108 return -1;
2112 case -2:
2109 case -2:
2113 return 0;
2110 return 0;
2114 default:
2111 default:
2115 return 1;
2112 return 1;
2116 }
2113 }
2117 }
2114 }
2118
2115
2119 static PyObject *index_m_has_node(indexObject *self, PyObject *args)
2116 static PyObject *index_m_has_node(indexObject *self, PyObject *args)
2120 {
2117 {
2121 int ret = index_contains(self, args);
2118 int ret = index_contains(self, args);
2122 if (ret < 0)
2119 if (ret < 0)
2123 return NULL;
2120 return NULL;
2124 return PyBool_FromLong((long)ret);
2121 return PyBool_FromLong((long)ret);
2125 }
2122 }
2126
2123
2127 static PyObject *index_m_rev(indexObject *self, PyObject *val)
2124 static PyObject *index_m_rev(indexObject *self, PyObject *val)
2128 {
2125 {
2129 char *node;
2126 char *node;
2130 int rev;
2127 int rev;
2131
2128
2132 if (node_check(val, &node) == -1)
2129 if (node_check(val, &node) == -1)
2133 return NULL;
2130 return NULL;
2134 rev = index_find_node(self, node, 20);
2131 rev = index_find_node(self, node, 20);
2135 if (rev >= -1)
2132 if (rev >= -1)
2136 return PyInt_FromLong(rev);
2133 return PyInt_FromLong(rev);
2137 if (rev == -2)
2134 if (rev == -2)
2138 raise_revlog_error();
2135 raise_revlog_error();
2139 return NULL;
2136 return NULL;
2140 }
2137 }
2141
2138
2142 typedef uint64_t bitmask;
2139 typedef uint64_t bitmask;
2143
2140
2144 /*
2141 /*
2145 * Given a disjoint set of revs, return all candidates for the
2142 * Given a disjoint set of revs, return all candidates for the
2146 * greatest common ancestor. In revset notation, this is the set
2143 * greatest common ancestor. In revset notation, this is the set
2147 * "heads(::a and ::b and ...)"
2144 * "heads(::a and ::b and ...)"
2148 */
2145 */
2149 static PyObject *find_gca_candidates(indexObject *self, const int *revs,
2146 static PyObject *find_gca_candidates(indexObject *self, const int *revs,
2150 int revcount)
2147 int revcount)
2151 {
2148 {
2152 const bitmask allseen = (1ull << revcount) - 1;
2149 const bitmask allseen = (1ull << revcount) - 1;
2153 const bitmask poison = 1ull << revcount;
2150 const bitmask poison = 1ull << revcount;
2154 PyObject *gca = PyList_New(0);
2151 PyObject *gca = PyList_New(0);
2155 int i, v, interesting;
2152 int i, v, interesting;
2156 int maxrev = -1;
2153 int maxrev = -1;
2157 bitmask sp;
2154 bitmask sp;
2158 bitmask *seen;
2155 bitmask *seen;
2159
2156
2160 if (gca == NULL)
2157 if (gca == NULL)
2161 return PyErr_NoMemory();
2158 return PyErr_NoMemory();
2162
2159
2163 for (i = 0; i < revcount; i++) {
2160 for (i = 0; i < revcount; i++) {
2164 if (revs[i] > maxrev)
2161 if (revs[i] > maxrev)
2165 maxrev = revs[i];
2162 maxrev = revs[i];
2166 }
2163 }
2167
2164
2168 seen = calloc(sizeof(*seen), maxrev + 1);
2165 seen = calloc(sizeof(*seen), maxrev + 1);
2169 if (seen == NULL) {
2166 if (seen == NULL) {
2170 Py_DECREF(gca);
2167 Py_DECREF(gca);
2171 return PyErr_NoMemory();
2168 return PyErr_NoMemory();
2172 }
2169 }
2173
2170
2174 for (i = 0; i < revcount; i++)
2171 for (i = 0; i < revcount; i++)
2175 seen[revs[i]] = 1ull << i;
2172 seen[revs[i]] = 1ull << i;
2176
2173
2177 interesting = revcount;
2174 interesting = revcount;
2178
2175
2179 for (v = maxrev; v >= 0 && interesting; v--) {
2176 for (v = maxrev; v >= 0 && interesting; v--) {
2180 bitmask sv = seen[v];
2177 bitmask sv = seen[v];
2181 int parents[2];
2178 int parents[2];
2182
2179
2183 if (!sv)
2180 if (!sv)
2184 continue;
2181 continue;
2185
2182
2186 if (sv < poison) {
2183 if (sv < poison) {
2187 interesting -= 1;
2184 interesting -= 1;
2188 if (sv == allseen) {
2185 if (sv == allseen) {
2189 PyObject *obj = PyInt_FromLong(v);
2186 PyObject *obj = PyInt_FromLong(v);
2190 if (obj == NULL)
2187 if (obj == NULL)
2191 goto bail;
2188 goto bail;
2192 if (PyList_Append(gca, obj) == -1) {
2189 if (PyList_Append(gca, obj) == -1) {
2193 Py_DECREF(obj);
2190 Py_DECREF(obj);
2194 goto bail;
2191 goto bail;
2195 }
2192 }
2196 sv |= poison;
2193 sv |= poison;
2197 for (i = 0; i < revcount; i++) {
2194 for (i = 0; i < revcount; i++) {
2198 if (revs[i] == v)
2195 if (revs[i] == v)
2199 goto done;
2196 goto done;
2200 }
2197 }
2201 }
2198 }
2202 }
2199 }
2203 if (index_get_parents(self, v, parents, maxrev) < 0)
2200 if (index_get_parents(self, v, parents, maxrev) < 0)
2204 goto bail;
2201 goto bail;
2205
2202
2206 for (i = 0; i < 2; i++) {
2203 for (i = 0; i < 2; i++) {
2207 int p = parents[i];
2204 int p = parents[i];
2208 if (p == -1)
2205 if (p == -1)
2209 continue;
2206 continue;
2210 sp = seen[p];
2207 sp = seen[p];
2211 if (sv < poison) {
2208 if (sv < poison) {
2212 if (sp == 0) {
2209 if (sp == 0) {
2213 seen[p] = sv;
2210 seen[p] = sv;
2214 interesting++;
2211 interesting++;
2215 } else if (sp != sv)
2212 } else if (sp != sv)
2216 seen[p] |= sv;
2213 seen[p] |= sv;
2217 } else {
2214 } else {
2218 if (sp && sp < poison)
2215 if (sp && sp < poison)
2219 interesting--;
2216 interesting--;
2220 seen[p] = sv;
2217 seen[p] = sv;
2221 }
2218 }
2222 }
2219 }
2223 }
2220 }
2224
2221
2225 done:
2222 done:
2226 free(seen);
2223 free(seen);
2227 return gca;
2224 return gca;
2228 bail:
2225 bail:
2229 free(seen);
2226 free(seen);
2230 Py_XDECREF(gca);
2227 Py_XDECREF(gca);
2231 return NULL;
2228 return NULL;
2232 }
2229 }
2233
2230
2234 /*
2231 /*
2235 * Given a disjoint set of revs, return the subset with the longest
2232 * Given a disjoint set of revs, return the subset with the longest
2236 * path to the root.
2233 * path to the root.
2237 */
2234 */
2238 static PyObject *find_deepest(indexObject *self, PyObject *revs)
2235 static PyObject *find_deepest(indexObject *self, PyObject *revs)
2239 {
2236 {
2240 const Py_ssize_t revcount = PyList_GET_SIZE(revs);
2237 const Py_ssize_t revcount = PyList_GET_SIZE(revs);
2241 static const Py_ssize_t capacity = 24;
2238 static const Py_ssize_t capacity = 24;
2242 int *depth, *interesting = NULL;
2239 int *depth, *interesting = NULL;
2243 int i, j, v, ninteresting;
2240 int i, j, v, ninteresting;
2244 PyObject *dict = NULL, *keys = NULL;
2241 PyObject *dict = NULL, *keys = NULL;
2245 long *seen = NULL;
2242 long *seen = NULL;
2246 int maxrev = -1;
2243 int maxrev = -1;
2247 long final;
2244 long final;
2248
2245
2249 if (revcount > capacity) {
2246 if (revcount > capacity) {
2250 PyErr_Format(PyExc_OverflowError,
2247 PyErr_Format(PyExc_OverflowError,
2251 "bitset size (%ld) > capacity (%ld)",
2248 "bitset size (%ld) > capacity (%ld)",
2252 (long)revcount, (long)capacity);
2249 (long)revcount, (long)capacity);
2253 return NULL;
2250 return NULL;
2254 }
2251 }
2255
2252
2256 for (i = 0; i < revcount; i++) {
2253 for (i = 0; i < revcount; i++) {
2257 int n = (int)PyInt_AsLong(PyList_GET_ITEM(revs, i));
2254 int n = (int)PyInt_AsLong(PyList_GET_ITEM(revs, i));
2258 if (n > maxrev)
2255 if (n > maxrev)
2259 maxrev = n;
2256 maxrev = n;
2260 }
2257 }
2261
2258
2262 depth = calloc(sizeof(*depth), maxrev + 1);
2259 depth = calloc(sizeof(*depth), maxrev + 1);
2263 if (depth == NULL)
2260 if (depth == NULL)
2264 return PyErr_NoMemory();
2261 return PyErr_NoMemory();
2265
2262
2266 seen = calloc(sizeof(*seen), maxrev + 1);
2263 seen = calloc(sizeof(*seen), maxrev + 1);
2267 if (seen == NULL) {
2264 if (seen == NULL) {
2268 PyErr_NoMemory();
2265 PyErr_NoMemory();
2269 goto bail;
2266 goto bail;
2270 }
2267 }
2271
2268
2272 interesting = calloc(sizeof(*interesting), ((size_t)1) << revcount);
2269 interesting = calloc(sizeof(*interesting), ((size_t)1) << revcount);
2273 if (interesting == NULL) {
2270 if (interesting == NULL) {
2274 PyErr_NoMemory();
2271 PyErr_NoMemory();
2275 goto bail;
2272 goto bail;
2276 }
2273 }
2277
2274
2278 if (PyList_Sort(revs) == -1)
2275 if (PyList_Sort(revs) == -1)
2279 goto bail;
2276 goto bail;
2280
2277
2281 for (i = 0; i < revcount; i++) {
2278 for (i = 0; i < revcount; i++) {
2282 int n = (int)PyInt_AsLong(PyList_GET_ITEM(revs, i));
2279 int n = (int)PyInt_AsLong(PyList_GET_ITEM(revs, i));
2283 long b = 1l << i;
2280 long b = 1l << i;
2284 depth[n] = 1;
2281 depth[n] = 1;
2285 seen[n] = b;
2282 seen[n] = b;
2286 interesting[b] = 1;
2283 interesting[b] = 1;
2287 }
2284 }
2288
2285
2289 /* invariant: ninteresting is the number of non-zero entries in
2286 /* invariant: ninteresting is the number of non-zero entries in
2290 * interesting. */
2287 * interesting. */
2291 ninteresting = (int)revcount;
2288 ninteresting = (int)revcount;
2292
2289
2293 for (v = maxrev; v >= 0 && ninteresting > 1; v--) {
2290 for (v = maxrev; v >= 0 && ninteresting > 1; v--) {
2294 int dv = depth[v];
2291 int dv = depth[v];
2295 int parents[2];
2292 int parents[2];
2296 long sv;
2293 long sv;
2297
2294
2298 if (dv == 0)
2295 if (dv == 0)
2299 continue;
2296 continue;
2300
2297
2301 sv = seen[v];
2298 sv = seen[v];
2302 if (index_get_parents(self, v, parents, maxrev) < 0)
2299 if (index_get_parents(self, v, parents, maxrev) < 0)
2303 goto bail;
2300 goto bail;
2304
2301
2305 for (i = 0; i < 2; i++) {
2302 for (i = 0; i < 2; i++) {
2306 int p = parents[i];
2303 int p = parents[i];
2307 long sp;
2304 long sp;
2308 int dp;
2305 int dp;
2309
2306
2310 if (p == -1)
2307 if (p == -1)
2311 continue;
2308 continue;
2312
2309
2313 dp = depth[p];
2310 dp = depth[p];
2314 sp = seen[p];
2311 sp = seen[p];
2315 if (dp <= dv) {
2312 if (dp <= dv) {
2316 depth[p] = dv + 1;
2313 depth[p] = dv + 1;
2317 if (sp != sv) {
2314 if (sp != sv) {
2318 interesting[sv] += 1;
2315 interesting[sv] += 1;
2319 seen[p] = sv;
2316 seen[p] = sv;
2320 if (sp) {
2317 if (sp) {
2321 interesting[sp] -= 1;
2318 interesting[sp] -= 1;
2322 if (interesting[sp] == 0)
2319 if (interesting[sp] == 0)
2323 ninteresting -= 1;
2320 ninteresting -= 1;
2324 }
2321 }
2325 }
2322 }
2326 } else if (dv == dp - 1) {
2323 } else if (dv == dp - 1) {
2327 long nsp = sp | sv;
2324 long nsp = sp | sv;
2328 if (nsp == sp)
2325 if (nsp == sp)
2329 continue;
2326 continue;
2330 seen[p] = nsp;
2327 seen[p] = nsp;
2331 interesting[sp] -= 1;
2328 interesting[sp] -= 1;
2332 if (interesting[sp] == 0)
2329 if (interesting[sp] == 0)
2333 ninteresting -= 1;
2330 ninteresting -= 1;
2334 if (interesting[nsp] == 0)
2331 if (interesting[nsp] == 0)
2335 ninteresting += 1;
2332 ninteresting += 1;
2336 interesting[nsp] += 1;
2333 interesting[nsp] += 1;
2337 }
2334 }
2338 }
2335 }
2339 interesting[sv] -= 1;
2336 interesting[sv] -= 1;
2340 if (interesting[sv] == 0)
2337 if (interesting[sv] == 0)
2341 ninteresting -= 1;
2338 ninteresting -= 1;
2342 }
2339 }
2343
2340
2344 final = 0;
2341 final = 0;
2345 j = ninteresting;
2342 j = ninteresting;
2346 for (i = 0; i < (int)(2 << revcount) && j > 0; i++) {
2343 for (i = 0; i < (int)(2 << revcount) && j > 0; i++) {
2347 if (interesting[i] == 0)
2344 if (interesting[i] == 0)
2348 continue;
2345 continue;
2349 final |= i;
2346 final |= i;
2350 j -= 1;
2347 j -= 1;
2351 }
2348 }
2352 if (final == 0) {
2349 if (final == 0) {
2353 keys = PyList_New(0);
2350 keys = PyList_New(0);
2354 goto bail;
2351 goto bail;
2355 }
2352 }
2356
2353
2357 dict = PyDict_New();
2354 dict = PyDict_New();
2358 if (dict == NULL)
2355 if (dict == NULL)
2359 goto bail;
2356 goto bail;
2360
2357
2361 for (i = 0; i < revcount; i++) {
2358 for (i = 0; i < revcount; i++) {
2362 PyObject *key;
2359 PyObject *key;
2363
2360
2364 if ((final & (1 << i)) == 0)
2361 if ((final & (1 << i)) == 0)
2365 continue;
2362 continue;
2366
2363
2367 key = PyList_GET_ITEM(revs, i);
2364 key = PyList_GET_ITEM(revs, i);
2368 Py_INCREF(key);
2365 Py_INCREF(key);
2369 Py_INCREF(Py_None);
2366 Py_INCREF(Py_None);
2370 if (PyDict_SetItem(dict, key, Py_None) == -1) {
2367 if (PyDict_SetItem(dict, key, Py_None) == -1) {
2371 Py_DECREF(key);
2368 Py_DECREF(key);
2372 Py_DECREF(Py_None);
2369 Py_DECREF(Py_None);
2373 goto bail;
2370 goto bail;
2374 }
2371 }
2375 }
2372 }
2376
2373
2377 keys = PyDict_Keys(dict);
2374 keys = PyDict_Keys(dict);
2378
2375
2379 bail:
2376 bail:
2380 free(depth);
2377 free(depth);
2381 free(seen);
2378 free(seen);
2382 free(interesting);
2379 free(interesting);
2383 Py_XDECREF(dict);
2380 Py_XDECREF(dict);
2384
2381
2385 return keys;
2382 return keys;
2386 }
2383 }
2387
2384
2388 /*
2385 /*
2389 * Given a (possibly overlapping) set of revs, return all the
2386 * Given a (possibly overlapping) set of revs, return all the
2390 * common ancestors heads: heads(::args[0] and ::a[1] and ...)
2387 * common ancestors heads: heads(::args[0] and ::a[1] and ...)
2391 */
2388 */
2392 static PyObject *index_commonancestorsheads(indexObject *self, PyObject *args)
2389 static PyObject *index_commonancestorsheads(indexObject *self, PyObject *args)
2393 {
2390 {
2394 PyObject *ret = NULL;
2391 PyObject *ret = NULL;
2395 Py_ssize_t argcount, i, len;
2392 Py_ssize_t argcount, i, len;
2396 bitmask repeat = 0;
2393 bitmask repeat = 0;
2397 int revcount = 0;
2394 int revcount = 0;
2398 int *revs;
2395 int *revs;
2399
2396
2400 argcount = PySequence_Length(args);
2397 argcount = PySequence_Length(args);
2401 revs = PyMem_Malloc(argcount * sizeof(*revs));
2398 revs = PyMem_Malloc(argcount * sizeof(*revs));
2402 if (argcount > 0 && revs == NULL)
2399 if (argcount > 0 && revs == NULL)
2403 return PyErr_NoMemory();
2400 return PyErr_NoMemory();
2404 len = index_length(self);
2401 len = index_length(self);
2405
2402
2406 for (i = 0; i < argcount; i++) {
2403 for (i = 0; i < argcount; i++) {
2407 static const int capacity = 24;
2404 static const int capacity = 24;
2408 PyObject *obj = PySequence_GetItem(args, i);
2405 PyObject *obj = PySequence_GetItem(args, i);
2409 bitmask x;
2406 bitmask x;
2410 long val;
2407 long val;
2411
2408
2412 if (!PyInt_Check(obj)) {
2409 if (!PyInt_Check(obj)) {
2413 PyErr_SetString(PyExc_TypeError,
2410 PyErr_SetString(PyExc_TypeError,
2414 "arguments must all be ints");
2411 "arguments must all be ints");
2415 Py_DECREF(obj);
2412 Py_DECREF(obj);
2416 goto bail;
2413 goto bail;
2417 }
2414 }
2418 val = PyInt_AsLong(obj);
2415 val = PyInt_AsLong(obj);
2419 Py_DECREF(obj);
2416 Py_DECREF(obj);
2420 if (val == -1) {
2417 if (val == -1) {
2421 ret = PyList_New(0);
2418 ret = PyList_New(0);
2422 goto done;
2419 goto done;
2423 }
2420 }
2424 if (val < 0 || val >= len) {
2421 if (val < 0 || val >= len) {
2425 PyErr_SetString(PyExc_IndexError, "index out of range");
2422 PyErr_SetString(PyExc_IndexError, "index out of range");
2426 goto bail;
2423 goto bail;
2427 }
2424 }
2428 /* this cheesy bloom filter lets us avoid some more
2425 /* this cheesy bloom filter lets us avoid some more
2429 * expensive duplicate checks in the common set-is-disjoint
2426 * expensive duplicate checks in the common set-is-disjoint
2430 * case */
2427 * case */
2431 x = 1ull << (val & 0x3f);
2428 x = 1ull << (val & 0x3f);
2432 if (repeat & x) {
2429 if (repeat & x) {
2433 int k;
2430 int k;
2434 for (k = 0; k < revcount; k++) {
2431 for (k = 0; k < revcount; k++) {
2435 if (val == revs[k])
2432 if (val == revs[k])
2436 goto duplicate;
2433 goto duplicate;
2437 }
2434 }
2438 } else
2435 } else
2439 repeat |= x;
2436 repeat |= x;
2440 if (revcount >= capacity) {
2437 if (revcount >= capacity) {
2441 PyErr_Format(PyExc_OverflowError,
2438 PyErr_Format(PyExc_OverflowError,
2442 "bitset size (%d) > capacity (%d)",
2439 "bitset size (%d) > capacity (%d)",
2443 revcount, capacity);
2440 revcount, capacity);
2444 goto bail;
2441 goto bail;
2445 }
2442 }
2446 revs[revcount++] = (int)val;
2443 revs[revcount++] = (int)val;
2447 duplicate:;
2444 duplicate:;
2448 }
2445 }
2449
2446
2450 if (revcount == 0) {
2447 if (revcount == 0) {
2451 ret = PyList_New(0);
2448 ret = PyList_New(0);
2452 goto done;
2449 goto done;
2453 }
2450 }
2454 if (revcount == 1) {
2451 if (revcount == 1) {
2455 PyObject *obj;
2452 PyObject *obj;
2456 ret = PyList_New(1);
2453 ret = PyList_New(1);
2457 if (ret == NULL)
2454 if (ret == NULL)
2458 goto bail;
2455 goto bail;
2459 obj = PyInt_FromLong(revs[0]);
2456 obj = PyInt_FromLong(revs[0]);
2460 if (obj == NULL)
2457 if (obj == NULL)
2461 goto bail;
2458 goto bail;
2462 PyList_SET_ITEM(ret, 0, obj);
2459 PyList_SET_ITEM(ret, 0, obj);
2463 goto done;
2460 goto done;
2464 }
2461 }
2465
2462
2466 ret = find_gca_candidates(self, revs, revcount);
2463 ret = find_gca_candidates(self, revs, revcount);
2467 if (ret == NULL)
2464 if (ret == NULL)
2468 goto bail;
2465 goto bail;
2469
2466
2470 done:
2467 done:
2471 PyMem_Free(revs);
2468 PyMem_Free(revs);
2472 return ret;
2469 return ret;
2473
2470
2474 bail:
2471 bail:
2475 PyMem_Free(revs);
2472 PyMem_Free(revs);
2476 Py_XDECREF(ret);
2473 Py_XDECREF(ret);
2477 return NULL;
2474 return NULL;
2478 }
2475 }
2479
2476
2480 /*
2477 /*
2481 * Given a (possibly overlapping) set of revs, return the greatest
2478 * Given a (possibly overlapping) set of revs, return the greatest
2482 * common ancestors: those with the longest path to the root.
2479 * common ancestors: those with the longest path to the root.
2483 */
2480 */
2484 static PyObject *index_ancestors(indexObject *self, PyObject *args)
2481 static PyObject *index_ancestors(indexObject *self, PyObject *args)
2485 {
2482 {
2486 PyObject *ret;
2483 PyObject *ret;
2487 PyObject *gca = index_commonancestorsheads(self, args);
2484 PyObject *gca = index_commonancestorsheads(self, args);
2488 if (gca == NULL)
2485 if (gca == NULL)
2489 return NULL;
2486 return NULL;
2490
2487
2491 if (PyList_GET_SIZE(gca) <= 1) {
2488 if (PyList_GET_SIZE(gca) <= 1) {
2492 return gca;
2489 return gca;
2493 }
2490 }
2494
2491
2495 ret = find_deepest(self, gca);
2492 ret = find_deepest(self, gca);
2496 Py_DECREF(gca);
2493 Py_DECREF(gca);
2497 return ret;
2494 return ret;
2498 }
2495 }
2499
2496
2500 /*
2497 /*
2501 * Invalidate any trie entries introduced by added revs.
2498 * Invalidate any trie entries introduced by added revs.
2502 */
2499 */
2503 static void index_invalidate_added(indexObject *self, Py_ssize_t start)
2500 static void index_invalidate_added(indexObject *self, Py_ssize_t start)
2504 {
2501 {
2505 Py_ssize_t i, len = PyList_GET_SIZE(self->added);
2502 Py_ssize_t i, len = PyList_GET_SIZE(self->added);
2506
2503
2507 for (i = start; i < len; i++) {
2504 for (i = start; i < len; i++) {
2508 PyObject *tuple = PyList_GET_ITEM(self->added, i);
2505 PyObject *tuple = PyList_GET_ITEM(self->added, i);
2509 PyObject *node = PyTuple_GET_ITEM(tuple, 7);
2506 PyObject *node = PyTuple_GET_ITEM(tuple, 7);
2510
2507
2511 nt_delete_node(&self->nt, PyBytes_AS_STRING(node));
2508 nt_delete_node(&self->nt, PyBytes_AS_STRING(node));
2512 }
2509 }
2513
2510
2514 if (start == 0)
2511 if (start == 0)
2515 Py_CLEAR(self->added);
2512 Py_CLEAR(self->added);
2516 }
2513 }
2517
2514
2518 /*
2515 /*
2519 * Delete a numeric range of revs, which must be at the end of the
2516 * Delete a numeric range of revs, which must be at the end of the
2520 * range.
2517 * range.
2521 */
2518 */
2522 static int index_slice_del(indexObject *self, PyObject *item)
2519 static int index_slice_del(indexObject *self, PyObject *item)
2523 {
2520 {
2524 Py_ssize_t start, stop, step, slicelength;
2521 Py_ssize_t start, stop, step, slicelength;
2525 Py_ssize_t length = index_length(self) + 1;
2522 Py_ssize_t length = index_length(self) + 1;
2526 int ret = 0;
2523 int ret = 0;
2527
2524
2528 /* Argument changed from PySliceObject* to PyObject* in Python 3. */
2525 /* Argument changed from PySliceObject* to PyObject* in Python 3. */
2529 #ifdef IS_PY3K
2526 #ifdef IS_PY3K
2530 if (PySlice_GetIndicesEx(item, length, &start, &stop, &step,
2527 if (PySlice_GetIndicesEx(item, length, &start, &stop, &step,
2531 &slicelength) < 0)
2528 &slicelength) < 0)
2532 #else
2529 #else
2533 if (PySlice_GetIndicesEx((PySliceObject *)item, length, &start, &stop,
2530 if (PySlice_GetIndicesEx((PySliceObject *)item, length, &start, &stop,
2534 &step, &slicelength) < 0)
2531 &step, &slicelength) < 0)
2535 #endif
2532 #endif
2536 return -1;
2533 return -1;
2537
2534
2538 if (slicelength <= 0)
2535 if (slicelength <= 0)
2539 return 0;
2536 return 0;
2540
2537
2541 if ((step < 0 && start < stop) || (step > 0 && start > stop))
2538 if ((step < 0 && start < stop) || (step > 0 && start > stop))
2542 stop = start;
2539 stop = start;
2543
2540
2544 if (step < 0) {
2541 if (step < 0) {
2545 stop = start + 1;
2542 stop = start + 1;
2546 start = stop + step * (slicelength - 1) - 1;
2543 start = stop + step * (slicelength - 1) - 1;
2547 step = -step;
2544 step = -step;
2548 }
2545 }
2549
2546
2550 if (step != 1) {
2547 if (step != 1) {
2551 PyErr_SetString(PyExc_ValueError,
2548 PyErr_SetString(PyExc_ValueError,
2552 "revlog index delete requires step size of 1");
2549 "revlog index delete requires step size of 1");
2553 return -1;
2550 return -1;
2554 }
2551 }
2555
2552
2556 if (stop != length - 1) {
2553 if (stop != length - 1) {
2557 PyErr_SetString(PyExc_IndexError,
2554 PyErr_SetString(PyExc_IndexError,
2558 "revlog index deletion indices are invalid");
2555 "revlog index deletion indices are invalid");
2559 return -1;
2556 return -1;
2560 }
2557 }
2561
2558
2562 if (start < self->length) {
2559 if (start < self->length) {
2563 if (self->ntinitialized) {
2560 if (self->ntinitialized) {
2564 Py_ssize_t i;
2561 Py_ssize_t i;
2565
2562
2566 for (i = start; i < self->length; i++) {
2563 for (i = start; i < self->length; i++) {
2567 const char *node = index_node_existing(self, i);
2564 const char *node = index_node_existing(self, i);
2568 if (node == NULL)
2565 if (node == NULL)
2569 return -1;
2566 return -1;
2570
2567
2571 nt_delete_node(&self->nt, node);
2568 nt_delete_node(&self->nt, node);
2572 }
2569 }
2573 if (self->added)
2570 if (self->added)
2574 index_invalidate_added(self, 0);
2571 index_invalidate_added(self, 0);
2575 if (self->ntrev > start)
2572 if (self->ntrev > start)
2576 self->ntrev = (int)start;
2573 self->ntrev = (int)start;
2577 } else if (self->added) {
2574 } else if (self->added) {
2578 Py_CLEAR(self->added);
2575 Py_CLEAR(self->added);
2579 }
2576 }
2580
2577
2581 self->length = start;
2578 self->length = start;
2582 if (start < self->raw_length) {
2579 if (start < self->raw_length) {
2583 if (self->cache) {
2580 if (self->cache) {
2584 Py_ssize_t i;
2581 Py_ssize_t i;
2585 for (i = start; i < self->raw_length; i++)
2582 for (i = start; i < self->raw_length; i++)
2586 Py_CLEAR(self->cache[i]);
2583 Py_CLEAR(self->cache[i]);
2587 }
2584 }
2588 self->raw_length = start;
2585 self->raw_length = start;
2589 }
2586 }
2590 goto done;
2587 goto done;
2591 }
2588 }
2592
2589
2593 if (self->ntinitialized) {
2590 if (self->ntinitialized) {
2594 index_invalidate_added(self, start - self->length);
2591 index_invalidate_added(self, start - self->length);
2595 if (self->ntrev > start)
2592 if (self->ntrev > start)
2596 self->ntrev = (int)start;
2593 self->ntrev = (int)start;
2597 }
2594 }
2598 if (self->added)
2595 if (self->added)
2599 ret = PyList_SetSlice(self->added, start - self->length,
2596 ret = PyList_SetSlice(self->added, start - self->length,
2600 PyList_GET_SIZE(self->added), NULL);
2597 PyList_GET_SIZE(self->added), NULL);
2601 done:
2598 done:
2602 Py_CLEAR(self->headrevs);
2599 Py_CLEAR(self->headrevs);
2603 return ret;
2600 return ret;
2604 }
2601 }
2605
2602
2606 /*
2603 /*
2607 * Supported ops:
2604 * Supported ops:
2608 *
2605 *
2609 * slice deletion
2606 * slice deletion
2610 * string assignment (extend node->rev mapping)
2607 * string assignment (extend node->rev mapping)
2611 * string deletion (shrink node->rev mapping)
2608 * string deletion (shrink node->rev mapping)
2612 */
2609 */
2613 static int index_assign_subscript(indexObject *self, PyObject *item,
2610 static int index_assign_subscript(indexObject *self, PyObject *item,
2614 PyObject *value)
2611 PyObject *value)
2615 {
2612 {
2616 char *node;
2613 char *node;
2617 long rev;
2614 long rev;
2618
2615
2619 if (PySlice_Check(item) && value == NULL)
2616 if (PySlice_Check(item) && value == NULL)
2620 return index_slice_del(self, item);
2617 return index_slice_del(self, item);
2621
2618
2622 if (node_check(item, &node) == -1)
2619 if (node_check(item, &node) == -1)
2623 return -1;
2620 return -1;
2624
2621
2625 if (value == NULL)
2622 if (value == NULL)
2626 return self->ntinitialized ? nt_delete_node(&self->nt, node)
2623 return self->ntinitialized ? nt_delete_node(&self->nt, node)
2627 : 0;
2624 : 0;
2628 rev = PyInt_AsLong(value);
2625 rev = PyInt_AsLong(value);
2629 if (rev > INT_MAX || rev < 0) {
2626 if (rev > INT_MAX || rev < 0) {
2630 if (!PyErr_Occurred())
2627 if (!PyErr_Occurred())
2631 PyErr_SetString(PyExc_ValueError, "rev out of range");
2628 PyErr_SetString(PyExc_ValueError, "rev out of range");
2632 return -1;
2629 return -1;
2633 }
2630 }
2634
2631
2635 if (index_init_nt(self) == -1)
2632 if (index_init_nt(self) == -1)
2636 return -1;
2633 return -1;
2637 return nt_insert(&self->nt, node, (int)rev);
2634 return nt_insert(&self->nt, node, (int)rev);
2638 }
2635 }
2639
2636
2640 /*
2637 /*
2641 * Find all RevlogNG entries in an index that has inline data. Update
2638 * Find all RevlogNG entries in an index that has inline data. Update
2642 * the optional "offsets" table with those entries.
2639 * the optional "offsets" table with those entries.
2643 */
2640 */
2644 static Py_ssize_t inline_scan(indexObject *self, const char **offsets)
2641 static Py_ssize_t inline_scan(indexObject *self, const char **offsets)
2645 {
2642 {
2646 const char *data = (const char *)self->buf.buf;
2643 const char *data = (const char *)self->buf.buf;
2647 Py_ssize_t pos = 0;
2644 Py_ssize_t pos = 0;
2648 Py_ssize_t end = self->buf.len;
2645 Py_ssize_t end = self->buf.len;
2649 long incr = v1_hdrsize;
2646 long incr = v1_hdrsize;
2650 Py_ssize_t len = 0;
2647 Py_ssize_t len = 0;
2651
2648
2652 while (pos + v1_hdrsize <= end && pos >= 0) {
2649 while (pos + v1_hdrsize <= end && pos >= 0) {
2653 uint32_t comp_len;
2650 uint32_t comp_len;
2654 /* 3rd element of header is length of compressed inline data */
2651 /* 3rd element of header is length of compressed inline data */
2655 comp_len = getbe32(data + pos + 8);
2652 comp_len = getbe32(data + pos + 8);
2656 incr = v1_hdrsize + comp_len;
2653 incr = v1_hdrsize + comp_len;
2657 if (offsets)
2654 if (offsets)
2658 offsets[len] = data + pos;
2655 offsets[len] = data + pos;
2659 len++;
2656 len++;
2660 pos += incr;
2657 pos += incr;
2661 }
2658 }
2662
2659
2663 if (pos != end) {
2660 if (pos != end) {
2664 if (!PyErr_Occurred())
2661 if (!PyErr_Occurred())
2665 PyErr_SetString(PyExc_ValueError, "corrupt index file");
2662 PyErr_SetString(PyExc_ValueError, "corrupt index file");
2666 return -1;
2663 return -1;
2667 }
2664 }
2668
2665
2669 return len;
2666 return len;
2670 }
2667 }
2671
2668
2672 static int index_init(indexObject *self, PyObject *args)
2669 static int index_init(indexObject *self, PyObject *args)
2673 {
2670 {
2674 PyObject *data_obj, *inlined_obj;
2671 PyObject *data_obj, *inlined_obj;
2675 Py_ssize_t size;
2672 Py_ssize_t size;
2676
2673
2677 /* Initialize before argument-checking to avoid index_dealloc() crash.
2674 /* Initialize before argument-checking to avoid index_dealloc() crash.
2678 */
2675 */
2679 self->raw_length = 0;
2676 self->raw_length = 0;
2680 self->added = NULL;
2677 self->added = NULL;
2681 self->cache = NULL;
2678 self->cache = NULL;
2682 self->data = NULL;
2679 self->data = NULL;
2683 memset(&self->buf, 0, sizeof(self->buf));
2680 memset(&self->buf, 0, sizeof(self->buf));
2684 self->headrevs = NULL;
2681 self->headrevs = NULL;
2685 self->filteredrevs = Py_None;
2682 self->filteredrevs = Py_None;
2686 Py_INCREF(Py_None);
2683 Py_INCREF(Py_None);
2687 self->ntinitialized = 0;
2684 self->ntinitialized = 0;
2688 self->offsets = NULL;
2685 self->offsets = NULL;
2689
2686
2690 if (!PyArg_ParseTuple(args, "OO", &data_obj, &inlined_obj))
2687 if (!PyArg_ParseTuple(args, "OO", &data_obj, &inlined_obj))
2691 return -1;
2688 return -1;
2692 if (!PyObject_CheckBuffer(data_obj)) {
2689 if (!PyObject_CheckBuffer(data_obj)) {
2693 PyErr_SetString(PyExc_TypeError,
2690 PyErr_SetString(PyExc_TypeError,
2694 "data does not support buffer interface");
2691 "data does not support buffer interface");
2695 return -1;
2692 return -1;
2696 }
2693 }
2697
2694
2698 if (PyObject_GetBuffer(data_obj, &self->buf, PyBUF_SIMPLE) == -1)
2695 if (PyObject_GetBuffer(data_obj, &self->buf, PyBUF_SIMPLE) == -1)
2699 return -1;
2696 return -1;
2700 size = self->buf.len;
2697 size = self->buf.len;
2701
2698
2702 self->inlined = inlined_obj && PyObject_IsTrue(inlined_obj);
2699 self->inlined = inlined_obj && PyObject_IsTrue(inlined_obj);
2703 self->data = data_obj;
2700 self->data = data_obj;
2704
2701
2705 self->ntlookups = self->ntmisses = 0;
2702 self->ntlookups = self->ntmisses = 0;
2706 self->ntrev = -1;
2703 self->ntrev = -1;
2707 Py_INCREF(self->data);
2704 Py_INCREF(self->data);
2708
2705
2709 if (self->inlined) {
2706 if (self->inlined) {
2710 Py_ssize_t len = inline_scan(self, NULL);
2707 Py_ssize_t len = inline_scan(self, NULL);
2711 if (len == -1)
2708 if (len == -1)
2712 goto bail;
2709 goto bail;
2713 self->raw_length = len;
2710 self->raw_length = len;
2714 self->length = len;
2711 self->length = len;
2715 } else {
2712 } else {
2716 if (size % v1_hdrsize) {
2713 if (size % v1_hdrsize) {
2717 PyErr_SetString(PyExc_ValueError, "corrupt index file");
2714 PyErr_SetString(PyExc_ValueError, "corrupt index file");
2718 goto bail;
2715 goto bail;
2719 }
2716 }
2720 self->raw_length = size / v1_hdrsize;
2717 self->raw_length = size / v1_hdrsize;
2721 self->length = self->raw_length;
2718 self->length = self->raw_length;
2722 }
2719 }
2723
2720
2724 return 0;
2721 return 0;
2725 bail:
2722 bail:
2726 return -1;
2723 return -1;
2727 }
2724 }
2728
2725
2729 static PyObject *index_nodemap(indexObject *self)
2726 static PyObject *index_nodemap(indexObject *self)
2730 {
2727 {
2731 Py_INCREF(self);
2728 Py_INCREF(self);
2732 return (PyObject *)self;
2729 return (PyObject *)self;
2733 }
2730 }
2734
2731
2735 static void _index_clearcaches(indexObject *self)
2732 static void _index_clearcaches(indexObject *self)
2736 {
2733 {
2737 if (self->cache) {
2734 if (self->cache) {
2738 Py_ssize_t i;
2735 Py_ssize_t i;
2739
2736
2740 for (i = 0; i < self->raw_length; i++)
2737 for (i = 0; i < self->raw_length; i++)
2741 Py_CLEAR(self->cache[i]);
2738 Py_CLEAR(self->cache[i]);
2742 free(self->cache);
2739 free(self->cache);
2743 self->cache = NULL;
2740 self->cache = NULL;
2744 }
2741 }
2745 if (self->offsets) {
2742 if (self->offsets) {
2746 PyMem_Free((void *)self->offsets);
2743 PyMem_Free((void *)self->offsets);
2747 self->offsets = NULL;
2744 self->offsets = NULL;
2748 }
2745 }
2749 if (self->ntinitialized) {
2746 if (self->ntinitialized) {
2750 nt_dealloc(&self->nt);
2747 nt_dealloc(&self->nt);
2751 }
2748 }
2752 self->ntinitialized = 0;
2749 self->ntinitialized = 0;
2753 Py_CLEAR(self->headrevs);
2750 Py_CLEAR(self->headrevs);
2754 }
2751 }
2755
2752
2756 static PyObject *index_clearcaches(indexObject *self)
2753 static PyObject *index_clearcaches(indexObject *self)
2757 {
2754 {
2758 _index_clearcaches(self);
2755 _index_clearcaches(self);
2759 self->ntrev = -1;
2756 self->ntrev = -1;
2760 self->ntlookups = self->ntmisses = 0;
2757 self->ntlookups = self->ntmisses = 0;
2761 Py_RETURN_NONE;
2758 Py_RETURN_NONE;
2762 }
2759 }
2763
2760
2764 static void index_dealloc(indexObject *self)
2761 static void index_dealloc(indexObject *self)
2765 {
2762 {
2766 _index_clearcaches(self);
2763 _index_clearcaches(self);
2767 Py_XDECREF(self->filteredrevs);
2764 Py_XDECREF(self->filteredrevs);
2768 if (self->buf.buf) {
2765 if (self->buf.buf) {
2769 PyBuffer_Release(&self->buf);
2766 PyBuffer_Release(&self->buf);
2770 memset(&self->buf, 0, sizeof(self->buf));
2767 memset(&self->buf, 0, sizeof(self->buf));
2771 }
2768 }
2772 Py_XDECREF(self->data);
2769 Py_XDECREF(self->data);
2773 Py_XDECREF(self->added);
2770 Py_XDECREF(self->added);
2774 PyObject_Del(self);
2771 PyObject_Del(self);
2775 }
2772 }
2776
2773
2777 static PySequenceMethods index_sequence_methods = {
2774 static PySequenceMethods index_sequence_methods = {
2778 (lenfunc)index_length, /* sq_length */
2775 (lenfunc)index_length, /* sq_length */
2779 0, /* sq_concat */
2776 0, /* sq_concat */
2780 0, /* sq_repeat */
2777 0, /* sq_repeat */
2781 (ssizeargfunc)index_get, /* sq_item */
2778 (ssizeargfunc)index_get, /* sq_item */
2782 0, /* sq_slice */
2779 0, /* sq_slice */
2783 0, /* sq_ass_item */
2780 0, /* sq_ass_item */
2784 0, /* sq_ass_slice */
2781 0, /* sq_ass_slice */
2785 (objobjproc)index_contains, /* sq_contains */
2782 (objobjproc)index_contains, /* sq_contains */
2786 };
2783 };
2787
2784
2788 static PyMappingMethods index_mapping_methods = {
2785 static PyMappingMethods index_mapping_methods = {
2789 (lenfunc)index_length, /* mp_length */
2786 (lenfunc)index_length, /* mp_length */
2790 (binaryfunc)index_getitem, /* mp_subscript */
2787 (binaryfunc)index_getitem, /* mp_subscript */
2791 (objobjargproc)index_assign_subscript, /* mp_ass_subscript */
2788 (objobjargproc)index_assign_subscript, /* mp_ass_subscript */
2792 };
2789 };
2793
2790
2794 static PyMethodDef index_methods[] = {
2791 static PyMethodDef index_methods[] = {
2795 {"ancestors", (PyCFunction)index_ancestors, METH_VARARGS,
2792 {"ancestors", (PyCFunction)index_ancestors, METH_VARARGS,
2796 "return the gca set of the given revs"},
2793 "return the gca set of the given revs"},
2797 {"commonancestorsheads", (PyCFunction)index_commonancestorsheads,
2794 {"commonancestorsheads", (PyCFunction)index_commonancestorsheads,
2798 METH_VARARGS,
2795 METH_VARARGS,
2799 "return the heads of the common ancestors of the given revs"},
2796 "return the heads of the common ancestors of the given revs"},
2800 {"clearcaches", (PyCFunction)index_clearcaches, METH_NOARGS,
2797 {"clearcaches", (PyCFunction)index_clearcaches, METH_NOARGS,
2801 "clear the index caches"},
2798 "clear the index caches"},
2802 {"get", (PyCFunction)index_m_get, METH_VARARGS, "get an index entry"},
2799 {"get", (PyCFunction)index_m_get, METH_VARARGS, "get an index entry"},
2803 {"get_rev", (PyCFunction)index_m_get, METH_VARARGS,
2800 {"get_rev", (PyCFunction)index_m_get, METH_VARARGS,
2804 "return `rev` associated with a node or None"},
2801 "return `rev` associated with a node or None"},
2805 {"has_node", (PyCFunction)index_m_has_node, METH_O,
2802 {"has_node", (PyCFunction)index_m_has_node, METH_O,
2806 "return True if the node exist in the index"},
2803 "return True if the node exist in the index"},
2807 {"rev", (PyCFunction)index_m_rev, METH_O,
2804 {"rev", (PyCFunction)index_m_rev, METH_O,
2808 "return `rev` associated with a node or raise RevlogError"},
2805 "return `rev` associated with a node or raise RevlogError"},
2809 {"computephasesmapsets", (PyCFunction)compute_phases_map_sets, METH_VARARGS,
2806 {"computephasesmapsets", (PyCFunction)compute_phases_map_sets, METH_VARARGS,
2810 "compute phases"},
2807 "compute phases"},
2811 {"reachableroots2", (PyCFunction)reachableroots2, METH_VARARGS,
2808 {"reachableroots2", (PyCFunction)reachableroots2, METH_VARARGS,
2812 "reachableroots"},
2809 "reachableroots"},
2813 {"headrevs", (PyCFunction)index_headrevs, METH_VARARGS,
2810 {"headrevs", (PyCFunction)index_headrevs, METH_VARARGS,
2814 "get head revisions"}, /* Can do filtering since 3.2 */
2811 "get head revisions"}, /* Can do filtering since 3.2 */
2815 {"headrevsfiltered", (PyCFunction)index_headrevs, METH_VARARGS,
2812 {"headrevsfiltered", (PyCFunction)index_headrevs, METH_VARARGS,
2816 "get filtered head revisions"}, /* Can always do filtering */
2813 "get filtered head revisions"}, /* Can always do filtering */
2817 {"issnapshot", (PyCFunction)index_issnapshot, METH_O,
2814 {"issnapshot", (PyCFunction)index_issnapshot, METH_O,
2818 "True if the object is a snapshot"},
2815 "True if the object is a snapshot"},
2819 {"findsnapshots", (PyCFunction)index_findsnapshots, METH_VARARGS,
2816 {"findsnapshots", (PyCFunction)index_findsnapshots, METH_VARARGS,
2820 "Gather snapshot data in a cache dict"},
2817 "Gather snapshot data in a cache dict"},
2821 {"deltachain", (PyCFunction)index_deltachain, METH_VARARGS,
2818 {"deltachain", (PyCFunction)index_deltachain, METH_VARARGS,
2822 "determine revisions with deltas to reconstruct fulltext"},
2819 "determine revisions with deltas to reconstruct fulltext"},
2823 {"slicechunktodensity", (PyCFunction)index_slicechunktodensity,
2820 {"slicechunktodensity", (PyCFunction)index_slicechunktodensity,
2824 METH_VARARGS, "determine revisions with deltas to reconstruct fulltext"},
2821 METH_VARARGS, "determine revisions with deltas to reconstruct fulltext"},
2825 {"append", (PyCFunction)index_append, METH_O, "append an index entry"},
2822 {"append", (PyCFunction)index_append, METH_O, "append an index entry"},
2826 {"partialmatch", (PyCFunction)index_partialmatch, METH_VARARGS,
2823 {"partialmatch", (PyCFunction)index_partialmatch, METH_VARARGS,
2827 "match a potentially ambiguous node ID"},
2824 "match a potentially ambiguous node ID"},
2828 {"shortest", (PyCFunction)index_shortest, METH_VARARGS,
2825 {"shortest", (PyCFunction)index_shortest, METH_VARARGS,
2829 "find length of shortest hex nodeid of a binary ID"},
2826 "find length of shortest hex nodeid of a binary ID"},
2830 {"stats", (PyCFunction)index_stats, METH_NOARGS, "stats for the index"},
2827 {"stats", (PyCFunction)index_stats, METH_NOARGS, "stats for the index"},
2831 {NULL} /* Sentinel */
2828 {NULL} /* Sentinel */
2832 };
2829 };
2833
2830
2834 static PyGetSetDef index_getset[] = {
2831 static PyGetSetDef index_getset[] = {
2835 {"nodemap", (getter)index_nodemap, NULL, "nodemap", NULL},
2832 {"nodemap", (getter)index_nodemap, NULL, "nodemap", NULL},
2836 {NULL} /* Sentinel */
2833 {NULL} /* Sentinel */
2837 };
2834 };
2838
2835
2839 PyTypeObject HgRevlogIndex_Type = {
2836 PyTypeObject HgRevlogIndex_Type = {
2840 PyVarObject_HEAD_INIT(NULL, 0) /* header */
2837 PyVarObject_HEAD_INIT(NULL, 0) /* header */
2841 "parsers.index", /* tp_name */
2838 "parsers.index", /* tp_name */
2842 sizeof(indexObject), /* tp_basicsize */
2839 sizeof(indexObject), /* tp_basicsize */
2843 0, /* tp_itemsize */
2840 0, /* tp_itemsize */
2844 (destructor)index_dealloc, /* tp_dealloc */
2841 (destructor)index_dealloc, /* tp_dealloc */
2845 0, /* tp_print */
2842 0, /* tp_print */
2846 0, /* tp_getattr */
2843 0, /* tp_getattr */
2847 0, /* tp_setattr */
2844 0, /* tp_setattr */
2848 0, /* tp_compare */
2845 0, /* tp_compare */
2849 0, /* tp_repr */
2846 0, /* tp_repr */
2850 0, /* tp_as_number */
2847 0, /* tp_as_number */
2851 &index_sequence_methods, /* tp_as_sequence */
2848 &index_sequence_methods, /* tp_as_sequence */
2852 &index_mapping_methods, /* tp_as_mapping */
2849 &index_mapping_methods, /* tp_as_mapping */
2853 0, /* tp_hash */
2850 0, /* tp_hash */
2854 0, /* tp_call */
2851 0, /* tp_call */
2855 0, /* tp_str */
2852 0, /* tp_str */
2856 0, /* tp_getattro */
2853 0, /* tp_getattro */
2857 0, /* tp_setattro */
2854 0, /* tp_setattro */
2858 0, /* tp_as_buffer */
2855 0, /* tp_as_buffer */
2859 Py_TPFLAGS_DEFAULT, /* tp_flags */
2856 Py_TPFLAGS_DEFAULT, /* tp_flags */
2860 "revlog index", /* tp_doc */
2857 "revlog index", /* tp_doc */
2861 0, /* tp_traverse */
2858 0, /* tp_traverse */
2862 0, /* tp_clear */
2859 0, /* tp_clear */
2863 0, /* tp_richcompare */
2860 0, /* tp_richcompare */
2864 0, /* tp_weaklistoffset */
2861 0, /* tp_weaklistoffset */
2865 0, /* tp_iter */
2862 0, /* tp_iter */
2866 0, /* tp_iternext */
2863 0, /* tp_iternext */
2867 index_methods, /* tp_methods */
2864 index_methods, /* tp_methods */
2868 0, /* tp_members */
2865 0, /* tp_members */
2869 index_getset, /* tp_getset */
2866 index_getset, /* tp_getset */
2870 0, /* tp_base */
2867 0, /* tp_base */
2871 0, /* tp_dict */
2868 0, /* tp_dict */
2872 0, /* tp_descr_get */
2869 0, /* tp_descr_get */
2873 0, /* tp_descr_set */
2870 0, /* tp_descr_set */
2874 0, /* tp_dictoffset */
2871 0, /* tp_dictoffset */
2875 (initproc)index_init, /* tp_init */
2872 (initproc)index_init, /* tp_init */
2876 0, /* tp_alloc */
2873 0, /* tp_alloc */
2877 };
2874 };
2878
2875
2879 /*
2876 /*
2880 * returns a tuple of the form (index, index, cache) with elements as
2877 * returns a tuple of the form (index, index, cache) with elements as
2881 * follows:
2878 * follows:
2882 *
2879 *
2883 * index: an index object that lazily parses RevlogNG records
2880 * index: an index object that lazily parses RevlogNG records
2884 * cache: if data is inlined, a tuple (0, index_file_content), else None
2881 * cache: if data is inlined, a tuple (0, index_file_content), else None
2885 * index_file_content could be a string, or a buffer
2882 * index_file_content could be a string, or a buffer
2886 *
2883 *
2887 * added complications are for backwards compatibility
2884 * added complications are for backwards compatibility
2888 */
2885 */
2889 PyObject *parse_index2(PyObject *self, PyObject *args)
2886 PyObject *parse_index2(PyObject *self, PyObject *args)
2890 {
2887 {
2891 PyObject *tuple = NULL, *cache = NULL;
2888 PyObject *tuple = NULL, *cache = NULL;
2892 indexObject *idx;
2889 indexObject *idx;
2893 int ret;
2890 int ret;
2894
2891
2895 idx = PyObject_New(indexObject, &HgRevlogIndex_Type);
2892 idx = PyObject_New(indexObject, &HgRevlogIndex_Type);
2896 if (idx == NULL)
2893 if (idx == NULL)
2897 goto bail;
2894 goto bail;
2898
2895
2899 ret = index_init(idx, args);
2896 ret = index_init(idx, args);
2900 if (ret == -1)
2897 if (ret == -1)
2901 goto bail;
2898 goto bail;
2902
2899
2903 if (idx->inlined) {
2900 if (idx->inlined) {
2904 cache = Py_BuildValue("iO", 0, idx->data);
2901 cache = Py_BuildValue("iO", 0, idx->data);
2905 if (cache == NULL)
2902 if (cache == NULL)
2906 goto bail;
2903 goto bail;
2907 } else {
2904 } else {
2908 cache = Py_None;
2905 cache = Py_None;
2909 Py_INCREF(cache);
2906 Py_INCREF(cache);
2910 }
2907 }
2911
2908
2912 tuple = Py_BuildValue("NN", idx, cache);
2909 tuple = Py_BuildValue("NN", idx, cache);
2913 if (!tuple)
2910 if (!tuple)
2914 goto bail;
2911 goto bail;
2915 return tuple;
2912 return tuple;
2916
2913
2917 bail:
2914 bail:
2918 Py_XDECREF(idx);
2915 Py_XDECREF(idx);
2919 Py_XDECREF(cache);
2916 Py_XDECREF(cache);
2920 Py_XDECREF(tuple);
2917 Py_XDECREF(tuple);
2921 return NULL;
2918 return NULL;
2922 }
2919 }
2923
2920
2924 static Revlog_CAPI CAPI = {
2921 static Revlog_CAPI CAPI = {
2925 /* increment the abi_version field upon each change in the Revlog_CAPI
2922 /* increment the abi_version field upon each change in the Revlog_CAPI
2926 struct or in the ABI of the listed functions */
2923 struct or in the ABI of the listed functions */
2927 2,
2924 2,
2928 index_length,
2925 index_length,
2929 index_node,
2926 index_node,
2930 HgRevlogIndex_GetParents,
2927 HgRevlogIndex_GetParents,
2931 };
2928 };
2932
2929
2933 void revlog_module_init(PyObject *mod)
2930 void revlog_module_init(PyObject *mod)
2934 {
2931 {
2935 PyObject *caps = NULL;
2932 PyObject *caps = NULL;
2936 HgRevlogIndex_Type.tp_new = PyType_GenericNew;
2933 HgRevlogIndex_Type.tp_new = PyType_GenericNew;
2937 if (PyType_Ready(&HgRevlogIndex_Type) < 0)
2934 if (PyType_Ready(&HgRevlogIndex_Type) < 0)
2938 return;
2935 return;
2939 Py_INCREF(&HgRevlogIndex_Type);
2936 Py_INCREF(&HgRevlogIndex_Type);
2940 PyModule_AddObject(mod, "index", (PyObject *)&HgRevlogIndex_Type);
2937 PyModule_AddObject(mod, "index", (PyObject *)&HgRevlogIndex_Type);
2941
2938
2942 nodetreeType.tp_new = PyType_GenericNew;
2939 nodetreeType.tp_new = PyType_GenericNew;
2943 if (PyType_Ready(&nodetreeType) < 0)
2940 if (PyType_Ready(&nodetreeType) < 0)
2944 return;
2941 return;
2945 Py_INCREF(&nodetreeType);
2942 Py_INCREF(&nodetreeType);
2946 PyModule_AddObject(mod, "nodetree", (PyObject *)&nodetreeType);
2943 PyModule_AddObject(mod, "nodetree", (PyObject *)&nodetreeType);
2947
2944
2948 if (!nullentry) {
2945 if (!nullentry) {
2949 nullentry =
2946 nullentry =
2950 Py_BuildValue(PY23("iiiiiiis#", "iiiiiiiy#"), 0, 0, 0, -1,
2947 Py_BuildValue(PY23("iiiiiiis#", "iiiiiiiy#"), 0, 0, 0, -1,
2951 -1, -1, -1, nullid, (Py_ssize_t)20);
2948 -1, -1, -1, nullid, (Py_ssize_t)20);
2952 }
2949 }
2953 if (nullentry)
2950 if (nullentry)
2954 PyObject_GC_UnTrack(nullentry);
2951 PyObject_GC_UnTrack(nullentry);
2955
2952
2956 caps = PyCapsule_New(&CAPI, "mercurial.cext.parsers.revlog_CAPI", NULL);
2953 caps = PyCapsule_New(&CAPI, "mercurial.cext.parsers.revlog_CAPI", NULL);
2957 if (caps != NULL)
2954 if (caps != NULL)
2958 PyModule_AddObject(mod, "revlog_CAPI", caps);
2955 PyModule_AddObject(mod, "revlog_CAPI", caps);
2959 }
2956 }
General Comments 0
You need to be logged in to leave comments. Login now