Mercurial > public > mercurial-scm > hg
annotate mercurial/mpatch.c @ 472:aa3d592df9b9
extensions: use stdint.h
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
extensions: use stdint.h
Not sure why I didn't do this the first time around. Hopefully still
builds everywhere.
manifest hash: 965582286a190728f8cc0dfb8e11ee56628a59a5
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
iD8DBQFCvfRgywK+sNU5EO8RAg9SAJ4/ZVpQZcDY5xovLDTZK2txEegEgwCdF2b+
lzSIP109qq8D+KIdUWsbEPc=
=+0Yy
-----END PGP SIGNATURE-----
| author | mpm@selenic.com |
|---|---|
| date | Sat, 25 Jun 2005 16:18:40 -0800 |
| parents | 50da4bb9cab6 |
| children | 098d1f039c18 e94cebc60d96 |
| rev | line source |
|---|---|
| 72 | 1 /* |
| 2 mpatch.c - efficient binary patching for Mercurial | |
| 3 | |
| 4 This implements a patch algorithm that's O(m + nlog n) where m is the | |
| 5 size of the output and n is the number of patches. | |
| 6 | |
| 7 Given a list of binary patches, it unpacks each into a hunk list, | |
| 8 then combines the hunk lists with a treewise recursion to form a | |
| 9 single hunk list. This hunk list is then applied to the original | |
| 10 text. | |
| 11 | |
| 12 The text (or binary) fragments are copied directly from their source | |
| 13 Python objects into a preallocated output string to avoid the | |
| 14 allocation of intermediate Python objects. Working memory is about 2x | |
| 15 the total number of hunks. | |
| 16 | |
| 17 Copyright 2005 Matt Mackall <mpm@selenic.com> | |
| 18 | |
| 19 This software may be used and distributed according to the terms | |
| 20 of the GNU General Public License, incorporated herein by reference. | |
| 21 */ | |
| 22 | |
| 23 #include <Python.h> | |
| 24 #include <stdlib.h> | |
| 25 #include <string.h> | |
| 472 | 26 #include <stdint.h> |
|
410
7c678976df3e
Make mpatch.c compilable under the other `OS'
mpm@selenic.com
parents:
384
diff
changeset
|
27 #ifdef _WIN32 |
|
411
9e9f7ab43ce2
Add 'other OS' bits to bdiff.c / style cleanups
mpm@selenic.com
parents:
410
diff
changeset
|
28 static uint32_t ntohl(uint32_t x) |
|
9e9f7ab43ce2
Add 'other OS' bits to bdiff.c / style cleanups
mpm@selenic.com
parents:
410
diff
changeset
|
29 { |
|
9e9f7ab43ce2
Add 'other OS' bits to bdiff.c / style cleanups
mpm@selenic.com
parents:
410
diff
changeset
|
30 return ((x & 0x000000ffUL) << 24) | |
|
9e9f7ab43ce2
Add 'other OS' bits to bdiff.c / style cleanups
mpm@selenic.com
parents:
410
diff
changeset
|
31 ((x & 0x0000ff00UL) << 8) | |
|
9e9f7ab43ce2
Add 'other OS' bits to bdiff.c / style cleanups
mpm@selenic.com
parents:
410
diff
changeset
|
32 ((x & 0x00ff0000UL) >> 8) | |
|
9e9f7ab43ce2
Add 'other OS' bits to bdiff.c / style cleanups
mpm@selenic.com
parents:
410
diff
changeset
|
33 ((x & 0xff000000UL) >> 24); |
|
410
7c678976df3e
Make mpatch.c compilable under the other `OS'
mpm@selenic.com
parents:
384
diff
changeset
|
34 } |
|
7c678976df3e
Make mpatch.c compilable under the other `OS'
mpm@selenic.com
parents:
384
diff
changeset
|
35 #else |
| 472 | 36 #include <netinet/in.h> |
|
410
7c678976df3e
Make mpatch.c compilable under the other `OS'
mpm@selenic.com
parents:
384
diff
changeset
|
37 #endif |
| 72 | 38 |
| 39 static char mpatch_doc[] = "Efficient binary patching."; | |
| 40 | |
| 41 struct frag { | |
| 42 int start, end, len; | |
| 43 char *data; | |
| 44 }; | |
| 45 | |
| 46 struct flist { | |
| 47 struct frag *base, *head, *tail; | |
| 48 }; | |
| 49 | |
| 50 static struct flist *lalloc(int size) | |
| 51 { | |
| 128 | 52 struct flist *a = NULL; |
| 72 | 53 |
| 54 a = malloc(sizeof(struct flist)); | |
| 128 | 55 if (a) { |
| 56 a->base = malloc(sizeof(struct frag) * size); | |
| 282 | 57 if (!a->base) { |
| 128 | 58 free(a); |
| 282 | 59 a = NULL; |
| 60 } else | |
| 128 | 61 a->head = a->tail = a->base; |
| 62 } | |
| 72 | 63 return a; |
| 64 } | |
| 65 | |
| 66 static void lfree(struct flist *a) | |
| 67 { | |
| 128 | 68 if (a) { |
| 69 free(a->base); | |
| 70 free(a); | |
| 71 } | |
| 72 | 72 } |
| 73 | |
| 74 static int lsize(struct flist *a) | |
| 75 { | |
| 76 return a->tail - a->head; | |
| 77 } | |
| 78 | |
| 79 /* move hunks in source that are less cut to dest, compensating | |
| 80 for changes in offset. the last hunk may be split if necessary. | |
| 81 */ | |
| 82 static int gather(struct flist *dest, struct flist *src, int cut, int offset) | |
| 83 { | |
| 84 struct frag *d = dest->tail, *s = src->head; | |
| 85 int postend, c, l; | |
| 86 | |
| 87 while (s != src->tail) { | |
| 88 if (s->start + offset >= cut) | |
| 82 | 89 break; /* we've gone far enough */ |
| 72 | 90 |
| 91 postend = offset + s->start + s->len; | |
| 92 if (postend <= cut) { | |
| 93 /* save this hunk */ | |
| 94 offset += s->start + s->len - s->end; | |
| 95 *d++ = *s++; | |
| 96 } | |
| 97 else { | |
| 98 /* break up this hunk */ | |
| 99 c = cut - offset; | |
| 100 if (s->end < c) | |
| 101 c = s->end; | |
| 102 l = cut - offset - s->start; | |
| 103 if (s->len < l) | |
| 104 l = s->len; | |
| 105 | |
| 106 offset += s->start + l - c; | |
| 107 | |
| 108 d->start = s->start; | |
| 109 d->end = c; | |
| 110 d->len = l; | |
| 111 d->data = s->data; | |
| 112 d++; | |
| 113 s->start = c; | |
| 114 s->len = s->len - l; | |
| 115 s->data = s->data + l; | |
| 116 | |
| 82 | 117 break; |
| 72 | 118 } |
| 119 } | |
| 120 | |
| 121 dest->tail = d; | |
| 122 src->head = s; | |
| 123 return offset; | |
| 124 } | |
| 125 | |
| 126 /* like gather, but with no output list */ | |
| 127 static int discard(struct flist *src, int cut, int offset) | |
| 128 { | |
| 129 struct frag *s = src->head; | |
| 130 int postend, c, l; | |
| 131 | |
| 132 while (s != src->tail) { | |
| 133 if (s->start + offset >= cut) | |
| 82 | 134 break; |
| 72 | 135 |
| 136 postend = offset + s->start + s->len; | |
| 137 if (postend <= cut) { | |
| 138 offset += s->start + s->len - s->end; | |
| 139 s++; | |
| 140 } | |
| 141 else { | |
| 142 c = cut - offset; | |
| 143 if (s->end < c) | |
| 144 c = s->end; | |
| 145 l = cut - offset - s->start; | |
| 146 if (s->len < l) | |
| 147 l = s->len; | |
| 148 | |
| 149 offset += s->start + l - c; | |
| 150 s->start = c; | |
| 151 s->len = s->len - l; | |
| 152 s->data = s->data + l; | |
| 153 | |
| 82 | 154 break; |
| 72 | 155 } |
| 156 } | |
| 157 | |
| 158 src->head = s; | |
| 159 return offset; | |
| 160 } | |
| 161 | |
| 162 /* combine hunk lists a and b, while adjusting b for offset changes in a/ | |
| 163 this deletes a and b and returns the resultant list. */ | |
| 164 static struct flist *combine(struct flist *a, struct flist *b) | |
| 165 { | |
| 128 | 166 struct flist *c = NULL; |
| 167 struct frag *bh, *ct; | |
| 72 | 168 int offset = 0, post; |
| 169 | |
| 128 | 170 if (a && b) |
| 171 c = lalloc((lsize(a) + lsize(b)) * 2); | |
| 172 | |
| 173 if (c) { | |
| 72 | 174 |
| 128 | 175 for (bh = b->head; bh != b->tail; bh++) { |
| 176 /* save old hunks */ | |
| 177 offset = gather(c, a, bh->start, offset); | |
| 72 | 178 |
| 128 | 179 /* discard replaced hunks */ |
| 180 post = discard(a, bh->end, offset); | |
| 72 | 181 |
| 128 | 182 /* insert new hunk */ |
| 183 ct = c->tail; | |
| 184 ct->start = bh->start - offset; | |
| 185 ct->end = bh->end - post; | |
| 186 ct->len = bh->len; | |
| 187 ct->data = bh->data; | |
| 188 c->tail++; | |
| 189 offset = post; | |
| 190 } | |
| 191 | |
| 192 /* hold on to tail from a */ | |
| 193 memcpy(c->tail, a->head, sizeof(struct frag) * lsize(a)); | |
| 194 c->tail += lsize(a); | |
| 72 | 195 } |
| 196 | |
| 197 lfree(a); | |
| 198 lfree(b); | |
| 199 return c; | |
| 200 } | |
| 201 | |
| 202 /* decode a binary patch into a hunk list */ | |
| 203 static struct flist *decode(char *bin, int len) | |
| 204 { | |
| 205 struct flist *l; | |
| 206 struct frag *lt; | |
| 207 char *end = bin + len; | |
|
384
a29decbf7475
mpatch: attempt to handle unpack alignment issues on Solaris
mpm@selenic.com
parents:
282
diff
changeset
|
208 char decode[12]; /* for dealing with alignment issues */ |
| 72 | 209 |
| 210 /* assume worst case size, we won't have many of these lists */ | |
| 211 l = lalloc(len / 12); | |
| 212 lt = l->tail; | |
| 213 | |
| 214 while (bin < end) { | |
|
384
a29decbf7475
mpatch: attempt to handle unpack alignment issues on Solaris
mpm@selenic.com
parents:
282
diff
changeset
|
215 memcpy(decode, bin, 12); |
|
a29decbf7475
mpatch: attempt to handle unpack alignment issues on Solaris
mpm@selenic.com
parents:
282
diff
changeset
|
216 lt->start = ntohl(*(uint32_t *)decode); |
|
a29decbf7475
mpatch: attempt to handle unpack alignment issues on Solaris
mpm@selenic.com
parents:
282
diff
changeset
|
217 lt->end = ntohl(*(uint32_t *)(decode + 4)); |
|
a29decbf7475
mpatch: attempt to handle unpack alignment issues on Solaris
mpm@selenic.com
parents:
282
diff
changeset
|
218 lt->len = ntohl(*(uint32_t *)(decode + 8)); |
| 72 | 219 lt->data = bin + 12; |
| 220 bin += 12 + lt->len; | |
| 221 lt++; | |
| 222 } | |
| 223 | |
| 224 l->tail = lt; | |
| 225 return l; | |
| 226 } | |
| 227 | |
| 228 /* calculate the size of resultant text */ | |
| 229 static int calcsize(int len, struct flist *l) | |
| 230 { | |
| 231 int outlen = 0, last = 0; | |
| 232 struct frag *f = l->head; | |
| 233 | |
| 234 while (f != l->tail) { | |
| 235 outlen += f->start - last; | |
| 236 last = f->end; | |
| 237 outlen += f->len; | |
| 238 f++; | |
| 239 } | |
| 240 | |
| 241 outlen += len - last; | |
| 242 return outlen; | |
| 243 } | |
| 244 | |
| 245 static void apply(char *buf, char *orig, int len, struct flist *l) | |
| 246 { | |
| 247 struct frag *f = l->head; | |
| 248 int last = 0; | |
| 249 char *p = buf; | |
| 250 | |
| 251 while (f != l->tail) { | |
| 252 memcpy(p, orig + last, f->start - last); | |
| 253 p += f->start - last; | |
| 254 memcpy(p, f->data, f->len); | |
| 255 last = f->end; | |
| 256 p += f->len; | |
| 257 f++; | |
| 258 } | |
| 259 memcpy(p, orig + last, len - last); | |
| 260 } | |
| 261 | |
| 262 /* recursively generate a patch of all bins between start and end */ | |
| 263 static struct flist *fold(PyObject *bins, int start, int end) | |
| 264 { | |
| 265 int len; | |
| 266 | |
| 267 if (start + 1 == end) { | |
| 268 /* trivial case, output a decoded list */ | |
| 269 PyObject *tmp = PyList_GetItem(bins, start); | |
| 128 | 270 if (!tmp) |
| 271 return NULL; | |
| 72 | 272 return decode(PyString_AsString(tmp), PyString_Size(tmp)); |
| 273 } | |
| 274 | |
| 275 /* divide and conquer, memory management is elsewhere */ | |
| 276 len = (end - start) / 2; | |
| 277 return combine(fold(bins, start, start + len), | |
| 278 fold(bins, start + len, end)); | |
| 279 } | |
| 280 | |
| 281 static PyObject * | |
| 282 patches(PyObject *self, PyObject *args) | |
| 283 { | |
| 284 PyObject *text, *bins, *result; | |
| 285 struct flist *patch; | |
| 286 char *in, *out; | |
| 287 int len, outlen; | |
| 288 | |
| 128 | 289 if (!PyArg_ParseTuple(args, "SO:mpatch", &text, &bins)) |
| 72 | 290 return NULL; |
| 291 | |
| 292 len = PyList_Size(bins); | |
| 293 if (!len) { | |
| 294 /* nothing to do */ | |
| 295 Py_INCREF(text); | |
| 296 return text; | |
| 297 } | |
| 298 | |
| 299 patch = fold(bins, 0, len); | |
| 128 | 300 if (!patch) |
| 301 return PyErr_NoMemory(); | |
| 302 | |
| 72 | 303 outlen = calcsize(PyString_Size(text), patch); |
| 304 result = PyString_FromStringAndSize(NULL, outlen); | |
| 128 | 305 if (result) { |
| 306 in = PyString_AsString(text); | |
| 307 out = PyString_AsString(result); | |
| 308 apply(out, in, PyString_Size(text), patch); | |
| 309 } | |
| 310 | |
| 72 | 311 lfree(patch); |
| 312 return result; | |
| 313 } | |
| 314 | |
| 315 static PyMethodDef methods[] = { | |
| 316 {"patches", patches, METH_VARARGS, "apply a series of patches\n"}, | |
| 317 {NULL, NULL} | |
| 318 }; | |
| 319 | |
| 320 PyMODINIT_FUNC | |
| 321 initmpatch(void) | |
| 322 { | |
| 323 Py_InitModule3("mpatch", methods, mpatch_doc); | |
| 324 } | |
| 325 |