Merge branch 'jt/repack-promisor-packs'
[git/git.git] / builtin / repack.c
1 #include "builtin.h"
2 #include "cache.h"
3 #include "config.h"
4 #include "dir.h"
5 #include "parse-options.h"
6 #include "run-command.h"
7 #include "sigchain.h"
8 #include "strbuf.h"
9 #include "string-list.h"
10 #include "argv-array.h"
11 #include "packfile.h"
12 #include "object-store.h"
13
14 static int delta_base_offset = 1;
15 static int pack_kept_objects = -1;
16 static int write_bitmaps;
17 static char *packdir, *packtmp;
18
19 static const char *const git_repack_usage[] = {
20 N_("git repack [<options>]"),
21 NULL
22 };
23
24 static const char incremental_bitmap_conflict_error[] = N_(
25 "Incremental repacks are incompatible with bitmap indexes. Use\n"
26 "--no-write-bitmap-index or disable the pack.writebitmaps configuration."
27 );
28
29
30 static int repack_config(const char *var, const char *value, void *cb)
31 {
32 if (!strcmp(var, "repack.usedeltabaseoffset")) {
33 delta_base_offset = git_config_bool(var, value);
34 return 0;
35 }
36 if (!strcmp(var, "repack.packkeptobjects")) {
37 pack_kept_objects = git_config_bool(var, value);
38 return 0;
39 }
40 if (!strcmp(var, "repack.writebitmaps") ||
41 !strcmp(var, "pack.writebitmaps")) {
42 write_bitmaps = git_config_bool(var, value);
43 return 0;
44 }
45 return git_default_config(var, value, cb);
46 }
47
48 /*
49 * Remove temporary $GIT_OBJECT_DIRECTORY/pack/.tmp-$$-pack-* files.
50 */
51 static void remove_temporary_files(void)
52 {
53 struct strbuf buf = STRBUF_INIT;
54 size_t dirlen, prefixlen;
55 DIR *dir;
56 struct dirent *e;
57
58 dir = opendir(packdir);
59 if (!dir)
60 return;
61
62 /* Point at the slash at the end of ".../objects/pack/" */
63 dirlen = strlen(packdir) + 1;
64 strbuf_addstr(&buf, packtmp);
65 /* Hold the length of ".tmp-%d-pack-" */
66 prefixlen = buf.len - dirlen;
67
68 while ((e = readdir(dir))) {
69 if (strncmp(e->d_name, buf.buf + dirlen, prefixlen))
70 continue;
71 strbuf_setlen(&buf, dirlen);
72 strbuf_addstr(&buf, e->d_name);
73 unlink(buf.buf);
74 }
75 closedir(dir);
76 strbuf_release(&buf);
77 }
78
79 static void remove_pack_on_signal(int signo)
80 {
81 remove_temporary_files();
82 sigchain_pop(signo);
83 raise(signo);
84 }
85
86 /*
87 * Adds all packs hex strings to the fname list, which do not
88 * have a corresponding .keep file. These packs are not to
89 * be kept if we are going to pack everything into one file.
90 */
91 static void get_non_kept_pack_filenames(struct string_list *fname_list,
92 const struct string_list *extra_keep)
93 {
94 DIR *dir;
95 struct dirent *e;
96 char *fname;
97
98 if (!(dir = opendir(packdir)))
99 return;
100
101 while ((e = readdir(dir)) != NULL) {
102 size_t len;
103 int i;
104
105 for (i = 0; i < extra_keep->nr; i++)
106 if (!fspathcmp(e->d_name, extra_keep->items[i].string))
107 break;
108 if (extra_keep->nr > 0 && i < extra_keep->nr)
109 continue;
110
111 if (!strip_suffix(e->d_name, ".pack", &len))
112 continue;
113
114 fname = xmemdupz(e->d_name, len);
115
116 if (!file_exists(mkpath("%s/%s.keep", packdir, fname)))
117 string_list_append_nodup(fname_list, fname);
118 else
119 free(fname);
120 }
121 closedir(dir);
122 }
123
124 static void remove_redundant_pack(const char *dir_name, const char *base_name)
125 {
126 const char *exts[] = {".pack", ".idx", ".keep", ".bitmap", ".promisor"};
127 int i;
128 struct strbuf buf = STRBUF_INIT;
129 size_t plen;
130
131 strbuf_addf(&buf, "%s/%s", dir_name, base_name);
132 plen = buf.len;
133
134 for (i = 0; i < ARRAY_SIZE(exts); i++) {
135 strbuf_setlen(&buf, plen);
136 strbuf_addstr(&buf, exts[i]);
137 unlink(buf.buf);
138 }
139 strbuf_release(&buf);
140 }
141
142 struct pack_objects_args {
143 const char *window;
144 const char *window_memory;
145 const char *depth;
146 const char *threads;
147 const char *max_pack_size;
148 int no_reuse_delta;
149 int no_reuse_object;
150 int quiet;
151 int local;
152 };
153
154 static void prepare_pack_objects(struct child_process *cmd,
155 const struct pack_objects_args *args)
156 {
157 argv_array_push(&cmd->args, "pack-objects");
158 if (args->window)
159 argv_array_pushf(&cmd->args, "--window=%s", args->window);
160 if (args->window_memory)
161 argv_array_pushf(&cmd->args, "--window-memory=%s", args->window_memory);
162 if (args->depth)
163 argv_array_pushf(&cmd->args, "--depth=%s", args->depth);
164 if (args->threads)
165 argv_array_pushf(&cmd->args, "--threads=%s", args->threads);
166 if (args->max_pack_size)
167 argv_array_pushf(&cmd->args, "--max-pack-size=%s", args->max_pack_size);
168 if (args->no_reuse_delta)
169 argv_array_pushf(&cmd->args, "--no-reuse-delta");
170 if (args->no_reuse_object)
171 argv_array_pushf(&cmd->args, "--no-reuse-object");
172 if (args->local)
173 argv_array_push(&cmd->args, "--local");
174 if (args->quiet)
175 argv_array_push(&cmd->args, "--quiet");
176 if (delta_base_offset)
177 argv_array_push(&cmd->args, "--delta-base-offset");
178 argv_array_push(&cmd->args, packtmp);
179 cmd->git_cmd = 1;
180 cmd->out = -1;
181 }
182
183 /*
184 * Write oid to the given struct child_process's stdin, starting it first if
185 * necessary.
186 */
187 static int write_oid(const struct object_id *oid, struct packed_git *pack,
188 uint32_t pos, void *data)
189 {
190 struct child_process *cmd = data;
191
192 if (cmd->in == -1) {
193 if (start_command(cmd))
194 die("Could not start pack-objects to repack promisor objects");
195 }
196
197 xwrite(cmd->in, oid_to_hex(oid), GIT_SHA1_HEXSZ);
198 xwrite(cmd->in, "\n", 1);
199 return 0;
200 }
201
202 static void repack_promisor_objects(const struct pack_objects_args *args,
203 struct string_list *names)
204 {
205 struct child_process cmd = CHILD_PROCESS_INIT;
206 FILE *out;
207 struct strbuf line = STRBUF_INIT;
208
209 prepare_pack_objects(&cmd, args);
210 cmd.in = -1;
211
212 /*
213 * NEEDSWORK: Giving pack-objects only the OIDs without any ordering
214 * hints may result in suboptimal deltas in the resulting pack. See if
215 * the OIDs can be sent with fake paths such that pack-objects can use a
216 * {type -> existing pack order} ordering when computing deltas instead
217 * of a {type -> size} ordering, which may produce better deltas.
218 */
219 for_each_packed_object(write_oid, &cmd,
220 FOR_EACH_OBJECT_PROMISOR_ONLY);
221
222 if (cmd.in == -1)
223 /* No packed objects; cmd was never started */
224 return;
225
226 close(cmd.in);
227
228 out = xfdopen(cmd.out, "r");
229 while (strbuf_getline_lf(&line, out) != EOF) {
230 char *promisor_name;
231 int fd;
232 if (line.len != 40)
233 die("repack: Expecting 40 character sha1 lines only from pack-objects.");
234 string_list_append(names, line.buf);
235
236 /*
237 * pack-objects creates the .pack and .idx files, but not the
238 * .promisor file. Create the .promisor file, which is empty.
239 */
240 promisor_name = mkpathdup("%s-%s.promisor", packtmp,
241 line.buf);
242 fd = open(promisor_name, O_CREAT|O_EXCL|O_WRONLY, 0600);
243 if (fd < 0)
244 die_errno("unable to create '%s'", promisor_name);
245 close(fd);
246 free(promisor_name);
247 }
248 fclose(out);
249 if (finish_command(&cmd))
250 die("Could not finish pack-objects to repack promisor objects");
251 }
252
253 #define ALL_INTO_ONE 1
254 #define LOOSEN_UNREACHABLE 2
255
256 int cmd_repack(int argc, const char **argv, const char *prefix)
257 {
258 struct {
259 const char *name;
260 unsigned optional:1;
261 } exts[] = {
262 {".pack"},
263 {".idx"},
264 {".bitmap", 1},
265 {".promisor", 1},
266 };
267 struct child_process cmd = CHILD_PROCESS_INIT;
268 struct string_list_item *item;
269 struct string_list names = STRING_LIST_INIT_DUP;
270 struct string_list rollback = STRING_LIST_INIT_NODUP;
271 struct string_list existing_packs = STRING_LIST_INIT_DUP;
272 struct strbuf line = STRBUF_INIT;
273 int i, ext, ret, failed;
274 FILE *out;
275
276 /* variables to be filled by option parsing */
277 int pack_everything = 0;
278 int delete_redundant = 0;
279 const char *unpack_unreachable = NULL;
280 int keep_unreachable = 0;
281 struct string_list keep_pack_list = STRING_LIST_INIT_NODUP;
282 int no_update_server_info = 0;
283 struct pack_objects_args po_args = {NULL};
284
285 struct option builtin_repack_options[] = {
286 OPT_BIT('a', NULL, &pack_everything,
287 N_("pack everything in a single pack"), ALL_INTO_ONE),
288 OPT_BIT('A', NULL, &pack_everything,
289 N_("same as -a, and turn unreachable objects loose"),
290 LOOSEN_UNREACHABLE | ALL_INTO_ONE),
291 OPT_BOOL('d', NULL, &delete_redundant,
292 N_("remove redundant packs, and run git-prune-packed")),
293 OPT_BOOL('f', NULL, &po_args.no_reuse_delta,
294 N_("pass --no-reuse-delta to git-pack-objects")),
295 OPT_BOOL('F', NULL, &po_args.no_reuse_object,
296 N_("pass --no-reuse-object to git-pack-objects")),
297 OPT_BOOL('n', NULL, &no_update_server_info,
298 N_("do not run git-update-server-info")),
299 OPT__QUIET(&po_args.quiet, N_("be quiet")),
300 OPT_BOOL('l', "local", &po_args.local,
301 N_("pass --local to git-pack-objects")),
302 OPT_BOOL('b', "write-bitmap-index", &write_bitmaps,
303 N_("write bitmap index")),
304 OPT_STRING(0, "unpack-unreachable", &unpack_unreachable, N_("approxidate"),
305 N_("with -A, do not loosen objects older than this")),
306 OPT_BOOL('k', "keep-unreachable", &keep_unreachable,
307 N_("with -a, repack unreachable objects")),
308 OPT_STRING(0, "window", &po_args.window, N_("n"),
309 N_("size of the window used for delta compression")),
310 OPT_STRING(0, "window-memory", &po_args.window_memory, N_("bytes"),
311 N_("same as the above, but limit memory size instead of entries count")),
312 OPT_STRING(0, "depth", &po_args.depth, N_("n"),
313 N_("limits the maximum delta depth")),
314 OPT_STRING(0, "threads", &po_args.threads, N_("n"),
315 N_("limits the maximum number of threads")),
316 OPT_STRING(0, "max-pack-size", &po_args.max_pack_size, N_("bytes"),
317 N_("maximum size of each packfile")),
318 OPT_BOOL(0, "pack-kept-objects", &pack_kept_objects,
319 N_("repack objects in packs marked with .keep")),
320 OPT_STRING_LIST(0, "keep-pack", &keep_pack_list, N_("name"),
321 N_("do not repack this pack")),
322 OPT_END()
323 };
324
325 git_config(repack_config, NULL);
326
327 argc = parse_options(argc, argv, prefix, builtin_repack_options,
328 git_repack_usage, 0);
329
330 if (delete_redundant && repository_format_precious_objects)
331 die(_("cannot delete packs in a precious-objects repo"));
332
333 if (keep_unreachable &&
334 (unpack_unreachable || (pack_everything & LOOSEN_UNREACHABLE)))
335 die(_("--keep-unreachable and -A are incompatible"));
336
337 if (pack_kept_objects < 0)
338 pack_kept_objects = write_bitmaps;
339
340 if (write_bitmaps && !(pack_everything & ALL_INTO_ONE))
341 die(_(incremental_bitmap_conflict_error));
342
343 packdir = mkpathdup("%s/pack", get_object_directory());
344 packtmp = mkpathdup("%s/.tmp-%d-pack", packdir, (int)getpid());
345
346 sigchain_push_common(remove_pack_on_signal);
347
348 prepare_pack_objects(&cmd, &po_args);
349
350 argv_array_push(&cmd.args, "--keep-true-parents");
351 if (!pack_kept_objects)
352 argv_array_push(&cmd.args, "--honor-pack-keep");
353 for (i = 0; i < keep_pack_list.nr; i++)
354 argv_array_pushf(&cmd.args, "--keep-pack=%s",
355 keep_pack_list.items[i].string);
356 argv_array_push(&cmd.args, "--non-empty");
357 argv_array_push(&cmd.args, "--all");
358 argv_array_push(&cmd.args, "--reflog");
359 argv_array_push(&cmd.args, "--indexed-objects");
360 if (repository_format_partial_clone)
361 argv_array_push(&cmd.args, "--exclude-promisor-objects");
362 if (write_bitmaps)
363 argv_array_push(&cmd.args, "--write-bitmap-index");
364
365 if (pack_everything & ALL_INTO_ONE) {
366 get_non_kept_pack_filenames(&existing_packs, &keep_pack_list);
367
368 repack_promisor_objects(&po_args, &names);
369
370 if (existing_packs.nr && delete_redundant) {
371 if (unpack_unreachable) {
372 argv_array_pushf(&cmd.args,
373 "--unpack-unreachable=%s",
374 unpack_unreachable);
375 argv_array_push(&cmd.env_array, "GIT_REF_PARANOIA=1");
376 } else if (pack_everything & LOOSEN_UNREACHABLE) {
377 argv_array_push(&cmd.args,
378 "--unpack-unreachable");
379 } else if (keep_unreachable) {
380 argv_array_push(&cmd.args, "--keep-unreachable");
381 argv_array_push(&cmd.args, "--pack-loose-unreachable");
382 } else {
383 argv_array_push(&cmd.env_array, "GIT_REF_PARANOIA=1");
384 }
385 }
386 } else {
387 argv_array_push(&cmd.args, "--unpacked");
388 argv_array_push(&cmd.args, "--incremental");
389 }
390
391 cmd.no_stdin = 1;
392
393 ret = start_command(&cmd);
394 if (ret)
395 return ret;
396
397 out = xfdopen(cmd.out, "r");
398 while (strbuf_getline_lf(&line, out) != EOF) {
399 if (line.len != 40)
400 die("repack: Expecting 40 character sha1 lines only from pack-objects.");
401 string_list_append(&names, line.buf);
402 }
403 fclose(out);
404 ret = finish_command(&cmd);
405 if (ret)
406 return ret;
407
408 if (!names.nr && !po_args.quiet)
409 printf("Nothing new to pack.\n");
410
411 /*
412 * Ok we have prepared all new packfiles.
413 * First see if there are packs of the same name and if so
414 * if we can move them out of the way (this can happen if we
415 * repacked immediately after packing fully.
416 */
417 failed = 0;
418 for_each_string_list_item(item, &names) {
419 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
420 char *fname, *fname_old;
421 fname = mkpathdup("%s/pack-%s%s", packdir,
422 item->string, exts[ext].name);
423 if (!file_exists(fname)) {
424 free(fname);
425 continue;
426 }
427
428 fname_old = mkpathdup("%s/old-%s%s", packdir,
429 item->string, exts[ext].name);
430 if (file_exists(fname_old))
431 if (unlink(fname_old))
432 failed = 1;
433
434 if (!failed && rename(fname, fname_old)) {
435 free(fname);
436 free(fname_old);
437 failed = 1;
438 break;
439 } else {
440 string_list_append(&rollback, fname);
441 free(fname_old);
442 }
443 }
444 if (failed)
445 break;
446 }
447 if (failed) {
448 struct string_list rollback_failure = STRING_LIST_INIT_DUP;
449 for_each_string_list_item(item, &rollback) {
450 char *fname, *fname_old;
451 fname = mkpathdup("%s/%s", packdir, item->string);
452 fname_old = mkpathdup("%s/old-%s", packdir, item->string);
453 if (rename(fname_old, fname))
454 string_list_append(&rollback_failure, fname);
455 free(fname);
456 free(fname_old);
457 }
458
459 if (rollback_failure.nr) {
460 int i;
461 fprintf(stderr,
462 "WARNING: Some packs in use have been renamed by\n"
463 "WARNING: prefixing old- to their name, in order to\n"
464 "WARNING: replace them with the new version of the\n"
465 "WARNING: file. But the operation failed, and the\n"
466 "WARNING: attempt to rename them back to their\n"
467 "WARNING: original names also failed.\n"
468 "WARNING: Please rename them in %s manually:\n", packdir);
469 for (i = 0; i < rollback_failure.nr; i++)
470 fprintf(stderr, "WARNING: old-%s -> %s\n",
471 rollback_failure.items[i].string,
472 rollback_failure.items[i].string);
473 }
474 exit(1);
475 }
476
477 /* Now the ones with the same name are out of the way... */
478 for_each_string_list_item(item, &names) {
479 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
480 char *fname, *fname_old;
481 struct stat statbuffer;
482 int exists = 0;
483 fname = mkpathdup("%s/pack-%s%s",
484 packdir, item->string, exts[ext].name);
485 fname_old = mkpathdup("%s-%s%s",
486 packtmp, item->string, exts[ext].name);
487 if (!stat(fname_old, &statbuffer)) {
488 statbuffer.st_mode &= ~(S_IWUSR | S_IWGRP | S_IWOTH);
489 chmod(fname_old, statbuffer.st_mode);
490 exists = 1;
491 }
492 if (exists || !exts[ext].optional) {
493 if (rename(fname_old, fname))
494 die_errno(_("renaming '%s' failed"), fname_old);
495 }
496 free(fname);
497 free(fname_old);
498 }
499 }
500
501 /* Remove the "old-" files */
502 for_each_string_list_item(item, &names) {
503 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
504 char *fname;
505 fname = mkpathdup("%s/old-%s%s",
506 packdir,
507 item->string,
508 exts[ext].name);
509 if (remove_path(fname))
510 warning(_("failed to remove '%s'"), fname);
511 free(fname);
512 }
513 }
514
515 /* End of pack replacement. */
516
517 reprepare_packed_git(the_repository);
518
519 if (delete_redundant) {
520 int opts = 0;
521 string_list_sort(&names);
522 for_each_string_list_item(item, &existing_packs) {
523 char *sha1;
524 size_t len = strlen(item->string);
525 if (len < 40)
526 continue;
527 sha1 = item->string + len - 40;
528 if (!string_list_has_string(&names, sha1))
529 remove_redundant_pack(packdir, item->string);
530 }
531 if (!po_args.quiet && isatty(2))
532 opts |= PRUNE_PACKED_VERBOSE;
533 prune_packed_objects(opts);
534 }
535
536 if (!no_update_server_info)
537 update_server_info(0);
538 remove_temporary_files();
539 string_list_clear(&names, 0);
540 string_list_clear(&rollback, 0);
541 string_list_clear(&existing_packs, 0);
542 strbuf_release(&line);
543
544 return 0;
545 }