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 "midx.h"
12#include "packfile.h"
13#include "object-store.h"
14
15static int delta_base_offset = 1;
16static int pack_kept_objects = -1;
17static int write_bitmaps = -1;
18static int use_delta_islands;
19static char *packdir, *packtmp;
20
21static const char *const git_repack_usage[] = {
22 N_("git repack [<options>]"),
23 NULL
24};
25
26static const char incremental_bitmap_conflict_error[] = N_(
27"Incremental repacks are incompatible with bitmap indexes. Use\n"
28"--no-write-bitmap-index or disable the pack.writebitmaps configuration."
29);
30
31
32static int repack_config(const char *var, const char *value, void *cb)
33{
34 if (!strcmp(var, "repack.usedeltabaseoffset")) {
35 delta_base_offset = git_config_bool(var, value);
36 return 0;
37 }
38 if (!strcmp(var, "repack.packkeptobjects")) {
39 pack_kept_objects = git_config_bool(var, value);
40 return 0;
41 }
42 if (!strcmp(var, "repack.writebitmaps") ||
43 !strcmp(var, "pack.writebitmaps")) {
44 write_bitmaps = git_config_bool(var, value);
45 return 0;
46 }
47 if (!strcmp(var, "repack.usedeltaislands")) {
48 use_delta_islands = git_config_bool(var, value);
49 return 0;
50 }
51 return git_default_config(var, value, cb);
52}
53
54/*
55 * Remove temporary $GIT_OBJECT_DIRECTORY/pack/.tmp-$$-pack-* files.
56 */
57static void remove_temporary_files(void)
58{
59 struct strbuf buf = STRBUF_INIT;
60 size_t dirlen, prefixlen;
61 DIR *dir;
62 struct dirent *e;
63
64 dir = opendir(packdir);
65 if (!dir)
66 return;
67
68 /* Point at the slash at the end of ".../objects/pack/" */
69 dirlen = strlen(packdir) + 1;
70 strbuf_addstr(&buf, packtmp);
71 /* Hold the length of ".tmp-%d-pack-" */
72 prefixlen = buf.len - dirlen;
73
74 while ((e = readdir(dir))) {
75 if (strncmp(e->d_name, buf.buf + dirlen, prefixlen))
76 continue;
77 strbuf_setlen(&buf, dirlen);
78 strbuf_addstr(&buf, e->d_name);
79 unlink(buf.buf);
80 }
81 closedir(dir);
82 strbuf_release(&buf);
83}
84
85static void remove_pack_on_signal(int signo)
86{
87 remove_temporary_files();
88 sigchain_pop(signo);
89 raise(signo);
90}
91
92static int has_pack_keep_file(void)
93{
94 struct packed_git *p;
95
96 for (p = get_all_packs(the_repository); p; p = p->next) {
97 if (p->pack_keep)
98 return 1;
99 }
100 return 0;
101}
102
103/*
104 * Adds all packs hex strings to the fname list, which do not
105 * have a corresponding .keep file. These packs are not to
106 * be kept if we are going to pack everything into one file.
107 */
108static void get_non_kept_pack_filenames(struct string_list *fname_list,
109 const struct string_list *extra_keep)
110{
111 DIR *dir;
112 struct dirent *e;
113 char *fname;
114
115 if (!(dir = opendir(packdir)))
116 return;
117
118 while ((e = readdir(dir)) != NULL) {
119 size_t len;
120 int i;
121
122 for (i = 0; i < extra_keep->nr; i++)
123 if (!fspathcmp(e->d_name, extra_keep->items[i].string))
124 break;
125 if (extra_keep->nr > 0 && i < extra_keep->nr)
126 continue;
127
128 if (!strip_suffix(e->d_name, ".pack", &len))
129 continue;
130
131 fname = xmemdupz(e->d_name, len);
132
133 if (!file_exists(mkpath("%s/%s.keep", packdir, fname)))
134 string_list_append_nodup(fname_list, fname);
135 else
136 free(fname);
137 }
138 closedir(dir);
139}
140
141static void remove_redundant_pack(const char *dir_name, const char *base_name)
142{
143 struct strbuf buf = STRBUF_INIT;
144 strbuf_addf(&buf, "%s/%s.pack", dir_name, base_name);
145 unlink_pack_path(buf.buf, 1);
146 strbuf_release(&buf);
147}
148
149struct pack_objects_args {
150 const char *window;
151 const char *window_memory;
152 const char *depth;
153 const char *threads;
154 const char *max_pack_size;
155 int no_reuse_delta;
156 int no_reuse_object;
157 int quiet;
158 int local;
159};
160
161static void prepare_pack_objects(struct child_process *cmd,
162 const struct pack_objects_args *args)
163{
164 argv_array_push(&cmd->args, "pack-objects");
165 if (args->window)
166 argv_array_pushf(&cmd->args, "--window=%s", args->window);
167 if (args->window_memory)
168 argv_array_pushf(&cmd->args, "--window-memory=%s", args->window_memory);
169 if (args->depth)
170 argv_array_pushf(&cmd->args, "--depth=%s", args->depth);
171 if (args->threads)
172 argv_array_pushf(&cmd->args, "--threads=%s", args->threads);
173 if (args->max_pack_size)
174 argv_array_pushf(&cmd->args, "--max-pack-size=%s", args->max_pack_size);
175 if (args->no_reuse_delta)
176 argv_array_pushf(&cmd->args, "--no-reuse-delta");
177 if (args->no_reuse_object)
178 argv_array_pushf(&cmd->args, "--no-reuse-object");
179 if (args->local)
180 argv_array_push(&cmd->args, "--local");
181 if (args->quiet)
182 argv_array_push(&cmd->args, "--quiet");
183 if (delta_base_offset)
184 argv_array_push(&cmd->args, "--delta-base-offset");
185 argv_array_push(&cmd->args, packtmp);
186 cmd->git_cmd = 1;
187 cmd->out = -1;
188}
189
190/*
191 * Write oid to the given struct child_process's stdin, starting it first if
192 * necessary.
193 */
194static int write_oid(const struct object_id *oid, struct packed_git *pack,
195 uint32_t pos, void *data)
196{
197 struct child_process *cmd = data;
198
199 if (cmd->in == -1) {
200 if (start_command(cmd))
201 die(_("could not start pack-objects to repack promisor objects"));
202 }
203
204 xwrite(cmd->in, oid_to_hex(oid), GIT_SHA1_HEXSZ);
205 xwrite(cmd->in, "\n", 1);
206 return 0;
207}
208
209static void repack_promisor_objects(const struct pack_objects_args *args,
210 struct string_list *names)
211{
212 struct child_process cmd = CHILD_PROCESS_INIT;
213 FILE *out;
214 struct strbuf line = STRBUF_INIT;
215
216 prepare_pack_objects(&cmd, args);
217 cmd.in = -1;
218
219 /*
220 * NEEDSWORK: Giving pack-objects only the OIDs without any ordering
221 * hints may result in suboptimal deltas in the resulting pack. See if
222 * the OIDs can be sent with fake paths such that pack-objects can use a
223 * {type -> existing pack order} ordering when computing deltas instead
224 * of a {type -> size} ordering, which may produce better deltas.
225 */
226 for_each_packed_object(write_oid, &cmd,
227 FOR_EACH_OBJECT_PROMISOR_ONLY);
228
229 if (cmd.in == -1)
230 /* No packed objects; cmd was never started */
231 return;
232
233 close(cmd.in);
234
235 out = xfdopen(cmd.out, "r");
236 while (strbuf_getline_lf(&line, out) != EOF) {
237 char *promisor_name;
238 int fd;
239 if (line.len != the_hash_algo->hexsz)
240 die(_("repack: Expecting full hex object ID lines only from pack-objects."));
241 string_list_append(names, line.buf);
242
243 /*
244 * pack-objects creates the .pack and .idx files, but not the
245 * .promisor file. Create the .promisor file, which is empty.
246 */
247 promisor_name = mkpathdup("%s-%s.promisor", packtmp,
248 line.buf);
249 fd = open(promisor_name, O_CREAT|O_EXCL|O_WRONLY, 0600);
250 if (fd < 0)
251 die_errno(_("unable to create '%s'"), promisor_name);
252 close(fd);
253 free(promisor_name);
254 }
255 fclose(out);
256 if (finish_command(&cmd))
257 die(_("could not finish pack-objects to repack promisor objects"));
258}
259
260#define ALL_INTO_ONE 1
261#define LOOSEN_UNREACHABLE 2
262
263int cmd_repack(int argc, const char **argv, const char *prefix)
264{
265 struct {
266 const char *name;
267 unsigned optional:1;
268 } exts[] = {
269 {".pack"},
270 {".idx"},
271 {".bitmap", 1},
272 {".promisor", 1},
273 };
274 struct child_process cmd = CHILD_PROCESS_INIT;
275 struct string_list_item *item;
276 struct string_list names = STRING_LIST_INIT_DUP;
277 struct string_list rollback = STRING_LIST_INIT_NODUP;
278 struct string_list existing_packs = STRING_LIST_INIT_DUP;
279 struct strbuf line = STRBUF_INIT;
280 int i, ext, ret, failed;
281 FILE *out;
282
283 /* variables to be filled by option parsing */
284 int pack_everything = 0;
285 int delete_redundant = 0;
286 const char *unpack_unreachable = NULL;
287 int keep_unreachable = 0;
288 struct string_list keep_pack_list = STRING_LIST_INIT_NODUP;
289 int no_update_server_info = 0;
290 int midx_cleared = 0;
291 struct pack_objects_args po_args = {NULL};
292
293 struct option builtin_repack_options[] = {
294 OPT_BIT('a', NULL, &pack_everything,
295 N_("pack everything in a single pack"), ALL_INTO_ONE),
296 OPT_BIT('A', NULL, &pack_everything,
297 N_("same as -a, and turn unreachable objects loose"),
298 LOOSEN_UNREACHABLE | ALL_INTO_ONE),
299 OPT_BOOL('d', NULL, &delete_redundant,
300 N_("remove redundant packs, and run git-prune-packed")),
301 OPT_BOOL('f', NULL, &po_args.no_reuse_delta,
302 N_("pass --no-reuse-delta to git-pack-objects")),
303 OPT_BOOL('F', NULL, &po_args.no_reuse_object,
304 N_("pass --no-reuse-object to git-pack-objects")),
305 OPT_BOOL('n', NULL, &no_update_server_info,
306 N_("do not run git-update-server-info")),
307 OPT__QUIET(&po_args.quiet, N_("be quiet")),
308 OPT_BOOL('l', "local", &po_args.local,
309 N_("pass --local to git-pack-objects")),
310 OPT_BOOL('b', "write-bitmap-index", &write_bitmaps,
311 N_("write bitmap index")),
312 OPT_BOOL('i', "delta-islands", &use_delta_islands,
313 N_("pass --delta-islands to git-pack-objects")),
314 OPT_STRING(0, "unpack-unreachable", &unpack_unreachable, N_("approxidate"),
315 N_("with -A, do not loosen objects older than this")),
316 OPT_BOOL('k', "keep-unreachable", &keep_unreachable,
317 N_("with -a, repack unreachable objects")),
318 OPT_STRING(0, "window", &po_args.window, N_("n"),
319 N_("size of the window used for delta compression")),
320 OPT_STRING(0, "window-memory", &po_args.window_memory, N_("bytes"),
321 N_("same as the above, but limit memory size instead of entries count")),
322 OPT_STRING(0, "depth", &po_args.depth, N_("n"),
323 N_("limits the maximum delta depth")),
324 OPT_STRING(0, "threads", &po_args.threads, N_("n"),
325 N_("limits the maximum number of threads")),
326 OPT_STRING(0, "max-pack-size", &po_args.max_pack_size, N_("bytes"),
327 N_("maximum size of each packfile")),
328 OPT_BOOL(0, "pack-kept-objects", &pack_kept_objects,
329 N_("repack objects in packs marked with .keep")),
330 OPT_STRING_LIST(0, "keep-pack", &keep_pack_list, N_("name"),
331 N_("do not repack this pack")),
332 OPT_END()
333 };
334
335 git_config(repack_config, NULL);
336
337 argc = parse_options(argc, argv, prefix, builtin_repack_options,
338 git_repack_usage, 0);
339
340 if (delete_redundant && repository_format_precious_objects)
341 die(_("cannot delete packs in a precious-objects repo"));
342
343 if (keep_unreachable &&
344 (unpack_unreachable || (pack_everything & LOOSEN_UNREACHABLE)))
345 die(_("--keep-unreachable and -A are incompatible"));
346
347 if (write_bitmaps < 0) {
348 write_bitmaps = (pack_everything & ALL_INTO_ONE) &&
349 is_bare_repository() &&
350 keep_pack_list.nr == 0 &&
351 !has_pack_keep_file();
352 }
353 if (pack_kept_objects < 0)
354 pack_kept_objects = write_bitmaps;
355
356 if (write_bitmaps && !(pack_everything & ALL_INTO_ONE))
357 die(_(incremental_bitmap_conflict_error));
358
359 packdir = mkpathdup("%s/pack", get_object_directory());
360 packtmp = mkpathdup("%s/.tmp-%d-pack", packdir, (int)getpid());
361
362 sigchain_push_common(remove_pack_on_signal);
363
364 prepare_pack_objects(&cmd, &po_args);
365
366 argv_array_push(&cmd.args, "--keep-true-parents");
367 if (!pack_kept_objects)
368 argv_array_push(&cmd.args, "--honor-pack-keep");
369 for (i = 0; i < keep_pack_list.nr; i++)
370 argv_array_pushf(&cmd.args, "--keep-pack=%s",
371 keep_pack_list.items[i].string);
372 argv_array_push(&cmd.args, "--non-empty");
373 argv_array_push(&cmd.args, "--all");
374 argv_array_push(&cmd.args, "--reflog");
375 argv_array_push(&cmd.args, "--indexed-objects");
376 if (repository_format_partial_clone)
377 argv_array_push(&cmd.args, "--exclude-promisor-objects");
378 if (write_bitmaps)
379 argv_array_push(&cmd.args, "--write-bitmap-index");
380 if (use_delta_islands)
381 argv_array_push(&cmd.args, "--delta-islands");
382
383 if (pack_everything & ALL_INTO_ONE) {
384 get_non_kept_pack_filenames(&existing_packs, &keep_pack_list);
385
386 repack_promisor_objects(&po_args, &names);
387
388 if (existing_packs.nr && delete_redundant) {
389 if (unpack_unreachable) {
390 argv_array_pushf(&cmd.args,
391 "--unpack-unreachable=%s",
392 unpack_unreachable);
393 argv_array_push(&cmd.env_array, "GIT_REF_PARANOIA=1");
394 } else if (pack_everything & LOOSEN_UNREACHABLE) {
395 argv_array_push(&cmd.args,
396 "--unpack-unreachable");
397 } else if (keep_unreachable) {
398 argv_array_push(&cmd.args, "--keep-unreachable");
399 argv_array_push(&cmd.args, "--pack-loose-unreachable");
400 } else {
401 argv_array_push(&cmd.env_array, "GIT_REF_PARANOIA=1");
402 }
403 }
404 } else {
405 argv_array_push(&cmd.args, "--unpacked");
406 argv_array_push(&cmd.args, "--incremental");
407 }
408
409 cmd.no_stdin = 1;
410
411 ret = start_command(&cmd);
412 if (ret)
413 return ret;
414
415 out = xfdopen(cmd.out, "r");
416 while (strbuf_getline_lf(&line, out) != EOF) {
417 if (line.len != the_hash_algo->hexsz)
418 die(_("repack: Expecting full hex object ID lines only from pack-objects."));
419 string_list_append(&names, line.buf);
420 }
421 fclose(out);
422 ret = finish_command(&cmd);
423 if (ret)
424 return ret;
425
426 if (!names.nr && !po_args.quiet)
427 printf_ln(_("Nothing new to pack."));
428
429 close_object_store(the_repository->objects);
430
431 /*
432 * Ok we have prepared all new packfiles.
433 * First see if there are packs of the same name and if so
434 * if we can move them out of the way (this can happen if we
435 * repacked immediately after packing fully.
436 */
437 failed = 0;
438 for_each_string_list_item(item, &names) {
439 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
440 char *fname, *fname_old;
441
442 if (!midx_cleared) {
443 clear_midx_file(the_repository);
444 midx_cleared = 1;
445 }
446
447 fname = mkpathdup("%s/pack-%s%s", packdir,
448 item->string, exts[ext].name);
449 if (!file_exists(fname)) {
450 free(fname);
451 continue;
452 }
453
454 fname_old = mkpathdup("%s/old-%s%s", packdir,
455 item->string, exts[ext].name);
456 if (file_exists(fname_old))
457 if (unlink(fname_old))
458 failed = 1;
459
460 if (!failed && rename(fname, fname_old)) {
461 free(fname);
462 free(fname_old);
463 failed = 1;
464 break;
465 } else {
466 string_list_append(&rollback, fname);
467 free(fname_old);
468 }
469 }
470 if (failed)
471 break;
472 }
473 if (failed) {
474 struct string_list rollback_failure = STRING_LIST_INIT_DUP;
475 for_each_string_list_item(item, &rollback) {
476 char *fname, *fname_old;
477 fname = mkpathdup("%s/%s", packdir, item->string);
478 fname_old = mkpathdup("%s/old-%s", packdir, item->string);
479 if (rename(fname_old, fname))
480 string_list_append(&rollback_failure, fname);
481 free(fname);
482 free(fname_old);
483 }
484
485 if (rollback_failure.nr) {
486 int i;
487 fprintf(stderr,
488 _("WARNING: Some packs in use have been renamed by\n"
489 "WARNING: prefixing old- to their name, in order to\n"
490 "WARNING: replace them with the new version of the\n"
491 "WARNING: file. But the operation failed, and the\n"
492 "WARNING: attempt to rename them back to their\n"
493 "WARNING: original names also failed.\n"
494 "WARNING: Please rename them in %s manually:\n"), packdir);
495 for (i = 0; i < rollback_failure.nr; i++)
496 fprintf(stderr, "WARNING: old-%s -> %s\n",
497 rollback_failure.items[i].string,
498 rollback_failure.items[i].string);
499 }
500 exit(1);
501 }
502
503 /* Now the ones with the same name are out of the way... */
504 for_each_string_list_item(item, &names) {
505 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
506 char *fname, *fname_old;
507 struct stat statbuffer;
508 int exists = 0;
509 fname = mkpathdup("%s/pack-%s%s",
510 packdir, item->string, exts[ext].name);
511 fname_old = mkpathdup("%s-%s%s",
512 packtmp, item->string, exts[ext].name);
513 if (!stat(fname_old, &statbuffer)) {
514 statbuffer.st_mode &= ~(S_IWUSR | S_IWGRP | S_IWOTH);
515 chmod(fname_old, statbuffer.st_mode);
516 exists = 1;
517 }
518 if (exists || !exts[ext].optional) {
519 if (rename(fname_old, fname))
520 die_errno(_("renaming '%s' failed"), fname_old);
521 }
522 free(fname);
523 free(fname_old);
524 }
525 }
526
527 /* Remove the "old-" files */
528 for_each_string_list_item(item, &names) {
529 for (ext = 0; ext < ARRAY_SIZE(exts); ext++) {
530 char *fname;
531 fname = mkpathdup("%s/old-%s%s",
532 packdir,
533 item->string,
534 exts[ext].name);
535 if (remove_path(fname))
536 warning(_("failed to remove '%s'"), fname);
537 free(fname);
538 }
539 }
540
541 /* End of pack replacement. */
542
543 reprepare_packed_git(the_repository);
544
545 if (delete_redundant) {
546 const int hexsz = the_hash_algo->hexsz;
547 int opts = 0;
548 string_list_sort(&names);
549 for_each_string_list_item(item, &existing_packs) {
550 char *sha1;
551 size_t len = strlen(item->string);
552 if (len < hexsz)
553 continue;
554 sha1 = item->string + len - hexsz;
555 if (!string_list_has_string(&names, sha1))
556 remove_redundant_pack(packdir, item->string);
557 }
558 if (!po_args.quiet && isatty(2))
559 opts |= PRUNE_PACKED_VERBOSE;
560 prune_packed_objects(opts);
561
562 if (!keep_unreachable &&
563 (!(pack_everything & LOOSEN_UNREACHABLE) ||
564 unpack_unreachable) &&
565 is_repository_shallow(the_repository))
566 prune_shallow(PRUNE_QUICK);
567 }
568
569 if (!no_update_server_info)
570 update_server_info(0);
571 remove_temporary_files();
572
573 if (git_env_bool(GIT_TEST_MULTI_PACK_INDEX, 0))
574 write_midx_file(get_object_directory());
575
576 string_list_clear(&names, 0);
577 string_list_clear(&rollback, 0);
578 string_list_clear(&existing_packs, 0);
579 strbuf_release(&line);
580
581 return 0;
582}