git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH 2/8] pack-write.c: prepare to write 'pack-*.rev' files

From
Taylor Blau <me@ttaylorr.com>
Date
Jan 8, 2021, 18:20 UTC
Message-ID
<88393e266292b022fba3ff41f6623486cb41f66d.1610129989.git.me@ttaylorr.com>
In-Reply-To
<cover.1610129989.git.me@ttaylorr.com>

This patch prepares for callers to be able to write reverse index files to disk.

It adds the necessary machinery to write a format-compliant .rev file from within 'write_rev_file()', which is called from 'finish_tmp_packfile()'.

Signed-off-by: Taylor Blau <me@ttaylorr.com>
---
 pack-write.c | 123 ++++++++++++++++++++++++++++++++++++++++++++++++++-
 pack.h       |   4 ++
 2 files changed, 126 insertions(+), 1 deletion(-)
diff --git a/pack-write.c b/pack-write.c
index 3513665e1e..68db5a9edf 100644
--- a/pack-write.c
+++ b/pack-write.c
@@ -166,6 +166,116 @@ const char *write_idx_file(const char *index_name, struct pack_idx_entry **objec
 	return index_name;
 }
 
+static int pack_order_cmp(const void *va, const void *vb, void *ctx)
+{
+	struct pack_idx_entry **objects = ctx;
+
+	off_t oa = objects[*(uint32_t*)va]->offset;
+	off_t ob = objects[*(uint32_t*)vb]->offset;
+
+	if (oa < ob)
+		return -1;
+	if (oa > ob)
+		return 1;
+	return 0;
+}
+
+#define RIDX_SIGNATURE 0x52494458 /* "RIDX" */
+#define RIDX_VERSION 1
+
+static void write_rev_header(struct hashfile *f)
+{
+	uint32_t oid_version;
+	switch (hash_algo_by_ptr(the_hash_algo)) {
+	case GIT_HASH_SHA1:
+		oid_version = 1;
+		break;
+	case GIT_HASH_SHA256:
+		oid_version = 2;
+		break;
+	default:
+		die("write_rev_header: unknown hash version");
+	}
+
+	hashwrite_be32(f, RIDX_SIGNATURE);
+	hashwrite_be32(f, RIDX_VERSION);
+	hashwrite_be32(f, oid_version);
+}
+
+static void write_rev_index_positions(struct hashfile *f,
+				      struct pack_idx_entry **objects,
+				      uint32_t nr_objects)
+{
+	uint32_t *pack_order;
+	uint32_t i;
+
+	ALLOC_ARRAY(pack_order, nr_objects);
+	for (i = 0; i < nr_objects; i++)
+		pack_order[i] = i;
+	QSORT_S(pack_order, nr_objects, pack_order_cmp, objects);
+
+	for (i = 0; i < nr_objects; i++)
+		hashwrite_be32(f, pack_order[i]);
+
+	free(pack_order);
+}
+
+static void write_rev_trailer(struct hashfile *f, const unsigned char *hash)
+{
+	hashwrite(f, hash, the_hash_algo->rawsz);
+}
+
+const char *write_rev_file(const char *rev_name,
+			   struct pack_idx_entry **objects,
+			   uint32_t nr_objects,
+			   const unsigned char *hash,
+			   unsigned flags)
+{
+	struct hashfile *f;
+	int fd;
+
+	if ((flags & WRITE_REV) && (flags & WRITE_REV_VERIFY))
+		die(_("cannot both write and verify reverse index"));
+
+	if (flags & WRITE_REV) {
+		if (!rev_name) {
+			struct strbuf tmp_file = STRBUF_INIT;
+			fd = odb_mkstemp(&tmp_file, "pack/tmp_rev_XXXXXX");
+			rev_name = strbuf_detach(&tmp_file, NULL);
+		} else {
+			unlink(rev_name);
+			fd = open(rev_name, O_CREAT|O_EXCL|O_WRONLY, 0600);
+			if (fd < 0)
+				die_errno("unable to create '%s'", rev_name);
+		}
+		f = hashfd(fd, rev_name);
+	} else if (flags & WRITE_REV_VERIFY) {
+		struct stat statbuf;
+		if (stat(rev_name, &statbuf)) {
+			if (errno == ENOENT) {
+				/* .rev files are optional */
+				return NULL;
+			} else
+				die_errno(_("could not stat: %s"), rev_name);
+		}
+		f = hashfd_check(rev_name);
+	} else
+		return NULL;
+
+	write_rev_header(f);
+
+	write_rev_index_positions(f, objects, nr_objects);
+	write_rev_trailer(f, hash);
+
+	if (rev_name && adjust_shared_perm(rev_name) < 0)
+		die(_("failed to make %s readable"), rev_name);
+
+	finalize_hashfile(f, NULL, CSUM_HASH_IN_STREAM | CSUM_CLOSE |
+				    ((flags & WRITE_IDX_VERIFY) ? 0 : CSUM_FSYNC));
+
+	return rev_name;
+}
+
 off_t write_pack_header(struct hashfile *f, uint32_t nr_entries)
 {
 	struct pack_header hdr;
@@ -341,7 +451,7 @@ void finish_tmp_packfile(struct strbuf *name_buffer,
 			 struct pack_idx_option *pack_idx_opts,
 			 unsigned char hash[])
 {
-	const char *idx_tmp_name;
+	const char *idx_tmp_name, *rev_tmp_name = NULL;
 	int basename_len = name_buffer->len;
 
 	if (adjust_shared_perm(pack_tmp_name))
@@ -352,6 +462,9 @@ void finish_tmp_packfile(struct strbuf *name_buffer,
 	if (adjust_shared_perm(idx_tmp_name))
 		die_errno("unable to make temporary index file readable");
 
+	rev_tmp_name = write_rev_file(NULL, written_list, nr_written, hash,
+				      pack_idx_opts->flags);
+
 	strbuf_addf(name_buffer, "%s.pack", hash_to_hex(hash));
 
 	if (rename(pack_tmp_name, name_buffer->buf))
@@ -365,5 +478,13 @@ void finish_tmp_packfile(struct strbuf *name_buffer,
 
 	strbuf_setlen(name_buffer, basename_len);
 
+	if (rev_tmp_name) {
+		strbuf_addf(name_buffer, "%s.rev", hash_to_hex(hash));
+		if (rename(rev_tmp_name, name_buffer->buf))
+			die_errno("unable to rename temporary reverse-index file");
+	}
+
+	strbuf_setlen(name_buffer, basename_len);
+
 	free((void *)idx_tmp_name);
 }
diff --git a/pack.h b/pack.h
index 9fc0945ac9..30439e0784 100644
--- a/pack.h
+++ b/pack.h
@@ -42,6 +42,8 @@ struct pack_idx_option {
 	/* flag bits */
 #define WRITE_IDX_VERIFY 01 /* verify only, do not write the idx file */
 #define WRITE_IDX_STRICT 02
+#define WRITE_REV 04
+#define WRITE_REV_VERIFY 010
 
 	uint32_t version;
 	uint32_t off32_limit;
@@ -87,6 +89,8 @@ off_t write_pack_header(struct hashfile *f, uint32_t);
 void fixup_pack_header_footer(int, unsigned char *, const char *, uint32_t, unsigned char *, off_t);
 char *index_pack_lockfile(int fd);
 
+const char *write_rev_file(const char *rev_name, struct pack_idx_entry **objects, uint32_t nr_objects, const unsigned char *hash, unsigned flags);
+
 /*
  * The "hdr" output buffer should be at least this big, which will handle sizes
  * up to 2^67.
-- 
2.30.0.138.g6d7191ea01
Previous: Taylor BlauNext: Taylor Blau
Message 2 of 82 in “pack-revindex: introduce on-disk '.rev' format”
  1. 0/8 pack-revindex: introduce on-disk '.rev' formatTaylor Blau, Jan 8, 2021
  2. 2/8 pack-write.c: prepare to write 'pack-*.rev' filesTaylor Blau, Jan 8, 2021
  3. 8/8 pack-revindex: ensure that on-disk reverse indexes are given precedenceTaylor Blau, Jan 8, 2021
  4. 3/8 builtin/index-pack.c: write reverse indexesTaylor Blau, Jan 8, 2021
  5. 1/8 packfile: prepare for the existence of '*.rev' filesTaylor Blau, Jan 8, 2021
  6. 4/8 builtin/pack-objects.c: respect 'pack.writeReverseIndex'Taylor Blau, Jan 8, 2021
  7. 7/8 t: support GIT_TEST_WRITE_REV_INDEXTaylor Blau, Jan 8, 2021
  8. Derrick StoleeJan 12, 2021
  9. Taylor BlauJan 12, 2021
  10. Ævar Arnfjörð BjarmasonJan 12, 2021
  11. Derrick StoleeJan 12, 2021
  12. Taylor BlauJan 12, 2021
  13. 6/8 t: prepare for GIT_TEST_WRITE_REV_INDEXTaylor Blau, Jan 8, 2021
  14. Ævar Arnfjörð BjarmasonJan 12, 2021
  15. Taylor BlauJan 12, 2021
  16. 5/8 Documentation/config/pack.txt: advertise 'pack.writeReverseIndex'Taylor Blau, Jan 8, 2021
  17. 0/8 pack-revindex: introduce on-disk '.rev' formatTaylor Blau, Jan 13, 2021
  18. 8/8 pack-revindex: ensure that on-disk reverse indexes are given precedenceTaylor Blau, Jan 13, 2021
  19. 6/8 t: prepare for GIT_TEST_WRITE_REV_INDEXTaylor Blau, Jan 13, 2021
  20. 1/8 packfile: prepare for the existence of '*.rev' filesTaylor Blau, Jan 13, 2021
  21. Junio C HamanoJan 14, 2021
  22. Derrick StoleeJan 14, 2021
  23. Jeff KingJan 14, 2021
  24. Taylor BlauJan 14, 2021
  25. Junio C HamanoJan 14, 2021
  26. Taylor BlauJan 14, 2021
  27. Junio C HamanoJan 14, 2021
  28. Jeff KingJan 22, 2021
  29. Taylor BlauJan 25, 2021
  30. Jeff KingJan 25, 2021
  31. Junio C HamanoJan 25, 2021
  32. Taylor BlauJan 25, 2021
  33. 5/8 Documentation/config/pack.txt: advertise 'pack.writeReverseIndex'Taylor Blau, Jan 13, 2021
  34. 4/8 builtin/pack-objects.c: respect 'pack.writeReverseIndex'Taylor Blau, Jan 13, 2021
  35. Jeff KingJan 22, 2021
  36. Jeff KingJan 23, 2021
  37. Taylor BlauJan 25, 2021
  38. Jeff KingJan 25, 2021
  39. 2/8 pack-write.c: prepare to write 'pack-*.rev' filesTaylor Blau, Jan 13, 2021
  40. Jeff KingJan 22, 2021
  41. Taylor BlauJan 25, 2021
  42. Jeff KingJan 26, 2021
  43. 7/8 t: support GIT_TEST_WRITE_REV_INDEXTaylor Blau, Jan 13, 2021
  44. 3/8 builtin/index-pack.c: write reverse indexesTaylor Blau, Jan 13, 2021
  45. Jeff KingJan 22, 2021
  46. Taylor BlauJan 25, 2021
  47. 00/10 pack-revindex: introduce on-disk '.rev' formatTaylor Blau, Jan 25, 2021
  48. 01/10 packfile: prepare for the existence of '*.rev' filesTaylor Blau, Jan 25, 2021
  49. Jeff KingJan 29, 2021
  50. Taylor BlauJan 29, 2021
  51. Jeff KingJan 30, 2021
  52. 08/10 t: support GIT_TEST_WRITE_REV_INDEXTaylor Blau, Jan 25, 2021
  53. Jeff KingJan 29, 2021
  54. 07/10 t: prepare for GIT_TEST_WRITE_REV_INDEXTaylor Blau, Jan 25, 2021
  55. Jeff KingJan 29, 2021
  56. Eric SunshineJan 29, 2021
  57. Taylor BlauJan 29, 2021
  58. Jeff KingJan 30, 2021
  59. Junio C HamanoJan 29, 2021
  60. 05/10 builtin/pack-objects.c: respect 'pack.writeReverseIndex'Taylor Blau, Jan 25, 2021
  61. 09/10 pack-revindex: ensure that on-disk reverse indexes are given precedenceTaylor Blau, Jan 25, 2021
  62. Jeff KingJan 29, 2021
  63. Taylor BlauJan 29, 2021
  64. Jeff KingJan 30, 2021
  65. 10/10 t5325: check both on-disk and in-memory reverse indexTaylor Blau, Jan 25, 2021
  66. Jeff KingJan 29, 2021
  67. Jeff KingJan 29, 2021
  68. Taylor BlauJan 29, 2021
  69. Jeff KingJan 30, 2021
  70. 06/10 Documentation/config/pack.txt: advertise 'pack.writeReverseIndex'Taylor Blau, Jan 25, 2021
  71. Jeff KingJan 29, 2021
  72. Taylor BlauJan 29, 2021
  73. Jeff KingJan 30, 2021
  74. 02/10 pack-write.c: prepare to write 'pack-*.rev' filesTaylor Blau, Jan 25, 2021
  75. 04/10 builtin/index-pack.c: write reverse indexesTaylor Blau, Jan 25, 2021
  76. 03/10 builtin/index-pack.c: allow stripping arbitrary extensionsTaylor Blau, Jan 25, 2021
  77. Jeff KingJan 29, 2021
  78. Taylor BlauJan 29, 2021
  79. Junio C HamanoJan 26, 2021
  80. Taylor BlauJan 26, 2021
  81. Jeff KingJan 29, 2021
  82. Taylor BlauJan 29, 2021

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.