@matheusbsbs / matheus-bagliano / commits / ba620d296a

reftable/block: simplify how we track restart points

Restart points record the location of reftable records that do not use prefix compression and are used to perform a binary search inside of a block. These restart points are encoded at the end of a block, between the record data and the footer of a table. The block structure contains three different variables related to these restart points: - The block length contains the length of the reftable block up to the restart points. - The restart count contains the number of restart points contained in the block. - The restart bytes variable tracks where the restart point data begins. Tracking all three of these variables is unnecessary though as the data can be derived from one another: the block length without restart points is the exact same as the offset of the restart count data, which we already track via the `restart_bytes` data. Refactor the code so that we track the location of restart bytes not as a pointer, but instead as an offset. This allows us to trivially get rid of the `block_len` variable as described above. This avoids having the confusing `block_len` variable and allows us to do less bookkeeping overall. Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Patrick Steinhardt committed Apr 7, 2025 at 15:16 UTC ba620d296ab7bcd93fcedfe13b265f84df1ed1eb

3 files changed +18 -17

reftable/block.c

+12 -13

@@ -216,10 +216,9 @@ int block_reader_init(struct block_reader *br, struct reftable_block *block,

216 uint32_t full_block_size = table_block_size;

217 uint8_t typ = block->data[header_off];

218 uint32_t sz = reftable_get_be24(block->data + header_off + 1);

219 - int err = 0;

220 - uint16_t restart_count = 0;

221 - uint32_t restart_start = 0;

222 - uint8_t *restart_bytes = NULL;

219 + uint16_t restart_count;

220 + uint32_t restart_off;

221 + int err;

222

223 block_source_return_block(&br->block);

224

@@ -300,8 +299,7 @@ int block_reader_init(struct block_reader *br, struct reftable_block *block,

299 }

300

301 restart_count = reftable_get_be16(block->data + sz - 2);

303 - restart_start = sz - 2 - 3 * restart_count;

304 - restart_bytes = block->data + restart_start;

302 + restart_off = sz - 2 - 3 * restart_count;

303

304 /* transfer ownership. */

305 br->block = *block;

@@ -309,11 +307,12 @@ int block_reader_init(struct block_reader *br, struct reftable_block *block,

307 block->len = 0;

308

309 br->hash_size = hash_size;

312 - br->block_len = restart_start;

310 + br->restart_off = restart_off;

311 br->full_block_size = full_block_size;

312 br->header_off = header_off;

313 br->restart_count = restart_count;

316 - br->restart_bytes = restart_bytes;

314 +

315 + err = 0;

316

317 done:

318 return err;

@@ -337,7 +336,7 @@ int block_reader_first_key(const struct block_reader *br, struct reftable_buf *k

336 int off = br->header_off + 4, n;

337 struct string_view in = {

338 .buf = br->block.data + off,

340 - .len = br->block_len - off,

339 + .len = br->restart_off - off,

340 };

341 uint8_t extra = 0;

342

@@ -354,13 +353,13 @@ int block_reader_first_key(const struct block_reader *br, struct reftable_buf *k

353

354 static uint32_t block_reader_restart_offset(const struct block_reader *br, size_t idx)

355 {

357 - return reftable_get_be24(br->restart_bytes + 3 * idx);

356 + return reftable_get_be24(br->block.data + br->restart_off + 3 * idx);

357 }

358

359 void block_iter_seek_start(struct block_iter *it, const struct block_reader *br)

360 {

361 it->block = br->block.data;

363 - it->block_len = br->block_len;

362 + it->block_len = br->restart_off;

363 it->hash_size = br->hash_size;

364 reftable_buf_reset(&it->last_key);

365 it->next_off = br->header_off + 4;

@@ -378,7 +377,7 @@ static int restart_needle_less(size_t idx, void *_args)

377 uint32_t off = block_reader_restart_offset(args->reader, idx);

378 struct string_view in = {

379 .buf = args->reader->block.data + off,

381 - .len = args->reader->block_len - off,

380 + .len = args->reader->restart_off - off,

381 };

382 uint64_t prefix_len, suffix_len;

383 uint8_t extra;

@@ -505,7 +504,7 @@ int block_iter_seek_key(struct block_iter *it, const struct block_reader *br,

504 else

505 it->next_off = br->header_off + 4;

506 it->block = br->block.data;

508 - it->block_len = br->block_len;

507 + it->block_len = br->restart_off;

508 it->hash_size = br->hash_size;

509

510 err = reftable_record_init(&rec, block_reader_type(br));

reftable/block.h

+5 -3

@@ -79,10 +79,12 @@ struct block_reader {

79 unsigned char *uncompressed_data;

80 size_t uncompressed_cap;

82 - /* size of the data, excluding restart data. */

83 - uint32_t block_len;

84 - uint8_t *restart_bytes;

82 + /*

83 + * Restart point data. Restart points are located after the block's

84 + * record data.

85 + */

86 uint16_t restart_count;

87 + uint32_t restart_off;

89 /* size of the data in the file. For log blocks, this is the compressed

90 * size. */

reftable/table.c

+1 -1

@@ -838,7 +838,7 @@ int reftable_table_print_blocks(const char *tablename)

838 printf("%s:\n", sections[i].name);

839

840 while (1) {

841 - printf(" - length: %u\n", ti.br.block_len);

841 + printf(" - length: %u\n", ti.br.restart_off);

842 printf(" restarts: %u\n", ti.br.restart_count);

843

844 err = table_iter_next_block(&ti);