greg/pcompress
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Fix segment offset sorting.

Browse source 
Get rid of incorrect duplicate checks in index.
This commit is contained in:
Moinak Ghosh 2013-05-05 18:50:52 +05:30
parent c6da2325e3
commit 6ecc400571
2 changed files with 6 additions and 7 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

11
rabin/global/index.c
View file

@ -485,10 +485,11 @@ db_lookup_insert_s(archive_config_t *cfg, uchar_t *sim_cksum, int interval,
pent = &(htab[htab_entry]);
ent = htab[htab_entry];
if (cfg->pct_interval == 0) { // Global dedupe with simple index
if (cfg->pct_interval == 0) { // Global dedupe with simple index.
assert(cfg->similarity_cksum_sz == cfg->chunk_cksum_sz);
while (ent) {
if (mycmp(sim_cksum, ent->cksum, cfg->similarity_cksum_sz) == 0 &&
ent->item_size == item_size && ent->item_offset != item_offset) {
ent->item_size == item_size) {
return (ent);
}
pent = &(ent->next);
@ -496,8 +497,7 @@ db_lookup_insert_s(archive_config_t *cfg, uchar_t *sim_cksum, int interval,
}
} else if (cfg->similarity_cksum_sz == 8) {// Fast path for 64-bit keys
while (ent) {
if (*((uint64_t *)sim_cksum) == *((uint64_t *)ent->cksum) &&
ent->item_offset != item_offset) {
if (*((uint64_t *)sim_cksum) == *((uint64_t *)ent->cksum)) {
return (ent);
}
pent = &(ent->next);
@ -505,8 +505,7 @@ db_lookup_insert_s(archive_config_t *cfg, uchar_t *sim_cksum, int interval,
}
} else {
while (ent) {
if (mycmp(sim_cksum, ent->cksum, cfg->similarity_cksum_sz) == 0 &&
ent->item_offset != item_offset) {
if (mycmp(sim_cksum, ent->cksum, cfg->similarity_cksum_sz) == 0) {
return (ent);
}
pent = &(ent->next);

2
rabin/rabin_dedup.c
View file

@ -989,7 +989,7 @@ process_blocks:
* file. Sort the offsets to avoid subsequent random access.
*/
tgt = src + 1;
isort_uint64((uint64_t *)tgt, k);
isort_uint64((uint64_t *)tgt, sub_i);
/*
* Now eliminate duplicate offsets and UINT64_MAX offset entries which