Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improves bsdiff performance by preventing excessive iterations when processing similar data blocks #1

Open
wants to merge 2 commits into
base: 2.x
Choose a base branch
from
+21 −2
Open

Improves bsdiff performance by preventing excessive iterations when processing similar data blocks #1

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
bf46af0
Port bsdiff optimizations for diffing large blocks that only differ b…
wfairclough Feb 22, 2025
f7fde21
Retrigger CI
wfairclough Feb 23, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 21 additions & 2 deletions Vendor/bsdiff/bsdiff.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -84,7 +84,7 @@ static off_t search(off_t *I, u_char *old, off_t oldsize,
}

x = st + (en - st)/2;
if (memcmp(old + I[x], new, (size_t)(MIN(oldsize - I[x], newsize))) < 0) {
if (memcmp(old + I[x], new, (size_t)(MIN(oldsize - I[x], newsize))) <= 0) {
return search(I, old, oldsize, new, newsize, x, en, pos);
} else {
return search(I, old, oldsize, new, newsize, st, x, pos);
Expand Down Expand Up @@ -212,8 +212,16 @@ int bsdiff(int argc, char *argv[])
lastoffset = 0;
while (scan < newsize) {
oldscore = 0;

/* If we come across a large block of data that only differs
* by less than 8 bytes, this loop will take a long time to
* go past that block of data. We need to track the number of
* times we're stuck in the block and break out of it. */
int num_less_than_eight = 0;
off_t prev_len, prev_oldscore, prev_pos;
for (scsc = scan += len; scan < newsize; scan++) {
prev_len = len;
prev_oldscore = oldscore;
prev_pos = pos;
/* 'oldscore' is the number of characters that match between the
* substrings 'old[lastoffset + scan:lastoffset + scsc]' and
* 'new[scan:scsc]'. */
Expand All @@ -240,6 +248,17 @@ int bsdiff(int argc, char *argv[])
if ((scan + lastoffset < oldsize) &&
(old[scan + lastoffset] == new[scan]))
oldscore--;

const off_t fuzz = 8;
if (prev_len - fuzz <= len && len <= prev_len &&
prev_oldscore - fuzz <= oldscore &&
oldscore <= prev_oldscore &&
prev_pos <= pos && pos <= prev_pos + fuzz &&
oldscore <= len && len <= oldscore + fuzz)
++num_less_than_eight;
else
num_less_than_eight=0;
if (num_less_than_eight > 100) break;
}

/* Skip this section if we found an exact match that would be
Expand Down
Loading