Mercurial > hg-stable
changeset 29014:f1ca249696ed stable
bdiff: balance recursion to avoid quadratic behavior (issue4704)
For highly structured files like JSON or XML dumps with large numbers
of duplicate lines (eg braces) and isolated matching lines, bdiff
could find large numbers of equally good spans. Because it prefers
earlier matches, this would result in pathologically unbalance
recursion that resulted in quadratic performance.
This patch makes it prefer matches closer to the middle that tend to
balance recursion. This change improves the speed of a pathological
test case from 1100s to 9s.
Included is a smaller test that has a roughly 50x safety margin on the
performance it accepts. It's likely to fail on pure builds because
difflib also has a recursion-balancing problem.
| author | Matt Mackall <mpm@selenic.com> |
|---|---|
| date | Thu, 21 Apr 2016 22:04:11 -0500 |
| parents | 9a8363d23419 |
| children | 87d4a6c5567e |
| files | mercurial/bdiff.c tests/test-issue4074.t |
| diffstat | 2 files changed, 33 insertions(+), 3 deletions(-) [+] |
line wrap: on
line diff
--- a/mercurial/bdiff.c Thu Apr 21 21:05:26 2016 -0500 +++ b/mercurial/bdiff.c Thu Apr 21 22:04:11 2016 -0500 @@ -148,7 +148,7 @@ static int longest_match(struct line *a, struct line *b, struct pos *pos, int a1, int a2, int b1, int b2, int *omi, int *omj) { - int mi = a1, mj = b1, mk = 0, mb = 0, i, j, k; + int mi = a1, mj = b1, mk = 0, mb = 0, i, j, k, half = (a1 + a2) / 2; for (i = a1; i < a2; i++) { /* skip all lines in b after the current block */ @@ -165,8 +165,9 @@ pos[j].pos = i; pos[j].len = k; - /* best match so far? */ - if (k > mk || (k == mk && i <= mi)) { + /* best match so far? we prefer matches closer + to the middle to balance recursion */ + if (k > mk || (k == mk && (i <= mi || i < half))) { mi = i; mj = j; mk = k;
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/tests/test-issue4074.t Thu Apr 21 22:04:11 2016 -0500 @@ -0,0 +1,29 @@ +#require no-pure + +A script to generate nasty diff worst-case scenarios: + + $ cat > s.py <<EOF + > import random + > for x in xrange(100000): + > print + > if random.randint(0, 100) >= 50: + > x += 1 + > print hex(x) + > EOF + + $ hg init a + $ cd a + +Check in a big file: + + $ python ../s.py > a + $ hg ci -qAm0 + +Modify it: + + $ python ../s.py > a + +Time a check-in, should never take more than 10 seconds user time: + + $ hg ci --time -m1 + time: real .* secs .user [0-9][.].* sys .* (re)