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.
$ hg init
$ cat > .hg/hgrc <<EOF
> [extensions]
> prefixfilter = prefix.py
> [encode]
> *.txt = stripprefix: Copyright 2046, The Masters
> [decode]
> *.txt = insertprefix: Copyright 2046, The Masters
> EOF
$ cat > prefix.py <<EOF
> from mercurial import error
> def stripprefix(s, cmd, filename, **kwargs):
> header = '%s\n' % cmd
> if s[:len(header)] != header:
> raise error.Abort('missing header "%s" in %s' % (cmd, filename))
> return s[len(header):]
> def insertprefix(s, cmd):
> return '%s\n%s' % (cmd, s)
> def reposetup(ui, repo):
> repo.adddatafilter('stripprefix:', stripprefix)
> repo.adddatafilter('insertprefix:', insertprefix)
> EOF
$ cat > .hgignore <<EOF
> .hgignore
> prefix.py
> prefix.pyc
> EOF
$ cat > stuff.txt <<EOF
> Copyright 2046, The Masters
> Some stuff to ponder very carefully.
> EOF
$ hg add stuff.txt
$ hg ci -m stuff
Repository data:
$ hg cat stuff.txt
Some stuff to ponder very carefully.
Fresh checkout:
$ rm stuff.txt
$ hg up -C
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat stuff.txt
Copyright 2046, The Masters
Some stuff to ponder very carefully.
$ echo "Very very carefully." >> stuff.txt
$ hg stat
M stuff.txt
$ echo "Unauthorized material subject to destruction." > morestuff.txt
Problem encoding:
$ hg add morestuff.txt
$ hg ci -m morestuff
abort: missing header "Copyright 2046, The Masters" in morestuff.txt
[255]
$ hg stat
M stuff.txt
A morestuff.txt