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.
#require test-repo slow docker
Ensure debuild doesn't run the testsuite, as that could get silly.
$ DEB_BUILD_OPTIONS=nocheck
$ export DEB_BUILD_OPTIONS
$ OUTPUTDIR=`pwd`
$ export OUTPUTDIR
$ cd "$TESTDIR"/..
$ make docker-debian-jessie > $OUTPUTDIR/build.log 2>&1
$ cd $OUTPUTDIR
$ ls *.deb
mercurial-common_*.deb (glob)
mercurial_*.deb (glob)
We check debian package contents with portable tools so that when
we're on non-debian machines we can still test the packages that are
built using docker.
main deb should have .so but no .py
$ ar x mercurial_*.deb
$ tar tf data.tar* | egrep '(localrepo|parsers)'
./usr/lib/python2.7/dist-packages/mercurial/parsers*.so (glob)
mercurial-common should have .py but no .so or .pyc
$ ar x mercurial-common_*.deb
$ tar tf data.tar* | egrep '(localrepo|parsers)'
./usr/lib/python2.7/dist-packages/mercurial/pure/parsers.py
./usr/lib/python2.7/dist-packages/mercurial/localrepo.py