bisect: avoid copying ancestor list for non-merge commits
During a bisection, hg needs to compute a list of all ancestors for every
candidate commit. This is accomplished via a bottom-up traversal of the set of
candidates, during which each revision's ancestor list is populated using the
ancestor list of its parent(s). Previously, this involved copying the entire
list, which could be very long in if the bisection range was large.
To help improve this, we can observe that each candidate commit is visited
exactly once, at which point its ancestor list is copied into its children's
lists and then dropped. In the case of non-merge commits, a commit's ancestor
list consists exactly of its parent's list plus itself. This means that we can
trivially reuse the parent's existing list for one of its non-merge children,
which avoids copying entirely if that commit is the parent's only child. This
makes bisections over linear ranges of commits much faster.
During some informal testing in the large publicly-available `mozilla-central`
repository, this noticeably sped up bisections over large ranges of history:
Setup:
$ cd mozilla-central
$ hg bisect --reset
$ hg bisect --good 0
$ hg log -r tip -T '{rev}\n'
628417
Test:
$ time hg bisect --bad tip --noupdate
Before:
real 3m35.927s
user 3m35.553s
sys 0m0.319s
After:
real 1m41.142s
user 1m40.810s
sys 0m0.285s
#!/bin/sh -e
#
# Build a Mercurial debian package from the current repo
#
# Tested on Jessie (stable as of original script authoring.)
. $(dirname $0)/packagelib.sh
ROOTDIR=$(cd $(dirname $0)/../.. > /dev/null; pwd)
BUILD=1
CLEANUP=1
DISTID=`(lsb_release -is 2> /dev/null | tr '[:upper:]' '[:lower:]') || echo debian`
CODENAME=`lsb_release -cs 2> /dev/null || echo unknown`
DEBFLAGS=-b
cleanup() {
if [ "$CLEANUP" ]; then
rm -r "$ROOTDIR/debian";
fi
}
while [ "$1" ]; do
case "$1" in
--distid )
shift
DISTID="$1"
shift
;;
--codename )
shift
CODENAME="$1"
shift
;;
--cleanup )
shift
BUILD=
;;
--build )
shift
CLEANUP=
;;
--source-only )
shift
DEBFLAGS=-S
;;
* )
echo "Invalid parameter $1!" 1>&2
exit 1
;;
esac
done
cd "$ROOTDIR"
trap 'cleanup' EXIT
set -u
if [ ! -d .hg ]; then
printf "You are inside %s, which is not the root of a Mercurial repository\n" $(pwd) 1>&2
exit 1
fi
gethgversion
debver="$version"
if [ -n "$type" ] ; then
debver="$debver~$type"
fi
if [ -n "$distance" ] ; then
debver="$debver+$distance-$CODENAME-$node"
elif [ "$DEBFLAGS" = "-S" ] ; then
# for building a ppa (--source-only) for a release (distance == 0), we need
# to version the distroseries so that we can upload to launchpad
debver="$debver~${CODENAME}1"
fi
control=debian/control
changelog=debian/changelog
if [ "$BUILD" ]; then
if [ -d debian ] ; then
printf "Error! debian control directory already exists at %s/debian\n" $(pwd)
exit 1
fi
cp -r "$ROOTDIR"/contrib/packaging/debian debian
sed -i.tmp "s/__VERSION__/$debver/" $changelog
sed -i.tmp "s/__DATE__/$(date --rfc-2822)/" $changelog
sed -i.tmp "s/__CODENAME__/$CODENAME/" $changelog
rm $changelog.tmp
# remove the node from the version string
SRCFILE="mercurial_$(echo $debver | sed "s,-$node,,").orig.tar.gz"
"$ROOTDIR/hg" archive $SRCFILE
mv $SRCFILE ..
debuild -us -uc -i -I $DEBFLAGS
if [ $? != 0 ]; then
echo 'debuild failed!'
exit 1
fi
fi
if [ "$CLEANUP" ] ; then
echo
OUTPUTDIR=${OUTPUTDIR:=packages/$DISTID-$CODENAME}
mkdir -p "$OUTPUTDIR"
find ../mercurial*.deb ../mercurial_*.build* ../mercurial_*.changes \
../mercurial*.dsc ../mercurial*.gz \
-type f -newer $control -print0 2>/dev/null | \
xargs -Inarf -0 mv narf "$OUTPUTDIR"
echo "Built packages for $debver:"
find "$OUTPUTDIR" -type f -newer $control -name '*.deb'
fi