add support for streaming clone.
existing clone code uses pull to get changes from remote repo. is very
slow, uses lots of memory and cpu.
new clone code has server write file data straight to client, client
writes file data straight to disk. memory and cpu used are very low,
clone is much faster over lan.
new client can still clone with pull, can still clone from older servers.
new server can still serve older clients.
#!/bin/sh
# This test tries to exercise the ssh functionality with a dummy script
cat <<'EOF' > dummyssh
#!/bin/sh
# this attempts to deal with relative pathnames
cd `dirname $0`
# check for proper args
if [ $1 != "user@dummy" ] ; then
exit -1
fi
# check that we're in the right directory
if [ ! -x dummyssh ] ; then
exit -1
fi
echo Got arguments 1:$1 2:$2 3:$3 4:$4 5:$5 >> dummylog
$2
EOF
chmod +x dummyssh
echo "# creating 'remote'"
hg init remote
cd remote
echo this > foo
hg ci -A -m "init" -d "1000000 0" foo
cd ..
echo "# clone remote via stream"
hg clone -e ./dummyssh ssh://user@dummy/remote local-stream 2>&1 | \
sed -e 's/[0-9][0-9.]*/XXX/g'
cd local-stream
hg verify
cd ..
echo "# clone remote via pull"
hg clone -e ./dummyssh --pull ssh://user@dummy/remote local
echo "# verify"
cd local
hg verify
echo "# empty default pull"
hg paths
hg pull -e ../dummyssh
echo "# local change"
echo bleah > foo
hg ci -m "add" -d "1000000 0"
echo "# updating rc"
echo "default-push = ssh://user@dummy/remote" >> .hg/hgrc
echo "[ui]" >> .hg/hgrc
echo "ssh = ../dummyssh" >> .hg/hgrc
echo "# find outgoing"
hg out ssh://user@dummy/remote
echo "# find incoming on the remote side"
hg incoming -R ../remote -e ../dummyssh ssh://user@dummy/local
echo "# push"
hg push
cd ../remote
echo "# check remote tip"
hg tip
hg verify
hg cat foo
echo z > z
hg ci -A -m z -d '1000001 0' z
cd ../local
echo r > r
hg ci -A -m z -d '1000002 0' r
echo "# push should fail"
hg push
echo "# push should succeed"
hg push -f
cd ..
cat dummylog