Avoid extra filelogs entries.
Right now, there are some situations in which localrepo.filecommit can
create filelog entries even though they're not needed. For example:
- permissions for a file have changed;
- qrefresh can create a filelog entry identical to its parent (see the
added test);
- convert-repo creates extra filelog entries in every merge where the
first parent has added files (for example, changeset ebebe9577a1a of
the kernel repo added extra filelog entries to files in the
arch/blackfin directory, even though the merge should only touch the
drivers/ata directory). This makes "hg log file" in a converted repo
less useful than it could be, since it may mention many merges that
don't actually touch that specific file.
They all come from the same basic problem: localrepo.commit (through
filecommit) creates new filelog entries for all files passed to it
(except for some cases during a merge).
Patch and test case provided by Benoit.
This should fix issue351.
#!/bin/sh
hg init a
mkdir a/d1
mkdir a/d1/d2
echo line 1 > a/a
echo line 1 > a/d1/d2/a
hg --cwd a ci -d '0 0' -Ama
echo line 2 >> a/a
hg --cwd a ci -u someone -d '1 0' -m'second change'
echo % import exported patch
hg clone -r0 a b
hg --cwd a export tip > tip.patch
hg --cwd b import ../tip.patch
echo % message should be same
hg --cwd b tip | grep 'second change'
echo % committer should be same
hg --cwd b tip | grep someone
rm -r b
echo % import of plain diff should fail without message
hg clone -r0 a b
hg --cwd a diff -r0:1 > tip.patch
hg --cwd b import ../tip.patch
rm -r b
echo % import of plain diff should be ok with message
hg clone -r0 a b
hg --cwd a diff -r0:1 > tip.patch
hg --cwd b import -mpatch ../tip.patch
rm -r b
echo % hg -R repo import
# put the clone in a subdir - having a directory named "a"
# used to hide a bug.
mkdir dir
hg clone -r0 a dir/b
hg --cwd a export tip > dir/tip.patch
cd dir
hg -R b import tip.patch
cd ..
rm -r dir
echo % import from stdin
hg clone -r0 a b
hg --cwd a export tip | hg --cwd b import -
rm -r b
echo % override commit message
hg clone -r0 a b
hg --cwd a export tip | hg --cwd b import -m 'override' -
hg --cwd b tip | grep override
rm -r b
cat > mkmsg.py <<EOF
import email.Message, sys
msg = email.Message.Message()
msg.set_payload('email commit message\n' + open('tip.patch').read())
msg['Subject'] = 'email patch'
msg['From'] = 'email patcher'
sys.stdout.write(msg.as_string())
EOF
echo % plain diff in email, subject, message body
hg clone -r0 a b
hg --cwd a diff -r0:1 > tip.patch
python mkmsg.py > msg.patch
hg --cwd b import ../msg.patch
hg --cwd b tip | grep email
rm -r b
echo % plain diff in email, no subject, message body
hg clone -r0 a b
grep -v '^Subject:' msg.patch | hg --cwd b import -
rm -r b
echo % plain diff in email, subject, no message body
hg clone -r0 a b
grep -v '^email ' msg.patch | hg --cwd b import -
rm -r b
echo % plain diff in email, no subject, no message body, should fail
hg clone -r0 a b
egrep -v '^(Subject|email)' msg.patch | hg --cwd b import -
rm -r b
echo % hg export in email, should use patch header
hg clone -r0 a b
hg --cwd a export tip > tip.patch
python mkmsg.py | hg --cwd b import -
hg --cwd b tip | grep second
rm -r b
# bug non regression test
# importing a patch in a subdirectory failed at the commit stage
echo line 2 >> a/d1/d2/a
hg --cwd a ci -u someoneelse -d '1 0' -m'subdir change'
echo % hg import in a subdirectory
hg clone -r0 a b
hg --cwd a export tip | sed -e 's/d1\/d2\///' > tip.patch
dir=`pwd`
cd b/d1/d2 2>&1 > /dev/null
hg import ../../../tip.patch
cd $dir
echo "% message should be 'subdir change'"
hg --cwd b tip | grep 'subdir change'
echo "% committer should be 'someoneelse'"
hg --cwd b tip | grep someoneelse
echo "% should be empty"
hg --cwd b status