view tests/test-merge-internal-tools-pattern @ 7732:3793802ea41b

Make util.find_exe alway returns existing file, fixing issue1459

It seems like the old behaviour with different handling for commands with and
without path was intended, but I think this behaviour of util.find_exe is
better:
* Always returns existing file
* or None if command not found - no default
* Windows: Returned file thus always ends with extension from PATHEXT

This fixes http://www.selenic.com/mercurial/bts/issue1459. The change might
fix other unintended behaviour too.

#!/bin/sh

# make sure that the internal merge tools (internal:fail, internal:local, and
# internal:other) are used when matched by a merge-pattern in hgrc

unset HGMERGE # make sure HGMERGE doesn't interfere with the test

hg init

echo "# initial file contents"
echo "line 1" > f
echo "line 2" >> f
echo "line 3" >> f
hg commit -Am "revision 0" -d "1000000 0"
cat f
echo "# branch 1: editing line 1"
sed 's/line 1/first line/' f > f.new
mv f.new f
hg commit -Am "edited first line" -d "1000000 0"

echo "# branch 2: editing line 3"
hg update 0
sed 's/line 3/third line/' f > f.new
mv f.new f
hg commit -Am "edited third line" -d "1000000 0"

echo "# merge using internal:fail tool"
echo "[merge-patterns]" > .hg/hgrc
echo "* = internal:fail" >> .hg/hgrc
hg merge
cat f
hg stat

echo "# merge using internal:local tool"
hg update -C 2
sed 's/internal:fail/internal:local/' .hg/hgrc > .hg/hgrc.new
mv .hg/hgrc.new .hg/hgrc
hg merge
cat f
hg stat

echo "# merge using internal:other tool"
hg update -C 2
sed 's/internal:local/internal:other/' .hg/hgrc > .hg/hgrc.new
mv .hg/hgrc.new .hg/hgrc
hg merge
cat f
hg stat

echo "# merge using default tool"
hg update -C 2
rm .hg/hgrc
hg merge
cat f
hg stat