merge-recursive: never leave index unmerged while recursing
When you are trying to come up with the final result (i.e. depth=0), you
want to record how the conflict arose by registering the state of the
common ancestor, your branch and the other branch in the index, hence you
want to do update_stages().
When you are merging with positive depth, that is because of a criss-cross
merge situation. In such a case, you would need to record the tentative
result, with conflict markers and all, as if the merge went cleanly, even
if there are conflicts, in order to write it out as a tree object later to
be used as a common ancestor tree.
update_file() calls update_file_flags() with update_cache=1 to signal that
the result needs to be written to the index at stage #0 (i.e. merged), and
the code should not clobber the index further by calling update_stages().
The codepath to deal with rename/delete conflict in a recursive merge
however left the index unmerged.
Signed-off-by: Dave Olszewski <cxreg@pobox.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-05-09 23:49:59 +02:00
|
|
|
#!/bin/sh
|
|
|
|
|
|
|
|
test_description='merge-recursive backend test'
|
|
|
|
|
|
|
|
. ./test-lib.sh
|
|
|
|
|
|
|
|
# A <- create some files
|
|
|
|
# / \
|
|
|
|
# B C <- cause rename/delete conflicts between B and C
|
|
|
|
# / \
|
|
|
|
# |\ /|
|
|
|
|
# | D E |
|
|
|
|
# | \ / |
|
|
|
|
# | X |
|
|
|
|
# | / \ |
|
|
|
|
# | / \ |
|
|
|
|
# |/ \|
|
|
|
|
# F G <- merge E into B, D into C
|
|
|
|
# \ /
|
|
|
|
# \ /
|
|
|
|
# \ /
|
|
|
|
# H <- recursive merge crashes
|
|
|
|
#
|
|
|
|
|
|
|
|
# initialize
|
|
|
|
test_expect_success 'setup repo with criss-cross history' '
|
|
|
|
mkdir data &&
|
|
|
|
|
|
|
|
# create a bunch of files
|
|
|
|
n=1 &&
|
|
|
|
while test $n -le 10
|
|
|
|
do
|
|
|
|
echo $n > data/$n &&
|
|
|
|
n=$(($n+1)) ||
|
2015-03-25 06:29:52 +01:00
|
|
|
return 1
|
merge-recursive: never leave index unmerged while recursing
When you are trying to come up with the final result (i.e. depth=0), you
want to record how the conflict arose by registering the state of the
common ancestor, your branch and the other branch in the index, hence you
want to do update_stages().
When you are merging with positive depth, that is because of a criss-cross
merge situation. In such a case, you would need to record the tentative
result, with conflict markers and all, as if the merge went cleanly, even
if there are conflicts, in order to write it out as a tree object later to
be used as a common ancestor tree.
update_file() calls update_file_flags() with update_cache=1 to signal that
the result needs to be written to the index at stage #0 (i.e. merged), and
the code should not clobber the index further by calling update_stages().
The codepath to deal with rename/delete conflict in a recursive merge
however left the index unmerged.
Signed-off-by: Dave Olszewski <cxreg@pobox.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-05-09 23:49:59 +02:00
|
|
|
done &&
|
|
|
|
|
|
|
|
# check them in
|
|
|
|
git add data &&
|
|
|
|
git commit -m A &&
|
|
|
|
git branch A &&
|
|
|
|
|
|
|
|
# a file in one branch
|
|
|
|
git checkout -b B A &&
|
|
|
|
git rm data/9 &&
|
|
|
|
git add data &&
|
|
|
|
git commit -m B &&
|
|
|
|
|
|
|
|
# with a branch off of it
|
|
|
|
git branch D &&
|
|
|
|
|
|
|
|
# put some commits on D
|
|
|
|
git checkout D &&
|
|
|
|
echo testD > data/testD &&
|
|
|
|
git add data &&
|
|
|
|
git commit -m D &&
|
|
|
|
|
|
|
|
# back up to the top, create another branch and cause
|
|
|
|
# a rename conflict with the file we deleted earlier
|
|
|
|
git checkout -b C A &&
|
|
|
|
git mv data/9 data/new-9 &&
|
|
|
|
git add data &&
|
|
|
|
git commit -m C &&
|
|
|
|
|
|
|
|
# with a branch off of it
|
|
|
|
git branch E &&
|
|
|
|
|
|
|
|
# put a commit on E
|
|
|
|
git checkout E &&
|
|
|
|
echo testE > data/testE &&
|
|
|
|
git add data &&
|
|
|
|
git commit -m E &&
|
|
|
|
|
|
|
|
# now, merge E into B
|
|
|
|
git checkout B &&
|
|
|
|
test_must_fail git merge E &&
|
|
|
|
# force-resolve
|
|
|
|
git add data &&
|
|
|
|
git commit -m F &&
|
|
|
|
git branch F &&
|
|
|
|
|
|
|
|
# and merge D into C
|
|
|
|
git checkout C &&
|
|
|
|
test_must_fail git merge D &&
|
|
|
|
# force-resolve
|
|
|
|
git add data &&
|
|
|
|
git commit -m G &&
|
|
|
|
git branch G
|
|
|
|
'
|
|
|
|
|
2018-08-04 01:09:23 +02:00
|
|
|
test_expect_success 'recursive merge between F and G does not cause segfault' '
|
merge-recursive: never leave index unmerged while recursing
When you are trying to come up with the final result (i.e. depth=0), you
want to record how the conflict arose by registering the state of the
common ancestor, your branch and the other branch in the index, hence you
want to do update_stages().
When you are merging with positive depth, that is because of a criss-cross
merge situation. In such a case, you would need to record the tentative
result, with conflict markers and all, as if the merge went cleanly, even
if there are conflicts, in order to write it out as a tree object later to
be used as a common ancestor tree.
update_file() calls update_file_flags() with update_cache=1 to signal that
the result needs to be written to the index at stage #0 (i.e. merged), and
the code should not clobber the index further by calling update_stages().
The codepath to deal with rename/delete conflict in a recursive merge
however left the index unmerged.
Signed-off-by: Dave Olszewski <cxreg@pobox.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-05-09 23:49:59 +02:00
|
|
|
git merge F
|
|
|
|
'
|
|
|
|
|
|
|
|
test_done
|