threads / patch / 29415

patchmergetool: Suppress stderr and fix the "both added" test

Subject: [PATCH] mergetool: Suppress stderr and fix the "both added" test

## tl;dr

3 messages between Jan 21, 2012 and Jan 21, 2012. Diffs are folded; open one to read it.

replies: 2people: 2as markdown or json

David Aguilar· Jan 21, 2012, 10:26 UTC · lore

Silence error messages when "git checkout-index" is used to checkout a stage that does not exist. This can happen now that mergetool calls checkout_staged_file() unconditionally when creating the temporary $BASE, $LOCAL, and $REMOTE files.

Fix the test so that it checks the contents of the "both added" file. The test was passing as a consequence of accidentally handing a bad path to "cat".

Signed-off-by: David Aguilar <davvid@gmail.com>
---
This applies on top of da/maint-mergetool-twoway in pu.
 git-mergetool.sh     |    4 +++-
 t/t7610-mergetool.sh |    5 +++--
 2 files changed, 6 insertions(+), 3 deletions(-)
Show changes to 2 files +6 −3

git-mergetool.sh, t/t7610-mergetool.sh

diff --git a/git-mergetool.sh b/git-mergetool.sh
index 24bedc5..a9f23f7 100755
--- a/git-mergetool.sh
+++ b/git-mergetool.sh
@@ -181,7 +181,9 @@ stage_submodule () {
 }
 
 checkout_staged_file () {
-    tmpfile=$(expr "$(git checkout-index --temp --stage="$1" "$2")" : '\([^	]*\)	')
+    tmpfile=$(expr \
+	    "$(git checkout-index --temp --stage="$1" "$2" 2>/dev/null)" \
+	    : '\([^	]*\)	')
 
     if test $? -eq 0 -a -n "$tmpfile" ; then
 	mv -- "$(git rev-parse --show-cdup)$tmpfile" "$3"
diff --git a/t/t7610-mergetool.sh b/t/t7610-mergetool.sh
index 2272743..3f261a3 100755
--- a/t/t7610-mergetool.sh
+++ b/t/t7610-mergetool.sh
@@ -465,8 +465,9 @@ test_expect_success 'directory vs modified submodule' '
 test_expect_success 'file with no base' '
     git checkout -b test13 branch1 &&
     test_must_fail git merge master &&
-    git mergetool --no-prompt --tool mybase -- base &&
-    test "$(cat "$MERGED")" = "" &&
+    git mergetool --no-prompt --tool mybase -- both &&
+    >expected &&
+    test_cmp both expected &&
     git reset --hard master >/dev/null 2>&1
 '
 
-- 
1.7.7.166.g1cd0c
Junio C Hamano· Jan 21, 2012, 21:27 UTC · re: David Aguilar · lore

Re: [PATCH] mergetool: Suppress stderr and fix the "both added" test

David Aguilar <davvid@gmail.com> writes:
Show 12 quoted lines
> Silence error messages when "git checkout-index" is used to
> checkout a stage that does not exist.  This can happen now that
> mergetool calls checkout_staged_file() unconditionally when
> creating the temporary $BASE, $LOCAL, and $REMOTE files.
>
> Fix the test so that it checks the contents of the "both added"
> file.  The test was passing as a consequence of accidentally
> handing a bad path to "cat".
>
> Signed-off-by: David Aguilar <davvid@gmail.com>
> ---
> This applies on top of da/maint-mergetool-twoway in pu.
Thanks.

It might make sense to squash this into the previous patch, which luckily hasn't hit 'next' yet, though---which I can do locally without need for re-send if you like.

David Aguilar· Jan 21, 2012, 23:58 UTC · re: Junio C Hamano · lore

Re: [PATCH] mergetool: Suppress stderr and fix the "both added" test

On Jan 21, 2012, at 1:27 PM, Junio C Hamano <gitster@pobox.com> wrote:
Show 20 quoted lines
> David Aguilar <davvid@gmail.com> writes:
>
>> Silence error messages when "git checkout-index" is used to
>> checkout a stage that does not exist.  This can happen now that
>> mergetool calls checkout_staged_file() unconditionally when
>> creating the temporary $BASE, $LOCAL, and $REMOTE files.
>>
>> Fix the test so that it checks the contents of the "both added"
>> file.  The test was passing as a consequence of accidentally
>> handing a bad path to "cat".
>>
>> Signed-off-by: David Aguilar <davvid@gmail.com>
>> ---
>> This applies on top of da/maint-mergetool-twoway in pu.
>
> Thanks.
>
> It might make sense to squash this into the previous patch, which luckily
> hasn't hit 'next' yet, though---which I can do locally without need for
> re-send if you like.
Yes, please squash this commit. This is certainly a fix up.

Thanks, David

← back to recent threads