Re[2]: [PATCH 3/3] Fixes bug: GIT_PS1_SHOWDIRTYSTATE is no not respect diff.ignoreSubmodules config variable

From: Алексей Крезов <zapped@mail.ru>
Date: Dec 26, 2010, 22:25 UTC
Message-ID: <1824011293.20101227012557@mail.ru>
In-Reply-To: <4D15E48A.9050805@web.de>

Hello!

I`m sorry, but I`m newbie in making and distributing of public patches. So, don't beat me much ))

JL> it would have been easier for me if the commit message would have JL> described the problem you tried to fix a bit more in detail ;-). My problem was in the following. I use Git on Windows XP under Cygwin. So its perfomance is slower than on *nix, I guess. My project has 40 submodules (external libs) and some of them could have untracked files (for some reasons). So when I run any command in Bash after its execution Bash "thought" for 2-3 seconds. That was annoying. I do not remember the version of Git I used at that moment but I remember it was an update from 1.6.x to early 1.7.x. So I decided to roll back to 1.6.x )) Then there was some updates of Git. But after updating the problem still happened. When I tried to discover the reason of such a behaviour I found that Git got status for all submodules including untracked content and so marked them with * Then I read manual and found diff.ignoreSubmodules and tried to set up for each submodule in a .gitmodules but nothing changed (as it seemed to me at that time). So I've found the easiest way to solve my problem - this patch ) Maybe after this patch there was some changes in Git solved this problem but I did not investigate it, sorry.

JL> 2) If diff.ignoreSubmodules is unset it leads to an error JL> every time the prompt is displayed: JL> 'fatal: bad --ignore-submodules argument:' Yeah, you're right

JL> Am 25.12.2010 02:20, schrieb Zapped:

Show 19 quoted lines

>> Signed-off-by: Zapped <zapped@mail.ru>
>> ---
>>  contrib/completion/git-completion.bash |    3 ++-
>>  1 files changed, 2 insertions(+), 1 deletions(-)
>> 
>> diff --git a/contrib/completion/git-completion.bash b/contrib/completion/git-completion.bash
>> index d3037fc..50fc385 100755
>> --- a/contrib/completion/git-completion.bash
>> +++ b/contrib/completion/git-completion.bash
>> @@ -280,7 +280,8 @@ __git_ps1 ()
>>               elif [ "true" = "$(git rev-parse --is-inside-work-tree 2>/dev/null)" ]; then
>>                       if [ -n "${GIT_PS1_SHOWDIRTYSTATE-}" ]; then
>>                               if [ "$(git config --bool bash.showDirtyState)" != "false" ]; then
>> -                                     git diff --no-ext-diff --quiet --exit-code || w="*"
>> +                                     is=$(git config diff.ignoreSubmodules)
>> +                                     git diff --no-ext-diff --quiet --exit-code --ignore-submodules=$is || w="*"
>>                                       if git rev-parse --quiet --verify HEAD >/dev/null; then
>>                                               git diff-index --cached --quiet HEAD -- || i="+"
>>                                       else

JL> Thanks for resubmitting this as an inline patch for review (although JL> it would have been easier for me if the commit message would have JL> described the problem you tried to fix a bit more in detail ;-).

JL> After testing this patch it looks like it has a few issues:

JL> 1) it will break any per-submodule configuration done via JL> the 'submodule.<name>.ignore' setting in .git/config or JL> .gitmodules, as using the --ignore-submodules option JL> overrides those while only setting 'diff.ignoreSubmodules' JL> should not do that.

JL> 2) If diff.ignoreSubmodules is unset it leads to an error JL> every time the prompt is displayed: JL> 'fatal: bad --ignore-submodules argument:'

JL> 3) And for me it didn't change the behavior at all:

JL> - The '*' in the prompt vanishes as I set diff.ignoreSubmodules JL> as expected with or without your patch. JL> Am I missing something here?

JL> - The real problem here is that the '+' never goes away even JL> when 'diff.ignoreSubmodules' is set to 'all'. This is due JL> to the fact that 'diff.ignoreSubmodules' is only honored by JL> "git diff", but not by "git diff-index".

JL> So the real issue here seems to be the "git diff-index" call, which JL> doesn't honor the 'diff.ignoreSubmodules' setting. In commit 37aea37 JL> Dscho (CCed) introduced this configuration setting while explicitly JL> stating that it only affects porcelain. As the other config options JL> always influence porcelain and plumbing, it looks like we would want JL> to have this option honored by plumbing too, no?

JL> So are there any reasons for the plumbing diff commands not to honor JL> the diff.ignoreSubmodules setting?