git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH] commit: configure submodules

From
Jens Lehmann <jens.lehmann@web.de>
Date
Sep 24, 2012, 20:10 UTC
Message-ID
<5060BE4F.2040506@web.de>
In-Reply-To
<CAGHpTBLs2tgUunDtFMhw7QMiuB2oXBtQJLKpLckhEyDp7Zr+Qg@mail.gmail.com>
Am 24.09.2012 21:59, schrieb Orgad Shaneh:
Show 83 quoted lines
> On Mon, Sep 24, 2012 at 9:56 PM, Jens Lehmann <Jens.Lehmann@web.de> wrote:
>> Am 24.09.2012 21:16, schrieb Orgad Shaneh:
>>> On Mon, Sep 24, 2012 at 9:06 PM, Junio C Hamano <gitster@pobox.com> wrote:
>>>> Orgad Shaneh <orgads@gmail.com> writes:
>>>>
>>>>> That is not correct. git-config is ignored as well for commit.
>>>>
>>>> What do you mean?  As far as I can tell, if you have
>>>>
>>>>     [submodule "var"]
>>>>         path = var
>>>>         ignore = dirty
>>>>
>>>> in $GIT_DIR/config, a work-tree-dirty submodule "var" is not
>>>> reported by "git status" and "git commit" without your patch, and
>>>> your patch does not seem to break that.  The only difference your
>>>> patch makes is that if you had the above three-line block in
>>>> the .gitmodules file and not in $GIT_DIR/config, "git status"
>>>> ignored the dirtyness in the working tree, but "git commit" did
>>>> notice and report it.
>>>>
>>>> What am I missing?
>>>>
>>>>
>>>
>>> I have:
>>> [submodule "mod"]
>>>         url = [...]
>>>         ignore = dirty
>>>
>>> in .git/config, and I removed the ignore part from .gitmodules to be even.
>>>
>>> I made a change inside mod, git status doesn't report its dirtiness,
>>> while git commit does.
>>>
>>> git status:
>>> # On branch master
>>> # Changes to be committed:
>>> #   (use "git reset HEAD <file>..." to unstage)
>>> #
>>> #       modified:   foo
>>> #
>>> # Changes not staged for commit:
>>> #   (use "git add <file>..." to update what will be committed)
>>> #   (use "git checkout -- <file>..." to discard changes in working directory)
>>> #
>>> #       modified:   .gitmodules
>>> #
>>>
>>> git commit:
>>> # Please enter the commit message for your changes. Lines starting
>>> # with '#' will be ignored, and an empty message aborts the commit.
>>> # On branch master
>>> # Changes to be committed:
>>> #   (use "git reset HEAD <file>..." to unstage)
>>> #
>>> #       modified:   foo
>>> #
>>> # Changes not staged for commit:
>>> #   (use "git add <file>..." to update what will be committed)
>>> #   (use "git checkout -- <file>..." to discard changes in working directory)
>>> #   (commit or discard the untracked or modified content in submodules)
>>> #
>>> #       modified:   .gitmodules
>>> #       modified:   mod (modified content)
>>> #
>>>
>>> Now I get it! That's because I don't have submodule.mod.path!
>>> config_name_for_path only gets initialized if path exists. Apparently
>>> git submodule init doesn't configure 'path', so it stays
>>> uninitialized.
>>
>> But submodule.mod.path should only be set in .gitmodules, not in
>> $GIT_DIR/config. Did you just remove the ignore setting from
>> .gitmodules or the path too?
> 
> Just the ignore, and my patch of course.
> 
> If it is not set in $GIT_DIR/config, then config_name_for_path is not
> initialized, and if it is not initialized, then
> set_diffopt_flags_from_submodule_config does nothing
> (handle_ignore_submodules_arg is not called). That is the main
> problem.

But config_name_for_path can only be set via .gitmodules. It is set in parse_submodule_config_option() called by submodule_config() which is called from gitmodules_config() ... but only if .gitmodules doesn't have a merge conflict.

So either your .gitmodules has a merge conflict or the logic setting gitmodules_is_unmerged in gitmodules_config() is buggy.

Previous: Orgad ShanehNext: Junio C Hamano
Message 13 of 14 in “commit: configure submodules”
  1. commit: configure submodulesOrgad Shaneh, Sep 23, 2012
  2. Junio C HamanoSep 23, 2012
  3. Jens LehmannSep 23, 2012
  4. Junio C HamanoSep 24, 2012
  5. Junio C HamanoSep 24, 2012
  6. Jens LehmannSep 24, 2012
  7. Junio C HamanoSep 24, 2012
  8. Orgad ShanehSep 24, 2012
  9. Junio C HamanoSep 24, 2012
  10. Orgad ShanehSep 24, 2012
  11. Jens LehmannSep 24, 2012
  12. Orgad ShanehSep 24, 2012
  13. Jens LehmannSep 24, 2012
  14. Junio C HamanoSep 24, 2012

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.