git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Stefan Beller <sbeller@google.com>
To: Jeff King <peff@peff.net>
Cc: Joey Hess <id@joeyh.name>, "git@vger.kernel.org" <git@vger.kernel.org>
Subject: Re: bug: git submodule add fails when .git is a symlink
Date: Wed, 20 Apr 2016 10:31:47 -0700	[thread overview]
Message-ID: <CAGZ79kZ2A6c9H7C4xPXJ1Lmh1TaVy+VgcjxvKrgcJb0-4LRvhA@mail.gmail.com> (raw)
In-Reply-To: <CAGZ79kbpyu34kz7gRgro-HZqNvpmRoCsyQatZRbxP19bk1gkfg@mail.gmail.com>

On Wed, Apr 20, 2016 at 9:41 AM, Stefan Beller <sbeller@google.com> wrote:
> On Wed, Mar 2, 2016 at 12:49 AM, Jeff King <peff@peff.net> wrote:
>> On Tue, Mar 01, 2016 at 07:17:20PM -0400, Joey Hess wrote:
>>
>>> Junio C Hamano wrote:
>>> > A more pertinent question may be which version of Git did the above
>>> > ever work, I guess.  We fairly liberally chdir around and I do not
>>> > think we deliberately avoid assuming that "cd .git && cd .." might
>>> > not come back to the original directory, for example, so I wouldn't
>>> > be surprised if it never worked.
>>>
>>> IIRC git used symlinks for .git in submodules before version 1.7.8, so I
>>> guess that older versions supported that pretty well.
>>>
>>> This one case is the only time I've seen a symlink for .git present a
>>> problem so far.
>>
>> Fortunately you provided a simple reproduction case, so it is easy to
>> bisect. It did work in v1.7.8, and broke in d75219b (submodules: always
>> use a relative path from gitdir to work tree, 2012-03-04). Not
>> surprising, I guess. It presumably worked before only because we were
>> using absolute paths.
>
> So I was looking into this bug again, as it was linked from another bug report.
>
> fatal: Could not chdir to '../../../sub': No such file or directory
>
> sounds like a path issue with the prefix thing.
>
> Using the " echo "gitdir: ../gitdir/.git" > .git" workaround does still work,
> I'll see if there is another way to fix it with actual links.

So I debugged into that using a test case

test_expect_success 'submodules are not confused by linked gitdir' '
        git init gitdir &&
        mkdir worktree &&
        (
                cd worktree &&
                #echo "gitdir: ../gitdir/.git" >.git &&
                ln -s ../gitdir/.git .git &&
                test_pause &&
                git submodule add ../ sub &&
                test_pause
        ) &&
        test_pause &&
        git status
'

My observation: the error comes up in `git submodule add`,
which consists of 2 parts: the cloning and then the checkout.
The cloning works fine using the new "submodule--helper update-clone",
the checkout however breaks. So in your original test case you can go
into the submodule and run

    git checkout -f -q

and you'll get the error message

    fatal: Could not chdir to '../../../sub': No such file or directory

Looking at sub/.git:

    $ cat .git
    gitdir: ../.git/modules/sub
    $ ls ../.git/modules/sub
    HEAD  branches config description  hooks  info  logs objects
packed-refs  ref

so that gitlink of the submodule works just fine. The problem is `git checkout`
or any other core command (I tested with `git status`) doesn't like
the gitlink pointing to a directory which is symlinked.

I think the issue is a wrongly configured "core.worktree"
in gitdir/.git/modules/sub/config which contains the "../../../sub".

So I think the fix needs to be in the vicinity of
builtin/submodule--helper.c:module_clone
near the end of the function, where core.worktree is set.

Thanks,
Stefan


>
>
>>
>> -Peff
>> --
>> To unsubscribe from this list: send the line "unsubscribe git" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html

      reply	other threads:[~2016-04-20 17:31 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-03-01 20:42 bug: git submodule add fails when .git is a symlink Joey Hess
2016-03-01 21:39 ` Stefan Beller
2016-03-01 21:49   ` Stefan Beller
2016-03-01 23:14     ` Joey Hess
2016-03-01 21:50   ` Junio C Hamano
2016-03-01 23:17     ` Joey Hess
2016-03-02  8:49       ` Jeff King
2016-04-20 16:41         ` Stefan Beller
2016-04-20 17:31           ` Stefan Beller [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: http://vger.kernel.org/majordomo-info.html

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAGZ79kZ2A6c9H7C4xPXJ1Lmh1TaVy+VgcjxvKrgcJb0-4LRvhA@mail.gmail.com \
    --to=sbeller@google.com \
    --cc=git@vger.kernel.org \
    --cc=id@joeyh.name \
    --cc=peff@peff.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://80x24.org/mirrors/git.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).