git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Junio C Hamano <gitster@pobox.com>
To: "Jin, Di" <di_jin@brown.edu>
Cc: git@vger.kernel.org,  Nikos Vasilakis <nikos_vasilakis@brown.edu>,
	michael@greenberg.science
Subject: Re: EXDEV when re-init with --separate-git-dir option
Date: Wed, 24 Jul 2024 08:39:49 -0700	[thread overview]
Message-ID: <xmqq1q3iztmi.fsf@gitster.g> (raw)
In-Reply-To: <CAKOkDnMGRfQoNygYLiAxPZB2q=VMYvw8kyu1dM=pM843-FH41Q@mail.gmail.com> (Di Jin's message of "Tue, 23 Jul 2024 22:47:19 -0400")

"Jin, Di" <di_jin@brown.edu> writes:

> We discovered that re-init with option --separate-git-dir will throw
> an EXDEV when the target directory is not on the same file system as
> the original directory.

Yup, it is hitting the limitation of your filesystem.  The code
wants to move the original .git directory together with its contents
to a new place, and it makes a single rename() system call to do so.
When the system cannot do so, you'd get the error message you are
seeing ("fatal: unable to move X to Y").

The code path could probably borrow some code to recursively "copy"
directory from the local "git clone" code path, and then invent a
new code to recursively remove the original ".git", and trigger that
new code when rename() fails.

But at that point, only as a fall-back measure, it might be simpler
and much less error prone to spawn a "mv src dst" as a subprocess
using the run_command() API.

It would make a good bite-sized #leftoverbits project for aspiring
new Git contributors.  Any takers?  ;-)

Thanks.


  reply	other threads:[~2024-07-24 15:39 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-07-24  2:47 EXDEV when re-init with --separate-git-dir option Jin, Di
2024-07-24 15:39 ` Junio C Hamano [this message]
2024-07-24 17:14   ` Eric Sunshine

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

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

  git send-email \
    --in-reply-to=xmqq1q3iztmi.fsf@gitster.g \
    --to=gitster@pobox.com \
    --cc=di_jin@brown.edu \
    --cc=git@vger.kernel.org \
    --cc=michael@greenberg.science \
    --cc=nikos_vasilakis@brown.edu \
    /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.
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 NNTP newsgroup(s).