unofficial mirror of bug-gnu-emacs@gnu.org 
 help / color / mirror / code / Atom feed
From: phillip.lord@russet.org.uk (Phillip Lord)
To: Dmitry Gutov <dgutov@yandex.ru>
Cc: 23769@debbugs.gnu.org
Subject: bug#23769: 25.0.95; Mode Line breakage in vc-git
Date: Wed, 15 Jun 2016 23:09:48 +0100	[thread overview]
Message-ID: <871t3y85oz.fsf@russet.org.uk> (raw)
In-Reply-To: <9856a1e4-be41-d781-bfbd-22bac360d18b@yandex.ru> (Dmitry Gutov's message of "Thu, 16 Jun 2016 00:02:12 +0300")

Dmitry Gutov <dgutov@yandex.ru> writes:

> On 06/15/2016 11:48 PM, Phillip Lord wrote:
>
>> I've noticed from instrumenting vc-git that it's not just the
>> symbolic-ref command that returns 128, but several others. So, it might
>> be that the best long term solution would be to make change vc-git--call
>> to set the default directory to the root, which would make GIT_DIR=.git
>> always correct.
>
> The patch looks wrong. Why does the problem script set GIT_DIR to '.git'?

Try it on a pre-commit hook. CWD is set to the root, and GIT_DIR is set
to .git. Other people have found this:

http://longair.net/blog/2011/04/09/missing-git-hooks-documentation/

AFAICT, it's not actually documented by git. Or it is but the
documentation is incomprehensible (hey, it's git, it could happen!).


> In all examples I've found, this variable is set to an absolute value. In
> general, its purpose, it seems, is to point to the '.git' directory when it's
> named otherwise and/or is situated somewhere outside of the current directory
> tree.
>
> The latter situation will break vc-git-root.

That situation does indeed happen. For example, I normally checkout
emacs into worktrees where you get this when commiting on emacs-25
branch which is a worktree off master.

/home/phillord/src/git/emacs-git/master/.git/worktrees/emacs-25

But in this case, I think vc-git-root will still work.

(defun vc-git-root (file)
  (or (vc-file-getprop file 'git-root)
      (vc-file-setprop file 'git-root (vc-find-root file ".git"))))

We still find ".git" because it is a *file* (not a directory) at the top
level of a worktree.

> As such, the submitted patch is only likely to work in the
> tautological case you've descried.

It's definately a risk. Git does many things, and people use it in many
ways.


> And it will add some performance penalty to each call, because
> vc-git-root, though usually fast, is not free.

Not free, but it is a constant time look up after the first. 


There is a simpler option. I am trying to solve the root cause of the
problem but, as you say, that might be fraught. Since the problem only
seems to cause an error with vc-git-mode-line-string, we could just
discard the error from vc-git-working-revision in this case.

Failing that, now I know what the problem is, at least I have a
workaround (unset GIT_DIR in the pre-commit hook).


Phil





  reply	other threads:[~2016-06-15 22:09 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-06-14 11:16 bug#23769: 25.0.95; Mode Line breakage in vc-git Phillip Lord
2016-06-14 16:29 ` Glenn Morris
2016-06-14 17:03   ` Phillip Lord
2016-06-14 21:52     ` Dmitry Gutov
2016-06-14 22:21       ` Phillip Lord
2016-06-15 20:48       ` Phillip Lord
2016-06-15 21:02         ` Dmitry Gutov
2016-06-15 22:09           ` Phillip Lord [this message]
2016-06-15 22:27             ` Noam Postavsky
2016-06-15 23:20               ` Dmitry Gutov
2016-06-16  7:11                 ` Phillip Lord
2016-06-16  7:45               ` Phillip Lord
2016-06-15 23:25             ` Dmitry Gutov
2016-06-16  7:41               ` Phillip Lord
2016-06-16 11:25                 ` Dmitry Gutov
2016-06-16 16:06                   ` Dmitry Gutov
2016-06-16 16:54                     ` Noam Postavsky
2016-06-16 16:59                       ` Dmitry Gutov
2016-06-16 21:47                       ` Phillip Lord
2016-06-16 22:13                         ` Dmitry Gutov
2016-06-16 22:23                           ` Phillip Lord
2016-06-16 22:31                             ` Dmitry Gutov
2016-06-16 22:42                               ` Phillip Lord
2016-06-17  2:41                                 ` Dmitry Gutov
2016-06-17  3:37                                   ` Noam Postavsky
2016-06-17  6:54                                     ` Eli Zaretskii
2016-06-17 12:06                                       ` Dmitry Gutov
2016-06-17 13:21                                         ` Eli Zaretskii
2016-06-17 13:54                                           ` Dmitry Gutov
2016-06-17 15:09                                         ` Phillip Lord
2016-06-16 21:47                     ` Phillip Lord
2016-06-16 22:04                       ` Dmitry Gutov
2016-06-21 16:46                         ` Dmitry Gutov
2016-06-21 20:45                           ` Phillip Lord
2016-06-21 23:06                             ` Dmitry Gutov
2016-06-15 21:56         ` John Wiegley
2016-06-14 21:30   ` Phillip Lord

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: https://www.gnu.org/software/emacs/

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

  git send-email \
    --in-reply-to=871t3y85oz.fsf@russet.org.uk \
    --to=phillip.lord@russet.org.uk \
    --cc=23769@debbugs.gnu.org \
    --cc=dgutov@yandex.ru \
    /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://git.savannah.gnu.org/cgit/emacs.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).