All of lore.kernel.org
 help / color / mirror / Atom feed
From: Mathias Lafeldt <misfire@debugon.org>
To: Jeff King <peff@peff.net>
Cc: Will Palmer <wmpalmer@gmail.com>,
	Eric Wong <normalperson@yhbt.net>, git <git@vger.kernel.org>
Subject: Re: [BUG] git-svn parses --pretty=medium log output, fails when log.decorate is true
Date: Tue, 05 Oct 2010 19:53:36 +0200	[thread overview]
Message-ID: <4CAB6620.4000909@debugon.org> (raw)
In-Reply-To: <20101005165830.GA13413@sigill.intra.peff.net>

On 10/05/2010 06:58 PM, Jeff King wrote:
> On Tue, Oct 05, 2010 at 11:18:46AM +0100, Will Palmer wrote:
> 
>> I started receiving this error during "git svn dcommit" today:
>>
>> Use of uninitialized value $hash in string eq
>> at /home/wpalmer/libexec/git-core/git-svn line 1534.
>>
>> Examining that section reveals that git-svn is running 
>> "git log --no-color --first-parent --pretty=medium"
>>
>> and parsing the output in order to find commit hashes and git-svn-id:
>> lines. This breaks when log.decorate is true.
>>
>> This could be patched-up by adding "--no-decorate" to the options
>> git-svn passes, but that seems to me like it would just be adding to the
>> pile, as "--pretty=medium" is a moving target. I assume the correct
>> solution is to specify the format exactly as it is expected.
> 
> The problem isn't necessarily --pretty=medium, but that we are using
> "git log" instead of "git rev-list" (though because "git log" does have
> some features that rev-list does not have, we do turn off most
> configurable features for "git log --pretty=raw").
> 
> So I think the simplest thing is just the patch below (which has only
> been lightly tested by me):
> 
> -- >8 --
> Subject: [PATCH] git-svn: use rev-list instead of log
> 
> We are parsing the output, so we don't want user
> configuration like color or decorations to appear in the
> output. The simplest way to accomplish this is to use the
> rev-list plumbing instead of the log porcelain.
> 
> Signed-off-by: Jeff King <peff@peff.net>
> ---
>  git-svn.perl |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/git-svn.perl b/git-svn.perl
> index d292224..03d93d8 100755
> --- a/git-svn.perl
> +++ b/git-svn.perl
> @@ -1514,7 +1514,7 @@ sub cmt_sha2rev_batch {
>  
>  sub working_head_info {
>  	my ($head, $refs) = @_;
> -	my @args = ('log', '--no-color', '--first-parent', '--pretty=medium');
> +	my @args = ('rev-list', '--first-parent', '--pretty=medium');
>  	my ($fh, $ctx) = command_output_pipe(@args, $head);
>  	my $hash;
>  	my %max;

You're right. Using rev-list looks like a much cleaner solution.

In addition, I think we can replace the call to git-log in git_svn_log_cmd() too.

-Mathias

  reply	other threads:[~2010-10-05 17:53 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-10-05 10:18 [BUG] git-svn parses --pretty=medium log output, fails when log.decorate is true Will Palmer
2010-10-05 11:48 ` Mathias Lafeldt
2010-10-05 16:58 ` Jeff King
2010-10-05 17:53   ` Mathias Lafeldt [this message]
2010-10-05 18:02     ` Jeff King

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=4CAB6620.4000909@debugon.org \
    --to=misfire@debugon.org \
    --cc=git@vger.kernel.org \
    --cc=normalperson@yhbt.net \
    --cc=peff@peff.net \
    --cc=wmpalmer@gmail.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.