git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jonas Fonseca <fonseca@diku.dk>
To: "Shawn O. Pearce" <spearce@spearce.org>
Cc: git@vger.kernel.org
Subject: Re: [PATCH] When renaming config sections delete conflicting sections
Date: Wed, 17 Oct 2007 12:37:26 +0200	[thread overview]
Message-ID: <20071017103726.GA23417@diku.dk> (raw)
In-Reply-To: <20071017005517.GJ13801@spearce.org>

Shawn O. Pearce <spearce@spearce.org> wrote Tue, Oct 16, 2007:
> Jonas Fonseca <fonseca@diku.dk> wrote:
> > The old behavior of keeping config sections matching the new name caused
> > problems leading to warnings being emitted by git-remote when renaming
> > branches where information about tracked remote branches differed. To
> > fix this any config sections that will conflict with the new name are
> > removed from the config file. Update test to check for this.
> ...
> >  This command sequence was causing problems for me:
> > 
> > 	git checkout -b test madcoder/next
> > 	git checkout -b test2 spearce/next
> > 	git branch -M test
> 
> Ouch.  But this may cause the user to lose what they might consider
> important settings relative to the old section named branch.test.

True, but to me the meaning of -M is "I know what I am doing".

> I think in the case you mention above where you are doing a
> `branch -M` the user really does want the basic branch properties
> to be forced over (branch.$name.remote, branch.$name.merge) but
> they probably do not want other branch properties to be removed
> or deleted.  Or maybe they do.

You never know, and sure if there is an option to gracefully avoid
lossing this information that is the right approach, but I don't see how
this can be done in this situation. Besides currently only
branch.$name.mergoptions will be lost, hardly a problem.

> Its really hard to second guess the user's intent here.  I think
> its too broad to whack an entire section when renaming. [...]
>
> So we don't blindly replace multi-valued keys just because the
> user asked us to.  I don't really see a section as being that much
> different to warrant a potentially lossy behavior by default.

Because it makes sense in this situation and erroring out is a good
choice, but we are running out of git-branch options based on the letter
'm'. And to me, -m is the default for renaming branches, and -M is a
shortcut for doing a lot of other stuff with well-defined implications.

Perhaps we can enable only this "lossy behavior" only for git-branch by
adding an extra argument to git_config_rename_section? Then we can later
add a new option to git-config along the lines of --overwrite-section.

-- 
Jonas Fonseca

      reply	other threads:[~2007-10-17 10:40 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-10-17  0:34 [PATCH] When renaming config sections delete conflicting sections Jonas Fonseca
2007-10-17  0:55 ` Shawn O. Pearce
2007-10-17 10:37   ` Jonas Fonseca [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

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

  git send-email \
    --in-reply-to=20071017103726.GA23417@diku.dk \
    --to=fonseca@diku.dk \
    --cc=git@vger.kernel.org \
    --cc=spearce@spearce.org \
    /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).