GNU bug report logs - #21383
Static revisions in vc-working-revision

Previous Next

Full log

Message #64 received at 21383-done <at> debbugs.gnu.org (full text, mbox):

From: Dmitry Gutov <dgutov <at> yandex.ru>
To: Stefan Monnier <monnier <at> IRO.UMontreal.CA>
Cc: 21383-done <at> debbugs.gnu.org, Jonathan H <pythonnut <at> gmail.com>
Subject: Re: bug#21383: Static revisions in vc-working-revision
Date: Thu, 3 Sep 2015 22:24:51 +0300

On 09/03/2015 07:04 PM, Stefan Monnier wrote:

> I think the problem is in the contract between VC itself and the
> backends, because VC mostly assumes that default-directory doesn't
> matter and uses absolute file names instead (that was the original
> design), whereas for many backends this is sometimes inconvenient so
> they occasionally rely on default-directory instead, which happens to
> work as well, tho it's mostly an accident.

Yes.

>> And are you talking about FILE arg to vc-status, or e.g. vc-git-status?
>
> vc-git-status (and other backend operations).

That might be viable. But the commands that don't use FILE currently 
only need the root, so we could avoid passing the argument in entirely.

> Right, we could fix the problem by keeping the original design and
> making sure the backends actually follow it, but I'm not sure it's the
> better design nowadays (and since using default-directory happens to
> work in 99% of the cases, it's hard to make sure we really fix all cases
> where we incorrectly rely on default-directory being the right parent of
> the absolute file names we get).

That is true. But could we abandon the current design for all backends? 
Some of the older ones still don't have vc-root implemented (because 
it's impossible for some of them?), and until it is, vc-state and 
friends won't know what to set default-directory to.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.