vc & cvs buglet

vc & cvs buglet

[Giorgos Keramidas]

Hi all,

It appears that some sort of internal initialization of vc.el is done a
bit late.  If I hit ``C-x v l'' while editing a version controlled file,
an empty log buffer comes up.  Re-running ``C-x v l'' a second time
brings up the correct log.

How should I go about tracing this?

Re: vc & cvs buglet

[Kevin Rodgers]

Giorgos Keramidas wrote:
> It appears that some sort of internal initialization of vc.el is done a
> bit late.  If I hit ``C-x v l'' while editing a version controlled file,
> an empty log buffer comes up.  Re-running ``C-x v l'' a second time
> brings up the correct log.
> 
> How should I go about tracing this?

Sometimes while using CVS there is a delay.  If you wait for a while,
does the correct log appear after the first `C-x v l'?

-- 
Kevin Rodgers

Re: vc & cvs buglet

[Romain Francoise]

Giorgos Keramidas <keramida@ceid.upatras.gr> writes:

> It appears that some sort of internal initialization of vc.el is done a
> bit late.  If I hit ``C-x v l'' while editing a version controlled file,
> an empty log buffer comes up.  Re-running ``C-x v l'' a second time
> brings up the correct log.

It works for me...  Could you provide more information?  Does the
*Messages* buffer mention loading vc.el?

Are you sure your CVS server isn't just long to respond?

-- 
Romain Francoise <romain@orebokech.com> | I've become someone else's
it's a miracle -- http://orebokech.com/ | nightmare...

Re: vc & cvs buglet

[Giorgos Keramidas]

On 2005-10-27 11:09, Kevin Rodgers <ihs_4664@yahoo.com> wrote:
> Giorgos Keramidas wrote:
> >It appears that some sort of internal initialization of vc.el is done a
> >bit late.  If I hit ``C-x v l'' while editing a version controlled file,
> >an empty log buffer comes up.  Re-running ``C-x v l'' a second time
> >brings up the correct log.
> >
> >How should I go about tracing this?
> 
> Sometimes while using CVS there is a delay.  If you wait for a while,
> does the correct log appear after the first `C-x v l'?

Nope.  VC seems to think that cvs is finished, but no output has been
copied in the *vc-change-log* buffer.  All I see is:

,-------------------------------------------------------------------------
| # Startup file for GNU bash on my FreeBSD laptop.
| #
| # Copyright (C) 2005, Giorgos Keramidas <keramida@FreeBSD.org>
| # All rights reserved.
| #
| # $GKER: dotfiles/freebsd/amd64/flame/dot.bashrc,v 1.19 2005/10/26 23:27
| 
| [...]
| -77-:---F1  dot.bashrc     Top (1,0)     CVS-1.19   (Fundamental)-------
| 
| 
| 
| -777:%%-F1  *vc-change-log*   All (1,0)       (Log-View from dot.bashrc)
|Running cvs in the background... done
`-------------------------------------------------------------------------

There are only 3 empty lines in *vc-change-log* (probably because that's
the starting size of the buffer's appearance, but the 'done' message
seems to imply that VC is not blocked waiting for input.  It may not be
a bug of VC, since I'm running FreeBSD 7.0-CURRENT #0: Mon Oct 24
23:56:24 EEST 2005, but I'm not sure how to trace the execution of
(vc-print-log) to find out what is going on.

- Giorgos

Re: vc & cvs buglet

[Giorgos Keramidas]

On 2005-10-27 19:29, Romain Francoise <romain@orebokech.com> wrote:
>Giorgos Keramidas <keramida@ceid.upatras.gr> writes:
>> It appears that some sort of internal initialization of vc.el is done a
>> bit late.  If I hit ``C-x v l'' while editing a version controlled file,
>> an empty log buffer comes up.  Re-running ``C-x v l'' a second time
>> brings up the correct log.
>
> It works for me...  Could you provide more information?  Does the
> *Messages* buffer mention loading vc.el?

Yes, it does.  The contents of the *Messages* buffer are:

    Loading help-fns...done
    (/home/keramida/opt/bin/emacs dot.bashrc)
    Loading disp-table...done
    Loading encoded-kb...done
    Loading cc-mode...done
    Loading hl-line...done
    For information about the GNU Project and its goals, type <f1> C-p.
    Loading vc-cvs...done
    Loading vc...done
    Running cvs in the background...
    Loading log-view...
    Loading easy-mmode...done
    Loading log-view...done
    Running cvs in the background... done

> Are you sure your CVS server isn't just long to respond?

The files are on a local CVS repository, under /home/cvs.  No
remote server enters the picture anywhere, so it's unlikely to
be a network-induced delay.

Re: vc & cvs buglet

[Romain Francoise]

That's weird.  Can you reproduce the bug when Emacs is started with -Q?
On what system is this?  And what's your version of CVS?

Does calling `cvs log <file>' from the command line work in all cases?

-- 
Romain Francoise <romain@orebokech.com> | All you've been is a piece of
it's a miracle -- http://orebokech.com/ | blood wrapped in skin.

Re: vc & cvs buglet

[Giorgos Keramidas]

On 2005-10-27 19:53, Romain Francoise <romain@orebokech.com> wrote:
> That's weird.  Can you reproduce the bug when Emacs is started with -Q?

yes.

> On what system is this?

A fairly recent version of FreeBSD-CURRENT:
% uname -v
FreeBSD 7.0-CURRENT #0: Mon Oct 24 23:56:24 EEST 2005

> And what's your version of CVS?

The base system version of FreeBSD, which is derived from:
Concurrent Versions System (CVS) 1.11.17-FreeBSD (client/server)

> Does calling `cvs log <file>' from the command line work in all cases?

Yep.  This is a file that has been checked out from a local repository,
and using cvs commands from the command line works fine.  It also works
fine with the 21.3.1 version of Emacs, installed from the ports.

I guess I'll go hunting for differences between the two Emacs versions
next.