VC acting strangely

VC acting strangely

[Jason Rumney]

Something strange has happened to vc.el recently.

The Version Control submenu is now empty when the current buffer is
under CVS. Strangely, it appears for other files.

Perhaps related is the fact that M-x vc-revert-buffer now brings up an
error message "Autoloading failed to define function vc-revert-buffer".
Indeed, the only reference to it is in ldefs-boot.el, but there is no
entry in NEWS or even the ChangeLog about its removal!

Re: VC acting strangely

[Dan Nicolaescu]

Jason Rumney <jasonr@gnu.org> writes:

  > Something strange has happened to vc.el recently.
  > 
  > The Version Control submenu is now empty when the current buffer is
  > under CVS. Strangely, it appears for other files.
  > 
  > Perhaps related is the fact that M-x vc-revert-buffer now brings up an
  > error message "Autoloading failed to define function vc-revert-buffer".
  > Indeed, the only reference to it is in ldefs-boot.el, but there is no
  > entry in NEWS or even the ChangeLog about its removal!

`vc-revert-buffer' seem to have been renamed to `vc-revert' in vc.el
version 1.434
ldefs-boot.el needs to be updated...

Re: VC acting strangely

[Jason Rumney]

Dan Nicolaescu wrote:
> `vc-revert-buffer' seem to have been renamed to `vc-revert' in vc.el
> version 1.434
> ldefs-boot.el needs to be updated...
>   

As a user command, I'd say it needs renaming back, or at least an alias
defined for it.

Re: VC acting strangely

[David Kastrup]

Jason Rumney <jasonr@gnu.org> writes:

> Dan Nicolaescu wrote:
>> `vc-revert-buffer' seem to have been renamed to `vc-revert' in vc.el
>> version 1.434
>> ldefs-boot.el needs to be updated...
>>   
>
> As a user command, I'd say it needs renaming back, or at least an alias
> defined for it.

Agreed.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: VC acting strangely

[David Kastrup]

David Kastrup <dak@gnu.org> writes:

> Jason Rumney <jasonr@gnu.org> writes:
>
>> Dan Nicolaescu wrote:
>>> `vc-revert-buffer' seem to have been renamed to `vc-revert' in vc.el
>>> version 1.434
>>> ldefs-boot.el needs to be updated...
>>>   
>>
>> As a user command, I'd say it needs renaming back, or at least an alias
>> defined for it.
>
> Agreed.

Actually, I think it needs to get renamed back.  There are vc-*
subsystems maintained outside of Emacs, and those will stop working
when their hook functions get called under a different name.

As an example, there is vc-git.el.  It is also distributed in git's
contrib directory for the sake of users of older Emacsen and XEmacs,
and keeping it compatible would become more of a nuisance.  Basically,
it would necessitate putting an alias into vc-git.el as well as
upstream vc.

Unless there is a really pressing reason to keep the renamed function,
I think we should really revert here.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: VC acting strangely

[Dan Nicolaescu]

David Kastrup <dak@gnu.org> writes:

  > David Kastrup <dak@gnu.org> writes:
  > 
  > > Jason Rumney <jasonr@gnu.org> writes:
  > >
  > >> Dan Nicolaescu wrote:
  > >>> `vc-revert-buffer' seem to have been renamed to `vc-revert' in vc.el
  > >>> version 1.434
  > >>> ldefs-boot.el needs to be updated...
  > >>>   
  > >>
  > >> As a user command, I'd say it needs renaming back, or at least an alias
  > >> defined for it.
  > >
  > > Agreed.
  > 
  > Actually, I think it needs to get renamed back.  There are vc-*
  > subsystems maintained outside of Emacs, and those will stop working
  > when their hook functions get called under a different name.

The vc backends maintained outside emacs already have problems with vc
in CVS HEAD: a few vc functions have been changed to take a list of
files as the first parameter instead of a single file.

For the external vc backends: the ones for free systems have time to
get in emacs until emacs-23 is released, the other ones, they'll have
to be adjusted anyway...

IMO the change is good, the other vc interactive commands don't have a
-buffer suffix, so this makes vc more consistent. 

So, IMO adding an alias is the best solution here.

  > As an example, there is vc-git.el.  It is also distributed in git's
  > contrib directory for the sake of users of older Emacsen and XEmacs,
  > and keeping it compatible would become more of a nuisance.  Basically,
  > it would necessitate putting an alias into vc-git.el as well as
  > upstream vc.

Alexandre Julliard, the vc-git author, said this (before vc-git was
added to the 22.2 branch): "If vc-git gets merged into the Emacs 22
branch then I'll remove it from git.git, otherwise I'll probably leave
it around for people who don't want to use Emacs from CVS."  
So I wouldn't worry about vc-git.



  > Unless there is a really pressing reason to keep the renamed function,
  > I think we should really revert here.
  > 
  > -- 
  > David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: VC acting strangely

[Richard Stallman]

    As a user command, I'd say it needs renaming back, or at least an alias
    defined for it.

I think the arguments are strong for bringing back the old name.
Can you please add vc-revert-buffer as an alias for vc-revert, then ack?

A separate question is whether the callers should be changed to call
vc-revert-buffer instead of vc-revert.  I have no opinion on that

Re: VC acting strangely

[David Kastrup]

Richard Stallman <rms@gnu.org> writes:

>     As a user command, I'd say it needs renaming back, or at least an alias
>     defined for it.
>
> I think the arguments are strong for bringing back the old name.
> Can you please add vc-revert-buffer as an alias for vc-revert, then ack?
>
> A separate question is whether the callers should be changed to call
> vc-revert-buffer instead of vc-revert.  I have no opinion on that

What is more relevant here is that vc-revert-buffer used to call
vc-cvs-revert-buffer and their ilk.  So quite a number of other
functions travel on that naming convention.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: VC acting strangely

[Jason Rumney]

Richard Stallman wrote:
>     As a user command, I'd say it needs renaming back, or at least an alias
>     defined for it.
>
> I think the arguments are strong for bringing back the old name.
> Can you please add vc-revert-buffer as an alias for vc-revert, then ack?
>
> A separate question is whether the callers should be changed to call
> vc-revert-buffer instead of vc-revert.  I have no opinion on that
>   

I added it as an obsolete alias. Another part of the same change removed
the norevert argument from vc-rollback, without actually removing all
uses within the function. I'm not sure whether the intention was to
remove it just from the backend interface since it is not needed there
or to remove it completely, but I added it back for now.

Re: VC acting strangely

[Jason Rumney]

Jason Rumney wrote:
> The Version Control submenu is now empty when the current buffer is
> under CVS. Strangely, it appears for other files.
>
> Perhaps related is the fact that M-x vc-revert-buffer now brings up an
> error message

I've now fixed both problems, which were unrelated.