bug#308: vc-delete-file

bug#308: vc-delete-file

[Sam Steingold]

GNU Emacs 23.0.60.7 (x86_64-unknown-linux-gnu, GTK+ Version 2.8.20)
 of 2008-05-22 on nyc-qws-005
 '--with-x-toolkit=gtk' '--prefix=/mnt/office/dev/opt'

first, vc-delete-file is available neither in any menu nor on any key.
despite your claim to the contrary, this is a common command, used at
least weakly in an active development environment.

second, when I invoke it with M-x, it requires that the file being
removed is still present on disk.
this requirement is absurd because when I decide to remove the file from
the VCS, it is already long gone because I have to remove it to test the
system without it (yes, it is NOT enough not to mention the file in
makefiles, because omake will use it if it finds it on disk, so you
want me to rename it for testing and then move it back so that vc can
remove it!)

third, when I do jump through the hoops, I get this error:

Debugger entered--Lisp error: (error "File myfilename no longer exists!")
  signal(error ("File myfilename no longer exists!"))
  error("File %s no longer exists!" "myfilename")
  revert-buffer(t t t)
  vc-revert-buffer-internal(t t)
  vc-resynch-window("myfilename" #<buffer convert_config.mli> t)
  vc-resynch-buffer("myfilename" #<buffer convert_config.mli> t)
  vc-delete-file("myfilename")
  call-interactively(vc-delete-file t nil)
  execute-extended-command(nil)
  call-interactively(execute-extended-command nil nil)

i.e., you just deleted it yourself - and now you are complaining that
the file does not exist!
I would rather expect kill-buffer instead of revert-buffer.


-- 
Sam Steingold (http://sds.podval.org/) on Fedora Core release 5 (Bordeaux)
http://honestreporting.com http://openvotingconsortium.org http://camera.org
http://iris.org.il http://pmw.org.il http://dhimmi.com http://truepeace.org
Don't use force -- get a bigger hammer.

vc-delete-file

[Sam Steingold]

GNU Emacs 23.0.60.7 (x86_64-unknown-linux-gnu, GTK+ Version 2.8.20)
 of 2008-05-22 on nyc-qws-005
 '--with-x-toolkit=gtk' '--prefix=/mnt/office/dev/opt'

first, vc-delete-file is available neither in any menu nor on any key.
despite your claim to the contrary, this is a common command, used at
least weakly in an active development environment.

second, when I invoke it with M-x, it requires that the file being
removed is still present on disk.
this requirement is absurd because when I decide to remove the file from
the VCS, it is already long gone because I have to remove it to test the
system without it (yes, it is NOT enough not to mention the file in
makefiles, because omake will use it if it finds it on disk, so you
want me to rename it for testing and then move it back so that vc can
remove it!)

third, when I do jump through the hoops, I get this error:

Debugger entered--Lisp error: (error "File myfilename no longer exists!")
  signal(error ("File myfilename no longer exists!"))
  error("File %s no longer exists!" "myfilename")
  revert-buffer(t t t)
  vc-revert-buffer-internal(t t)
  vc-resynch-window("myfilename" #<buffer convert_config.mli> t)
  vc-resynch-buffer("myfilename" #<buffer convert_config.mli> t)
  vc-delete-file("myfilename")
  call-interactively(vc-delete-file t nil)
  execute-extended-command(nil)
  call-interactively(execute-extended-command nil nil)

i.e., you just deleted it yourself - and now you are complaining that
the file does not exist!
I would rather expect kill-buffer instead of revert-buffer.


-- 
Sam Steingold (http://sds.podval.org/) on Fedora Core release 5 (Bordeaux)
http://honestreporting.com http://openvotingconsortium.org http://camera.org
http://iris.org.il http://pmw.org.il http://dhimmi.com http://truepeace.org
Don't use force -- get a bigger hammer.

bug#308: vc-delete-file

[Stefan Monnier]

> first, vc-delete-file is available neither in any menu nor on any key.
> despite your claim to the contrary, this is a common command, used at
> least weakly in an active development environment.

I'm not sure who is "you", but for what it's worth, I agree it should be
available via some key-binding.  Even more so in vc-dir.

> second, when I invoke it with M-x, it requires that the file being
> removed is still present on disk.

That's a bug, indeed.  But I don't see any place in the code where
that's done on purpose.  Could you provide a recipe to reproduce the
problem?  It might depend on the backend.

> this requirement is absurd because when I decide to remove the file from
> the VCS, it is already long gone because I have to remove it to test the
> system without it

No need to argue that it's a bug, I agree.
[ For what it's worth, I'd recommend you vc-delete-file before doing
  your tests, tho.  After all, you can always undo it before committing
  if it turns out it was a bad idea.  ]

> Debugger entered--Lisp error: (error "File myfilename no longer exists!")
>   signal(error ("File myfilename no longer exists!"))
>   error("File %s no longer exists!" "myfilename")
>   revert-buffer(t t t)
>   vc-revert-buffer-internal(t t)
>   vc-resynch-window("myfilename" #<buffer convert_config.mli> t)
>   vc-resynch-buffer("myfilename" #<buffer convert_config.mli> t)
>   vc-delete-file("myfilename")
>   call-interactively(vc-delete-file t nil)
>   execute-extended-command(nil)
>   call-interactively(execute-extended-command nil nil)

> i.e., you just deleted it yourself - and now you are complaining that
> the file does not exist!
> I would rather expect kill-buffer instead of revert-buffer.

I've just changed the code to kill the buffer, indeed.
Thank you,


        Stefan

bug#308: vc-delete-file

[Dan Nicolaescu]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

  > No need to argue that it's a bug, I agree.
  > [ For what it's worth, I'd recommend you vc-delete-file before doing
  >   your tests, tho.  After all, you can always undo it before committing
  >   if it turns out it was a bad idea.  ]
  > 
  > > Debugger entered--Lisp error: (error "File myfilename no longer exists!")
  > >   signal(error ("File myfilename no longer exists!"))
  > >   error("File %s no longer exists!" "myfilename")
  > >   revert-buffer(t t t)
  > >   vc-revert-buffer-internal(t t)
  > >   vc-resynch-window("myfilename" #<buffer convert_config.mli> t)
  > >   vc-resynch-buffer("myfilename" #<buffer convert_config.mli> t)
  > >   vc-delete-file("myfilename")
  > >   call-interactively(vc-delete-file t nil)
  > >   execute-extended-command(nil)
  > >   call-interactively(execute-extended-command nil nil)
  > 
  > > i.e., you just deleted it yourself - and now you are complaining that
  > > the file does not exist!
  > > I would rather expect kill-buffer instead of revert-buffer.
  > 
  > I've just changed the code to kill the buffer, indeed.

That's not quite right, the *vc-dir* buffers where this file might have
been displayed need to be updated.   I checked in an alternative fix
that I've had in my tree and seems to DTRT.

bug#308: vc-delete-file

[Stefan Monnier]

> That's not quite right, the *vc-dir* buffers where this file might have
> been displayed need to be updated.   I checked in an alternative fix
> that I've had in my tree and seems to DTRT.

Thanks,


        Stefan