cvs-mode-diff on marked files

cvs-mode-diff on marked files

[Juri Linkov]

I wonder what are the reasons for cvs-mode-diff to ignore selected
files?  It is very convenient to mark all relevant files in the *cvs*
buffer and type `=' to generate a composite patch.  But currently
`cvs-invert-ignore-marks' contains "diff" by default, thus preventing
`=' to operate on all selected files.  It operates only on the current
file at point.  Even the documentation string of `cvs-mode-diff' says
"Diff the selected files against the repository" where "selected files"
is in the plural form.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: cvs-mode-diff on marked files

[Stefan Monnier]

> I wonder what are the reasons for cvs-mode-diff to ignore selected
> files?  It is very convenient to mark all relevant files in the *cvs*
> buffer and type `=' to generate a composite patch.  But currently
> `cvs-invert-ignore-marks' contains "diff" by default, thus preventing
> `=' to operate on all selected files.  It operates only on the current
> file at point.  Even the documentation string of `cvs-mode-diff' says
> "Diff the selected files against the repository" where "selected files"
> is in the plural form.

With the prefix T you can get what you want: T =
As for why, it's mostly historical (that's how it was originally in PCL-CVS
when I took over maintenance and it may have been originally due to
limitations of early CVS).  I actually like it because I often select
files to commit and then do diff one file at a time (either to fill the
ChangeLog or to decide whether to add the file to set of selected files for
commit).


        Stefan

Re: cvs-mode-diff on marked files

[Juri Linkov]

> With the prefix T you can get what you want: T =
> As for why, it's mostly historical (that's how it was originally in PCL-CVS
> when I took over maintenance and it may have been originally due to
> limitations of early CVS).  I actually like it because I often select
> files to commit and then do diff one file at a time (either to fill the
> ChangeLog or to decide whether to add the file to set of selected files for
> commit).

Thanks, now I see.  The same problem exists in Dired where `!' always
operates on marked files.  I use the special command that operates on
the file at point.

(defun my-dired-do-shell-command-on-current-file (&optional arg)
  "Run a shell command on the current file instead of marked files."
  (interactive)
  (let ((dired-marker-char ?M))         ; ?M is unused marker char
    (call-interactively 'dired-do-shell-command)))
(define-key dired-mode-map "%!" 'my-dired-do-shell-command-on-current-file)
(define-key dired-mode-map [(control meta ?!)] 'my-dired-do-shell-command-on-current-file)

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: cvs-mode-diff on marked files

[Andreas Schwab]

Juri Linkov <juri@jurta.org> writes:

> Thanks, now I see.  The same problem exists in Dired where `!' always
> operates on marked files.

Unless you pass a prefix argument.

> I use the special command that operates on the file at point.

M-1 !

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: cvs-mode-diff on marked files

[Juri Linkov]

>> Thanks, now I see.  The same problem exists in Dired where `!' always
>> operates on marked files.
>
> Unless you pass a prefix argument.

Thanks, I missed this too.  It would be good to think after the
release about making PCL-CVS more similar to Dired.  Ideally, PCV-CVS
should be a sub-mode of Dired mode with CVS status information
displayed instead of permission bits, group and owner names, and with
some its keys bound to CVS commands.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: cvs-mode-diff on marked files

[Stefan Monnier]

> Thanks, I missed this too.  It would be good to think after the
> release about making PCL-CVS more similar to Dired.  Ideally, PCV-CVS
> should be a sub-mode of Dired mode with CVS status information
> displayed instead of permission bits, group and owner names, and with
> some its keys bound to CVS commands.

I won't comment on this idea until I see patches, but I'm not
too enthusiastic to waste screen real estate on info such as group
and owner name that I use, like, once every other year, maybe?

Maybe you'd be better served with VC-dired ?


        Stefan

Re: cvs-mode-diff on marked files

[Juri Linkov]

> I won't comment on this idea until I see patches, but I'm not
> too enthusiastic to waste screen real estate on info such as group
> and owner name that I use, like, once every other year, maybe?
>
> Maybe you'd be better served with VC-dired ?

I tried out VC-dired, but it doesn't seem to be so convenient and
feature-rich as PCL-CVS.  OTOH, its dired-like appearance is quite good.

BTW, after trying VC-dired, it saved vc-dired-mode buffers to the
desktop file which failed during restoring.  Does this bug mean that
nobody uses VC-dired?  OK, more precisely, this means nobody uses
VC-dired with desktop.el :-)  I could install the following fix if
there are no objections.

Index: lisp/desktop.el
===================================================================
RCS file: /sources/emacs/emacs/lisp/desktop.el,v
retrieving revision 1.95
diff -c -r1.95 desktop.el
*** lisp/desktop.el	21 Oct 2005 08:48:53 -0000	1.95
--- lisp/desktop.el	12 Dec 2005 07:44:44 -0000
***************
*** 411,417 ****
  
  (defcustom desktop-minor-mode-table
    '((auto-fill-function auto-fill-mode)
!     (vc-mode nil))
    "Table mapping minor mode variables to minor mode functions.
  Each entry has the form (NAME RESTORE-FUNCTION).
  NAME is the name of the buffer-local variable indicating that the minor
--- 411,418 ----
  
  (defcustom desktop-minor-mode-table
    '((auto-fill-function auto-fill-mode)
!     (vc-mode nil)
!     (vc-dired-mode nil))
    "Table mapping minor mode variables to minor mode functions.
  Each entry has the form (NAME RESTORE-FUNCTION).
  NAME is the name of the buffer-local variable indicating that the minor

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: cvs-mode-diff on marked files

[Lars Hansen]

Juri Linkov wrote:

>BTW, after trying VC-dired, it saved vc-dired-mode buffers to the
>desktop file which failed during restoring.  Does this bug mean that
>nobody uses VC-dired?  OK, more precisely, this means nobody uses
>VC-dired with desktop.el :-)  I could install the following fix if
>there are no objections.
>  
>
Please do.