Option to customize the -m message flag of cvs?

Option to customize the -m message flag of cvs?

[Juanma Barranquero]

Hi.

For a long time I've been using CVSNT (http://march-hare.com/cvsnt/),
which is a free CVS derivative with better Windows compatibility,
though CVSNT is not Windows-specific at all (it works on GNU/Linux,
Windows, OS X, etc.).

The most recent release (2.5.04 (Zen) Build 3236) has been changed so
that commit messages added with -m get an additional text:

"
Committed on the Free edition of March Hare Software CVSNT Server.
Upgrade to CVS Suite for more features and support:
http://march-hare.com/cvsnt/"

There's been already a couple of such commits done against the Emacs
repository (by accident, of course).

The ad can be circumvented by using the new option -M instead of -m.
Unfortunately, "-m" is hardcoded in at least three places of
vc-cvs.el.

Would it be possible to add some kind of option to customize the
message flag passed to CVS?

    Juanma

Re: Option to customize the -m message flag of cvs?

[martin rudalics]

 > The ad can be circumvented by using the new option -M instead of -m.
 > Unfortunately, "-m" is hardcoded in at least three places of
 > vc-cvs.el.
 >
 > Would it be possible to add some kind of option to customize the
 > message flag passed to CVS?

I recently switched to CVSNT because I was not able to run `vc-dir' for
Emacs directories with my cvs 1.11.19 executable.  However, `vc-dir'
seems a prerequisite for implementing Miles' request on emacs-devel from
2008-11-20:

    I'd suggest tracking the git mirror instead; it does a pretty good job
    of packaging up related file changes into single commits, as long as the
    user committed them all at once with CVS.  Many people _don't_ do that,
    of course, but that's largely for historical reasons, and hopefully they
    can be convinced to start doing so.

So on Windows there's not only historical reasons but also the practical
one of having a CVS executable able to handle updates and commits as
requested.

To avoid the advert, I had to replace all occurrences of "-m" with "-M"
in vc-cvs.el.  An option for customizing this would be therefore greatly
appreciated.

Thanks in advance, martin.

Re: Option to customize the -m message flag of cvs?

[Eli Zaretskii]

> Date: Mon, 29 Dec 2008 18:10:08 +0100
> From: martin rudalics <rudalics@gmx.at>
> 
> To avoid the advert, I had to replace all occurrences of "-m" with "-M"
> in vc-cvs.el.

What is the "new -M option", and where it is documented?  I cannot
find it in the CVS manuals to which I have access.

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

On Mon, Dec 29, 2008 at 20:14, Eli Zaretskii <eliz@gnu.org> wrote:

> What is the "new -M option", and where it is documented?  I cannot
> find it in the CVS manuals to which I have access.

It's a CVSNT-specific option.

I don't have CVSNT 2.0.54 in this computer so I cannot show you the
precise help, but -M is just -m, with the only difference that it does
not add anything to the message.

It has been discussed in this thread of the CVSNT list:

http://www.mail-archive.com/cvsnt@cvsnt.org/msg03258.html

    Juanma

Re: Option to customize the -m message flag of cvs?

[Eli Zaretskii]

> Date: Mon, 29 Dec 2008 20:57:46 +0100
> From: "Juanma Barranquero" <lekktu@gmail.com>
> Cc: "martin rudalics" <rudalics@gmx.at>, emacs-devel@gnu.org
> 
> On Mon, Dec 29, 2008 at 20:14, Eli Zaretskii <eliz@gnu.org> wrote:
> 
> > What is the "new -M option", and where it is documented?  I cannot
> > find it in the CVS manuals to which I have access.
> 
> It's a CVSNT-specific option.

Yes, and for a CVSNT-specific problem.

What I don't understand is why would someone use CVSNT instead of some
other Windows port that doesn't have such a problem in -m, and thus
does not need any changes in VC.

Re: Option to customize the -m message flag of cvs?

[martin rudalics]

 > What I don't understand is why would someone use CVSNT instead of some
 > other Windows port that doesn't have such a problem in -m, and thus
 > does not need any changes in VC.

I can't speak for others.  Here,


Concurrent Versions System (CVS) 1.11.19 (client)

Copyright (C) 2005 Free Software Foundation, Inc.

Senior active maintainers include Larry Jones, Derek R. Price,
and Mark D. Baushke.  Please see the AUTHORS and README files from the CVS
distribution kit for a complete list of contributors and copyrights.

CVS may be copied only under the terms of the GNU General Public License,
a copy of which can be found with the CVS distribution kit.


is not able to handle `vc-dir'.

martin

Re: Option to customize the -m message flag of cvs?

[Eli Zaretskii]

> Date: Mon, 29 Dec 2008 22:18:12 +0100
> From: martin rudalics <rudalics@gmx.at>
> CC: Juanma Barranquero <lekktu@gmail.com>, emacs-devel@gnu.org
> 
> is not able to handle `vc-dir'.

Why not? what feature(s) is/are missing?

Re: Option to customize the -m message flag of cvs?

[martin rudalics]

 >> is not able to handle `vc-dir'.
 >
 > Why not? what feature(s) is/are missing?

That's what I asked _you_ in a mail on 2008-12-08.  Unfortunately, I
never got a response.

The problem is that when I do `vc-dir' in my Emacs directory I get - on
Windows XP - a *vc-dir* buffer looking like

--------------------------------------------------------
VC backend : CVS
Working dir: c:/Programme/Emacs-trunk/
Repository : m061211@cvs.savannah.gnu.org:/sources/emacs
Module     : emacs


                          ./
--------------------------------------------------------

and nothing that would populate that buffer with modified, unknown or
up-to-date entries.  cvs.exe starts doing something but dies quickly.
Doing `g' in the *vc-dir* buffer doesn't do anything.  I traced the
problem back to `vc-do-command' where

   (let* ((files
       (mapcar (lambda (f) (file-relative-name (expand-file-name f)))
           (if (listp file-or-list) file-or-list (list file-or-list))))

sets `files' to ("./").  Apparently, my cvs.exe doesn't like that.  I
found a comment in `cvs-mode-run' saying

   ;; Some w32 versions of CVS don't like an explicit . too much.

so I conclude that's the problem.  Indeed, when in `vc-do-command' I
assign `squeezed' as follows

     (when (and files (not (equal files (list "./"))))
       (setq squeezed (nconc squeezed files)))

`vc-dir' works as expected.

Since changing `vc-do-command' as sketched above doesn't strike me as
very digestible for most people here, I finally decided to get myself a
more subservient cvs client.  The one I found is CVSNT which has just
one drawback.  With "-m" it inserts an advert in my commits ...

martin

Re: Option to customize the -m message flag of cvs?

[Eli Zaretskii]

> Date: Tue, 30 Dec 2008 08:26:26 +0100
> From: martin rudalics <rudalics@gmx.at>
> CC: emacs-devel@gnu.org
> 
>  >> is not able to handle `vc-dir'.
>  >
>  > Why not? what feature(s) is/are missing?
> 
> That's what I asked _you_ in a mail on 2008-12-08.  Unfortunately, I
> never got a response.

It should be clear now that I never responded because I didn't know
the answer.  And I still don't.

>    (let* ((files
>        (mapcar (lambda (f) (file-relative-name (expand-file-name f)))
>            (if (listp file-or-list) file-or-list (list file-or-list))))
> 
> sets `files' to ("./").  Apparently, my cvs.exe doesn't like that.

What is the command line passed to cvs.exe in this case?

Re: Option to customize the -m message flag of cvs?

[martin rudalics]

 > It should be clear now that I never responded because I didn't know
 > the answer.  And I still don't.

I'd probably have felt less silly if you had told me that immediately.

 >
 >>    (let* ((files
 >>        (mapcar (lambda (f) (file-relative-name (expand-file-name f)))
 >>            (if (listp file-or-list) file-or-list (list file-or-list))))
 >>
 >> sets `files' to ("./").  Apparently, my cvs.exe doesn't like that.
 >
 > What is the command line passed to cvs.exe in this case?

Emacs passes

cvs -f status "./"

and in a DOS window I get, for example,

------------------------------------------------
Microsoft Windows XP [Version 5.1.2600]
(C) Copyright 1985-2001 Microsoft Corp.

c:\Programme\Emacs-trunk\src>cvs -f status "./"
cvs status: Examining .

c:\Programme\Emacs-trunk\src>
------------------------------------------------

with similar results for many other dot/slash/backslash combinations I
tried.  cvs -f status works as expected.  Does `vc-dir' work for you?

martin

Re: Option to customize the -m message flag of cvs?

[Eli Zaretskii]

> Date: Wed, 31 Dec 2008 10:15:11 +0100
> From: martin rudalics <rudalics@gmx.at>
> CC: emacs-devel@gnu.org
> 
>  > It should be clear now that I never responded because I didn't know
>  > the answer.  And I still don't.
> 
> I'd probably have felt less silly if you had told me that immediately.

Sorry.

>  >>    (let* ((files
>  >>        (mapcar (lambda (f) (file-relative-name (expand-file-name f)))
>  >>            (if (listp file-or-list) file-or-list (list file-or-list))))
>  >>
>  >> sets `files' to ("./").  Apparently, my cvs.exe doesn't like that.
>  >
>  > What is the command line passed to cvs.exe in this case?
> 
> Emacs passes
> 
> cvs -f status "./"
> 
> and in a DOS window I get, for example,
> 
> ------------------------------------------------
> Microsoft Windows XP [Version 5.1.2600]
> (C) Copyright 1985-2001 Microsoft Corp.
> 
> c:\Programme\Emacs-trunk\src>cvs -f status "./"
> cvs status: Examining .
> 
> c:\Programme\Emacs-trunk\src>
> ------------------------------------------------
> 
> with similar results for many other dot/slash/backslash combinations I
> tried.  cvs -f status works as expected.

So the conclusion is that cvs -f status works correctly from the shell
prompt, but not from inside Emacs, is that it?  Could it be that
cmdproxy does something bad with the "./" argument?

> Does `vc-dir' work for you?

I don't use it.

Re: Option to customize the -m message flag of cvs?

[martin rudalics]

 > So the conclusion is that cvs -f status works correctly from the shell
 > prompt, but not from inside Emacs, is that it?  Could it be that
 > cmdproxy does something bad with the "./" argument?

No.  cvs -f status works correctly from the shell prompt _and_ from
inside Emacs.  cvs -f status "./" doesn't work neither from the shell
prompt nor from inside Emacs.  With CVSNT everything works correctly.

martin

Re: Option to customize the -m message flag of cvs?

[martin rudalics]

[-- Attachment #1: Type: text/plain, Size: 202 bytes --]

 >> Does `vc-dir' work for you?
 >
 > I don't use it.

Apparently only Juanma and me are presently using `vc-dir' on Windows.
So if there are no objections I'd like to apply the attached patch.

martin

[-- Attachment #2: vc-cvs.el.diff --]
[-- Type: text/plain, Size: 2275 bytes --]

*** vc-cvs.el.~1.163.~	2009-01-05 10:00:27.031250000 +0100
--- vc-cvs.el	2009-01-05 10:19:29.265625000 +0100
***************
*** 180,185 ****
--- 180,194 ----
    :version "22.1"
    :group 'vc)
  
+ (defcustom vc-cvs-message-switch "-m"
+   "String passed as message switch to CVS.
+ By default this is \"-m\".  With more recent versions of CVSNT
+ you might want to set this to \"-M\" to avoid inserting adverts
+ in commit logs."
+   :type 'string
+   :version "23.1"
+   :group 'vc)
+ 
  ;;;
  ;;; Internal variables
  ;;;
***************
*** 288,294 ****
    (apply 'vc-cvs-command nil 0 files
           "add"
           (and comment (string-match "[^\t\n ]" comment)
!               (concat "-m" comment))
           (vc-switches 'CVS 'register)))
  
  (defun vc-cvs-responsible-p (file)
--- 297,303 ----
    (apply 'vc-cvs-command nil 0 files
           "add"
           (and comment (string-match "[^\t\n ]" comment)
!               (concat vc-cvs-message-switch comment))
           (vc-switches 'CVS 'register)))
  
  (defun vc-cvs-responsible-p (file)
***************
*** 325,331 ****
  	    files)))
    (let ((status (apply 'vc-cvs-command nil 1 files
  		       "ci" (if rev (concat "-r" rev))
! 		       (concat "-m" comment)
  		       (vc-switches 'CVS 'checkin))))
      (set-buffer "*vc*")
      (goto-char (point-min))
--- 334,340 ----
  	    files)))
    (let ((status (apply 'vc-cvs-command nil 1 files
  		       "ci" (if rev (concat "-r" rev))
! 		       (concat vc-cvs-message-switch comment)
  		       (vc-switches 'CVS 'checkin))))
      (set-buffer "*vc*")
      (goto-char (point-min))
***************
*** 485,491 ****
  (defun vc-cvs-modify-change-comment (files rev comment)
    "Modify the change comments for FILES on a specified REV.
  Will fail unless you have administrative privileges on the repo."
!   (vc-cvs-command nil 0 files "admin" (concat "-m" rev ":" comment)))
  
  ;;;
  ;;; History functions
--- 494,501 ----
  (defun vc-cvs-modify-change-comment (files rev comment)
    "Modify the change comments for FILES on a specified REV.
  Will fail unless you have administrative privileges on the repo."
!   (vc-cvs-command
!    nil 0 files "admin" (concat vc-cvs-message-switch rev ":" comment)))
  
  ;;;
  ;;; History functions

Re: Option to customize the -m message flag of cvs?

[Dan Nicolaescu]

martin rudalics <rudalics@gmx.at> writes:

  > >> Does `vc-dir' work for you?
  > >
  > > I don't use it.
  > 
  > Apparently only Juanma and me are presently using `vc-dir' on Windows.
  > So if there are no objections I'd like to apply the attached patch.

I'd like to strongly object to this type of change.

The same goal can be accomplished by using a wrapper script for the cvs
command that changes -m to -M, or by changing the CVSNT sources not to
do this.

This patch is proposed because CVSNT chose to annoy they users into
paying for a license by inserting a commercial in the commit message.
And asking users to use the non-standard -M flag in order to avoid
that annoyance.

There is no guarantee that the same thing will not be done in the future
to other commands: like log, diff, annotate, etc.  Which would mean that
we'd have to continuously change both vc-cvs.el and PCL-CVS (which is
not done here, but it's as needed) to keep up.

The user needs to explicitly do something anyway, so better just have
the user either avoid CVSNT, change the CVSNT source (and recompile)or
use a wrapper script instead of littering the emacs code with this type
of needless changes.

  > martin
  > *** vc-cvs.el.~1.163.~	2009-01-05 10:00:27.031250000 +0100
  > --- vc-cvs.el	2009-01-05 10:19:29.265625000 +0100
  > ***************
  > *** 180,185 ****
  > --- 180,194 ----
  >     :version "22.1"
  >     :group 'vc)
  >   
  > + (defcustom vc-cvs-message-switch "-m"
  > +   "String passed as message switch to CVS.
  > + By default this is \"-m\".  With more recent versions of CVSNT
  > + you might want to set this to \"-M\" to avoid inserting adverts
  > + in commit logs."
  > +   :type 'string
  > +   :version "23.1"
  > +   :group 'vc)
  > + 
  >   ;;;
  >   ;;; Internal variables
  >   ;;;
  > ***************
  > *** 288,294 ****
  >     (apply 'vc-cvs-command nil 0 files
  >            "add"
  >            (and comment (string-match "[^\t\n ]" comment)
  > !               (concat "-m" comment))
  >            (vc-switches 'CVS 'register)))
  >   
  >   (defun vc-cvs-responsible-p (file)
  > --- 297,303 ----
  >     (apply 'vc-cvs-command nil 0 files
  >            "add"
  >            (and comment (string-match "[^\t\n ]" comment)
  > !               (concat vc-cvs-message-switch comment))
  >            (vc-switches 'CVS 'register)))
  >   
  >   (defun vc-cvs-responsible-p (file)
  > ***************
  > *** 325,331 ****
  >   	    files)))
  >     (let ((status (apply 'vc-cvs-command nil 1 files
  >   		       "ci" (if rev (concat "-r" rev))
  > ! 		       (concat "-m" comment)
  >   		       (vc-switches 'CVS 'checkin))))
  >       (set-buffer "*vc*")
  >       (goto-char (point-min))
  > --- 334,340 ----
  >   	    files)))
  >     (let ((status (apply 'vc-cvs-command nil 1 files
  >   		       "ci" (if rev (concat "-r" rev))
  > ! 		       (concat vc-cvs-message-switch comment)
  >   		       (vc-switches 'CVS 'checkin))))
  >       (set-buffer "*vc*")
  >       (goto-char (point-min))
  > ***************
  > *** 485,491 ****
  >   (defun vc-cvs-modify-change-comment (files rev comment)
  >     "Modify the change comments for FILES on a specified REV.
  >   Will fail unless you have administrative privileges on the repo."
  > !   (vc-cvs-command nil 0 files "admin" (concat "-m" rev ":" comment)))
  >   
  >   ;;;
  >   ;;; History functions
  > --- 494,501 ----
  >   (defun vc-cvs-modify-change-comment (files rev comment)
  >     "Modify the change comments for FILES on a specified REV.
  >   Will fail unless you have administrative privileges on the repo."
  > !   (vc-cvs-command
  > !    nil 0 files "admin" (concat vc-cvs-message-switch rev ":" comment)))
  >   
  >   ;;;
  >   ;;; History functions

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

On Tue, Jan 6, 2009 at 17:08, Dan Nicolaescu <dann@ics.uci.edu> wrote:

> The user needs to explicitly do something anyway, so better just have
> the user either avoid CVSNT, change the CVSNT source (and recompile)or
> use a wrapper script instead of littering the emacs code with this type
> of needless changes.

Emacs is not our toy, but a tool for many people. Of course we can say
users of CVSNT "fix the problem yourselves", but there's quite a
difference between setting a variable, and writing a wrapping script
(which, to be robust, is not just blindingly changing -m to -M), not
to say downloading and setting up all the dependencies and tools
needed to rebuild CVSNT. All for the single bit that separates "m"
from "M".

    Juanma

Re: Option to customize the -m message flag of cvs?

[Eli Zaretskii]

> Date: Tue, 6 Jan 2009 18:06:03 +0100
> From: "Juanma Barranquero" <lekktu@gmail.com>
> Cc: "martin rudalics" <rudalics@gmx.at>, "Eli Zaretskii" <eliz@gnu.org>, 
> 	emacs-devel@gnu.org
> 
> Emacs is not our toy, but a tool for many people. Of course we can say
> users of CVSNT "fix the problem yourselves", but there's quite a
> difference between setting a variable, and writing a wrapping script
> (which, to be robust, is not just blindingly changing -m to -M), not
> to say downloading and setting up all the dependencies and tools
> needed to rebuild CVSNT. All for the single bit that separates "m"
> from "M".

Is the vc-dir and "cvs -f status" problem the only reason why you
cannot switch to the recent port of CVS offered by
http://ftp.gnu.org/non-gnu/cvs/binary/ ?  If there are other reasons,
what are they?

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

On Tue, Jan 6, 2009 at 21:49, Eli Zaretskii <eliz@gnu.org> wrote:

> Is the vc-dir and "cvs -f status" problem the only reason why you
> cannot switch to the recent port of CVS offered by
> http://ftp.gnu.org/non-gnu/cvs/binary/ ?  If there are other reasons,
> what are they?

As I said, CVSNT has been exceptionally stable for me, doesn't require
setting up SSH or PuTTY, and in the past (not recently) I used it's
server features. So yes, I can switch, but I'm not sure why I should.
CVSNT is still free, after all. Adding an ad to a message and
providing a way to avoid it can hardly be reason enough to consider it
non-free.

Anyway, what do you propose about the [non-Emacs-developer] users' out
there, for which this whole -m/-M issue will be an inconvenience? It's
not like they're guilty of using proprietary software.

    Juanma

Re: Option to customize the -m message flag of cvs?

[Eli Zaretskii]

> Date: Tue, 6 Jan 2009 21:59:25 +0100
> From: "Juanma Barranquero" <lekktu@gmail.com>
> Cc: rudalics@gmx.at, emacs-devel@gnu.org
> 
> Anyway, what do you propose about the [non-Emacs-developer] users' out
> there, for which this whole -m/-M issue will be an inconvenience?

That they use the official port from ftp.gnu.org.

Re: Option to customize the -m message flag of cvs?

[martin rudalics]

 > Is the vc-dir and "cvs -f status" problem the only reason why you
 > cannot switch to the recent port of CVS offered by
 > http://ftp.gnu.org/non-gnu/cvs/binary/ ?  If there are other reasons,
 > what are they?

The "cvs -f status" problem is an initial obstacle.  I don't have the
slightest idea what other problems I might encounter when using that
port.  Currently, I'm mostly harassed by the problems that

(1) reverting a buffer under VC may change its line endings from Unix
     to DOS, and

(2) `vc-diff' gets me a Diff buffer with ^M appended on each line.

My knowledge of vc issues is far too limited to get me a usable, working
environment here.  Adding the problem that I can't convince people to
include a simple option in vc-cvs.el, I meanwhile deeply regret having
considered `vc-dired' for managing my commits.  In the future, I'll be
more prudent when reading text like

      Hmmm, so all the more reason for people to stop committing stuff
      separately... :-/

      -Miles

martin

Re: Option to customize the -m message flag of cvs?

[Eli Zaretskii]

> Date: Wed, 07 Jan 2009 09:18:55 +0100
> From: martin rudalics <rudalics@gmx.at>
> CC: emacs-devel@gnu.org
> 
>  > Is the vc-dir and "cvs -f status" problem the only reason why you
>  > cannot switch to the recent port of CVS offered by
>  > http://ftp.gnu.org/non-gnu/cvs/binary/ ?  If there are other reasons,
>  > what are they?
> 
> The "cvs -f status" problem is an initial obstacle.  I don't have the
> slightest idea what other problems I might encounter when using that
> port.

There's no need to assume there will be other problems, without trying
it first.  Can you try and see how well it goes?

> Currently, I'm mostly harassed by the problems that
> 
> (1) reverting a buffer under VC may change its line endings from Unix
>      to DOS, and
> 
> (2) `vc-diff' gets me a Diff buffer with ^M appended on each line.

These two sound unrelated to vc-dir (or even VC in general), but
rather bugs related to setting up the encoding/decoding stuff when
these commands run.

> In the future, I'll be more prudent when reading text like
> 
>       Hmmm, so all the more reason for people to stop committing stuff
>       separately... :-/
> 
>       -Miles

Yes, people who live on GNU/Linux exclusively frequently underestimate
the plight of using their recipes on other platforms.

Re: Option to customize the -m message flag of cvs?

[Stefan Monnier]

> I'd like to strongly object to this type of change.

I don't know about "strongly" but I tend to agree: -m is the right
option to use, and if it inserts advertisement, well, that's on purpose,
so it's a feature rather than a bug.


        Stefan


PS: We could go one step further and automatically detect that it's
CVSNT and it supports -M and automatically use -M over -m, thus
defeating CVSNT's advertisement.  But this may simply lead to CVSNT
using some other means of getting its advertisement message in some
future version.

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

On Tue, Jan 6, 2009 at 19:26, Stefan Monnier <monnier@iro.umontreal.ca> wrote:

> I don't know about "strongly" but I tend to agree: -m is the right
> option to use, and if it inserts advertisement, well, that's on purpose,
> so it's a feature rather than a bug.

I understand that to mean that it's not a problem if I simply continue
using CVSNT and the advertisements go into the Emacs CVS logs.

> PS: We could go one step further and automatically detect that it's
> CVSNT and it supports -M and automatically use -M over -m, thus
> defeating CVSNT's advertisement.  But this may simply lead to CVSNT
> using some other means of getting its advertisement message in some
> future version.

It's the CVSNT people who added -M and explained it's use. They have a
FAQ about how to avoid the message, both in the client and the CVSNT
server; they even distribute (or are going to) a custom TortoiseCVS
that already uses -M. This is not a war, just a way (misguided or not)
of increasing visibility for their chosen business model.

    Juanma

Re: Option to customize the -m message flag of cvs?

[Sam Steingold]

Stefan Monnier wrote:
> 
> PS: We could go one step further and automatically detect that it's
> CVSNT and it supports -M and automatically use -M over -m, thus
> defeating CVSNT's advertisement. 

with all due respect, I find the whole idea of polluting emacs logs with spam 
so objectionable, that I suggest that emacs detect cvsnt and _refuse_ to work 
with it outright - the same way emacs might refuse to work with any other malware.

> But this may simply lead to CVSNT
> using some other means of getting its advertisement message in some
> future version.

exactly!
a firm position that spamware is not supported would be better, IMO.

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

n Mon, Jan 12, 2009 at 23:18, Sam Steingold <sds@gnu.org> wrote:

> that I suggest that emacs detect cvsnt and _refuse_
> to work with it outright

Yes, we should go and say to Emacs users that they can use Emacs with
whatever free VCS they choose, as long as it's on our Holy Software
List.

> the same way emacs might refuse to work with any
> other malware.

Rhetoric is not going to turn CVSNT non-free.

    Juanma

Re: Option to customize the -m message flag of cvs?

[dhruva]

If there is a company mail server automatically, would emacs stop
sending mails from that server?

On 1/13/09, Juanma Barranquero <lekktu@gmail.com> wrote:
> n Mon, Jan 12, 2009 at 23:18, Sam Steingold <sds@gnu.org> wrote:
>
>> that I suggest that emacs detect cvsnt and _refuse_
>> to work with it outright
>
> Yes, we should go and say to Emacs users that they can use Emacs with
> whatever free VCS they choose, as long as it's on our Holy Software
> List.
>
>> the same way emacs might refuse to work with any
>> other malware.
>
> Rhetoric is not going to turn CVSNT non-free.
>
>     Juanma
>
>
>


-- 
Contents reflect my personal views only!

Re: Option to customize the -m message flag of cvs?

[dhruva]

I meant, if a company mail server automatically adds a signature message.

On 1/13/09, dhruva <dhruvakm@gmail.com> wrote:
> If there is a company mail server automatically, would emacs stop
> sending mails from that server?
>
> On 1/13/09, Juanma Barranquero <lekktu@gmail.com> wrote:
>> n Mon, Jan 12, 2009 at 23:18, Sam Steingold <sds@gnu.org> wrote:
>>
>>> that I suggest that emacs detect cvsnt and _refuse_
>>> to work with it outright
>>
>> Yes, we should go and say to Emacs users that they can use Emacs with
>> whatever free VCS they choose, as long as it's on our Holy Software
>> List.
>>
>>> the same way emacs might refuse to work with any
>>> other malware.
>>
>> Rhetoric is not going to turn CVSNT non-free.
>>
>>     Juanma
>>
>>
>>
>
>
> --
> Contents reflect my personal views only!
>


-- 
Contents reflect my personal views only!

Re: Option to customize the -m message flag of cvs?

[Sam Steingold]

Juanma Barranquero wrote:
> n Mon, Jan 12, 2009 at 23:18, Sam Steingold <sds@gnu.org> wrote:
> 
>> that I suggest that emacs detect cvsnt and _refuse_
>> to work with it outright
> 
> Yes, we should go and say to Emacs users that they can use Emacs with
> whatever free VCS they choose, as long as it's on our Holy Software
> List.

there is a difference between that and what I am proposing.
I am saying "support all software which has not been proven to be spamware".
do NOT support known spamware however easy to support it might be.

>> the same way emacs might refuse to work with any
>> other malware.
> 
> Rhetoric is not going to turn CVSNT non-free.

spamware should be avoided whether it is free or not.

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

On Tue, Jan 13, 2009 at 19:51, Sam Steingold <sds@gnu.org> wrote:

> there is a difference between that and what I am proposing.
> I am saying "support all software which has not been proven to be spamware".
> do NOT support known spamware however easy to support it might be.

Obviously, your definition of spamware is quite extreme.

    Juanma

Re: Option to customize the -m message flag of cvs?

[mail]

Dan Nicolaescu <dann@ics.uci.edu> writes:
> martin rudalics <rudalics@gmx.at> writes:
>
>   > >> Does `vc-dir' work for you?
>   > >
>   > > I don't use it.
>   > 
>   > Apparently only Juanma and me are presently using `vc-dir' on Windows.
>   > So if there are no objections I'd like to apply the attached patch.
>
> I'd like to strongly object to this type of change.
>
> The same goal can be accomplished by using a wrapper script for the cvs
> command that changes -m to -M, or by changing the CVSNT sources not to
> do this.

+1. CVSNT's decision to add advertising to -m is obnoxious, working
around it will only encourage them to add more silly annoyances. This
should be fixed on the CVSNT side (either by compiling your own version
or pleading with the developers)

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

On Tue, Jan 6, 2009 at 23:32,  <mail@justinbogner.com> wrote:

> CVSNT's decision to add advertising to -m is obnoxious, working
> around it will only encourage them to add more silly annoyances.

Yes, that's surely the reason they explained how to work around it, so
they have an excuse to add more silly annoyances.

> (either by compiling your own version or pleading with the developers)

I don't think it's realistic to propose that to everybody who uses CVSNT.

    Juanma

Re: Option to customize the -m message flag of cvs?

[Eli Zaretskii]

> Date: Tue, 06 Jan 2009 11:57:24 +0100
> From: martin rudalics <rudalics@gmx.at>
> Cc: emacs-devel@gnu.org
> 
> Apparently only Juanma and me are presently using `vc-dir' on Windows.
> So if there are no objections I'd like to apply the attached patch.

I'd rather we fix vc-dir to not trigger the problem in the free
versions of ported CVS.  That is, instead of "cvs -f status /." invoke
it as "cvs -f status".  Then you could toss CVSNT and use the free
port again.  Could you propose a patch along these lines?

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

On Tue, Jan 6, 2009 at 21:46, Eli Zaretskii <eliz@gnu.org> wrote:

> I'd rather we fix vc-dir to not trigger the problem in the free
> versions of ported CVS.  That is, instead of "cvs -f status /." invoke
> it as "cvs -f status".

That would be good.

> Then you could toss CVSNT and use the free
> port again.

CVSNT is a free port.

    Juanma

Re: Option to customize the -m message flag of cvs?

[Eli Zaretskii]

> Date: Tue, 6 Jan 2009 21:53:17 +0100
> From: "Juanma Barranquero" <lekktu@gmail.com>
> Cc: "martin rudalics" <rudalics@gmx.at>, emacs-devel@gnu.org
> 
> > Then you could toss CVSNT and use the free
> > port again.
> 
> CVSNT is a free port.

Well, the ``other'' ports, then.

(A port that shoves some advertisement into my face is not free in my
book, but that's me.)

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

On Tue, Jan 6, 2009 at 22:01, Eli Zaretskii <eliz@gnu.org> wrote:

> (A port that shoves some advertisement into my face is not free in my
> book, but that's me.)

I don't like the advertisement, either. But they've been quite
straightforward about it and have taken the pain to add ways (and even
a FAQ) to get rid of it, so I think we can cut them some slack.

    Juanma

Re: Option to customize the -m message flag of cvs?

[Eli Zaretskii]

> Date: Tue, 6 Jan 2009 22:06:39 +0100
> From: "Juanma Barranquero" <lekktu@gmail.com>
> Cc: rudalics@gmx.at, emacs-devel@gnu.org
> 
> On Tue, Jan 6, 2009 at 22:01, Eli Zaretskii <eliz@gnu.org> wrote:
> 
> > (A port that shoves some advertisement into my face is not free in my
> > book, but that's me.)
> 
> I don't like the advertisement, either. But they've been quite
> straightforward about it and have taken the pain to add ways (and even
> a FAQ) to get rid of it, so I think we can cut them some slack.

It's still an annoyance, and being straightforward about it doesn't
make it less so.  I wish they'd understand that they are hindering
every piece of Free Software out there that invokes CVS as its
back-end.  But since you or Martin already said that it's impractical
to ask them to revert that annoyance, I'm suggesting to fix VC instead
so that the official gnu.org ports could be used instead.

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

On Tue, Jan 6, 2009, Eli Zaretskii <eliz@gnu.org> wrote:

> That they use the official port from ftp.gnu.org.

That's unrealistic. They could be using CVSNT because of its many
features over standard CVS.

Moreover, switching from CVSNT to CVS is not necessarily trivial. For
my setup, for example, I would have to checkout Emacs again or modify
every single CVS/root file, which currently is like

   :ssh;key='\path\to\my\key':username@cvs.savannah.gnu.org:/cvsroot/emacs

> It's still an annoyance, and being straightforward about it doesn't
> make it less so.

I can not deny that it is annoying. But I don't think we should be
judging free software as less worthy of our support because of some
minor annoyance.

> I'm suggesting to fix VC instead
> so that the official gnu.org ports could be used instead.

I'm trying CVS 1.12.13a from http://ftp.gnu.org/non-gnu/cvs/binary/,
and, in a ironic twist, that's part of the output of cvs -H:

  The Concurrent Versions System (CVS) is a tool for version control.
  For CVS updates and additional information, see
      the CVS home page at http://www.nongnu.org/cvs/ or
      the CVSNT home page at http://www.cvsnt.org/

    Juanma

Re: Option to customize the -m message flag of cvs?

[Stefan Monnier]

>> I'd rather we fix vc-dir to not trigger the problem in the free
>> versions of ported CVS.  That is, instead of "cvs -f status /." invoke
>> it as "cvs -f status".
> That would be good.

Agreed.

>> Then you could toss CVSNT and use the free
>> port again.
> CVSNT is a free port.

Agreed.


        Stefan "Feeling AOLish today"

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

On Mon, Dec 29, 2008 at 21:52, Eli Zaretskii <eliz@gnu.org> wrote:

> What I don't understand is why would someone use CVSNT instead of some
> other Windows port that doesn't have such a problem in -m, and thus
> does not need any changes in VC.

In my experience, as a CVS client on Windows it is more stable, well
integrated and bug free that other implementations I've tried. It also
has nice features, like built-in support for ssh (I mean, you don't
have to install any ssh or PuTTY package, just CVSNT). As a server it
has lots of useful features; there's a list at
http://www.march-hare.com/cvsnt/newfeatures/.

As for the -m "problem", it is recent, and it would be more of a
problem if they had not provided ways to circumvent it; as it is now,
it's a minor inconvenience.

    Juanma

Re: Option to customize the -m message flag of cvs?

[Juanma Barranquero]

On Mon, Dec 29, 2008 at 21:52, Eli Zaretskii <eliz@gnu.org> wrote:

> What I don't understand is why would someone use CVSNT instead of some
> other Windows port that doesn't have such a problem in -m, and thus
> does not need any changes in VC.

Also, leaving aside the reasons I might have for using CVSNT, it is
quite popular, and not only on Windows; if you can trust their data,
there are 1.4 million downloads per year. It's not unreasonable to add
a couple of lines of code to help users of a free software tool, IMHO.

    Juanma