Re: FW: commit problem with cvs mode

Re: FW: commit problem with cvs mode

[Stefan Monnier]

> I do  not think  it's so important  because adding  a line in  my .emacs
> solves the problem, so I will not  fill a bug against emacs nor jde, but

And you've bumped into this problem exactly because people before
reacted just like you're reacting now.  If people don't complain
about the problems they encounter, they never get fixed.


        Stefan

Re: commit problem with cvs mode

[Arnaud Vandyck]

"Stefan Monnier" <monnier+gnu.emacs.help/news/@flint.cs.yale.edu> wrote:
> > I do not think it's so  important because adding a line in my .emacs
> > solves the problem, so I will  not fill a bug against emacs nor jde,
> > but
> 
> And you've bumped into this problem exactly because people before
> reacted just like you're reacting now.  If people don't complain
> about the problems they encounter, they never get fixed.

Everybody does not  use CVS_RSH=ssh. I submit enough bug  to know what I
have to do.  I warn about the problem,  but I do not think it's  a bug I
have to fill against the Debian package... or maybe a wishlist.

-- Arnaud Vandyck, STE fi, ULg
   Formateur Cellule Programmation.

Re: commit problem with cvs mode

[Kai Großjohann]

Arnaud Vandyck <arnaud.vandyck@ulg.ac.be> writes:

> Everybody does not  use CVS_RSH=ssh. I submit enough bug  to know what I
> have to do.  I warn about the problem,  but I do not think it's  a bug I
> have to fill against the Debian package... or maybe a wishlist.

The problem is that you put "export CVS=ssh" into some shell init file
(~/.bash_profile or ~/.profile, I guess) and that it didn't take
effect for programs started under Gnome or whatever you're using.

At least that's my understanding.
-- 
This line is not blank.

Re: commit problem with cvs mode

[Stefan Monnier]

> Everybody does not  use CVS_RSH=ssh. I submit enough bug  to know what I

The problem has nothing to do with CVS_RSH, it's got to do with
envvars not being inherited everywhere.  It can screw up any program
in any number of ways.  Typically you get problems like

   "I can start `foo' just fine from the shell, but Netscape says it
   can't find it"


-- Stefan

envvars not being inherited everywhere (was: Re: commit problem with cvs mode)

[Arnaud Vandyck]

[-- Attachment #1: Type: Text/Plain, Size: 1015 bytes --]

"Stefan Monnier" <monnier+gnu.emacs.help/news/@flint.cs.yale.edu> wrote:
> > Everybody does not  use CVS_RSH=ssh. I submit enough bug  to know what I
> 
> The problem has nothing to do with CVS_RSH, it's got to do with
> envvars not being inherited everywhere.  It can screw up any program
> in any number of ways.  Typically you get problems like
> 
>    "I can start `foo' just fine from the shell, but Netscape says it
>    can't find it"

OK, sorry  if I've been rude, now  I understand the problem  and I think
you  both (Kai and  Stefan) are  right. Do  you have  some link  or more
information so I can send to the package maintainer? 

Maybe help me to describe the  problem with enough precision. I like the
way Stefan points it: 

                     "it's got  to do  with envvars not  being inherited
                     everywhere.   It can  screw up  any program  in any
                     number of ways".

Thanks for your help

-- Arnaud Vandyck, STE fi, ULg
   Formateur Cellule Programmation.

[-- Attachment #2: Type: Message/Rfc822, Size: 1017 bytes --]

From: kai.grossjohann@gmx.net (Kai Großjohann)
To: help-gnu-emacs@gnu.org
Subject: Re: commit problem with cvs mode
Date: Wed, 18 Jun 2003 18:26:38 +0200
Message-ID: <84brwv4aip.fsf@lucy.is.informatik.uni-duisburg.de>

Arnaud Vandyck <arnaud.vandyck@ulg.ac.be> writes:

> Everybody does not  use CVS_RSH=ssh. I submit enough bug  to know what I
> have to do.  I warn about the problem,  but I do not think it's  a bug I
> have to fill against the Debian package... or maybe a wishlist.

The problem is that you put "export CVS=ssh" into some shell init file
(~/.bash_profile or ~/.profile, I guess) and that it didn't take
effect for programs started under Gnome or whatever you're using.

At least that's my understanding.
-- 
This line is not blank.

[-- Attachment #3: Type: Message/Rfc822, Size: 831 bytes --]

From: "Stefan Monnier" <monnier+gnu.emacs.help/news/@flint.cs.yale.edu>
To: help-gnu-emacs@gnu.org
Subject: Re: commit problem with cvs mode
Date: 18 Jun 2003 14:49:11 -0400
Message-ID: <5lvfv39q6w.fsf@rum.cs.yale.edu>

> Everybody does not  use CVS_RSH=ssh. I submit enough bug  to know what I

The problem has nothing to do with CVS_RSH, it's got to do with
envvars not being inherited everywhere.  It can screw up any program
in any number of ways.  Typically you get problems like

   "I can start `foo' just fine from the shell, but Netscape says it
   can't find it"


-- Stefan

[-- Attachment #4: Type: text/plain, Size: 151 bytes --]

_______________________________________________
Help-gnu-emacs mailing list
Help-gnu-emacs@gnu.org
http://mail.gnu.org/mailman/listinfo/help-gnu-emacs

Re: commit problem with cvs mode

[Sam Halliday]

Arnaud Vandyck wrote:
> I think I'am having some troubles with  the commit in cvs mode. I have
> a project on sourceforge and it  takes times and times to commit
> ("waiting process to die" I think) but when I do it in a console, it's
> rather fast and ok!? 

i know for savannah commits, you need this in your ~/.emacs

  (setenv "CVS_RSH" "ssh")

to use ssh login (i think sourceforge is the same). maybe you forgot to
do this and cvs is trying to use rsh instead of ssh; but then again, i
would really expect cvs to moan if that was the case. try that setenv
anyway.

can you tell us about any relevant parts of your ~/.emacs file?

Re: commit problem with cvs mode

[Arnaud Vandyck]

Sam Halliday <devnull@example.com> wrote:
> (setenv "CVS_RSH" "ssh")

That solves my problem! many thanks ;)

-- Arnaud Vandyck, STE fi, ULg
   Formateur Cellule Programmation.

Re: commit problem with cvs mode

[Artur Hefczyc]

Sam Halliday wrote:

> Arnaud Vandyck wrote:
>> I think I'am having some troubles with  the commit in cvs mode. I have
>> a project on sourceforge and it  takes times and times to commit
>> ("waiting process to die" I think) but when I do it in a console, it's
>> rather fast and ok!?
> i know for savannah commits, you need this in your ~/.emacs
>   (setenv "CVS_RSH" "ssh")
It is good to set this variable on shell level so it can be used
not only yb emacs (and pcl-cvs) but also by command line
or other cvs clients.

for example for bash shell add to your .bashrc following

export CVS_RSH=ssh

Artur
-- 
Artur Hefczyc
Open Source Developer
http://www.geotools.org/
http://wttools.sourceforge.net/

Re: commit problem with cvs mode

[Sam Halliday]

Artur Hefczyc wrote:
> Sam Halliday wrote:
> > Arnaud Vandyck wrote:
> >> I think I'am having some troubles with  the commit in cvs mode. I
> >have> a project on sourceforge and it  takes times and times to
> >commit> ("waiting process to die" I think) but when I do it in a
> >console, it's> rather fast and ok!?
> > i know for savannah commits, you need this in your ~/.emacs
> >   (setenv "CVS_RSH" "ssh")
> It is good to set this variable on shell level so it can be used
> not only yb emacs (and pcl-cvs) but also by command line
> or other cvs clients.
> 
> for example for bash shell add to your .bashrc following
> 
> export CVS_RSH=ssh

in conjunction with setting it in ~/.emacs of course. my emacs is
spawned from within a blackbox menu, and it doesnt pick up shell
variables. thats probably what happened with Arnaud as well, because cvs
commits seemed to work from the command line.

Re: commit problem with cvs mode

[Stefan Monnier]

> in conjunction with setting it in ~/.emacs of course. my emacs is
> spawned from within a blackbox menu, and it doesnt pick up shell
> variables. thats probably what happened with Arnaud as well, because cvs
> commits seemed to work from the command line.

Please report such brain dead setups to whoever is responsible (probably
something like RedHat, Debian, ...).  There should be one place where
you can set all your envvars and have them inherited everywhere.
This place is typically ~/.bash_login or ~/.login but could be something
else.  ~/.bashrc is obviously wrong, although it might work.


        Stefan

Re: commit problem with cvs mode

[Reiner Steib]

On Tue, Jun 17 2003, Stefan Monnier wrote:

> There should be one place where you can set all your envvars and
> have them inherited everywhere.  This place is typically
> ~/.bash_login or ~/.login but could be something else.

I'd use `~/.profile' instead of `~/.bash_login'.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo--- PGP key available via WWW   http://rsteib.home.pages.de/

Re: commit problem with cvs mode

[Artur Hefczyc]

Stefan Monnier wrote:

>> in conjunction with setting it in ~/.emacs of course. my emacs is
>> spawned from within a blackbox menu, and it doesnt pick up shell
>> variables. thats probably what happened with Arnaud as well, because cvs
>> commits seemed to work from the command line.
> 
> Please report such brain dead setups to whoever is responsible (probably
> something like RedHat, Debian, ...).  There should be one place where
> you can set all your envvars and have them inherited everywhere.
> This place is typically ~/.bash_login or ~/.login but could be something
> else.  ~/.bashrc is obviously wrong, although it might work.
Oh, yes you are right.

At the moment I don't use RH so I can't check it to ensure however
as I remember or many boxes ".bash_profile" is read after login.
So maybe the best place is to put this variables in this file.
Or maybe on some linux distributions this is ".profile" file.

Check this, I am sure one of above should work with window managers.

Artur
-- 
Artur Hefczyc
Open Source Developer
http://www.geotools.org/
http://wttools.sourceforge.net/

Re: commit problem with cvs mode

[Sam Halliday]

Stefan Monnier wrote:
> > in conjunction with setting it in ~/.emacs of course. my emacs is
> > spawned from within a blackbox menu, and it doesnt pick up shell
> > variables. thats probably what happened with Arnaud as well, because
> > cvs commits seemed to work from the command line.
> Please report such brain dead setups to whoever is responsible

the fault is mine in this case i am afraid and not debians... i hadn't
set my window manager to load under a --login shell. silly brain dead
me... however, having emacs set this variable is more important to me
than in the shell, since i only ever use cvs from within emacs and i
have logins on several machines, all with a syncronised ~/.emacs file,
but i dont bother syncronising my .profiles.

Re: commit problem with cvs mode

[Kai Großjohann]

Arnaud Vandyck <arnaud.vandyck@ulg.ac.be> writes:

> I think I'am having some troubles with  the commit in cvs mode. I have a
> project on sourceforge and it  takes times and times to commit ("waiting
> process to die" I think) but when I do it in a console, it's rather fast
> and ok!? 

PCL-CVS doesn't provide error messages by default.  But when you hit
C-c C-c to interrupt the cvs subprocess, then you usually find some
message in the *cvs* buffer.

What do you see?  Does the message tell you what's wrong?
-- 
This line is not blank.

Re: commit problem with cvs mode

[Arnaud Vandyck]

kai.grossjohann@gmx.net (Kai Großjohann) wrote:
> Arnaud Vandyck <arnaud.vandyck@ulg.ac.be> writes:
> 
> > I think  I'am having some  troubles with the  commit in cvs  mode. I
> > have a project on sourceforge and it takes times and times to commit
> > ("waiting process  to die" I think) but  when I do it  in a console,
> > it's rather fast and ok!?
> 
> PCL-CVS doesn't provide  error messages by default.  But  when you hit
> C-c C-c  to interrupt the cvs  subprocess, then you  usually find some
> message in the *cvs* buffer.
> 
> What do you see?  Does the message tell you what's wrong?

Problem solved with Sam's mail, many thanks to both of you :)

-- Arnaud Vandyck, STE fi, ULg
   Formateur Cellule Programmation.

commit problem with cvs mode

[Arnaud Vandyck]

Hi all! 

I think I'am having some troubles with  the commit in cvs mode. I have a
project on sourceforge and it  takes times and times to commit ("waiting
process to die" I think) but when I do it in a console, it's rather fast
and ok!? 

Any idea of what happens? 

Here are my system informations (same problem with a Debian/ppc)

-- System Information:
Debian Release: testing/unstable
Architecture: i386
Kernel: Linux vbstefi30.fapse.ulg.ac.be 2.4.20-1-686 #1 Sat Mar 22 13:16:21 EST 2003 i686
Locale: LANG=fr_BE, LC_CTYPE=fr_BE (ignored: LC_ALL set)

Versions of packages emacs21 depends on:
ii  emacs21-common            21.3-1         The GNU Emacs editor's common infr
ii  libc6                     2.3.1-17       GNU C Library: Shared libraries an
ii  libjpeg62                 6b-7           The Independent JPEG Group's JPEG 
ii  libncurses5               5.3.20030510-1 Shared libraries for terminal hand
ii  libpng12-0                1.2.5.0-3      PNG library - runtime
ii  libtiff3g                 3.5.7-2        Tag Image File Format library
ii  xaw3dg                    1.5-23         Xaw3d widget set
ii  xlibs                     4.2.1-6        X Window System client libraries
ii  zlib1g                    1:1.1.4-12     compression library - runtime

Thanks,

-- Arnaud Vandyck, STE fi, ULg
   Formateur Cellule Programmation.