Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 2910 - cvs commit and no message
Summary: cvs commit and no message
Status: RESOLVED FIXED
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] Development (show other bugs)
Hardware: All Linux
: High normal (vote)
Assignee: Bjoern Brauel (RETIRED)
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2002-05-22 04:52 UTC by Take Vos
Modified: 2003-10-17 11:24 UTC (History)
4 users (show)

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Take Vos 2002-05-22 04:52:24 UTC
When using cvs commit it askes for a message to log. 
If you keep this message empty it returns to cvs which says no message 
specified, continue, abort, etc. 
If you enter continue, it starts vi again for a message, but the input and 
output for vi are not coupled to the terminal (vi tells this). and you can't 
leave vi expect using ^C. Which leaves a lock on the cvs repository. 
 
my EDITOR variable is set to vi and vi is a symlink to vim
Comment 1 John Davis (zhen) (RETIRED) gentoo-dev 2002-12-15 21:13:30 UTC
Well, vi and cvs save the log to a temp file in /temp, and upon closure of vi 
via ":wq", cvs automagically uses that temp file.
//ZhEN
Comment 2 Take Vos 2002-12-16 04:55:00 UTC
I don't think you understand my bug report.  If you don't enter a comment, cvs will start vi a second time.  except this time vi does not have a controling terminal, so that you can't type anything. 
Comment 3 John Davis (zhen) (RETIRED) gentoo-dev 2002-12-17 16:34:36 UTC
I guess my second comment would be, why arent you entering a comment?
//ZhEN
Comment 4 J Robert Ray 2002-12-17 20:25:09 UTC
Take, are you connecting to a remote CVS repository over ssh?

The problem seems to come from vi being launched on the server end, not on your
local machine.  With strace I can see the error messages being relayed through
the ssh connection to the local cvs process, which then prints them out.  I do
not see vi being exec'd a second time.

This could be a bug in CVS.  I don't see this problem when using a local
repository.  I agree the best workaround at this point is to avoid giving a
blank comment, or chose 'abort' or 'edit' instead of 'continue'.
Comment 5 Take Vos 2002-12-18 03:25:11 UTC
I sometimes don't enter a comment, because some diffs speak for them selfs 
and because I commit quite a lot. 
 
I use pserver and I see the same behavior, vi is started at the remote end. 
Comment 6 John Davis (zhen) (RETIRED) gentoo-dev 2002-12-19 16:23:10 UTC
This may be a cvs bug as stated by jrray, I will check on this. If this is a cvs
bug, you will have to move this upstream to the cvs developers. Let me crunch on
this some before you move upstream though.

Thanks,
//ZhEN
Comment 7 John Davis (zhen) (RETIRED) gentoo-dev 2002-12-19 18:16:25 UTC

*** This bug has been marked as a duplicate of 11136 ***
Comment 8 John Davis (zhen) (RETIRED) gentoo-dev 2002-12-19 20:22:20 UTC
dup issue
Comment 9 John Davis (zhen) (RETIRED) gentoo-dev 2002-12-20 19:37:09 UTC
Bjb:
I saw your name in the CVS ebuild, so I am reassigning this to you ;-)

Cheers,
//ZhEN
Comment 10 Andrew Cooks (RETIRED) gentoo-dev 2003-10-17 04:04:39 UTC
I think this has been fixed. I'm using cvs over ssh and vim. CVS version:
1.11.6 , ssh version: 3.7.1p2, vim version: 6.2-r3
Comment 11 Martin Holzer (RETIRED) gentoo-dev 2003-10-17 11:24:10 UTC
closing