[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: need help with broken cvs trigger
From: |
Paul Sander |
Subject: |
Re: need help with broken cvs trigger |
Date: |
Mon, 2 Oct 2000 00:35:51 -0700 |
Are the nested sub-shells really desired? I find that that particular
construct does not work (though I admit I'm not a bash user). What happens
if you move that stuff into a separate shell script and invoke it from your
loginfo file instead?
Also, is there no chance that you've somehow changed the shell that
executes your loginfo commands?
--- Forwarded mail from address@hidden
hi guys. i've spent the last several hours falling in hate with cvs and
could use some help getting that loving feeling back. i had my repository
all set up, everything had been working nicely for months. i had a trigger
set up through loginfo to automatically do some tasks upon checkin as
described in the cvs manual - the file looks basically like so:
ALL $CVSROOT/CVSROOT/log_accum.pl %s
^wz (date; cat; (sleep 2; /usr/local/cvs/bin/checkin wz) &)
>> $CVSROOT/CVSROOT/updatelog 2>&1
(second two lines appear on the same line, of course)
that is to say, all projects fired up the log_accum.pl script upon checkin
to handle the mail notification, while the wz project also had the
/usr/local/cvs/bin/checkin script executed - which, among other things,
checked out a copy of the project to a working directory on the
server. now, instead of getting the output from the checkin program, my
cvs commits report:
Checking in foo8.html;
/usr/local/cvs/virtual/wz/htdocs/foo8.html,v <--
foo8.html
new revision: 1.2; previous revision: 1.1
done
Mailing the commit message...
: Ambiguous redirect
the 'Mailing the commit message...' line comes from the log_accum.pl
script, so i know it's working. the problem is in the checkin line for the
wz repository - for some reason, it's resulting in ': Ambiguous redirect'
instead of its normal output. problem is, that expression parses fine to
the eye and to bash, it's copied straight from the cvs manual, and it
_used_ to work!
the only thing 'weird' that i did to the server was import a new project,
then nuke it and import it again with a different directory structure. i
can't see any way that would hurt anything.
so can someone please give me a hand here? i'm at wit's end. for the
record, i'm using cvs-1.10.6-2 on a redhat linux box. thanks.
--- End of forwarded message from address@hidden