Messages in this thread | | | Date | Wed, 26 Sep 2012 09:21:45 -0700 | From | Greg KH <> | Subject | Re: ktest.pl always returns 0? |
| |
On Tue, Sep 25, 2012 at 07:48:30PM -0400, Steven Rostedt wrote: > I just forced a build failure to see what ktest would show. This is my > result: > > cp /home/rostedt/work/git/configs/ixf/config-use /home/rostedt/work/git/nobackup/ixf/trace/.config ... SUCCESS > touch /home/rostedt/work/git/nobackup/ixf/trace/.config ... SUCCESS > Applying minimum configurations into /home/rostedt/work/git/nobackup/ixf/trace/.config.new > mv /home/rostedt/work/git/nobackup/ixf/trace/.config.new /home/rostedt/work/git/nobackup/ixf/trace/.config ... SUCCESS > GCC_VERSION=4.6.0 distmake-64 O=/home/rostedt/work/git/nobackup/ixf/trace oldnoconfig ... SUCCESS > GCC_VERSION=4.6.0 distmake-64 O=/home/rostedt/work/git/nobackup/ixf/trace -j40 ... FAILED! > CRITICAL FAILURE... failed build > See /home/rostedt/work/git/nobackup/ixf/ixf.log for more info. > failed build > [rostedt@fedora git]$ echo $? > 25 > > > So it seems that if DIE_ON_FAILURE is set, it returns non-zero error.
Ugh, you are right, my mistake, it was a bug in my script that I was not checking the correct return value (for those wondering, my bash script was writing out a bash script that then got run, I forgot to escape the "$?" properly, so it was being intrepreted when written out and not when run.
So sorry for the noise, all is now working just fine.
greg "I should stick to C" k-h
| |