David.Comay at Sun.COM wrote: >> Nit that you should update the copyright in build_dist.bash, but >> otherwise looks OK. Looked at a couple of the postrun scripts, didn't >> see anything that looked odd. Out of curiosity, are we getting fewer >> failures from them now? > > Yes, I'll update the copyright. > > We're actually still seeing the same (or similar) failures with some of > the postruns. One thing I noticed is the exec_postrun script is > throwing away the output (include stderr) - I don't know if it's > intentional but I suspect not since there's the following line in the > case of a non-zero exist status > > cat $OUTPUT >2 > > which of course puts the command output in a file called "2" :-) > > I've changed this as follows - let me know if this makes sense or have > I missed the intent here? > > --- /tmp/exec_postrun Fri Jan 25 21:52:45 2008 > +++ ./exec_postrun Fri Jan 25 20:59:49 2008 > @@ -13,10 +13,10 @@ > > exec_name=`echo *.$num` > echo "Executing $exec_name" > - /bin/sh $exec_name > $OUTPUT 2>& 1 > + /bin/sh $exec_name 2> $OUTPUT > /dev/null
I think Karen was attempting to capture both stdout and stderr so we'd have all the output possible to report in the failure case below. > if [ "$?" != "0" ] ; then > echo "$exec_name failed" > - cat $OUTPUT >2 > + cat $OUTPUT Which was obviously wrong, thanks for noticing that. Dave
