Re: Interpreting gcc-3.3 "make check" results

[Dan Kegel <dank at kegel dot com>]

Doug Evans wrote:
> Ken Rose writes:
>  > Dan Kegel wrote:
>  > > 
>  > > Ken Rose wrote:
>  > > >>I'm testing gcc-3.3/glibc-2.3.2, and I forgot to use the -k option
>  > > >>to 'make check', so it terminated on the first failure.
>  > > >
>  > > > -k is a new one for me.  When did it appear?
>  > > 
>  > > 1985 or so.  It just means 'keep running make even if one target fails'.
>  > > It's used in e.g. contrib/regression/btest-gcc.sh
>  > 
>  > Oh, right, make.  I was thinking of runtest.
>  > 
>  > Still, it shouldn't bail if a testcase fails, just if runtest fails,
>  > that is, exits abnormally.  I've had test runs produce thousands of
>  > failures, and then "make check" goes on and tests the next compiler
>  > anyway.
>
> I think by "first failure" Dan meant the first runtest failed
> (and by failure here I mean of course "non-zero exit code").
>
> i.e. suppose you did a make check in a binutils release
> and both ld and gas have testsuite failures (and for the sake of
> discussion suppose gas gets tested first).
> "make check" will run the gas testsuite but not the ld testsuite.
> "make -k check" will run both testsuites (plus all the others
> in the release of course).

Yes, I was writing imprecisely.  I used make's -k flag to avoid aborting
after "make check-gcc" even though I still wanted to do "make check-target-libstdc++-v3".
Or something like that.  Sorry for the very public confusion.

Anyway, I'm happily on to figuring out why dejagnu isn't finding standard.exp
for me (see the dejagnu mailing list...)
- Dan


--
Dan Kegel
http://www.kegel.com
http://counter.li.org/cgi-bin/runscript/display-person.cgi?user=78045


------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sources.redhat.com