On Wed, Jul 04, 2001 at 09:45:00PM +0000, Ton Hospel wrote: > In article <20010703092027.F19034@chaos.wustl.edu>, > Jarkko Hietaniemi <jhi@iki.fi> writes: > > On Tue, Jul 03, 2001 at 04:03:04PM +0100, H.Merijn Brand wrote: > >> On Tue 03 Jul 2001 15:44, Jarkko Hietaniemi <jhi@iki.fi> wrote: > >> > > Failures: > >> > > > >> > > HP-UX 11.00 stdio/perlio -Uuseperlio > >> > > HP-UX 11.00 stdio/perlio -DDEBUGGING -Uuseperlio > >> > > HP-UX 11.00 stdio/perlio -Duseperlio > >> > > HP-UX 11.00 stdio/perlio -DDEBUGGING -Duseperlio > >> > > ext/Storable/t/lock................FAILED at test 3 > >> > > >> > What the ... 11060 did change Storable (1.0.12 update from Raphael, > >> > but wasn't the previous smoke report from 11061 and it was all oks for > >> > the vanilla config?) > >> > >> It was for ext/Storable tests for sure. 10.20 still had the op/write fail, but > >> I tackled that one. > > > > Hmmm. Another possibility is that I broke the fcntl probe somehow? > > (Change #11093, I added an alarm() to the test because I saw > > fcntl(..., F_SETLK, ...) to hang...) > > > > to be honest, that sounds like a dubious change anyways since it doing > anything at all depends on the sigaction settings. a F_SETLK should only > hang if the file is really locked (or some weird not per related failures > like doing the test over nfs and not being able to reach the remote lockd) > on the sigaction settings. > > I suspect it is symptom fixing instead of bug fixing. I'm not going to be able to fix the NFS code of the particular UNIX implementation. -- $jhi++; # http://www.iki.fi/jhi/ # There is this special biologist word we use for 'stable'. # It is 'dead'. -- Jack CohenThread Previous | Thread Next