Re: Results for egcs-2.91.54 19980816 (gcc2 ss-980609 experimental) testsuite on mips-sgi-irix5.2

[Jeffrey A Law <law at cygnus dot com>]

  In message <orsoimosg5.fsf@iguacu.dcc.unicamp.br>you write:
  > I guess I should stop posting results for this host.  It's a rather
  > old host, and it seems to be randomly failing.  I usually have to
  > restart the egcs bootstrap a few times to get it build correctly. :-(
  > 
  > What would you prefer?  Should I stop caring about it or keep posting
  > possibly incorrect results?
Good question.  Cygnus has had similar reliability issues with our SGI
machines.    They've been *extremely* heat sensitive.   Some of our
customers have experienced intermittant problems of this nature too.


  > >> FAIL: gcc.c-torture/compile/combine-hang.c,  -Os  
  > > Hmm.  I didn't get this on irix5.3 and it's a regression since your
  > > last report.  Can you take a peek at the log file and see why it
  > > failed?
  > 
  > It was a signal 11: memory or swap disk problem
Can you try just compiling combine-hang.c with -Os with the 19980816
compiler and see if it still gets a sig 11?  If it does, then we've
probably got a reason to investigate further.  If it does not it might
be time to deep six irix5.2.

jeff