Results for egcs-2.91.52 19980729 (gcc2 ss-980609 experimental) testsuite...

[Manfred Hollstein]

On Sat, 1 August 1998, 09:56:33, N8TM@aol.com wrote:

 > In a message dated 7/31/98 8:24:40 AM Pacific Daylight Time, manfred@s-
 > direktnet.de writes:
 > 
 > > XPASS: g++.robertl/eb132.C (test for excess errors)
 > 
 > When I tested with the 19980727 snapshot, I didn't get this one, but I had a
 > total of 13 more unexpected failures.  The only obvious difference is that I
 > built with default options throughout, except for using --with-bin-as for the
 > snapshot (not for 1.0.3a, because it fails to build objc with that option),
 > thinking that this would produce a more valid comparison between egcs-1.0.3a
 > and the snapshot. I'm assuming that relatively naive users like myself would
 > leave the options alone without any documented reason to change them.  I would
 > think that a more obvious variation would be to profile the compiler with
 > standard options and -Os, and use that wherever it didn't show a performance
 > reduction or produce an error.  More hard work for someone.
 > 

Yep, I'm afraid you're right; I've gotten failures in libstdc++ which
nobody else reported - I've been using -fomit-frame-pointer which
caused a problem to show up, which wasn't seen without.

manfred