Re: egcs-1.1 release issues

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

  In message <9808261603.AA28812@marc.watson.ibm.com>you write:
  > >>>>> Jeffrey A Law writes:
  > 
  > Jeff> Regression testing:
  > 
  > Jeff> There's still a few regressions on the rs6000/powerpc AIX targets.
  > Jeff> For various reasons I can't perform any tests on them right now,
  > Jeff> so it is important for someone to test them with the latest
  > Jeff> snapshot.
  > 
  > 	Regressions relative to what?  There are some testsuite failures
  > which still are not fixed, but these are things like the small-structure
  > passing bug which requires a big change to argument passing and the
  > meaning / necessity of MUST_PASS_IN_STACK.  [execute/931004-12.c]
They were g++ issues.  I couldn't reproduce them a week or two ago,
when I tried, but we've gotten reports from two folks which show the
same regressions.  I was going to dive further into it myself before
asking you to get involved.


For rs6000-ibm-aix4.2.0.0 2.91.51 19980803 Rodney Brown reported:
FAIL: g++.jason/thunk2.C (test for excess errors)
FAIL: g++.mike/eh30.C (test for excess errors)
FAIL: g++.mike/p4750.C (test for excess errors)
FAIL: g++.other/partspec1.C (test for excess errors)
FAIL: g++.pt/static3.C (test for excess errors)
FAIL: g++.robertl/eb101.C (test for excess errors)
FAIL: g++.robertl/eb106.C (test for excess errors)
XPASS: g++.robertl/eb132.C (test for excess errors)
FAIL: g++.robertl/eb71.C (test for excess errors)
FAIL: g++.robertl/eb91.C (test for excess errors)

For rs6000-ibm-aix4.2.0.0 2.91.54 19980816 Rodney Brown reported:
FAIL: g++.jason/thunk2.C (test for excess errors)
FAIL: g++.mike/eh30.C (test for excess errors)
FAIL: g++.mike/p4750.C (test for excess errors)
FAIL: g++.other/partspec1.C (test for excess errors)
FAIL: g++.pt/static3.C (test for excess errors)
FAIL: g++.robertl/eb101.C (test for excess errors)
FAIL: g++.robertl/eb106.C (test for excess errors)
XPASS: g++.robertl/eb132.C (test for excess errors)
FAIL: g++.robertl/eb71.C (test for excess errors)
FAIL: g++.robertl/eb91.C (test for excess errors)


For rs6000-ibm-aix4.1.4.0 2.51.54 19980816 David Schuler reported:
FAIL: g++.jason/thunk2.C (test for excess errors)
FAIL: g++.mike/eh30.C (test for excess errors)
FAIL: g++.mike/p4750.C (test for excess errors)
FAIL: g++.pt/static3.C (test for excess errors)
FAIL: g++.robertl/eb101.C (test for excess errors)
FAIL: g++.robertl/eb106.C (test for excess errors)
XPASS: g++.robertl/eb132.C (test for excess errors)
FAIL: g++.robertl/eb71.C (test for excess errors)
FAIL: g++.robertl/eb91.C (test for excess errors)

For rs6000-ibm-aix4.2.1.0  2.51.54 19980816 David Schuler reported:
FAIL: g++.jason/thunk2.C (test for excess errors)
FAIL: g++.law/profile1.C (test for excess errors)
FAIL: g++.mike/eh30.C (test for excess errors)
FAIL: g++.mike/p4750.C (test for excess errors)
FAIL: g++.pt/static3.C (test for excess errors)
FAIL: g++.robertl/eb101.C (test for excess errors)
FAIL: g++.robertl/eb106.C (test for excess errors)
XPASS: g++.robertl/eb132.C (test for excess errors)
FAIL: g++.robertl/eb71.C (test for excess errors)
FAIL: g++.robertl/eb91.C (test for excess errors)


Now if we look at the egcs-1.0.3a results we find the following
tests are regressions:
PASS: g++.jason/thunk2.C (test for excess errors)
PASS: g++.law/profile1.C (test for excess errors)
PASS: g++.mike/eh30.C (test for excess errors)
PASS: g++.mike/p4750.C (test for excess errors)
PASS: g++.robertl/eb106.C (test for excess errors)

For for gcc-2.8.1 we find the following additional regression:
PASS: g++.other/partspec1.C (test for excess errors)



Between the power failures in Sunnyvale and a couple dead disks I
haven't been able to get a ppc toolchain built so that I could
look at the regressions.

jeff