new FAILs in the g++ testsuite?

new FAILs in the g++ testsuite?

[Peter Schmid]

While examining the file g++.log for the FAILs of the g++ testsuite I
found strange logging messages.
There are FAIls without an explanation.
They do not count in the g++ summary.

Peter Schmid

PASS: g++.law/bit-fields2.C (test for excess errors)
spawn [open ...]
FAIL
PASS: g++.law/bit-fields2.C  Execution test
PASS: g++.law/operators1.C (test for excess errors)
spawn [open ...]
FAIL
PASS: g++.law/operators1.C  Execution test
PASS: g++.law/operators7.C (test for excess errors)
spawn [open ...]
FAIL
PASS: g++.law/operators7.C  Execution test
PASS: g++.law/temps4.C (test for excess errors)
spawn [open ...]
FAIL
PASS: g++.law/temps4.C  Execution test

		=== g++ Summary ===

# of expected passes		3611
# of unexpected failures	2
# of expected failures		81
# of untested testcases		6
Executing on host: /home/schmid/ftp/egcs/gcc/testsuite/../xgcc -v   
/home/schmid/ftp/egcs/gcc/testsuite/../xgcc version egcs-2.91.05 980122 
(gcc-2.8.0 release)

Re: new FAILs in the g++ testsuite?

[Manfred.Hollstein]

On Thu, 29 January 1998, 18:43:34, schmid@ltoi.iap.physik.tu-darmstadt.de wrote:

 > While examining the file g++.log for the FAILs of the g++ testsuite I
 > found strange logging messages.
 > There are FAIls without an explanation.
 > They do not count in the g++ summary.
 > 

Even  worse:  on  my   m88k-motorola-sysv3  system various  libio  and
libstdc++  tests  failed  with  dejagnu-971222;  running   those tests
manually showed there is _no_  error! Switching back to dejagnu-971028
cured this problem for me.

BTW, the 971028's .log file format was IMHO much better and obvious.

Manfred