Last known to work with: "Mon Aug 8 09:42:20 UTC 2005".
Known to fail with: "Mon Aug 8 09:42:20 UTC 2005".
With LAST_UPDATED: "Tue Aug 9 09:32:29 UTC 2005" I still get failures,
with the messages in the .log being:
FAIL: g++.dg/ext/packed3.C (test for errors, line 22)
FAIL: g++.dg/ext/packed3.C (test for errors, line 24)
PASS: g++.dg/ext/packed3.C (test for excess errors)
PASS: g++.dg/ext/packed4.C (test for excess errors)
FAIL: g++.dg/ext/packed4.C execution test
FAIL: g++.dg/ext/packed8.C (test for errors, line 18)
PASS: g++.dg/ext/packed8.C (test for excess errors)
All these tests are about the attribute packed.
For cris-*, the ABI says everything already *is* packed; no
alignment or padding of structures or members.
I could of course skip these tests for cris-* myself,
but I think the failures reveal either bugs in the tests or in the code.
(Perhaps the presence of the attribute in the source should matter
for the "binding" referred to, not the actual effect.)
And if the tests should be skipped target-dependently anyway because of no
alignment/padding, it should be done by some testsuite machinery testing
for presence of padding, like testing for validity of the -fPIC/-fpic options.
Else we'll just start growing target skip lists in each test.
Author of only suspected patch (as well as the tests!) CCed.
--
Summary: [4.1 regression] testsuite failures:
g++.dg/ext/packed3.C, packed4.C, packed8.c and
g++.dg/other/crash-4.C
Product: gcc
Version: 4.1.0
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: c++
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: hp at gcc dot gnu dot org
CC: gcc-bugs at gcc dot gnu dot org,nathan at codesourcery
dot com
GCC host triplet: x86_64-unknown-linux-gnu
GCC target triplet: cris-elf and cris-axis-linux-gnu
BugsThisDependsOn: 21166
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23304
