This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Current libstdc++-v3 testsuite failures
- From: Loren James Rittle <rittle at latour dot rsch dot comm dot mot dot com>
- To: libstdc++ at gcc dot gnu dot org
- Date: Wed, 11 Feb 2004 01:57:36 -0600 (CST)
- Subject: Current libstdc++-v3 testsuite failures
- Reply-to: rittle at labs dot mot dot com
Great work everyone! I know there is still one assigned libstdc++-v3
PR in "Mark's 3.4 release show-stoppers" URL (Paolo, want help solving
it? ;-) however...
i386-unknown-freebsd4.9, no failures on mainline
i386-unknown-freebsd4.9, no failures on 3.3.3 or 3.4 (last checked)
sparc64-unknown-freebsd5.2, no interesting failures on 3.4 (today)
Some FAILs related to:
"cc1plus: no iconv implementation, cannot convert from iso-8859-1 to UTF-8"
FAIL: 22_locale/ctype/is/wchar_t/1.cc execution test
FAIL: 22_locale/ctype/scan/wchar_t/1.cc execution test
FAIL: 27_io/basic_filebuf/open/char/9507.cc execution test
FAIL: 27_io/basic_filebuf/showmanyc/char/9533-1.cc execution test
alpha-unknown-freebsd5.2, no interesting failures on 3.4 (today)
All above FAILs plus:
FAIL: 22_locale/time_get/get_year/wchar_t/3.cc execution test
No wchar or iconv-related failure is a regression from any past
release (support for wchar is still coming in FreeBSD5). The
27_io/basic_filebuf/* failures are poll() hangs in the thread library
(i.e. they go away when I run the tests without -pthread; completely
new implementations under FreeBSD5, still baking...).
I pulled these "test case only" fixes over to 3.4 (with update -j1.):
* testsuite/ext/stdio_filebuf/char/10063-2.cc: Treat unbuffered.
* testsuite/ext/stdio_filebuf/char/10063-3.cc: New test. Like -2 but
use __gnu_cxx::stdio_sync_filebuf<char> instead; allow buffered stream.
* testsuite/27_io/ios_base/storage/11584.cc: Correct new and
delete declarations, add include and test variable.
Regards,
Loren