This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Small problem in testsuite/27_io/basic_filebuf/overflow/wchar_t/11305-[1-4].cc


Pétur Runólfsson wrote:

Hummmm. However, for some reason the original testcase gave problems on 64 bit platforms, like x86-64:

11305-1.exe: /usr/src/aj/cvs/gcc/libstdc++-v3/testsuite/27_io/basic_filebuf/overflow/wchar_t/11305-1.cc:38: void test01(): Assertion `n1 != (0xffffffffu)' failed.
FAIL: 27_io/basic_filebuf/overflow/wchar_t/11305-1.cc execution test


Any idea why?


One possible explanation is that the C library is failing to convert
these characters. They are also used in some tests in
22_locale/codecvt, are there any failures there?

Doesn't seem so... I think the baset thing to do is restoring the proper sputc() arguments and then investigate on which 64 platforms, with which libc, the testcases fail.

We can try restoring only the sputc() argument of your original testcase and see what happens...


Since wfilebuf::int_type is the same type as wint_t and
wfilebuf::traits_type::eof() == WEOF the testcase should fail (if it
doesn't, there is a bug somewhere else).

I agree.

Paolo.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]