Re: [musl] POSIX compliance testing of strftime

[Ken Mays <kmays2000@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 2174 bytes --]

I posted a bug report with some detail here:  #17884 (TMR: strftime
conversion specifier failed during test) – Haiku (haiku-os.org)
<https://dev.haiku-os.org/ticket/17884>

The default locale for me is: LC_* = en.UTF-8

The failure for test 1-1 is: X Bytes 0 - Test Failed: %X doesn't equal a
least 8 bytes

Thanks for letting me know it passes on your setup. Any known issues with
BSD-compatible setups?

~K


On Mon, Sep 12, 2022 at 2:47 PM Rich Felker <dalias@libc.org> wrote:

> On Mon, Sep 12, 2022 at 01:32:28PM -0700, Ken Mays wrote:
> > Hello,
> >
> > I am doing POSIX compliance testing of strftime on HAIKU (w/MUSL)
> > using the Open POSIX Test Suite v1.5.2 located here:
> >
> > Open POSIX Test Suite - Browse Files at SourceForge.net
> > <https://sourceforge.net/projects/posixtest/files/>
> >
> > Specifically:
> >
> > open_posix_testsuite/1-1.c at master · haiku/open_posix_testsuite
> > (github.com) <
> https://github.com/haiku/open_posix_testsuite/blob/master/conformance/interfaces/strftime/1-1.c
> >
> >
> > and
> >
> > open_posix_testsuite/2-1.c at master · haiku/open_posix_testsuite
> > (github.com) <
> https://github.com/haiku/open_posix_testsuite/blob/master/conformance/interfaces/strftime/2-1.c
> >
> >
> > Issue: MUSL's strftime only fails one minor test for both test 1-1 and
> 2-1.
> >
> > I'd like to know if a MUSL developer could review the mentioned tests?
> >
> > Are there known issues with strftime with POSIX compliance testing?
> >
> > Thanks,
> >
> > ~K
>
> When reporting test failures, could you report *which test* (not just
> which files, but which particular assertion they're trying to make) is
> failing? It would have saved a lot of time not having to get, compile,
> and run them. But now after doing that, I can't even reproduce any
> failures; I just get all PASS. So in order to investigate at all I
> really need to know what specific failures you hit.
>
> FWIW these are rather dubious tests since they depend on the current
> time when the test is run and the user's locale. Perhaps the failures
> you saw were dependent on those factors?
>
> Rich
>

[-- Attachment #2: Type: text/html, Size: 4266 bytes --]