On 2 February 2013 18:34, Andreas Färber <andreas.faer...@web.de> wrote:
> Am 02.02.2013 18:17, schrieb Peter Maydell:
>> This is the only clang warning on MacOSX apart from the ones about
>> deprecated features in audio/coreaudio.c and ui/cocoa.m
> On v10.5.8 I'm seeing some redundant declaration warnings from
> disas/s390.c and assumed we wouldn't want to fix/workaround the disas
> stuff. Same for some exynos4 file.
>
> Also some uninitialized buf2 in the Realtek NIC or so.
I don't see any of these with 10.6.8 (with either clang or
the Apple gcc). Do they appear to be legitimate warnings?
> And test-thread-pool breaks due to undefined ___sync_fetch_and_add and
> ___sync_val_compare_and_swap.
Hmm, I hadn't tried running make check. I see:
(a) tests randomly hang (possibly only if built with gcc? hmm)
(b) test-string-input-visitor(59002) malloc: *** error for object
0x12068: pointer being freed was not allocated
(c) test-thread-pool seems to hang
(d) make distclean doesn't delete the test executables in test/ :-)
(e) lots of format-security warnings from check-qjson.c:
tests/check-qjson.c:53:33: warning: format string is not a string
literal (potentially insecure) [-Wformat-security]
obj = qobject_from_json(test_cases[i].encoded);
^~~~~~~~~~~~~~~~~~~~~
I'll prod a bit further on (b) at least...
-- PMM
