Bruno Haible writes:
> If "Android" was to be renamed to something else, or Bionic adopted by
> a different project, __BIONIC__ and __ANDROID__ would end up undefined
> here. So we would need a fix at this place again anyway. => Testing
> __BIONIC__ here is redundant. I'm applying this:
Thanks,
Marc-André Lureau wrote:
> +2012-01-24 Marc-André Lureau
> +
> + * lib/malloca.c: Use uintptr_t to convert pointers to
> + integers. Fix warnings with MinGW64 x64.
Thanks for the patch. However, use of uintptr_t requires inclusion of the
header that defines this type. In POSIX, or can
One more tiny change of quoting in comments:
2012-01-24 Bruno Haible
havelib: Modern quoting.
* build-aux/config.rpath: Quote 'like this', not `like this', as per
the recent change to the GNU coding standards.
--- build-aux/config.rpath.orig Tue Jan 24 13:06:16 2012
+
Simon Josefsson wrote:
> I don't have time nor interest to work on fixing this, I was just
> curious how far away from building on Android we were.
Any volunteers then? By doing this port, the volunteer can learn a lot
about gnulib's mechanisms.
> human.c:178: error: 'const struct lconv' has no m
Simon Josefsson wrote:
> It failed directly in accept4.c with this backtrace:
>
> depbase=`echo accept4.o | sed 's|[^/]*$|.deps/&|;s|\.o$||'`;\
> arm-linux-androideabi-gcc --sysroot
> /home/jas/android-ndk-r7/platforms/android-14/arch-arm -std=gnu99
> -DHAVE_CONFIG_H -DEXEEXT=\"\" -DEXEEXT
