Re: [Buildroot] [PATCH 1/1] package/apr: fix musl build

[Thomas Petazzoni via buildroot <buildroot@buildroot.org>]

On Mon, 22 Jul 2024 12:48:59 +0200
Fabrice Fontaine <fontaine.fabrice@gmail.com> wrote:

> strerror_r on musl always returns an int since its addition back in 2011
> with
> https://git.musl-libc.org/cgit/musl/commit/src/string/strerror_r.c?id=0b44a0315b47dd8eced9f3b7f31580cf14bbfc01
> 
> As a result, setting ac_cv_strerror_r_rc_int to no results in the
> following build failure since bump to version 1.7.2 in commit
> 783cd8d90d37b5e0b59d6f0bfca6667855b2b9e1:
> 
> misc/unix/errorcodes.c: In function 'native_strerror':
> misc/unix/errorcodes.c:385:9: error: assignment to 'const char *' from 'int' makes pointer from integer without a cast [-Wint-conversion]
>   385 |     msg = strerror_r(statcode, buf, bufsize);
>       |         ^
> 
> Fixes: 783cd8d90d37b5e0b59d6f0bfca6667855b2b9e1
>  - http://autobuild.buildroot.org/results/9a42a4427ff64d47da61c731abb99d7585781cdd
> 
> Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>

Actually, wouldn't it make sense to modify the APR_CHECK_STRERROR_R_RC
test so that it doesn't need AC_TRY_RUN() ? I guess by building a
program that expects strerror_r() to return an int and see if it
compiles without errors (with -Werror), it should allow finding out if
strerror_r() returns an int or a char* without AC_TRY_RUN. This would
allow the test to work for cross-compilation, without having to
explicitly feed a ac_cv_strerror_r_rc_int value.

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot