Re: [Buildroot] [PATCH v2 01/17] package/python-iptables: fix _find_library()

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

On Fri,  9 Jun 2023 01:28:43 +0200
Romain Naour <romain.naour@gmail.com> wrote:

> While the commit [1] already fixed some runtime issue with
> uClibc toolchain, the same test TestPythonPy3Iptables fail
> with Glibc toolchain.
> 
>   xtables_version = 12
>   if xtables_version:
>       _searchlib = "libxtables.so.%s" % (xtables_version,)
>   else:
>       _searchlib = "xtables"
>   _lib_xtables, xtables_version = find_library(_searchlib)
> 
>   _lib_xtables and xtables_version are null with glibc
> 
> The implementation of find_library() rely on the custom
> _find_library() that try to use the IPTABLES_LIBDIR
> environment variable (that does not exist in the context
> of Buildroot).
> 
> Within the scope of buildroot we can determine what
> IPTABLES_LIBDIR should be at build time and replace the
> calls to os.environ.get('IPTABLES_LIBDIR', None) with
> the correct value.
> 
> [1] d341ec035067120b456f669a826e26eb594c0237
> 
> Signed-off-by: Romain Naour <romain.naour@gmail.com>
> ---
> I don't know why TestPythonPy3Iptables pass succesfully
> with a uClibc toolchain but not with a glibc toolchain.
> The current python code searching for xtables, modules
> and libraries seems fragile...
> ---
>  package/python-iptables/python-iptables.mk | 1 +
>  1 file changed, 1 insertion(+)

There hasn't been much feedback on the overall principle, but I found
it reasonable: we're using glibc as the default C library now, so it
kind of makes sense to use it primarily in our test cases.

So I've applied the entire series, thanks for this work! 

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