[PATCH v2] Add TEST_CC and TEST_CXX support

[H.J. Lu]

On Fri, Dec 6, 2024 at 5:41 PM Sam James <sam@gentoo.org> wrote:
>
> "H.J. Lu" <hjl.tools@gmail.com> writes:
>
> > On Fri, Dec 6, 2024 at 7:35 AM H.J. Lu <hjl.tools@gmail.com> wrote:
> >>
> >> Support testing glibc build with a different C compiler or a different
> >> C++ compiler with
> >>
> >> $ ../glibc-VERSION/configure TEST_CC="gcc-6.4.1" TEST_CXX="gi++-6.4.1"
>
> gi++ -> g++

Fixed.

>
> >>
> >> 1. Also check compiler options with TEST_CC and TEST_CXX.
> >> 2. Add check and xcheck targets to Makefile.in and override build compiler
> >> options with ones from TEST_CC and TEST_CXX.
> >>
> >
> > Here is the v2 patch.  Tested with GCC 14.2.1 build and
> > GCC 11.2.1 test on x86-64.  I got
> >
> > FAIL: math/test-float64x-fpclassify
> > FAIL: math/test-float64x-isinf
> > FAIL: math/test-float64x-isnan
> > FAIL: math/test-ldouble-fpclassify
> > FAIL: math/test-ldouble-isinf
> > FAIL: math/test-ldouble-isnan
> >
> > More changes in configure.ac under sysdeps are
> > needed to support older GCCs.
>
> Please fix '==' use in configure.ac; configure is run with #!/bin/sh and
> hence must be POSIX compatible (i.e. change it to '=', or better, use
> AS_IF).

Fixed.

>
> Ideally please quote arguments to the macros too like AC_CHECK_PROGS.

I thought all necessary macros are quoted.   Which ones did I miss?

> I like the idea.

Thanks.

I didn't get these failures:

FAIL: math/test-float64x-fpclassify
FAIL: math/test-float64x-isinf
FAIL: math/test-float64x-isnan
FAIL: math/test-ldouble-fpclassify
FAIL: math/test-ldouble-isinf
FAIL: math/test-ldouble-isnan

when using GCC 11.2.1 to build and test.   I will investigate.

-- 
H.J.