Re: [PATCH] [testsuite] [i386] require fpic for pr111497.C

2024-04-16 Thread Richard Biener
On Tue, Apr 16, 2024 at 5:43 AM Alexandre Oliva  wrote:
>
>
> Fix another test that uses -fPIC without requiring fpic support.
>
> Regstrapped on x86_64-linux-gnu.  Also tested with gcc-13 on arm-,
> aarch64-, x86- and x86_64-vxworks7r2.  Ok to install?

OK.

> PS: This is neither the first nor the last such patch.  Maybe the test
> harness could detect -fPIC et al in compile options and react
> intelligently to them, whether by warning if dg-require-effective-target
> fpic is missing, or adding it implicitly.  We could have more such
> smarts in the testsuite machinery.  WDYT?

Might be a possibility I guess.  There's other options like
-fprofile-* which have matching effective targets.

>
> for  gcc/testsuite/ChangeLog
>
> * g++.target/i386/pr111497.C: Require fpic support.
> ---
>  gcc/testsuite/g++.target/i386/pr111497.C |1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/gcc/testsuite/g++.target/i386/pr111497.C 
> b/gcc/testsuite/g++.target/i386/pr111497.C
> index a645bb95907ee..30e2e0409ad0e 100644
> --- a/gcc/testsuite/g++.target/i386/pr111497.C
> +++ b/gcc/testsuite/g++.target/i386/pr111497.C
> @@ -1,5 +1,6 @@
>  // { dg-do compile { target ia32 } }
>  // { dg-options "-march=i686 -mtune=generic -fPIC -O2 -g" }
> +// { dg-require-effective-target fpic }
>
>  class A;
>  struct B { const char *b1; int b2; };
>
> --
> Alexandre Oliva, happy hackerhttps://FSFLA.org/blogs/lxo/
>Free Software Activist   GNU Toolchain Engineer
> More tolerance and less prejudice are key for inclusion and diversity
> Excluding neuro-others for not behaving ""normal"" is *not* inclusive


[PATCH] [testsuite] [i386] require fpic for pr111497.C

2024-04-15 Thread Alexandre Oliva


Fix another test that uses -fPIC without requiring fpic support.

Regstrapped on x86_64-linux-gnu.  Also tested with gcc-13 on arm-,
aarch64-, x86- and x86_64-vxworks7r2.  Ok to install?

PS: This is neither the first nor the last such patch.  Maybe the test
harness could detect -fPIC et al in compile options and react
intelligently to them, whether by warning if dg-require-effective-target
fpic is missing, or adding it implicitly.  We could have more such
smarts in the testsuite machinery.  WDYT?


for  gcc/testsuite/ChangeLog

* g++.target/i386/pr111497.C: Require fpic support.
---
 gcc/testsuite/g++.target/i386/pr111497.C |1 +
 1 file changed, 1 insertion(+)

diff --git a/gcc/testsuite/g++.target/i386/pr111497.C 
b/gcc/testsuite/g++.target/i386/pr111497.C
index a645bb95907ee..30e2e0409ad0e 100644
--- a/gcc/testsuite/g++.target/i386/pr111497.C
+++ b/gcc/testsuite/g++.target/i386/pr111497.C
@@ -1,5 +1,6 @@
 // { dg-do compile { target ia32 } }
 // { dg-options "-march=i686 -mtune=generic -fPIC -O2 -g" }
+// { dg-require-effective-target fpic }
 
 class A;
 struct B { const char *b1; int b2; };

-- 
Alexandre Oliva, happy hackerhttps://FSFLA.org/blogs/lxo/
   Free Software Activist   GNU Toolchain Engineer
More tolerance and less prejudice are key for inclusion and diversity
Excluding neuro-others for not behaving ""normal"" is *not* inclusive