Hi Martin!

On 3/25/20 9:27 AM, Martin Maechler wrote:
> Thanks a lot, Adrian!
> 
> { When I had added  __PPC64__ it was because it "seemed clear"
>   (by the OP, I don't have access to these architectures) that the
>   clause should only be used for powerpc architectures where
>     -mlong-double-128
>   is available / in used .. and so I would have "concluded"
>   would only apply to 64-bit pppc ..
> }

As mentioned in my earlier mail, you can get free access to a POWER
machine (and many other architectures) through the GCC compile farm,
you just need to apply for an account.

> https://gcc.gnu.org/wiki/CompileFarm

> But it seems using the macro instead of the static variable is
> safer (albeit possibly somewhat slower when the code is run, but
> that will probably also depend much on the compilation
> optimizations ..).
> 
> I will change to use  __powerpc__ now   and wait some hours (on
> a possible reply from you !) before committing the change to the
> upstream sources.

I can only say that the workaround is needed on 32-bit PowerPC
as well, so it should match against "__powerpc__" and not just
"__PPC64__", the former is defined on all PowerPC targets.

Thanks for fixing the issue upstream, much appreciated.

Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaub...@debian.org
`. `'   Freie Universitaet Berlin - glaub...@physik.fu-berlin.de
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913

Reply via email to