ruby22 buillds with gcc38 on powerpc64

[Mark Millard]

On Thu May 5 18:53:18 UTC 2016 regis.etourmy at free.fr wrote:

> Sorry, I meant gcc48, not 38....
> 
> ----- Mail original -----
> De: "regis etourmy" <
> regis.etourmy at free.fr
> >
> À: 
> ruby at FreeBSD.org
> 
> Cc: 
> freebsd-ports at freebsd.org
> 
> Envoyé: Jeudi 5 Mai 2016 20:28:52
> Objet: ruby22 buillds with gcc38 on powerpc64
> 
> Hi,
> 
> I tried to build ruby22 with gcc38 on my powermac G5 and it built (it didn't with clang). I have not tested yet if it runs well.
> 
> Thanks for all your work,
> 
> Régis


Relative to clang 3.8.0 (for example). . .

https://llvm.org/bugs/show_bug.cgi?id=25780 [a meta-list of powerpc64 and powerpc clang reports that block use by freebsd world and kernel] lists various powerpc and powerpc64 code-generation issues for clang 3.8.0 vs. FreeBSD. As I remember each of the following includes examples of powerpc64 code-generation issues, not just powerpc (non-64) ones.

https://llvm.org/bugs/show_bug.cgi?id=26970
https://llvm.org/bugs/show_bug.cgi?id=26856
https://llvm.org/bugs/show_bug.cgi?id=26844
https://llvm.org/bugs/show_bug.cgi?id=26761

One of the types of things that is broken on powerpc64 and powerpc is C++ exception handling --and such is broken by bad code generation, independent of involved support libraries possibly adding even more issues. For powerpc (non-64) there are also stack-handling ABI violations that can, for example, mess up the stack contents when signal delivery happens.

As far as I can tell depending on clang/clang++ for powerpc64 or for powerpc is risky or requires analysis that things are actually working for all the specific uses being made. But so far as I can tell clang 3.8.0 is an improvement over prior clang vintages for powerpc64 and for powerpc.

===
Mark Millard
markmi at dsl-only.net