This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] enable gcc.dg/struct-ret-1.c on all targets

On 8/7/19 5:15 PM, Martin Sebor wrote:
> This test is reported as UNSUPPORTED when it runs on x86_64
> and I expect everywhere else except hppa-*-*.  There's nothing
> PA-RISC specific in it that I can see and it runs successfully,
> so I'm thinking I'll enable it everywhere just to get rid of
> the UNSUPPORTED result.
> Jeff, it's a test you added back in 1997.  If you can think
> of a reason not to enable it please let me know, otherwise
> I'll go ahead with it as obvious.
Actually I added it in 1995, 1997 is when we moved to a real VCS for GCC
development :-)

+Thu Jun  1 00:06:19 1995  Jeff Law  (
+       * gcc.c-torture/compile/950512-1.c: New test.
+       * gcc.c-torture/execute/strct-pack-4.c, strct-pack-5.c}: Likewise.
+       * gcc.dg/struct-ret-1.c: Likewise.

Given the date and the contents of the test I suspect it won't
necessarily work on PCC_STATIC_STRUCT_RETURN targets which we still
cared about that the time.  Structure returns in general were
problematical and I was probably just being very conservative.

One could argue that PCC_STATIC_STRUCT_RETURN should just go away.
There's only two ports using it.  One is the m68k-openbsd port which I
think has been dead for several years and could probably just be
deprecated and removed.  The other is the vax -- I have no idea how hard
it would be for them to drop PCC_STATIC_STRUCT_RETURN since I don't know
their ABI stability concerns or need for compatibility with ancient vax

So, no objections to the change.  Worst case is the vax guys would see a
failure for that test.  If they complain we can open the discussion


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]