This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: backport of Revision 120684 for Macintel into gcc 4.2.1?
Andreas,
I have posted my results for backporting r120684, 120801,
120853, 120874 and 120977 to gcc 4.2.0 onto gcc-testresults
for i686-apple-darwin8 and powerpc-apple-darwin8...
http://gcc.gnu.org/ml/gcc-testresults/2007-05/msg00822.html
http://gcc.gnu.org/ml/gcc-testresults/2007-05/msg00849.html
I only saw 25 libjava failures on i686-apple-darwin8 with
the backports (all of which already exist in gcc 4.2.0
so no regressions occur). Interestingly, in both tests
I have done with the r120684 backport, this also seems
to eliminate the Thread_Sleep_2 libjava testsuite
failures that we see on powerpc-apple-darwin8 with
stock gcc 4.2.0...
http://gcc.gnu.org/ml/gcc-testresults/2007-05/msg00756.html
Can you backport r120684, 120801, 120853, 120874, 120977
to gcc 4.2.1 branch please? Mike and Eric both think this
is a good idea. Thanks in advance.
Jack
On Thu, May 17, 2007 at 09:23:22PM +0200, Andreas Tobler wrote:
>
> I have backported r120684, 120801, 120853, 120874, 120977 and I do not
> see any failure on Process_#.
>
> I have the same failures as you have (except Process_#). The Throw_2 is
> due to missing signal handling stuff in gcc and libjava. The others I
> did not investigate. I gave up testing libjava on i?86-darwin since I
> lack a working ld64....
>
> These are the results:
>
> === libjava Summary ===
>
> # of expected passes 6962
> # of unexpected failures 26
> # of expected failures 12
> # of untested testcases 26
>
> Andreas
>