[committed] Change default for enable_hash_synchronization_default on PA-RISC

[Michael Koch]

On Tue, Jun 13, 2006 at 09:51:55PM -0400, John David Anglin wrote:
> I've noticed that we get occassional failures of process intensive
> tests in the libjava testsuite (e.g., Process_3).  Carlos O'Donell
> indicated that hashing locks might be unsafe because of the special
> way locks are handled in parisc-linux.  I've also seen a few failures
> in the libjava suite on hpux.  So, I've decided that it would be
> better to be conservative in this regard.
> 
> With the patch installed, I've seen no libjava testsuite failures
> on hppa-unknown-linux-gnu on the trunk in several runs.
> 
> Committed to trunk.  If the patch works out, I will probably backport.
> 
> Dave
> -- 
> J. David Anglin                                  dave.anglin@nrc-cnrc.gc.ca
> National Research Council of Canada              (613) 990-0752 (FAX: 952-6602)
> 
> 2006-06-13  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
> 
> 	* configure.host (hppa*-*): Set 'enable_hash_synchronization_default' to
> 	no.
> 
> Index: configure.host
> ===================================================================
> --- configure.host	(revision 114624)
> +++ configure.host	(working copy)
> @@ -130,7 +130,7 @@
>    hppa*-*)
>  	sysdeps_dir=pa
>  	libgcj_interpreter=yes
> -	enable_hash_synchronization_default=yes
> +	enable_hash_synchronization_default=no
>  	;;
>    powerpc64*-*)
>  	sysdeps_dir=powerpc
> 

I would really welcome a backport tp 4.1 branch. This would help me a
lot to make GCJ reliable on Debian hppa.


Cheers,
Michael
-- 
http://www.worldforge.org/