Richard,
A test build of r120684 applied to gcc 4.2.0 on powerpc-apple-darwin8
doesn't show the same Process_5 failure but rather a single Process_1
failure. Interestingly all the Thread_Sleep_2 failures I normally
see disappeared. Hopefully some of the libjava developers can shed
some light on the Process_# failures (perhaps we need to backport
additional patches from gcc trunk).
Jack
ps I'll do a build on i386 and x86_64 linux to see if the Process_#
regressions are Darwin specific.
On Thu, May 17, 2007 at 10:51:58AM +0200, Richard Guenther wrote:
>
> Introducing a regression on the branch doesn't sound reasonable. Also
> are those regressions? Or is the patch referenced strictly specifc to
> Darwin?
>
> Richard.
