RE: Patch: Boehm GC 6.6 merge

["Boehm, Hans" <hans dot boehm at hp dot com>]

I ended up merging the attached, much smaller, patch into the upstream
sources.  Many of the remaining differences seemed dependent on the gcc
configuration mechanism, which is not currently shared.  The Darwin
handling in gcconfig.h was replaced by another patch in the meantime.  I
couldn't quickly make sense out of the MIPS ABI test issue, and it
wasn't clear to me that the patch was safe outside a gcc context.  I
also wasn't comfortable with the switch from assembly code to
intrinsics, since the upstream code has to handle old gcc versions.

But this should get rid of the completely spurious differences.

Hans

> -----Original Message-----
> From: java-patches-owner@gcc.gnu.org 
> [mailto:java-patches-owner@gcc.gnu.org] On Behalf Of Bryce McKinlay
> Sent: Tuesday, January 24, 2006 6:59 PM
> To: java-patches@gcc.gnu.org
> Cc: Boehm, Hans
> Subject: Patch: Boehm GC 6.6 merge
> 
> 
> This patch merges the GC 6.6 sources into the libgcj trunk. 
> Two patches 
> follow: the first patch is what is being applied against the current 
> code in trunk. The second patch is a diff against the clean, 
> upstream GC 
> 6.6 distribution showing what changes remain in libgcj's version 
> (configury changes have been omitted).
> 
> The most obvious changes I noticed which are in gcc/boehm-gc 
> but not the 
> upstream GC 6.6 sources are:
> 
> IA64 locking changes
> Changes for PPC64 support - see 
> http://gcc.gnu.org/ml/gcc-patches/2005-03/msg02050.html
> 
> No testsuite regressions. I'm checking this in.
> 
> Bryce
> 
>

Attachment: gcc_patch
Description: gcc_patch