xscale-elf testsuite results

[Andrew Haley]

Rainer Orth writes:
 > Anthony Green <green@redhat.com> writes:
 > 
 > > Here's my analisys of the failures.  A number of them appear to be issues
 > > with the front-end...
 > > 
 > > FAIL: final_initialization_in_ctor compilation from source
 > > FAIL: final_initialization_in_ctor -O compilation from source
 > > 
 > > I have no idea what's going on here:
 > 
 > same here: I'm seeing the same problem when using -fcheck-references on
 > i386-pc-solaris2.8 and sparc-sun-solaris2.8 (breaks bootstrap):
 > 
 > 	http://gcc.gnu.org/ml/java/2002-03/msg00493.html

I will investigate this.

 > Given that the use of -fcheck-references affects java only, while the other
 > suggested solution (configuring with --enable-sjlj-exceptions) affects EH
 > performance across all languages, I'd prefer to use the former for now.

Okay, but at the risk of becoming boring I'd like to point out that
(at least for the time being) this isn't a real solution because SEGVs
will still happen in CNI code and not be trapped.

Andrew.