Done with Gcj? Send us a link or a story!

Tue Jun 27 22:41:00 GMT 2000

Matt Welsh writes:
> Yeah - I'm done with GCJ - it's too unstable for me to use with complex 
> multithreaded applications.
>> Although I suspect this was not the meaning you had intended :-)

No. Good point.
> I spent a LOT of time trying to nail down a stable version of the
> compiler and runtime that had the features I wanted *without* all of
> the bugs. The multiplicity of versions, patches, and
> incompatibilities were quite a nightmare to untangle. Why is GCJ
> still relying upon separate patches (from folks like Bryce) to keep
> the new runtime working with the old compilers? It's really a mess.

Yes, it's still a work in progress. The recent 1.1 compiler extensions
weren't (and still aren't) right at the first commit. And it's not
like the rest of gcc is changing its infrastructure with cautious
considerations for the Java bits. And the runtime also had upgrades on
its own with the classpath merge. I've been really excited to see all
these bug reports coming in but sad not being able to spend as much
time as needed in order to fix them.
The future should bring us GC in the front-end, which will probably
bring its own problems. There are more inherited structural things
that I'd like to work on (global variables, qualified expressions
resolutions, etc...) and more optimization could be added (inlining,
etc...)
We do our best to keep things stable. Gcj is sometimes unstable
because it's moving it forward, it's unfortunate it's not always
necessarily in a direction that matters to you or that the price to
pay to move along is too high.
./A