Things to do for 1.1.7 (Was: Re: [kaffe] CVS kaffe (robilad): Merged in zlib)

[Dalibor Topic]

Kaffe CVS wrote:
> PatchSet 7063 
> Date: 2005/12/28 20:43:48
> Author: robilad
> Branch: HEAD
> Tag: (none) 
> Log:
> Merged in zlib
> 
> 2005-12-28 Dalibor Topic  <robilad at kaffe.org>
> 
>         * libraries/clib/zlib/ : Merged in zlib 1.2.3 from gcc.
> 
>         * THIRDPARTY: Added zlib.
> 
>         * configure.ac: (--with-internal-zlib) New option.
> 
>         * libraries/clib/zip/Makefile.am: Adapted to allow use
>         of internal zlib.
> 
>         * libraries/clib/Makefile.am: Use internal zlib if selected.

this is a first step towards merging in fastjar to replace kaffe's jar,
and remove the dependency on zip for class library builds. A nice small
project for a volunteer could be to rip out kaffe/kaffevm/inflate.[ch]
and rewrite code using them to use zlib directly.

I may still have to reshuffle the code around a bit, and to change the
configure option's name to make integration with gcc's code simpler, but
that's the plan. This should be the last 'feature' before 1.1.7.

It became necessary because of Debian builds of subversion breaking on
Kaffe 1.1.6's pure java JAR tool on those platforms where Kaffe is not
running that well yet :/ Given that fastjar is maintained by gcc
developers, there is a higher chance we'll get its bugs flushed out and
fixed on various platforms together, then with a Kaffe-specific jar tool.

Further remaining changes on my list for 1.1.7 are
* moving ahead various configure options from gnu classpath into kaffe's
configure.ac to make sure oour users also see them

Other things (volunteers welcome)
* improving win32/cygwin docs
* looking at various swing related bug reports in debian/on list
* fixing the test -ef bug in gen-classlist.sh
* merging in libffi for libffi supported platforms

I'd like to have a release candidate by the end of week, if I can make
the fastjar merge happen till then. The latest locking fixes have
improved the things a lot for me on amd64-linux/smp, and the test
results I got this week on x86-freebsd5.4, x86-linux, amd64-linux and
powerpc-linux with the various available configurations
(intrp/jit/jit3)*(jthread/pthread) have been very good (no failures).

cheers,
dalibor topic