Re: commons-collections in main
Arnaud Vandyck wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> Nathanael Nerode <firstname.lastname@example.org> writes:
>> Arnaud Vandyck wrote:
>>> Thanks, note that with this package in main, I'm able to move two other
>>> java packages to main... and with these two, more (and more, and
>>> more...) ;-)
>> May I ask how you're dealing with the architecture coordination problem?
>> Otherwise all the java packages are probably going to have lots of
>> getting into 'testing'. :-(
> I now focus on my packages and if you see I'm an uploader of kaffe, it's
> a mistake. Some said I tried to hi-jack the package and after that, I
> did not touch the package anymore. There is a lot of things to do with
> my package and with the project I did propose: 'moving java to main'
> Also, I'm working on GNUJAXP (I'm on of the upstream maintainer) to get
> transformation (xslt). I can now say that it's finished. So the next
> thing is to have a working javadoc alternative (gjdoc) fully in main (I
> think it will be done before the end of the next week).
> I'm not telling you the story of my life, but I also have some other
> plans for java, free java and free java in debian. The main thing I can
> tell is that I'm thinking about integrating native compiled java
> programs in Debian.
I approve! :-)
> AFAIK, Only RedHat/Fedora do this (but they are at
> the source! A lot of RedHat devs are working on gcj). If we can do this,
> we'll be one of the few operating system (not just linux distribution!)
> to be able to ship java applications/libraries compiled to native!
I work on GCC upstream -- if I can help with getting necessary GCJ patches
from CVS HEAD, 3.4, or elsewhere, into 3.3 branch, please tell me. :-)
>> Kaffe currently builds only for i386, mips, mipsel, powerpc (but is still
>> uploaded as an 'any' package). It would probably be better to upload it
>> as a restricted-architecture package for the time being, so that new
>> could get into testing. Ean Schuessler has mysteriously been unwilling
>> to do this.
<snip horror story>
> The best thing with all the whistles is that now, doogie takes
> care of kaffe! Thanks to him.
:-O Well, that's good anyway. It looks like currently SableVM and GCJ
don't run on mips/mipsel, so we may need to get kaffe for mips/mipsel into
'testing' before some Java packages can get into 'testing'. :-/
>> gij-3.3 works (to some extent) for alpha, arm, hppa, i386, ia64, m68k,
>> powerpc, s390, sparc (not mips, mipsel).
> Yep, also, note that I sent some bugs to sablevm (on irc or on the list)
> to improve it, and SableVM is still in testing! So we have two others
> free alternative to kaffe. Maybe it's a way to have the java packages in
>> If you want to get stuff into 'testing' which depends on one or the other
>> (only), I hope you have a plan for it. (It seems that any package which
>> works on *both* kaffe and gij is currently OK for all architectures,
>> which is nice.)
> The three free VM's (sablevm, kaffe, gij) are based on classpath. So
> something that works with one of the VM _should_ work on others (there
> are some differences but the vast majority should be fine).
> Thanks for pointing me that. I'll set up free VM's alternatives to my
> packages, but I also have to talk to Stefan Gybas about the debian-java
> policy on the java libraries.
So the plan is that the Java packages will depend on the appropriate list of
alternative free VMs, so that they'll be available for all architectures
even when no individual VM is? That sounds good. :-)
> Nathanael, I also wanna thank you for your work on the RC-bugs and
> reviewing the packages and bugs. Please continue doing this job, we
> really need it and you do it fine.
Aw shucks. ;-) Thank you!
Make sure your vote will count.