Re: [PROPOSAL] New Virtual Packages and way to handle Classpath
Hallo Daniel,
* Daniel Bonniot wrote:
>I see two options to solve this:
>1) Strict interpretation of javaN-runtime
>Indeed, probably no free JVM will be able to provide them at this point.
>However, it is still possible for packagers to make sure at least one
>free JVM works for their package, and depend on, say, kaffe | javaN-runtime
>2) Looser interpretation of javaN-runtime
>JVMs would be allowed to provide them, as long as they include some/most
>of the runtime. Then bug reports can be opened when a specific feature
>does not work. This bugs can be forwarded upstream to provide feedback
>about where work is most urgently needed. People willing to help free
>Java in Debian can also use these bugs reports to find tasks, and submit
>patches.
IMO, the second should be enough (in the first, kaffe wouln't allow to
setup /usr/bin/java-* alternatives). The problem is, where to draw the
line and say, that it not anymore qualify as a java2-runtime-*? We
want to have as many packages in main as possible, but I don't really
want to be kaffe|.. maintainer, with all the bugreports saying 'what a
mess, it does not work'.
Also, if a packager knows, that his package will not work with a free
VM, there should be a way to make this happen as well.
Debian has a reputation, that you can install a package 'and it
works'. Currently, this isn't so with java packages (you need at least
a BD mirror and some magic in the start-script). And I think we
will never get as far that this is true with all packages :(
Jan
--
Jan Schulz jasc@gmx.net
"Wer nicht fragt, bleibt dumm."
Reply to: