--- Ben Burton <[EMAIL PROTECTED]> wrote: > > > >I'm quite happy with this suggestion as well (must -> may for non-free > > >JVM dependencies). If at least one of the dependencies is satisfied > > >- even if this is selected from a list of only free JVMs - then the > > >app will presumably run successfully and so there's no problem if > > >non-free JVMs are omitted from this dependency list. > > > > Sorry, I'm not doing this. If a package can be used with unfree VM, > > then this package 'should' also include this VM in the search path. > > > > If you talk about the 'choice of the user', then there should also be > > a choice to run on a unfree VM. Debian is not about *forcing* the user > > to use free software. > > Nobody's forcing the user to do anything. I'm presuming that the app > startup scripts will allow a user to specify their own JVM with which to > run an app (and I wouldn't mind putting *that* requirement in policy). > > By doing this, users can run an app under whatever JVM they like. It's > just a matter of the dependency system requiring them to install one of > the "officially supported" JVMs, whether or not they choose to use it in > the end. > > I honestly do not think you're going to be able to force package > maintainers to install non-free software on their systems for testing > their DFSG-free apps. Some of these non-free packages have > eyebrow-raising clauses in their licenses, and I won't blame some > maintainers for refusing to test with them.
And I don't think a policy should force maintainers to claim their packages will work with some non-free software without being able to test it. It goes against Jan's goal of improving the quality of java experience in debian if the maintainers are forced to claim their packages run on something they can't test. I think it's a matter of who you want to force to do what: Jan's proposal forces maintainers to either install non-free software to test their packages with it or make dishonest claims about the java environments their package can run on in order to support users who want to run the packages on non-free software. We agree that some people may want to do that. He wants to make a requirement that packagers support those people, I don't. My proposal, to explicitely depend on the VMs that the maintainer can honestly say the package runs on, forces the user to download one of the environments that the maintainer says the software will run on, if none such environment is installed. This goes a long way to make sure the user doesn't experience problems running the software. It does not prohibit the maintainer from also depending on non-free VMs if he wants to, but it doesn't force him either. The missing link is some magical 'byte code runtime lookup' script, that allows the user some flexibility in the choice of runtimes supported by the maintainer, but doesn't prevent him/her from trying to run the software with another, not officially supported by the maintainer, runtime if he wishes so. The maintainer should not be required to support all java runtimes in debian, but encouraged by the policy to test his package on as many runtimes as possible, with a priority on the free ones, as the user is more likely to have those installed on his system (they are either part of main, or will enter it after testing, after all). That's where Jan's and my ideas again are not very far apart: he already tried to specify something similar in his ant environment proposal, to provide some flexibility for java runtimes capable of runing ant. Philosophical differences aside, I don't think we are that far apart from each other. I believe the discussion on debian-java has been quite fruitful so far in exposing some weaknesses of the proposals and paving the road to improvements (and Jan has done a great job on improving the policy proposal when the arguments from the other side were convincing enough for him). As I said before, it may take a little bit longer, but I think it's worth it. cheers, dalibor topic __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
