On Thu, 2015-10-29 at 09:38 +0100, Arne Babenhauserheide wrote: > Am Donnerstag, 29. Oktober 2015, 08:40:50 schrieb Arne > Babenhauserheide: > > > I'm not sure I understand why you'd get a vote on what will > > > happen. > > > The part I’m talking about is the not-changing if not needed. > > Also I don’t see why I have to get a vote to voice an opinion. >
This is precisely why: you've no idea of what you're talking about. The "project" of splitting freenet-ext.jar up has been ongoing for over half a decade now... yet you don't seem to be aware of it. > - Who writes the guide for plugin authors on bundling db4o? Why should we? The only guide that needs writing is the one that says that from build X onwards, plugins will get a virgin classloader. Plugin authors should *never* rely on what Fred depends on. There's a separate (now stalled) project of providing ways for plugin developers to handle dependencies and "share/export" libraries. There's even a completely different approach being considered (sace's usgi). > - Who ensures that the plugins will get updated? The plugin authors and the versioning system > - Who releases new versions of plugins which bundle db4o? > The plugin authors (isn't that obvious?) > Breakage like purge-db4o should not happen again. > I'm not sure of which breakage you're talking about. There was no breakage, just a poorly documented plugin API change. We haven't solved the problem there yet, meaning that next plugin API change will also break stuff. > And in general, if you feel that only those who work on code get a > vote, then please broaden your view. There are a lot of people > creating value for Freenet without writing code. An answer like > yours > essentially tells all of them that they are not wanted here. > This is a development mailing list (it's in the name!). Yes, I do think that we can't afford making decision by consensus if the said consensus involves non-active developers. There's very little we can do about $pluginDeveloper relying on a freenet dependency that we won't be providing going forward. Keeping on bundling/shipping the said library is unilaterally recognized as something not worth the effort (except by you). I'm not sure which part of that you don't understand. > As long as Freenet has only around 10k people it can be technically > perfect all it wants without providing real freedom of speech — > except > for the few people who never raised their opinions in public at all > (or rather: never even write in public). So what we need is adoption. > > And too little content along with guides which no longer work because > we changed something they described are at least as important for > adoption of Freenet as updating dependencies (or any other > medium-sized change). Not to forget all the people who work on > promoting Freenet. > If you have a specific concern voice it... what you've done is not that. What you've done is state some general ideological principle "don't break stuff, maintain backward compatibility forever!" without measuring the cost associated with it nor producing a rational that can be argued against. Start by the beginning, how many plugins will that change break? I've told you that neither the maintainers, nor the packagers nor the plugin authors are happy with the current situation -> that's why we need change. If you have a rational argument that indicates otherwise I'm all hears... but I won't let you derail the 5y+ process based on some vague non-technical, irrelevant principles. It's great that we currently have people interested in working on that (complex) project; I won't let you derail it. Florent PS: ask xor what he thinks; odds are he wants to update db4o but he can't since we don't provide any way for him to get a virgin classloader he could load the updated db4o code from
signature.asc
Description: This is a digitally signed message part
_______________________________________________ Devl mailing list Devl@freenetproject.org https://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl