Ulrich Gemkow wrote: > is there any chance that also the stable svn-branch(es) are exported to > the public svn server? If not - could you give some reason for this > decision?
This is a 100% inofficial answer containing my personal opinion (and some personal experience from the last year, after going open source): this suggestion would mean that people would use even more different versions. More precisely far too many users would treat the "stable" branch like releases - as it is right now almost no one bothers to mention which virtualbox version they're using - which is especially annoying if they use some random self-compiled weeks old revision of the development code (or even some incorrectly packaged version done by some unknown third party). And this will just get worse by having more code floating around. Adding time-bombs to the source code like some other projects do isn't helpful either, as it encourages people to remove it (just like some Linux distributions remove the registration dialog in the packages). So exporting a stable branch of the source code doesn't improve the current situation, it'll make it worse. From our point of view there is no such thing as a completely stable branch, just the code for releases (which we provide as .tar.gz). I agree that things need to be tested, but testing doesn't just happen automatically, it's hard work. It would be great if more people were testing VirtualBox of course - but in reality the exact opposite would happen: users would run some not yet finished and tested code assuming it's a release, and spread the impression of quality they got. Also there is a technical reason - it needs more work from our side to maintain the public code repository. It becomes more difficult to sensure that no code we're not allowed to publish leaks into the repository. Sorry if this sounds negative. The overall situation is very positive, though. VirtualBox has seen great contributions from the community, and the ever growing stream of feedback (bug reports, feature requests, usability) has helped to shape VirtualBox releases. So in a sense the biggest positive effects that can be expected from the community are already in effect - and continue to improve VirtualBox. > Access also to the stable branches would help testing and would probably > give more bug reports: When using VBox in production systems, using the > trunk is not possible. So bugs and problems will often reported only > _after_ a release from the stable branch. In an ideal world this would be true. But from my personal view this can't be effective if the code would be just dropped, as I explained above. I can imagine a slightly more controlled environment where testing is the main focus. "Controlled" doesn't mean that there should be a high barrier - in fact it should be kept relatively low. Also providing sources is probably not ideal either, as it requires skills and tools (and some code e.g. in the Windows additions area simply cannot be released). What would be easier to achieve is to provide e.g. nightly builds (which I mentioned above) to "beta testers". These are just my personal thoughts, and shouldn't been taken as a definitive answer on this topic. This is still open for discussion, and that's the reason why I haven't explicitly put labels such as "this won't ever happen" or "this will likely happen" to the suggestions, both yours and mine. And I promise not to write any more such lengthy mails in the discussion. -- Dr. Klaus Espenlaub innotek GmbH, http://www.innotek.de _______________________________________________ vbox-dev mailing list [email protected] http://vbox.innotek.de/mailman/listinfo/vbox-dev
