I also think it makes perfect sense for developers to use :latest to see what breakage (like this) might be coming but we should stick to recommending :quarterly for the users. Also we'll have to consider at some point what FreeBSD's changed support policy means for us. Since they are going to EOL 11.0 a short time (one month?) after 11.1 is out, the current approach of GhostBSD (taking it's time to build a new release, add new features to the installer and such) is depending heavily on the old model of the previous release being supported for a considerable amount of time.ericbsd wrote:This why I will stay with latest until we get our own repository and even there depending how the ports are going we might stay with something more stable.
IMO we shouldn't really touch CURRENT. While TrueOS seems to do it successfully, they have far more people working on it - and they have BEs. But my vision is this: We keep to provide fixed releases for the regular users, same thing as currently (with our own package repo, though, to get more independent of FreeBSD's port option and version choices). In addition to that we create another branch, e.g. "GhostBSD rolling" which will be based on -STABLE and :latest instead of -RELEASE and :quarterly. This would be what the developers run on their main machine and we could also recommend it to more experienced FreeBSD users who either want newer software or probably require backported graphice drivers - things like that. We'd thus all the time be more close to what will eventually become 11.1, 11.2, and so on, and my hope is that this will enable us to provide PREALPHA, ALPHA, BETA and RC probably something like two weeks after FreeBSD does it for the next release. That way we wouldn't leave our users on an unsupported base for long even if problems are identified. That name is just a suggestion that I came up with right now, it's not something that I've thought about a lot. However I'd very much like to avoid the "stable" term since for people who are not that familiar with FreeBSD that is often wrongly thought to be the "more stable" release even though it's name just comes from keeping the ABI stable!
What do you think about that? What should we do after 11.0 so that 11.1 can be adopted rather quickly when we need to? One problem that I see with my suggestion is that right now that "rolling" approach would mean updating from source since freebsd-update does not support -STABLE and we certainly don't want to add that kind of thing feature. However since "packaged base" is meant to become the default in FreeBSD 11.1 (do we have any plans regarding this after 11.0 is out?) this would enable us to provide the "rolling" version of GhostBSD for those who like it.
It's a pitty that it didn't work out - but I'm glad that it could be fixed by reverting to the old way... Even though that's what Alphas are for, the currently released ALPHA1 is pretty much worthless for a lot of people (and I hope not too many downloaded it...).The problem I have with X right now it is not related to the problem from Xorg, I was trying to make something to configure card to fix a problem for user but it did cause more problem, I was not able to install mate in any of my computer, I did remove that code and let boot like it was booting in 10.3 and that fix the random problem of X and MATE for me.
Eric: We should really, really think about a way to organize things a bit to allow at least for some minimal testing internally before releasing things... I had a discussion on a German tech news site the last couple of days after they published an article about the FreeBSD foundation's latest quarterly status report. There are mainly Linux people around there but some seem to be interested in *BSD as well. The question arose if FreeBSD could actually be used as a desktop which I confirmed of course. Among the questions were things like "is it stable enough?". Of course I'd recommend GhostBSD - but after confirming that FreeBSD is extremely well engineered... Well let's just say that it wasn't cool that I had to admit that anybody interested in trying things out should NOT get the current Alpha of GhostBSD because that's currently broken ("sure, internal builds break - but why do you release it to the public then?"). I would totally appreciate if we could do something about this kind of issue (which is not even a technical one!)...