FreeBSD 11.1 and FreeBSD 10.4 are Release is comming.

[ericbsd]

Hi everyone

I would like to know what is your thought on skipping 11.0 and going for 11.1 instead . In the mean time I could release a 10.3 SE in a week or so to give some security update to users, that proposed by ASX.

Also what is your thought on having 10.4 released I am sure that people are happy with 10 and also if i am not mistaken 10.4 long term release.

[kraileth]

I was also going to propose skipping 11.0 and going to 11.1 directly since that version is already in beta and the release should be available in about 1.5 months.

I've updated my main GhostBSD workstation to 11.1-BETA1 a week ago. Then I deleted all cached data and the package repo and used Synth to rebuild all packages (and did a forced reinstall of everything). So far I've noticed only one problem - graphic glitches. I'm not sure what they are related to since I'm only experiencing them in certain applications (watching videos in Firefox, FreeBSD booting in VirtualBox, ...). Also I had problems related to VirtualBox in general - sometimes using VirtualBox results in an unexpected sudden reboot of the host machine...

I'll try this out on another PC, soon.

When it comes to 10.x I don't have a strong opinion. I kind of would prefer to move on, especially since we're a desktop distribution and benefit greatly from additional hardware drivers and other features in newer FreeBSD versions. But to be honest, 10.3 has been an excellent experience for me, too, and probably some people prefer sticking with it. With our own packages in place, we might really consider putting out a 10.3 SE with our own packages. However it might also be a good time to introduce a new naming scheme. What about basing the new release on 10-STABLE instead? We would only need a name that suggests that it's bigger than 10.3. Something like "GhostBSD 10X" or "GhostBSD 10+" or even "GhostBSD 10-LTS" perhaps? Just thinking aloud.

[ASX]

I agree tos kip 11.0 and go for 11.1.

About 10.x, I would like to have a 10.4 release, that mainly will make use of our own repos, at this point I would skip enirely the idea of 10.3 SE.

[kraileth]

On second thought, I don't think that 10-STABLE is a good idea. One of the prerequisites of going with -STABLE is packaged base - otherwise we'd have to look into generating the diff patches for freebsd-update to use (and we certainly don't want to do that).

The problem that I see with 10.4 is that it is about 3.5 month away. That's a long time! The idea of putting out a "10.3 SE" or something before too long so that we can then concentrate on making 11.1 work as good as possible does make some sense to me. That "10.3 SE" could be a test bed for our own packages (and that would be the big new feature for this "SE"). Otherwise I don't think there's a lot of gain for another 10 release once we have 11.1 out.

If we do decide to brush off the dust of 10.3 and make a second release it would buy us some time to go to 11.1 and perhaps include another important feature: Switching to packaged base. TrueOS already does this and while it has been in 11.0 as a tech preview, it should be fully supported on 11.1 and becoming the new standard way of updating FreeBSD. We will have to adapt it anyway at some point and preferably not between the releases in one release cycle. So it'd either be 11 or 12; and since 11 has been delayed until now it might be a good idea to make the best out of the situation and jump the chance to do it now. That is if we agree upon some kind of "10 reloaded", because otherwise it would likely just take too much more time before the overdue next release.

There's also another thing that we should talk about IMO. We're obviously dependent on FreeBSD as the upstream base for our desktop OS. Starting with 11.0, FreeBSD dramatically changed the release policy, narrowing down the time the previous point release is supported after a new one is released. While point releases were really important for the release numbers in 10.x, I feel that they are much less important with 11.x. Previously 10.1 was an extended release and you could have stayed on it a while after 10.2 and even 10.3 was released, you're now expected to upgrade your 11.0 boxes to 11.1 within three month after the new point release is out. So my question is: Does it really make a lot of sense to claim "My version is 11.0" or "I'm using 11.1"? Technically speaking, yes, of course. But for the end user? Probably not. If that user runs GhostBSD and updates it properly, saying "I'm on 11" is just fine. We know when we released the update bringing the system up to the latest point release after all!

So my suggestion that i put up for discussion is to drop the point release number in the future GhostBSD version scheme. We are not going to support any 11.x for a longer time (due to FreeBSD's new release model). Why not reflect that change by calling the coming version just "GhostBSD 11"? This would also fit well into the current trend of simplifying the version numbering in a lot of projects lately. Or since Eric also considered additional -STABLE releases, we could have "GhostBSD 11R" and "GhostBSD 11S" or something. There are probably better names for the two branches, but this is just a conceptual idea without carving naming details in stone.

[ASX]

On second thought, I don't think that 10-STABLE is a good idea. One of the prerequisites of going with -STABLE is packaged base - otherwise we'd have to look into generating the diff patches for freebsd-update to use (and we certainly don't want to do that).

The idea to use xx-STABLE is about speedup development only, and therefore there will be no ISO for public use, only for developers (and eventually for testers). By testing on xx-STABLE we should be able to release practically in sync with FreeBSD, or at most a few days later.

The problem that I see with 10.4 is that it is about 3.5 month away. That's a long time! The idea of putting out a "10.3 SE" or something before too long so that we can then concentrate on making 11.1 work as good as possible does make some sense to me. That "10.3 SE" could be a test bed for our own packages (and that would be the big new feature for this "SE"). Otherwise I don't think there's a lot of gain for another 10 release once we have 11.1 out.

What we gain is that we can make happy people who are happy with 10.x, and don't want to have problems intrduced with 11.x (think at the graphics glitches you are experiencing).

If we do decide to brush off the dust of 10.3 and make a second release it would buy us some time to go to 11.1 and perhaps include another important feature: Switching to packaged base. TrueOS already does this and while it has been in 11.0 as a tech preview, it should be fully supported on 11.1 and becoming the new standard way of updating FreeBSD. We will have to adapt it anyway at some point and preferably not between the releases in one release cycle. So it'd either be 11 or 12; and since 11 has been delayed until now it might be a good idea to make the best out of the situation and jump the chance to do it now. That is if we agree upon some kind of "10 reloaded", because otherwise it would likely just take too much more time before the overdue next release.

I would not introduce any "new feature" at this point in time, only focus to make things works, else we are going to introduce further delays.

There's also another thing that we should talk about IMO. We're obviously dependent on FreeBSD as the upstream base for our desktop OS. Starting with 11.0, FreeBSD dramatically changed the release policy, narrowing down the time the previous point release is supported after a new one is released. While point releases were really important for the release numbers in 10.x, I feel that they are much less important with 11.x. Previously 10.1 was an extended release and you could have stayed on it a while after 10.2 and even 10.3 was released, you're now expected to upgrade your 11.0 boxes to 11.1 within three month after the new point release is out. So my question is: Does it really make a lot of sense to claim "My version is 11.0" or "I'm using 11.1"? Technically speaking, yes, of course. But for the end user? Probably not. If that user runs GhostBSD and updates it properly, saying "I'm on 11" is just fine. We know when we released the update bringing the system up to the latest point release after all!

Disagree here, we need the "point" specification, it is a requirement for our "ghostbsd-build".
Eventually we have to work for a smoth upgrade to the next point release, i.e. 10.3 -> 10.4, and I suspect it will be easier to do that using the packaged base, so lets plan this for 11.2 or 12.0

So my suggestion that i put up for discussion is to drop the point release number in the future GhostBSD version scheme. We are not going to support any 11.x for a longer time (due to FreeBSD's new release model). Why not reflect that change by calling the coming version just "GhostBSD 11"? This would also fit well into the current trend of simplifying the version numbering in a lot of projects lately. Or since Eric also considered additional -STABLE releases, we could have "GhostBSD 11R" and "GhostBSD 11S" or something. There are probably better names for the two branches, but this is just a conceptual idea without carving naming details in stone.

we can support 10.x or 11.x for all time it will be supported upstream, without the need to drop the "point" release attribute.

I think we agree around the main concept (long term support for 10.x or 11.x), but we have different opinion about naming convention.

[ericbsd]

Here I agree with ASX no feature would be added in 11.1 and 10.4, because there is a lot of issue that need to be fix before going further.

A fix that I will like to see is ZFS with the boot loader, that we can finally install ZFS with out grub, but this is a fix that need to be done in pc-sysinstall.

[ericbsd]

I have not been much active on the development side this week. Hopefully I will manage to finish all setting for mate version this weekend and get back to XFCE.

I did update the download page and remove the download for 11.0 and 10.1 and add link that people can go to old version.

[NevilleGoddard]

Personally, I agree to skipping 11.0 and go to 11.1 and 10.4. By the time 11 is ready it will be out of date.
If it is not too much work, it would be better to go straight to 11.1. Same with GhostBSD 10 and going to 10.4.