Is there some sort of shakeup going on?

OH said:
FreeBSD 8.3 was barely released as preparations for 9.1 got underway (as a result of the shortened release cycle that has us supporting two production releases?) and at the same time we're introducing pkgng, optionsng, new Makefile headers, replacing CVS with SVN, not to mention the ongoing work on clang (and to a lesser extent ZFS).

Correct but that is always the case. There are always 2 production branches that need to be supported. At least that is the case since FreeBSD 6.0-RELEASE.

Pkgng. A very nice and promising tool. Originally it was expected to be part of FreeBSD 9.1-RELEASE. This has changed and there are some thoughts for either 9.2 or 10.0. Why?

The switch from CVS to SVN. The idea is not bad. However, timing is everything. There was absolutely no reason to force this without any planning. Look at the result. Nobody knows what will happen for the 8 & 9 branches yet. Even worse, people who used to update their sources using csup() were not able to do that anymore for 9.1-beta and 9.1-RC1 but they can use it for 9.1-RC2!!!

Clang. It didn't work for 9.0-RELEASE and now it is planned for 10.0-RELEASE Again, why? And I really don't want to hear anymore about the licensing issues. At least not until 2014 when 10 will be released!

Things are really developing fast. Look at the competition:

Virtualization... Yes, there is a on going project for a native hypervisor expected for FreeBSD 10. blah, blah
High Availability... HAST is here but I don't see any rapid evolvement. Where is the asynchronous replication?
IPv6... Why on earth do I still have to compile IPSEC in the kernel?
The list can go on...

OH said:
For a project, as wblock@ correctly states, handled mainly by volunteers, that seems like quite the extra workload for what I assume to be a virtually equal amount of developers / committers.

You are right. That is why I would never blame developers or commiters. It is somebody else's job to support them and support the project as well.

OH said:
Just before submitting I see gkontos' reply. I just have to add that I'm not bitter (but then again I haven't tried setting up LT2P VPN with radius either, net/mpd5 FTW).
This is just what I'm seeing, I don't want to put words in anyone's mouth, so I'll stop right here and get myself some proper whisky.

I think that I am more frustrated than bitter. And that is because I love FreeBSD and I would really hate to see all this effort going wasted.

Btw. My comments are not personal, just the result of pure frustration!
 
Niatross said:
He can't leave. He has to go through me before he leaves.

Besides all that, I just learned portmaster and don't want to have to re-learn somebody else's utility.

Get him back over here so I can reprimand him.

portupgrade isn't that bad( I used it for about 4 years before switching to portmaster ). IIRC it's switches are the same as pormaster. The big deal about portmaster was that it's built on a pure posix shell so no dependencies. It's also quicker than portupgrade for the the reason that it relies on ash and not ruby+berkelydb.

Either way whether it's pure ports; portmaster; portupgrade, or pkgng it's good to have choices. I'm glad to hear that portmaster was picked up.

Also if Doug is reading this; I also want to thank you for the project.
 
gkontos said:
Journal UFS2. It was finally released in FreeBSD 9.0-RELEASE. But wait, there is a bug and you can not backup a UFS2+J fs!!!

FWIW
[I use as a filesystem and backup...]

SUJ. I cannot say enough good things about it. rsync... --bwlimit=1000 (if the paths are correct and pretested and working.) Many times it has saved data here, and/or
speeded up
Code:
 fsck_ffs -y /dev/...
(Sometimes without the -y if a SUJ fsck freezes).


Agree with some of the rest of the post but I've posted it elsewhere...
 
gkontos said:
Correct but that is always the case. There are always 2 production branches that need to be supported. At least that is the case since FreeBSD 6.0-RELEASE.

I'll quickly explain my point with the use of the wayback machine:
And now both 9.1 and 8.3 are production releases and 7.4 remains the legacy release
 
OH said:
I'll quickly explain my point with the use of the wayback machine:

And now both 9.1 and 8.3 are production releases and 7.4 remains the legacy release

If you go back you will see that when 8.1 was released 6.4 was the legacy release. At some point there is always work done on 2 different releases.

This time there was a significant delay in FreeBSD 9.0-RELEASE causing 8.3 to delay and 8.2 EOL to be adjusted.
 
gkontos said:
If you go back you will see that when 8.1 was released 6.4 was the legacy release. At some point there is always work done on 2 different releases.

This time there was a significant delay in FreeBSD 9.0-RELEASE causing 8.3 to delay and 8.2 EOL to be adjusted.

I can see that 6.4 wasn't EOL (which happened nov 30, 2010) when both 8.0 and 8.1 were released, but on the frontpage 6.4 was no longer advertised as an available release soon after 8.0 was released on nov 25, 2009.

Similarly, with 7.4 EOL in five months, I see your point that advertising three available branches does not mean that there is a bigger workload than when two branches were advertised. A better question is perhaps, why 7.4 is still on the frontpage, knowing that it'll be EOL in five months. (I vaguely remember reading about that somewhere, but can't find it now.)
 
OH said:
I can see that 6.4 wasn't EOL (which happened nov 30, 2010) when both 8.0 and 8.1 were released, but on the frontpage 6.4 was no longer advertised as an available release soon after 8.0 was released on nov 25, 2009.

Yes, it looks that way. Why was I under the impression that always 2 branches are actively maintained :q
Maybe I had too much Jack Daniels ;)

OH said:
Similarly, with 7.4 EOL in five months, I see your point that advertising three available branches does not mean that there is a bigger workload than when two branches were advertised. A better question is perhaps, why 7.4 is still on the frontpage, knowing that it'll be EOL in five months. (I vaguely remember reading about that somewhere, but can't find it now.)

For the same reason why 8.2 was the official production release even though its EOL was approaching. I still remember some questions about 8.2 and EOL warnings! Things were really pushed way back with 9.0-RELEASE.
 
Crivens said:
Even these posings only show a fragment of the whole picture
Probably taking that quote out of context, [more in the context of my previous post]... but it may be that a few FreeBSD particulars, which were previously "let me illustrate that for you..." (cvsup ...) and which are now "let us each read the manual [wiki] ..." IOW less explainable in just a few minutes to someone brand new to it all. [A chance of added complexity inadvertently not preceded by documentation that is as readily available as used to be the case... ]
 
Back
Top