Date: Tue, 10 Apr 2001 15:33:23 -0400 From: "Dan Langille" <dan@langille.org> To: "Rodney W. Grimes" <freebsd@gndrsh.dnsmgr.net> Cc: freebsd-stable@FreeBSD.ORG Subject: Re: Releases Message-ID: <200104101933.f3AJXYe81872@ns1.unixathome.org> In-Reply-To: <200104101534.IAA35432@gndrsh.dnsmgr.net> References: <200104101336.JAA16571@sjt-u10.cisco.com> from Steve Tremblett at "Apr 10, 2001 09:36:40 am"
next in thread | previous in thread | raw e-mail | index | archive | help
On 10 Apr 2001, at 8:34, Rodney W. Grimes wrote some very fine stuff: > Ahhh... well... in my 8 years of being around FreeBSD when we enter > the -BETA phase on the -STABLE branch the tree goes to hell in a > hand basket for about 2 weeks. So please don't make the handbook > state the above. Perhaps more like: > > Every few months in preperation for the next -RELEASE from the > -STABLE branch the system will call itself -BETA, this is to reflect > the fact that lots of changes are occuring in the tree, mainly merges > of well tested code from the developers branch (-CURRENT). The tag > -BETA is used to denoate that these events are occuring and you may > enconter minor problems. This is not like a -BETA product from most > software companies, as the code being brought in has general had a > rather extensive test period, but the project has no way to test all > code in all situations and breakage is bound to happen. This would go a long way to both reducing -questions traffic and increasing understanding. > Though many have called this a bikeshed, and at times I have agreed > that a lot of what is being said is a bikeshed (names are names, people who > attatch permanent fixed meaning to names are going to have problems > understanding lots of things, the -STABLE, -RELEASE, -BETA, -CURRENT tags > being one of them.) But I have seen 2 things come up in this last round of > this 8 year old thread that could actually use some fixing: > > a) Rename the standard-supfile, that one in itself has caused a lot > of grief and is a trivial change with minimal impact to the masses as > it really should only be used by -developers, who know how to deal with > all this. Renaming is a fine idea. What about one more step. Remove it altogether. The people that need it already know how to create it. How does that sound? > b) Correct everyone who says ``FreeBSD -BETA is really just as stable > as the normal -STABLE. It is not! During this phase of a branch > things get borked all over the place, from not being able to build the > tree due to partial MFC's or botched commits, to kernels that crash and > burn due to subtle bugs and corner cases that didn't get tested in > -CURRENT due to the smaller user base and narrower scope of hardware > being tested on. > > One way to correct this would be to actually branch for -BETA, then > once the group of us -STABLE users who actually have a clue about how > to build and test this stuff have had a fair chance (2 weeks?) to pound > it a bit merge it back into -STABLE in one big sweep and call it -RC. > This is a short lived branch, but may have a rather high impact on > cvsup resources due to the tagging operations required to do this type > of operation cleanly. > > The masses would never see -BETA, but it would be done by people > who actually know how to be beta testers :-) Well, the branch makes a good deal of sense. It stops beta from getting out to the people who track -stable and allows changes to be even more isolated. This will also eliminate the inevitable posts to - stable and -questions reagarding "how come I got BETA code?". Do others think this is a good idea? -- Dan Langille pgpkey - finger dan@unixathome.org | http://unixathome.org/finger.php got any work? I'm looking for some. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200104101933.f3AJXYe81872>