On 09-Dec-2003 Jon Noack wrote: > On 12/8/2003 9:31 PM, Scott Long wrote: >> Jon Noack wrote: >> >>> On 12/8/2003 2:29 PM, Doug White wrote: >>> >>>> On Sun, 7 Dec 2003, Jon Noack wrote: >>>> >>>> >>>>> I ask this for 5.2 because it never happened for 5.1: >>>>> Will src/share/examples/cvsup/standard-supfile be updated to point to >>>>> the "RELENG_5_2" tag instead of "." for 5.2? >>>> >>>> >>>> >>>> >>>> Doubtful -- standard-supfile is for grabbing -current. If you want a >>>> specific tag, you need to specify it. I just copy the same cvsupfile >>>> around to different machines as I build them so I don't forget :) >>>> >>>> I agree that stable-supfile should be updated, though. But 5.X isn't >>>> -stable yet. :) >>> >>> >>> >>> Copying re_at_ on this... >>> >>> I respectfully disagree. Here's an open bug report from someone else >>> who thinks the same way I do: >>> http://www.freebsd.org/cgi/query-pr.cgi?pr=conf/53197 >>> >>> Even if you disagree with me, check out the CVS commits to >>> standard-supfile: >>> http://www.freebsd.org/cgi/cvsweb.cgi/src/share/examples/cvsup/standard-supfile >>> >>> >>> A very common entry is something to the effect of: >>> "The 'standard-supfile' should track its own branch." >>> >>> (As Colin Percival just point out:) >>> From the inception of the security release branch with RELENG_4_3, >>> every release *but* 5.1 has had standard-supfile point to the security >>> release branch. That's 8 releases in my favor vs. 1 release in your >>> favor. I win ;-). >>> >>> I'd wager a lot of folks used to 4.x giving 5.x a try would get bitten >>> by this, accidentally upgrading to -CURRENT and possibly hosing their >>> systems as a result. >>> >>> In any case, the only color for the shed is midnight blue. >>> Jon >>> >>> >> >> There was discussion about this after 5.1 too. Basically, we need to >> create another cvsup example file, one for RELENG_4, one for RELENG_5_x, >> and one for HEAD. Does this lead us down the road to having even more >> example files? What about one for RELENG_4_9? I guess I'm not opposed >> to this. If someone will submit a patch, I'll consider it. >> >> Scott > > This is what I envision: > current-supfile gets you -CURRENT > stable-supfile get you -STABLE > standard-supfile gets you updates to what you have > > The following should accomplish this without breaking prior functionality: > > For all: > current-supfile tracks -CURRENT ("." tag) > stable-supfile tracks -STABLE (currently "RELENG_4" tag) > > For -CURRENT: > standard-supfile tracks -CURRENT ("." tag) > > For -STABLE: > standard-supfile tracks -STABLE (currently "RELENG_4" tag) > > For releases: > standard-supfile tracks the release branch ("RELENG_5_2" tag for 5.2) > > The attached "supfile-current.diff" adds current-supfile (based on the > old standard-supfile). I didn't touch the CVS info at the top of the > file -- this is automatically updated when the file is checked in, right? > > The attached "supfile-5.2.diff" updates the standard-supfile for 5.2 to > reflect this scheme by changing the CVS tag to RELENG_5_2 and modifying > a comment for accuracy. > > The attached "supfile-README.diff" updates the README to reflect this > scheme (I pulled part of the text from the RELENG_4_9 README). It > should be fine for everything (releases, -CURRENT, and -STABLE). This sounds very sane to me. -- John Baldwin <jhb_at_FreeBSD.org> <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/Received on Tue Dec 09 2003 - 09:44:26 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:33 UTC