Date: Sun, 03 Aug 1997 16:10:23 -0700 From: "Jordan K. Hubbard" <jkh@time.cdrom.com> To: Wolfgang Helbig <helbig@MX.BA-Stuttgart.De> Cc: andreas@klemm.gtn.com, current@FreeBSD.ORG Subject: Re: Current is currently really a mess (was: Re: Tk/Tcl broken(?)) Message-ID: <4492.870649823@time.cdrom.com> In-Reply-To: Your message of "Mon, 04 Aug 1997 00:43:04 %2B0200." <199708032243.AAA01485@helbig.informatik.ba-stuttgart.de>
next in thread | previous in thread | raw e-mail | index | archive | help
[Are we sick of this thread yet? ;-)] I promised to stay out of this, but there was just one misconception that I wanted to clear up here given that both Wolfgang and Richard S. seem to have latched onto it. I never suggested that support for the ports collection would be completely dropped for whatever release that -current eventually becomes, simply that the development of this would be done differently. Rather than trying for a daily "match" between /usr/src and /usr/ports on a -current system, you'd work instead on the model where you only sync'd up the two when it came close to release time for -current, whatever the value for "-current" might be. With over a year or so between "roll-overs" (e.g. -current switches to the next major release # and the old -current becoemes -stable), I don't see how this is such a tremendous hardship - there will be ample time to sort the issues out. And to those who would argue that not having ports for the duration of -current's run is such a terrible thing, I might respectfully suggest that you have your priorities exactly reversed, and not because "users are more important than developers" (as Garrett accused me of believing in a private email) but rather because of the user ratios we have. Most people run the release branches and most people are *PISSED OFF* that it's been our long-standing policy to support -current and not -releng, leaving the users of the last release high and dry with whatever ancient ports snapshot was bundled with their release. Is that somehow better? Why is no one indignant about that? It seems to me that failing to support your release users would be considered almost hallucinogenically weird by anyone in the commercial software industry, and I've certainly taken my share of annoyed emails over the issue. Satoshi has been petititioned more times than I can count to support the 2.2.x folks and he's answered each time that trying to maintain an active ports tree for *two* branches is just too much work. Now given that, who does it make more sense to keep ports "active" for - the -current users or the -stable users? Given the comparative rates of change in each branch, which makes the most *sense* to support? Given the size of the user populations for each, which is the most logical? I'd say -stable for both answers and I have a very hard time beliving why anyone would pick -current (and remember - you only get to pick *one* branch due to workload concerns, something which has been clarified on many many occasions and not worthy of yet another debate at this time). > yes, -current is a testing ground--for the ports. If developing > the ports collection on -current is discontinued now, I don't see > *any* ports running on the future RELENG_3_0_RELEASE. And I don't I'm not so pessimistic as this, given the long release cycles we have. Once -current actually shows signs of becoming a released product, and I don't see that happening anywhere before the end of the year, people can take whatever was active in the RELENG_2_2 branch and retrofit it into -current. Jordan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4492.870649823>