From mboxrd@z Thu Jan 1 00:00:00 1970 From: Nicolas Goaziou Subject: Re: Bug: Please supply stable releases on ELPA or MELPA Stable [8.3.4 (8.3.4-dist @ /usr/local/share/emacs/25.1.50/site-lisp/org-mode/)] Date: Wed, 02 Nov 2016 00:43:58 +0100 Message-ID: <871syuepfl.fsf@nicolasgoaziou.fr> References: <87mvhltqpw.fsf@sc3d.org> <87oa21tmp0.fsf@gmail.com> <87ins8ftyv.fsf@nicolasgoaziou.fr> <87pomfv6kf.fsf@bzg.fr> <87k2cne9ud.fsf@nicolasgoaziou.fr> <87r36v85r0.fsf@Rainer.invalid> Mime-Version: 1.0 Content-Type: text/plain Return-path: Received: from eggs.gnu.org ([2001:4830:134:3::10]:42247) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1c1iil-0000tO-3W for emacs-orgmode@gnu.org; Tue, 01 Nov 2016 19:44:00 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1c1iij-0001mf-TL for emacs-orgmode@gnu.org; Tue, 01 Nov 2016 19:43:59 -0400 Received: from relay3-d.mail.gandi.net ([2001:4b98:c:538::195]:51529) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1c1iij-0001m6-MX for emacs-orgmode@gnu.org; Tue, 01 Nov 2016 19:43:57 -0400 In-Reply-To: <87r36v85r0.fsf@Rainer.invalid> (Achim Gratz's message of "Tue, 01 Nov 2016 18:33:23 +0100") List-Id: "General discussions about Org-mode." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: emacs-orgmode-bounces+geo-emacs-orgmode=m.gmane.org@gnu.org Sender: "Emacs-orgmode" To: Achim Gratz Cc: emacs-orgmode@gnu.org Hello, Achim Gratz writes: > No, we shouldn't. It's either automatic or a release, it can't be > both. I can't see why. We already "release" automatically new packages. > Also, the releases are distinct from the ELPA packages and you seem to > mix up the two. They are not the same thing. By release, I mean the act of releasing a new version, not the type of output (tar.gz, ELPA package). In any case, I think it is confusing to have ELPA packages differ from source releases. Installing through ELPA could be made equivalent to installing latest stable sources, i.e., both output could be generated at the same time. It means that bugfix releases should happen more often and ELPA packages less often. >> 1. org-YYYYMMDD could be renamed org-MAJOR-MINOR-BUGFIX# where MAJOR >> MINOR are never modified automatically, and BUGFIX# is (1+ last >> BUGFIX#). > > Org's ELPA package has its own versioning scheme and unless package.el > grows some functionality to sanely deal with discontinuities in > versioning, we need to stick to it. It may be. Not providing org-YYYYMMDD versions anymore could be a step in the direction of replacing them. In the worst scenario, we can keep them as packages, and still provide org-X.Y.Z alongside. >> 2. Conditions to make a new automated release ought to change. We could >> wait for a full "idle" week after a commit before releasing (IOW, >> wait for one week after a commit but every new commit during that >> period resets the counter). "next Monday" rule has bitten us already. >> The new rule is not perfect either, but is more secure. If one full >> week is too long, we may reduce it to 4 days. > > Again, if you talk about releases, the way to do that is to tag what > should be released, preferrably sign the tag as well. Whether you roll > the release automatically or script up somthing that picks up new tags > and does it for you is a separate issue. Per above, the script should also generate a new ELPA package. > BTW, if the Monday ELPA package is bad you can always trigger another > release to ELPA manually How? > and don't need to wait for next Monday to have cron run the release > script. This is the opposite of what I'd like to see. Release script should be run less frequently, not more. Regards, -- Nicolas Goaziou