From mboxrd@z Thu Jan 1 00:00:00 1970 From: Junio C Hamano Subject: Re: What's cooking in git.git (Sep 2011, #06; Wed, 21) Date: Fri, 23 Sep 2011 11:00:26 -0700 Message-ID: <7vvcsjuojp.fsf@alter.siamese.dyndns.org> References: <7vaa9xyxpf.fsf@alter.siamese.dyndns.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: git@vger.kernel.org To: Scott Chacon X-From: git-owner@vger.kernel.org Fri Sep 23 20:00:36 2011 Return-path: Envelope-to: gcvg-git-2@lo.gmane.org Received: from vger.kernel.org ([209.132.180.67]) by lo.gmane.org with esmtp (Exim 4.69) (envelope-from ) id 1R7A38-0008K7-VR for gcvg-git-2@lo.gmane.org; Fri, 23 Sep 2011 20:00:35 +0200 Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750776Ab1IWSAa (ORCPT ); Fri, 23 Sep 2011 14:00:30 -0400 Received: from b-pb-sasl-quonix.pobox.com ([208.72.237.35]:51594 "EHLO smtp.pobox.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750717Ab1IWSA3 (ORCPT ); Fri, 23 Sep 2011 14:00:29 -0400 Received: from smtp.pobox.com (unknown [127.0.0.1]) by b-sasl-quonix.pobox.com (Postfix) with ESMTP id BA0966296; Fri, 23 Sep 2011 14:00:28 -0400 (EDT) DKIM-Signature: v=1; a=rsa-sha1; c=relaxed; d=pobox.com; h=from:to:cc :subject:references:date:in-reply-to:message-id:mime-version :content-type; s=sasl; bh=B5PMbzbX/8v9lSliHnaC2d3M6Tw=; b=ug8s97 typAMnGHMoN4bZ8VhaoRhnNnyf9h0bjmTHfBzDV8V421qegFVi0QtLzJYo5bnnWR fm+U/ne279FXEvjMKZxxVxKFQP0dvo3iPGCoIkcggH5S106y2HJfBDyfeeQ0KakD SoEn2HydNvPhAverndkuwbx0wxm1ke60Y3WIA= DomainKey-Signature: a=rsa-sha1; c=nofws; d=pobox.com; h=from:to:cc :subject:references:date:in-reply-to:message-id:mime-version :content-type; q=dns; s=sasl; b=PN/pABFsbTkthPpyuU66byPxDiz8wkk/ vu+OD8b4j4oKuLlURxq/0NBxlltvRez7aCBxFeLQXCK/qVsNAEcFKWgL7QDs65L1 7bK1cBPMdV5qyh/K9Dx508NfguPIhL/YSzPkwjY+ktMrISz+8kEl9IwFe5YzvcUm /2Ld2I4Xtxs= Received: from b-pb-sasl-quonix.pobox.com (unknown [127.0.0.1]) by b-sasl-quonix.pobox.com (Postfix) with ESMTP id B267E6295; Fri, 23 Sep 2011 14:00:28 -0400 (EDT) Received: from pobox.com (unknown [76.102.170.102]) (using TLSv1 with cipher DHE-RSA-AES128-SHA (128/128 bits)) (No client certificate requested) by b-sasl-quonix.pobox.com (Postfix) with ESMTPSA id 1A3906294; Fri, 23 Sep 2011 14:00:28 -0400 (EDT) In-Reply-To: (Scott Chacon's message of "Fri, 23 Sep 2011 07:44:13 -0700") User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.2 (gnu/linux) X-Pobox-Relay-ID: EB810F58-E60D-11E0-BE04-9DB42E706CDE-77302942!b-pb-sasl-quonix.pobox.com Sender: git-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: git@vger.kernel.org Archived-At: Scott Chacon writes: > It appears that you're no longer pushing the 'html' branch to the > GitHub repo,... It is not even "no longer". Unlike the other public mirrors that are designed to mimic what my repository at k.org looks like, https://github.com/gitster/git is primarily meant to mirror my internal development repository, with the full set of topic branches. As I do not have 'html' as a branch in my development repository, the result "git push --mirror" there never had refs/heads/html, either. My workflow used to be: - push 4 integration branches from the development repository to k.org, and have its update hook update html and man branches there; - optionally push todo from a different private repository to k.org; - fetch back refs/heads/* from k.org to refs/remotes/ko/*; - verify that my refs/heads/* and refs/remotes/ko/* match for 4 integration branches; - push stuff under refs/remotes/ko/* to public mirrors (sourceforge does not get next and pu); - push --mirror to https://github.com/gitster/git. Now the interim workflow, until k.org comes back, is tentatively: - in a separate local "formatting" repository, pull 'master' from my development repository, build html and man there, and push them back to refs/docs/{html,man} to my development repository; - push refs/heads/{master,maint,next,pu} and refs/docs/{html,man} to public mirrors (sourceforge does not get next and pu); - push --mirror to https://github.com/gitster/git. I am hoping that I do not have to keep doing this and refs/docs/ can eventually go away. I specifically do _not_ want the administrative non-source branches html/man/todo to be my local branches. > Since the k.org hosting of the man pages is down, I was working on > putting them on git-scm.com, where you could select which version of > the docs you want to see and could search through them and whatnot. What is the expected the end-user page-flow? Would people come to https://github.com/gitster/git first, find "doc" link there, and start browsing the html version of the documentation? Or do people come to http://github.com/, find a link "documentation for the scm this site supports", your machinery kicks in and serves the blobs from https://github.com/gitster/git under the hood, without letting users know where the pages they are looking at really come from? What I am getting at is that if these users do not need to know about my repository, then you do not _need_ to see html branch in my repository, either. You only need to find it _somewhere_, and that somewhere does not have to be https://github.com/gitster/git repository. You may think you could use refs/docs/html but I would rather want to avoid letting you guys rely on refs/docs/ I started using just a few days ago as an interim measure. I may later change my mind and get rid of that from my development repository, but that would break your machinery to show preformatted documentation pages. I could add a _separate_ repository at github, and treat it just like other public mirrors like repo.or.cz, so that you can find html branch there. That would be consistent with my wish to use "push --mirror" to at least one place (https://github.com/gitster/git), and would be easier to manage for me. Would that work?