From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeff King Subject: Re: [PATCH v2] credential-osxkeychain: support more protocols Date: Mon, 27 May 2013 12:05:11 -0400 Message-ID: <20130527160511.GA9235@sigill.intra.peff.net> References: <1369641431-44504-1-git-send-email-quanxunzhen@gmail.com> <1369665359-48242-1-git-send-email-quanxunzhen@gmail.com> <20130527150855.GA8250@sigill.intra.peff.net> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Cc: git@vger.kernel.org, John Szakmeister To: Xidorn Quan X-From: git-owner@vger.kernel.org Mon May 27 18:05:27 2013 Return-path: Envelope-to: gcvg-git-2@plane.gmane.org Received: from vger.kernel.org ([209.132.180.67]) by plane.gmane.org with esmtp (Exim 4.69) (envelope-from ) id 1UgzvG-0003lI-3l for gcvg-git-2@plane.gmane.org; Mon, 27 May 2013 18:05:22 +0200 Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753258Ab3E0QFP (ORCPT ); Mon, 27 May 2013 12:05:15 -0400 Received: from cloud.peff.net ([50.56.180.127]:44842 "EHLO peff.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752341Ab3E0QFO (ORCPT ); Mon, 27 May 2013 12:05:14 -0400 Received: (qmail 32016 invoked by uid 102); 27 May 2013 16:05:52 -0000 Received: from c-71-62-74-146.hsd1.va.comcast.net (HELO sigill.intra.peff.net) (71.62.74.146) (smtp-auth username relayok, mechanism cram-md5) by peff.net (qpsmtpd/0.84) with ESMTPA; Mon, 27 May 2013 11:05:52 -0500 Received: by sigill.intra.peff.net (sSMTP sendmail emulation); Mon, 27 May 2013 12:05:11 -0400 Content-Disposition: inline In-Reply-To: Sender: git-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: git@vger.kernel.org Archived-At: On Mon, May 27, 2013 at 11:46:24PM +0800, Xidorn Quan wrote: > > We may also eventually want IMAP for git-imap-send, but we have not yet > > implemented credential-helper support there. We may also want http/socks > > proxy authentication, but we also have not implemented the git side of > > that yet. So I think both of those can wait for now. > > Hope the helpers will be implemented soon. IMO, we can add IMAP and > SOCKS for now since the protocol names are clear, while it is unclear > what protocol name will be used for HTTP/HTTPS proxy. I guess that > some may prefer using http/https as protocol name instead of something > specific. What do you think? Yes, I think git will need to munge the protocol to indicate to the helper that it is a proxy (or alternatively, add a new key "proxy=1" or something). So it would make sense to wait until the git side of the code materializes so that the helpers knows what to expect. There was some work done in this area last May, but there are a lot of corner cases with knowing exactly what the proxy URL is (because in some cases, curl pulls it from the environment without git's knowledge). There was a patch series under consideration for libcurl that would let us supply an "authentication callback" to curl that would be called when credentials were needed. But I have not kept track of the state of that patch and whether it ever got merged upstream into libcurl. -Peff