From mboxrd@z Thu Jan 1 00:00:00 1970 From: Achim Gratz Subject: Re: mixed orgmode installation Date: Sun, 08 Sep 2013 20:39:54 +0200 Message-ID: <87wqmr5f9h.fsf@Rainer.invalid> References: <8761ugs8as.fsf@Rainer.invalid> <87d2ojs1t4.fsf@Rainer.invalid> Mime-Version: 1.0 Content-Type: text/plain Return-path: Received: from eggs.gnu.org ([2001:4830:134:3::10]:44579) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VIjuB-00058s-Lk for emacs-orgmode@gnu.org; Sun, 08 Sep 2013 14:40:21 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VIju5-0000d6-H5 for emacs-orgmode@gnu.org; Sun, 08 Sep 2013 14:40:15 -0400 Received: from plane.gmane.org ([80.91.229.3]:34234) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VIju5-0000cs-A4 for emacs-orgmode@gnu.org; Sun, 08 Sep 2013 14:40:09 -0400 Received: from list by plane.gmane.org with local (Exim 4.69) (envelope-from ) id 1VIju1-00066R-PT for emacs-orgmode@gnu.org; Sun, 08 Sep 2013 20:40:05 +0200 Received: from pd9eb263f.dip0.t-ipconnect.de ([217.235.38.63]) by main.gmane.org with esmtp (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Sun, 08 Sep 2013 20:40:05 +0200 Received: from Stromeko by pd9eb263f.dip0.t-ipconnect.de with local (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Sun, 08 Sep 2013 20:40:05 +0200 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-bounces+geo-emacs-orgmode=m.gmane.org@gnu.org To: emacs-orgmode@gnu.org John Hendy writes: > I'm with you so far. But if all of Org lives in /path/to/org.git/lisp, > what's to go wrong if it's there vs. /system/path/site-lisp? It is only "there" when you've built Org and whenever you do something in Git, it's "gone", only that you might not see that. Having Org installed in some other place decouples it from what you do in the work tree. That makes it less likely that you load up an Emacs session, do something in Org, then do something in Git and then go back to sour Emacs session and load some other parts of Org that won't fit with the version you've started with. > I'm not sure I follow this one. Does `make up2` look for changed paths > (contrib/lisp vs lisp/) since the last `make up2` ? If not, how would > I know to do `make clean-install` vs. just `make install`? You should know if you changed something, I suppose. > I'm talking about your original comment that running out of a git repo > can lead to: > - it being just to easy to mess up with the autoloads Yes, if you forget to re-make them after a change to the source code. > - have stale byte-compiled files I forgot about somewhere Yes, because Emacs prefers the byte-compiled files over the sources, even when it knows the sources are newer. So when you update from Git, but don't byte-compile, you will load an older version of Org rather than the one you think you are using. If you are running from a Git tree, you should always keep Org uncompiled for this reason (that's why that make target exists). > P.S. And yes, I derailed from the mixed install case due to your > comment as I thought it was worth looking into. I'm doing what you > advise against and I wanted to know the risks and more details about > what I might run into. You can do whatever you want as long as you can deal with the resulting problems. Depending on how careful you are you may never encounter one, but the most frequent reasons for mixed installs are forgetting to generate the autoload files or using an init sequence that loads Org from two different places. Regards, Achim. -- +<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+ SD adaptation for Waldorf Blofeld V1.15B11: http://Synth.Stromeko.net/Downloads.html#WaldorfSDada