X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient: geda-user AT delorie DOT com
X-Sender: qpaz
From: "al davis (ad252 AT freeelectron DOT net) [via geda-user AT delorie DOT com]" <geda-user AT delorie DOT com>
To: geda-user AT delorie DOT com
Subject: Re: [geda-user] A lesson from gnet-makefile
Date: Wed, 21 Oct 2015 09:54:46 -0400
User-Agent: KMail/1.13.7 (Linux/3.2.0-4-amd64; KDE/4.8.4; x86_64; ; )
References: <1042003D-82E2-40F0-AB60-8186580C46AD AT noqsi DOT com> <34B17816-9EA5-45FD-BFB4-9D623A8D3D87 AT noqsi DOT com> <CAM2RGhR+K+dvDdXsbk0Y6LN=-7RhhG5wvtG4i0k4+uMgzd=H0w AT mail DOT gmail DOT com>
In-Reply-To: <CAM2RGhR+K+dvDdXsbk0Y6LN=-7RhhG5wvtG4i0k4+uMgzd=H0w@mail.gmail.com>
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="utf-8"
Content-Transfer-Encoding: 7bit
Message-Id: <201510210954.46552.ad252@freeelectron.net>
Reply-To: geda-user AT delorie DOT com

On Tuesday 13 October 2015, Evan Foss (evanfoss AT gmail DOT com) [via 
geda-user AT delorie DOT com] wrote:
> We could prototype it via a plugin but in the long term it
> should really be in the core. 

Maybe, but maybe you should rethink  plugins.

Gnucap takes the approach of putting as much as possible in 
plugins.  Anything that can be a plugin is required to be a 
plugin.  A set of plugins is distributed with core, but they are 
still plugins.



> To be honest I find your
> "don't touch the core you will break something" attitude
> kind of insulting.

Don't touch core if you can do it in a plugin is good policy, 
but core needs to develop too.  

There needs to be some discipline in how core changes are done.  
Having a bunch of developers all messing with "master" leads to 
a big mess.  

In Gnucap, all work on core is done in branches.  A branch is 
considered ready to merge when it is shown to work correctly, 
has test cases, is formatted correctly, announced and discussed 
on the developer list, and its branch can be merged to master or 
unstable as a fast-forward merge.  When ready, the branch is 
pushed to unstable for final review and then to master after a 
few weeks.  So, master is always "considered stable".