X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient: geda-user AT delorie DOT com
MIME-version: 1.0
Content-type: text/plain; charset=ISO-8859-1; format=flowed
Message-id: <4EAB1F68.6060307@unige.ch>
Date: Fri, 28 Oct 2011 23:32:24 +0200
From: Juergen Harms <Juergen DOT Harms AT unige DOT ch>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.23) Gecko/20110930
 Mageia/3.1.15-1.mga1 (1) Thunderbird/3.1.15
To: geda-user AT delorie DOT com
Subject: Re: [geda-user] failed tests
References: <4EA96FC8 DOT 1080305 AT unige DOT ch> <20111027160001 DOT GA987 AT malakian DOT lan>
 <4EA9AD37 DOT 8070204 AT unige DOT ch> <20111027201819 DOT GC987 AT malakian DOT lan>
 <4EAADBAB DOT 6090803 AT unige DOT ch> <4EAB00EC DOT 8060003 AT mucharuina DOT com>
In-reply-to: <4EAB00EC.8060003@mucharuina.com>
Content-transfer-encoding: 7bit
Reply-To: geda-user AT delorie DOT com

Well, I think I get nearer to see what happens: I am trying to manually 
do what the tests do, running the pcb program created by make in the 
tarball with the test arguments: done that way, the gerber files created 
look as they should (they also look ok if I do it with the installed pcb 
and export to gerber - it is just the tests that - make apear? - that 
the result is wrong).

A likely hypotheses is therefore that the problem lies in the way the 
scripts in the tarball create the environment and the arguments for 
running the test - that is quite intricate (I do not yet fully 
understand how this is done) and sufficiently complicated to maybe work 
on one OS and maybe not on another.

With a little bit of chance, I might get to the bottom (I would never 
have dared to dig into how the graphics are generated, understanding how 
these test scripts work is not evident, but somehow in reach)