X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient: geda-user AT delorie DOT com
Date: Fri, 5 May 2017 04:23:16 +0200 (CEST)
X-X-Sender: igor2 AT igor2priv
To: "James Battat (jbattat AT wellesley DOT edu) [via geda-user AT delorie DOT com]" <geda-user AT delorie DOT com>
X-Debug: to=geda-user AT delorie DOT com from="gedau AT igor2 DOT repo DOT hu"
From: gedau AT igor2 DOT repo DOT hu
Subject: Re: [geda-user] pads changing size in gEDA PCB
In-Reply-To: <83e14b55-a199-e28a-168d-32fb38308277@wellesley.edu>
Message-ID: <alpine.DEB.2.00.1705050413050.27212@igor2priv>
References: <201705041302 DOT v44D2lHv001103 AT delorie DOT com> <590B9BEC DOT 9060509 AT xs4all DOT nl> <83e14b55-a199-e28a-168d-32fb38308277 AT wellesley DOT edu>
User-Agent: Alpine 2.00 (DEB 1167 2008-08-23)
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Reply-To: geda-user AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: geda-user AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk



On Thu, 4 May 2017, James Battat (jbattat AT wellesley DOT edu) [via geda-user AT delorie DOT com] wrote:

> I reported a similar issue (same symptom, maybe the same cause) to this list 
> on 3/14/16 (with further discussion on 5/19/16) -- subject line was:  "pad 
> dimensions shrink when footprint flips to other side of board"...
>
> Also documented here:
> http://academics.wellesley.edu/Physics/jbattat/geda/issues/flipbug.html
>
> Looking back through my email archive, I see that pcb-rnd did not have this 
> issue, so the problem was likely introduced after the pcb-rnd fork.
>

I've also tested the files: pcb-4.0.0 with gtk/gdk does have the 
issue, while pcb-rnd (with gtk and lesstif) does not. I agree with James, 
this bug probably got introduced after the fork (but it may be worth 
looking through pcb-rnd svn logs, we may have fixed it after the fork, 
before James tested - I can't recall, I've fixed so many bugs during the 
years).

Also looked at the patch and the source. I think this is not the right 
fix: the bug is somewhere in the pad rendering when coordinates are 
flipped. The patch tries toB make sure coordinates don't get flipped by 
mirroring - but what if a script generated a file with flipped pad coords, 
or an "old file" has such a pad or another operation also flips? -> pcb 
will render the pad incorrectly.

Another hint: the pad renders correctly even flipped, if it has round cap 
(press 'q').

So I recommend looking at the pad rendering code instead.

Regards,

Igor2