X-Authentication-Warning:	delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient:	geda-user AT delorie DOT com
Date:	Thu, 7 Jul 2016 07:01:47 +0200 (CEST)
X-X-Sender:	igor2 AT igor2priv
To:	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:	[geda-user] pcb: polygon "twin hole" bug
Message-ID:	<alpine.DEB.2.00.1607070651120.7286@igor2priv>
User-Agent:	Alpine 2.00 (DEB 1167 2008-08-23)
MIME-Version:	1.0
Reply-To:	geda-user AT delorie DOT com

Hi all,

Evan found a bug in pcb-rnd which turned out to be a bug in mainline as 
well. It is really easy to reproduce from scratch:

1. draw a poly
2. draw a hole in it (strange: first click ignored?)
3. keep on drawing the second hole (strange: marker is not moved?)
4. bug: the poly of step 1 is duplicated (copied in place), and the new 
hole is created on the new poly

Important: do not switch tool between step 2 and 3! This, together with 
the fact that it can be reproduced with both HIDs suggests it's a bug in 
core (the strangenesses are maybe related to handling of the mode 
variable, I have no idea how the poly gets dupped).

Detecting the duplication polys:

- save before and after step 4, compare the two save files; instead of one 
poly object with two holes there are two poly objects with one hole each

- move the poly after step 4; expected: moving one poly; what happens 
instead: moving the top poly, leaving the bottom poly in place

- in step 3 draw the second hole so that it has an intersection with the 
first; that intersection is the only copper hole at the end, because it 
is not covered by any of the polygons

In pcb-rnd I am still knee-deep in the conf rewrite, but I plan to look at 
this bug after I finish. In case someone debugs this in mainline, I'm 
interested in the result. Else I can share my findings/patches when I 
get to it, so mainline developers can fix this in mainline.

Regards,

Igor2

- Raw text -