X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f X-Recipient: geda-user AT delorie DOT com Message-ID: <4FC8AA09.5010001@innocent.com> Date: Fri, 01 Jun 2012 07:39:53 -0400 From: Gus Fantanas User-Agent: Mozilla/5.0 (X11; Linux i686; rv:12.0) Gecko/20120430 Thunderbird/12.0.1 MIME-Version: 1.0 To: geda-user AT delorie DOT com Subject: Re: [geda-user] Odd Behavior Around Vias References: <4FC6FDFA DOT 1020001 AT innocent DOT com> <20120531003002 DOT 38be5d00 AT svelte> <20120531062719 DOT 7a615e3f AT svelte> <4FC84D65 DOT 2060302 AT innocent DOT com> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Reply-To: geda-user AT delorie DOT com On 06/01/2012 03:32 AM, Stephen Ecob wrote: > I've trimmed Gus's "buggy.pcb" down to the attached "bug.pcb" that > seems to be the smallest file that still shows the bug. > > The observed problem is related to a somewhat poorly formed polygon: > > Polygon("clearpoly,fullpoly") > ( > [713.00mil 1557.00mil] [662.00mil 1557.00mil] [662.00mil 1589.00mil] > [654.00mil 1589.00mil] [654.00mil 1665.00mil] > [665.00mil 1676.00mil] [674.00mil 1676.00mil] [674.00mil 1708.00mil] > [674.00mil 1706.00mil] [524.00mil 1706.00mil] > [524.00mil 1705.00mil] [566.00mil 1663.00mil] [566.00mil 1557.00mil] > ) > > The polygon is somewhat poorly formed because the vertex [713.00mil > 1557.00mil] forms a zero width protrusion from the vertex [662.00mil > 1557.00mil] > > If the via is removed the polygon is fine, but the combination of the > zero length extension with the via in close proximity triggers a bug > in the polygon handling code. > > I'd guess the main aspect of the bug is a geometry error in the > polygon handling code. > A secondary aspect of the bug is that PCB allows a poorly formed > polygon to be created from the UI (I assume it was created from the UI > !) > Yes, it was created from the UI. Gus