X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient: geda-user AT delorie DOT com
Message-ID: <51C06B1F.8000601@sonic.net>
Date: Tue, 18 Jun 2013 07:13:51 -0700
From: Dave Curtis <davecurtis AT sonic DOT net>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:16.0) Gecko/20121028 Thunderbird/16.0.2
MIME-Version: 1.0
To: geda-user AT delorie DOT com
Subject: Re: [geda-user] gschem text measure
References: <51BE355E DOT 5040600 AT sonic DOT net> <51BE73B7 DOT 4020905 AT sonic DOT net> <alpine DOT DEB DOT 2 DOT 10 DOT 1306181147510 DOT 4029 AT home DOT local>
In-Reply-To: <alpine.DEB.2.10.1306181147510.4029@home.local>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Reply-To: geda-user AT delorie DOT com

OK, that really isn't my question, though.

I want to mimic gschem's computations, broken or not, closely enough 
that my tool can lay out symbol text with reasonable accuracy.

-dave


On 06/18/2013 03:17 AM, Roland Lutz wrote:
> Hi,
>
> In my opinion, the way text measurements are handled in gEDA is 
> somewhat broken.  The bounding box of an object is usually retrieved 
> via world_get_single_object_bounds.  This, however, depends on the 
> TOPLEVEL context because libgeda by itself is unable to tell the 
> bounding box of a text object, effectively asking the GUI to do so.  
> So there is data stored in the OBJECT struct which actually belongs to 
> a TOPLEVEL object, making it impossible to share objects between 
> contexts.
>
> This is a problem in the FLAG_TEXT_OUTLINE rendering mode where the 
> object's bounding box should be drawn but there is no TOPLEVEL context 
> available in the renderer, so the objects' cached (and maybe invalid) 
> bounding boxes are drawn instead.
>
> One possible solution to address this would be to add a TOPLEVEL 
> reference to the renderer.
>
> Roland
>