delorie.com/archives/browse.cgi   search  
Mail Archives: geda-help/2017/04/20/09:50:10

X-Authentication-Warning: delorie.com: mail set sender to geda-help-bounces using -f
X-Recipient: geda-help AT delorie DOT com
Date: Thu, 20 Apr 2017 15:56:10 +0200 (CEST)
X-X-Sender: igor2 AT igor2priv
To: "Smilie (smilie AT posteo DOT de) [via geda-help AT delorie DOT com]" <geda-help AT delorie DOT com>
X-Debug: to=geda-help AT delorie DOT com from="gedah AT igor2 DOT repo DOT hu"
From: gedah AT igor2 DOT repo DOT hu
Subject: Re: [geda-help] gschem Feature request
In-Reply-To: <20170420152505.1365d58e@debian>
Message-ID: <alpine.DEB.2.00.1704201542580.27212@igor2priv>
References: <20170420152505 DOT 1365d58e AT debian>
User-Agent: Alpine 2.00 (DEB 1167 2008-08-23)
MIME-Version: 1.0
Reply-To: geda-help AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: geda-help AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com

Hi,

On Thu, 20 Apr 2017, Smilie (smilie AT posteo DOT de) [via geda-help AT delorie DOT com] wrote:

> Hello Guys,
>
> i have a idea.
>
> If i select a symbol in gschem, then could be written to pcb-file the
> option "selected" and possible also vis versa.

/me hears the sound of a can opener; then the sound of 1000 worms.

IN short: it's not totally impossible, but is difficult:

- it potentially needs some sort of a project manager that runs both 
gschem and pcb(-rnd); else if you have multiple instances of either editor 
running (I often do), the other editor wouldn't know which one to talk to 
(e.g. 2 pcb-rnds, 1 gschem; click in gschem; how does it know which one to 
control?); alternatively we need a project concept (pcb-rnd already has 
that)

- gschem -> pcb is not a direct path, it goes through gnetlist; the final 
refdes values are invented by gnetlist. Thus all such notification would 
need to go through gnetlist.

- pcb -> gschem would need to travel back through such a reverse-gnetlist 
layer too, which we simply don't have at all

- this also implies gschem would need to have better, bidirectional 
communication with the backends (gnetlist), which we also don't have afaik

- and then we have tons of corner cases: what if you click on a large fpga 
in pcb that is 10 symbols in gschem? What if 10 symbols on 4 pages? from 
which 2 are not even open...

(- and once we have these, the natural next step would be to do the same 
with nets; all the above problems apply to nets too, so once they are 
solved for elements they could be solved for nets too. But nets have an 
unique property: gschem doesn't want to know about them. So even if we got 
back the right netname info to gschem, it wouldn't be able to easily 
highlight or select or show that net)

(- some of these problems are already solved/covered/worked-around by the 
pcb-rnd->gschem back annotation, but that thing works mainly because it 
doesn't assume gschem can or need to reliably identify elements or nets, 
it rather just assists the user and its the user's intelligence at the end 
that solves the problem; in other words, it'd be rather easy to tell the 
user in gschem that in pcb-rnd U12 is selected, but it wouldn't be easy to 
auto-find U12 in gschem)

So all in all, don't have high hopes about this.

Regards,

Igor2



- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019