X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient: geda-user AT delorie DOT com
Content-Type: text/plain; charset=us-ascii
Mime-Version: 1.0 (Apple Message framework v1085)
Subject: Re: [geda-user] gnetlist hierarchy uref order setting
From: John Doty <jpd AT noqsi DOT com>
In-Reply-To: <CACNnPRmn1rLA8enjo6epUhyb2OTzoyzYzrj7S4qxHsHdfAeh-g@mail.gmail.com>
Date: Thu, 17 Jan 2013 09:26:47 -0700
Message-Id: <CABE3E91-466B-4769-99A5-85ACDE7E6A5C@noqsi.com>
References: <CACNnPRkowensyyy0pneO0VJeDaMpaMqms-74wCwM=Rtd-bDQHg AT mail DOT gmail DOT com> <644026BD-F6E6-4533-B450-A91CAF83937A AT noqsi DOT com> <CACNnPRnyahDE8Hhhyc6+UFE6=55zmWVP1f8m=E-p4CJWOb=2=g AT mail DOT gmail DOT com> <6C1F8807-DFAD-4256-8F70-BD4AF4EF3A3E AT noqsi DOT com> <CACNnPRmn1rLA8enjo6epUhyb2OTzoyzYzrj7S4qxHsHdfAeh-g AT mail DOT gmail DOT com>
To: geda-user AT delorie DOT com
X-Mailer: Apple Mail (2.1085)
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by delorie.com id r0HGR0GS028265
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 Jan 16, 2013, at 11:21 PM, Abhijit Kshirsagar wrote:

>> Interesting. Stuart seems not to have implemented refdes "munging" for all of the components he lists at http://www.brorson.com/gEDA/SPICE/x631.html.
> 
> So does that count as a bug?

I think so. Mismatch between code and documentation always counts as a bug as far as I'm concerned.

> 
>> Put (hierarchy-traversal "disabled") in your gnetlistrc. Then I netlist subcircuits separately and
>> let SPICE do the hierarchy expansion. See
>> http://www.brorson.com/gEDA/SPICE/intro.html.
> 
> I've looked through this - and it seems to be the same as the RF
> example that ships with the gEDA distribution.
> As i understand, I need to generate the cir files from the
> sub-schematics first, then netlist the top level schematics.

Yes. The spice-sdb back end normally does the inclusion of model files for you. Even if you turn this off with -I, it inspects model files to figure out whether they are macro models (.SUBCKT) or primitive model parameters (.MODEL). Thus, all model files that a schematic page uses must exist prior to invoking the netlister. I consider this a misfeature.

Since SPICE has no scoping of subcircuit or model names, it is useless to include the same model file repeatedly in different contexts, but that's what spice-sdb does, unfortunately. I prefer to have a Makefile use "cat" to assemble the final netlist from the individual subcircuit netlists.

> 
> Do most people use something like a makefile, when systems become
> large? Is there a makefile generator out there?
> 

I just write a Makefile. No generator.

> I was hoping to have everything (including the SPICE simulation
> commands) in schematics only - so the end user only needs to keep
> gSCHEM open and navigate the hierarchy to see the entire design.

Well, it makes more sense to me to edit things like .CONTROL sections in a text editor rather than a graphics editor. And, as noted above, I prefer to assemble the simulation file with "cat" rather than gnetlist: this gives me much better control. Also, I often have very large test stimulus files, megabytes of PWL commands, generated by separate scripts. It's not practical to put *those* in a schematic. But if you want, you can put your extra stuff in spice-directive symbols. You can even put multi-line code into the value= attribute by editing the attribute as text ("ex" rather than "ee").

John Doty              Noqsi Aerospace, Ltd.
http://www.noqsi.com/
jpd AT noqsi DOT com