X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient: geda-user AT delorie DOT com
X-TCPREMOTEIP: 72.130.189.202
X-Authenticated-UID: jpd AT noqsi DOT com
Content-Type: text/plain; charset=windows-1252
Mime-Version: 1.0 (Mac OS X Mail 7.3 \(1878.6\))
Subject: Re: [geda-user] gschem refactoring ideas  -- overall architecture document.
From: John Doty <jpd AT noqsi DOT com>
In-Reply-To: <mblqsg$pth$1@ger.gmane.org>
Date: Fri, 13 Feb 2015 19:40:19 -1000
Message-Id: <8E30E768-A95D-4AFA-9BD3-759C7BFEEBE2@noqsi.com>
References: <54DBDFF1 DOT 1010409 AT ecosensory DOT com> <FE92BD5F-0CB3-4CAE-B3BC-D63D22B579DA AT noqsi DOT com> <mbgul5$j8i$1 AT ger DOT gmane DOT org> <220C1787-45BF-459E-B217-29686DC25DF2 AT noqsi DOT com> <mblqsg$pth$1 AT ger DOT gmane DOT org>
To: geda-user AT delorie DOT com
X-Mailer: Apple Mail (2.1878.6)
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by delorie.com id t1E5eR9R005224
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 Feb 13, 2015, at 11:46 AM, Kai-Martin Knaak <knaak AT iqo DOT uni-hannover DOT de> wrote:

> John Doty wrote:
> 
>> On Feb 11, 2015, at 3:19 PM, Kai-Martin Knaak
>> <knaak AT iqo DOT uni-hannover DOT de> wrote:
>> 
>>> A unambiguous sort of symbols on save would finally solve a still
>>> standing issue you may remember: gnetlist behaves differently depending
>>> on the order symbols were added.

I don’t think canonicalization is really a fix. It might even make it harder to figure out which symbol is in control.

>> 
>> I agree this is an annoyance.
> 
> It messes with components that are represented by more than one symbol. 
> Consider an opamp which is represented by the usual  triangle and a 
> separate box for the power supply. The power supply symbol contains the 
> footprint attribute.
> This is good style and recommended for clarity. It reduces the clutter but 
> still shows all the pins. 
> 

Yes. I use your symbols for this. Thank you.

> The netlister looks only at the first symbol of a group when it determines 
> the footprint.

By default, using the old API, yes. I started calling this the "attribute censorship bug” over five years ago. But censor-fix.scm changes this behavior. It compares all of the attribute values of a given name associated with the package, and complains of mismatches. A back end can also use (gnetlist:get-all-package-attributes) directly in geda-gaf 1.8 or later to see all of the values.

> If the power rectangle was placed after the triangle, the 
> netlister thinks, acts like this is a component without a footprint. This 
> frequently leads to confusion among students I teach the fundamentals of 
> geda.

I wish the code from censor-fix.scm had made it into 1.8. I tried. When Bas Gieltjes coded (gnetlist:get-all-package-attributes) on the C side, I rewrote (gnetlist:get-package-attribute) as a Scheme wrapper around it, with the bug fixed. But some developers believed that some users could be depending on this bug, and rewrote the wrapper to reproduce it. That’s what wound up in 1.8. The part of the code that actually fixes the bug is in censor-fix.scm.

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