X-Authentication-Warning: delorie.com: mail set sender to geda-help-bounces using -f
X-Recipient: geda-help AT delorie DOT com
Resent-Message-Id: <201210161918 DOT q9GJIQ0t030298 AT atl4mhob15 DOT myregisteredsite DOT com>
Subject: Re: [geda-help] Adding new gschem symbols?
Mime-Version: 1.0 (Apple Message framework v1085)
Content-Type: text/plain; charset=us-ascii
From: John Doty <jpd AT noqsi DOT com>
In-Reply-To: <20121016081200.587F08117B4C@turkos.aspodata.se>
Resent-From: John Doty <jpd AT noqsi DOT com>
Date: Tue, 16 Oct 2012 12:14:18 -0600
Cc: karl AT aspodata DOT se
Resent-Date: Tue, 16 Oct 2012 13:18:24 -0600
Resent-To: geda-help AT delorie DOT com
References: <alpine DOT BSO DOT 2 DOT 00 DOT 1210110132340 DOT 24638 AT wolfman DOT devio DOT us> <1349966191 DOT 2412 DOT 33 DOT camel AT AMD64X2 DOT fritz DOT box> <alpine DOT BSO DOT 2 DOT 00 DOT 1210112245150 DOT 13854 AT wolfman DOT devio DOT us> <20121012100446 DOT 93D648096B1E AT turkos DOT aspodata DOT se> <alpine DOT BSO DOT 2 DOT 00 DOT 1210122243220 DOT 2692 AT wolfman DOT devio DOT us> <20121013174902 DOT 160C98117B42 AT turkos DOT aspodata DOT se> <CA+DWcQacCk-QG03iwsZJUXvTumXc+FnHvzZeXf5PjNAgph2hiw AT mail DOT gmail DOT com> <F3E56F79-7653-4E48-B247-058128BF022E AT noqsi DOT com> <20121016081200 DOT 587F08117B4C AT turkos DOT aspodata DOT se>
Message-Id: <35CC5D35-73DC-42C5-91D1-D7F0DA3E3654@noqsi.com>
To: geda-help 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 q9GJIbsR032081
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
Precedence: bulk


On Oct 16, 2012, at 2:11 AM, Karl Hammar wrote:

> John Doty:
>> On Oct 14, 2012, at 2:27 PM, Benjamin Bergman wrote:
>>> On Oct 13, 2012 12:50 PM, "Karl Hammar" <karl AT aspodata DOT se> wrote:
>>>> Well you could possible use the md5sum of the sym file, what about:
>>>> $ md5sum git/openhw/share/gschem/diode.sym
>>>> cc2da042b5ea5afd65f4153bfff79b92  git/openhw/share/gschem/diode.sym
>>>> And in the .sch file, this file is referenced by:
>>>> C 18600 19900 1 0 0 diode.sym md5=cc2da042b5ea5afd65f4153bfff79b92
>>> I think this is the way to go. This is essentially how all content
>>> is tracked and referenced in git and adds a layer of corruption
>>> detection, while costing very little in terms of resources or
>>> infrastructure.
>> 
>> Shudder. That assumes that symbols rarely change.
> 
> That is a false assumptions, and I have seen symbols dissapear from
> the cvs also. But that doesn't hinder you to have a directory
> structure like
> 
> ver1/diode.sym
> ver2/diode.sym
> ver3/diode.sym
> 
> and regardless of the order which thoose appear in the lib-browser,
> a md5-tag could find the right diode.sym for you.

Well, when I use gEDA, I do not have promiscuous settings for library search. Indeed, for big projects, I've become fond of the (reset-component-library) function. This allows me to have complete control of the symbols a project uses. The browser then only sees symbols in the project's controlled set. For broader search, I'll use a command line like "locate .sym | grep -i diode". That finds just about every diode symbol on the entire machine: standard gEDA, gedasymbols, and every gEDA project I've worked on (over 10 years!). If I find what I need, I can copy to one of the project's symbol directories.

You can see the beginnings of an elaborate project of this kind at github.com/noqsi/VideoASIC. Not too much of the gEDA part is there yet, but it'll be coming soon now that the background math is done.

> And why not allow
> a cvs-tag/git-commit-id to extract that old sym file from the
> version handler?
> 
>> But a sensible way to use gEDA is to have project-local symbols
>> that change as packaging or other part selection attributes change.
> 
> That is what works today, but that shouldn't hinder us to try out
> other ideas.

The problem with your idea is that it puts restrictions on the tools.

> 
> ///
> 
> It is nice to include the cvs sym's in the library browser, at
> least to find what's out there. Currently one cannot use the second
> or third of a duplicate sym (unless you copy or embed it); I think
> it would be nice to solve that little problem.
> 
> ///
> 
> Having a md5-thing as exemplified above could alert us when we have
> forgotten to update our sch's when we made changes to the project
> local sym's.

Well, the way I use the toolkit I don't normally have to worry about that. But a fairly common case is to change something like the default resistor footprint. That's an advantage of project-specific symbols now. I'll have a "resistor.sym" in the project. For the "breadboard model" that'll have a reasonably big footprint, maybe 0805. Then, when the customer's happy with the design, I can change the package to whatever the area constraints for the production board demand. To do this, all I need to do is change one symbol. Nothing else needs updating. But your way I might have to update hundreds of symbols across dozens of schematics.

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