Mail Archives: geda-user/2015/09/12/17:34:02
On Sep 12, 2015, at 3:08 PM, Evan Foss (evanfoss AT gmail DOT com) [via geda-user AT delorie DOT com] <geda-user AT delorie DOT com> wrote:
> On Sat, Sep 12, 2015 at 8:27 PM, John Doty <jpd AT noqsi DOT com> wrote:
>>
>> On Sep 12, 2015, at 2:07 PM, Evan Foss (evanfoss AT gmail DOT com) [via geda-user AT delorie DOT com] <geda-user AT delorie DOT com> wrote:
>>
>>> On Sat, Sep 12, 2015 at 5:01 PM, Nicklas Karlsson
>>> (nicklas DOT karlsson17 AT gmail DOT com) [via geda-user AT delorie DOT com]
>>> <geda-user AT delorie DOT com> wrote:
>>>>>> Would it be hard ta add the netname all connected segments then it is added or changed?
>>>>>
>>>>> I would have issues with this. One of the features I wanted to add to
>>>>> gnetlist was the ability to output specific properties for parts of
>>>>> nets.
>>>>> (ex. trace width for PCB.)
>>>>
>>>> I also want. As is now connections are stored as lines and an attribute could be added to one of the lines, I guess it would be possible to achieve this with current file format although there may be a need split lines in two sometimes to form segments.
>>>>
>>>> What would be needed in gschem would a function to select all line segments belonging to a net and a function selecting all lines belonging to a segment which possibly also sometimes have to split a line in the middle to form a segment. Then an attribute is added or changed it will be attached to only one off the lines and then I guess it will work with gnetlist.
>>>
>>> Igor2 and I have been fixated on the idea of placing additional
>>> attributes on parts of nets so that you can have things like trace
>>> width passed to PCB. I am looking at having gnetlist understand when
>>> someone wants a star ground by looking at how nets are angled when
>>> they meet.
>>>
>>> There is the question of how much of gnetlist should be moved into
>>> libgeda
>>
>> None. Scheme scripting is flexible, libgeda is rigid. Moving things into libgeda is the wrong direction.
>
> 1. People fought me when I wanted a second plugin interface in
> parallel to scheme.
Not me. But I don’t think you realize how hard it is.
> 2. People want back annotation and that requires gschem being able to
> import some net related info.
Nope. It requires a tool that can annotate a schematic. Gschem, as an interactive editor, is not that tool. https://github.com/xcthulhu/lambda-geda is such a tool, and I have used it, but scripting in Haskell is probably not to your liking.
> How would you do it? If gschem not going
> to import back annotation info in the form of a netlist what is the
> logical input for it?
*Back* annotation is to me a bad idea in general. Annotation is good. If you want back annotation, simply copy the result of annotation to your source schematics.
> 3. I have been a lot more open to your ideas than a *lot* of other
> people on this list. Simply stamping "wrong direction" on someone
> elses idea is pretty closed minded. You give no technical reasons and
> no alternative solution.
>
>>> since I don't want to do this in scheme.
>>
>> Then do it in xorn-netlist. Options are a wonderful thing to have. I’d be happy to help you with doing it in Scheme, but the gnetlist API cannot access connection details at segment level.
>
> Yes and I can add to the gnetlist API with out breaking things.
If you want to do that, I’ll support it as a Scheme programmer, as I supported the refactoring that enabled the attribute censorship bug fix. But I thought you didn’t want a Scheme solution. Xorn might very well be what you want.
John Doty Noqsi Aerospace, Ltd.
http://www.noqsi.com/
jpd AT noqsi DOT com
- Raw text -