Mail Archives: geda-user/2015/07/16/17:45:45
On Jul 16, 2015, at 3:23 PM, DJ Delorie <dj AT delorie DOT com> wrote:
>
>> Nope. I sometimes have to make multiple copies of a subcircuit
>> schematic: same topology but different components. Having more
>> control over hierarchy expansion in gnetlist would be good.
>
> I was thinking about the m4 problem in pcb and thought "I wonder if we
> could put a generic expression parser in where we expect simple
> values". I don't think it would help much for pcb (without a looping
> construct, at least), but it might help for heirarchical schematics if
> a component could have a value that was an expression referencing
> inherited attributes from the parent.
Inherited attributes would be nice. Needs a new mechanism in code few understand, though.
> The schematic flattener (if
> you're flattening it for netlisting, for example) would expand the
> expression.
>
> Part of me says "put a guile expression there" but that opens up the
> problem of having a schematic with executable code in it.
I�ve considered, for similar purposes, making a tiny recursive top-down interpreter for ordinary arithmetic expressions in Guile. Such things are not difficult to write. That would keep the executable code genie in the bottle.
> Given the
> nature of electronics, a mathematical (TeX? Mathematica?) expression
> would be more natural, but bring its own can of worms�
>
If you�re specifying discrete part values, one tricky issue is specifying the catalog of acceptable values.
John Doty Noqsi Aerospace, Ltd.
http://www.noqsi.com/
jpd AT noqsi DOT com
- Raw text -