gEDA-dev: Broken slotting

[Peter Clifton]

On Wed, 2007-09-12 at 06:39 -0400, Stuart Brorson wrote:
> IIRC, the original problem was this:  gnetlist did not provide a way
> to get the net attached to a pin using pinseq for slots 2, 3, 4....
> Spice-sdb needs to get the nets using pinseq since it needs to emit
> the pins in a particular order dictated by the ordering of the net
> connections in SPICE.  That is, spice-sdb loops over pinseq, and asks
> gnetlist for the name of the net attached to pinseq when outputting
> netnames.
> 
> The code I wrote provided a simple way to get the pin's net by asking
> for pinseq modulo the slot number.  That is, if the slotted device
> had 14 pins, and the SPICE model required 5 connections, I could ask
> for pinseq=7 to get slot=2, pinseq=2.  Or something like that.
> 
> If you have a better way to do it, that's fine.  But I'm not sure that
> this scheme is broken in any way.  (I am happy to be shown wrong, of
> course.)
> 
> Stuart

I have no strong feeling either way, and I guess the real issue is
defining the semantics for the pinseq= attribute.

I may brave digging into gnetlist at some point in the future, and see
if there is anything which can be improved there for increased
flexibility in the backends.

I already have to venture into it to figure out why it doesn't like
spaces in output file-names.

Peter