Mail Archives: geda-user/2015/08/31/10:29:30
On Mon, 31 Aug 2015, Evan Foss (evanfoss AT gmail DOT com) [via geda-user AT delorie DOT com] wrote:
> This might be a good place to throw GPMI in again. Like I have said in
> other threads scheme needs to be phased out.
The most problematic part of my back annotation plans is how to get gschem
to do what I want. I was considering using GPMI in "parasite" mode:
injecting the whole thing as an LD_PRELOAD lib at the GTK API.
Pros: 0 modification in gschem, still can hook init, uninit, the main
loop and random gtk API, separate code base.
Cons: I can not hook in function calls where gschem code is calling gschem
C functions or scheme code calling gschem C functions. One of the minor
features I'd need to add would be a way to add an extra mark or highlight
on some object, like pins, to mark they are in netlist mismatch due to
back annotation. I could easily find the part of the pin drawing code
where I'd need to add a 2 or 3 lines. I could easily wrap this function in
another that calls it and then does the extra steps for the extra
highlighting. But this method wouldn't work for internal gschem->gschem
calls.
So GPMI alone wouldn't solve my main problem here; once my main problem is
solved, the rest could be done in plain C, with or without GPMI.
Regards,
Igor2
- Raw text -