Re: gEDA-dev: New diagram (attempt at UML)

[Steve Meier <smeier-at-NOSPAMalchemyresearch_dot_com>]

Dan McMahill wrote:
>
> I think that may be premature. First a solid framework for walking
> through a hierarchy and producing a hierarchical netlist (of any
> format) needs to be in place.  Then spice, verilog, vhdl, whatever
> should be fairly easy.
>

If this was the only criteria then I would be there. In reality I am
working on converting my changes back into a new version of libgeda and
a new version of gschem that uses the new library. I now have my
netlister using the new library.

The library itself holds all of the internal geda data structure
manipulation methods and I am working on segregating out all of the
graphics and adding something similar to an hid. There is no reason that
gschem couldn't run portions of the netlister methods from the library.
For example to highlite a specific net.

The netlister itself has been reduced to 6 source files (not including
the scheme scripts) all the rest are part of the library.

src/globals.c    src/g_rc.c        src/i_vars.c       src/parsecmd.c
src/g_netlist.c  src/g_register.c  src/mra_netlist.c  src/vams_misc.c

The main other issue that I am exploring is back anotation.

For back anotation which supports hierarchical designs, I am considering
being able to automatically fork shematics and symbols then applying the
back anotation by adding  an attribute to the schematic attached to a
complex object that describes a pin swap. For embeded complex objects
the pins could be literaly swapped.

Steve Meier



_______________________________________________
geda-dev mailing list
geda-dev@moria.seul.org
http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev