On Friday 19 July 2013, al davis wrote: > The reason obsolete_callback exists is that there can be > expressions of commons, with possibilities that don't fit > the parameter=value model, so these files were much more > difficult than the others.
To clarify ... The real reason obsolete_callback exists should be evident from the word "obsolete". It's old code, from before any consideration of multiple simulation languages, that hasn't been converted yet. That is not the only place in gnucap where there is old yucky code in need of replacement. The most flagrant ones are the connectmodule code in e_node* and d_logic*, the output code, and the list of if's that enables shortcut command names. All of these are examples of stuff, my own work, that I would not accept in new work. There is also an issue with the difference in "model" code, stemming back to the spice notion of a "model" card, essentially made obsolete by the notion of cloning a master device. It seemed to make sense at the time. Now it is unneeded extra complexity. This one will need to change to have full support of the Verilog "paramset" as defined by the standard. _______________________________________________ Gnucap-devel mailing list [email protected] https://lists.gnu.org/mailman/listinfo/gnucap-devel
