On Jul 29, 2009, at 7:54 AM, Bill Gatliff wrote: > John Doty wrote: >> That's one symptom of the bug, yes. But the real problem is that the >> front end reduces what is logically a list of attributes to a single >> attribute by arbitrarily picking one. That's wrong. Give them all to >> the back end. Simpler, cleaner, and no barrier to any back end that >> might need the whole list. >> > > Why would the back-end want to be able to discriminate between devices > with the same refdes?
Well, that wasn't really what I was thinking, but consider a heterogeneous resistor pack with different resistances in different slots. I was thinking more in the line of multiple attributes with the same name, not necessarily associated with particular pieces of the package. Might have multiple documentation= attributes, for example, and making a list of documents to merge into the manual is a reasonable back end function. Then there's source=, which is often multiple. Right now, it's handled by the front end, but that's a bug itself. It would be better to handle hierarchy in the back end. Remember that there's common back end code in gnetlist.scm and gnetlist-post.scm (why two separate files?). That means that we can present a backward-compatible interface to the back ends even if we move functionality out of the front end. > > > b.g. > > -- > Bill Gatliff > [email protected] > > > > _______________________________________________ > geda-user mailing list > [email protected] > http://www.seul.org/cgi-bin/mailman/listinfo/geda-user > John Doty Noqsi Aerospace, Ltd. http://www.noqsi.com/ [email protected] _______________________________________________ geda-user mailing list [email protected] http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
