Re: gEDA-user: gnetlist BUG?

[Ales Hvezda <ahvezda@seul.org>]

Hi,

>I'm using version 20010304
>
>There seems to be a bug in it, it no longer handles buses (it doesn't just
>ignore them).


	Well both behaviors are bugs (handling them and ignoring them).
However, ignoring buses seems to be the lesser of the two evils right 
now.  So I have gotten off of my lazy bum and fixed gnetlist so that it 
ignores buses again.  :-|

	This bug fix is in cvs but if you have 200010304's source
handy you can fix it now by applying this patch:

*** s_traverse.c.ORIG   Fri Jun 15 21:11:30 2001
--- s_traverse.c        Fri Jun 15 21:12:48 2001
***************
*** 434,440 ****
  #endif
  
        if (!c_current->other_object->visited &&
!           c_current->other_object != o_current) {
  
          nets =
            s_traverse_net(pr_current, object, nets, c_current->other_object,
--- 434,441 ----
  #endif
  
        if (!c_current->other_object->visited &&
!           c_current->other_object != o_current &&
!           c_current->other_object->type != OBJ_BUS) {
  
          nets =
            s_traverse_net(pr_current, object, nets, c_current->other_object,


in gnetlist/src to s_traverse.c.  Put the above into a file and do:
patch < patchfilename

	
>
>When two nets are connected to a bus it shorts them together...
>
>Is there a fix, except to write a program that removes all bus entry's from
>a schematic file, and has it been written? (else I will :)


	That's the other more drastic solution.  I have been looking at 
gnetlist's code and seeing how bad getting "real" bus support into the
existing code will really be.  I will probably have to rewrite my traverse
routines quite a bit.  

	Either way try the about patch out and see if works for you.
I have a good chunk of this upcoming weekend reserved for work on gaf,
so I might get some other stuff done too. :)

								-Ales