Re: gEDA-user: Comments after successfully running gnetman . . . .

[sdb@cloud9.net (Stuart Brorson)]

Hi --

> >*  Bill attached the "source" attribute to the symbol, while gschem
> >only allows the "source" attribute to be attached at the schematic
> >level.  Therefore, gschem is not handling Bill's attribute correctly
> >-- "source" isn't showing up in the symbol at the schematic level.  I
> >propose to fix this by making "source" a "symbol and schematic"
> >attribute.
> 
> 	source= is a symbol attribute.	I'm confused, how can you specify
> underlying schematics from within an uninstanciated symbol?  Unless,
> source= is being used for something else?  Are these symbol source=
> attributes meant to be some sort of default?  I haven't found a moment
> to really look at gnetman yet.

Bill's hierarchical schematic examples use "source" within a .sym file
to point to a schematic holding a SPICE subcircuit.  The idea is that
it makes sense to associate (i.e. hardcode) a pointer to the
subcircuit within the symbol file.  Right now the user is forced to
attach the source file name as an attribute every time the symbol is
placed on the schematic, which can be a PITA if you instantiate the
same subcircuit multiple times.  

As long as the symbol is specific to a particular part (or
subcircuit), carrying the subcircuit pointer on an uninstantiated part
is sensible.  Since most IC symbols are device specific (e.g I tend to
have separate symbols -- with unique "device" attributes -- for an
OP177, an LF356, and a uA741, even though they are all op-amps),
hard-coding the source into the symbol file is sensible.   

At least, that's my reading of the situation.  I'd appreciate it if
others chimed in here.

> >*  As for accessing the underlying files, being able to do it through
> >the right mouse button in addition to the hierarch menu would be a
> >good thing.  (Yes, it is also already available through a keyboard
> >shortcut, which is very nice.)  I can look at gschem and try to figure
> >out how this works . . . .
> 
> 	Right now the right mouse button popup menu is hard coded in
> gschem/noweb/x_menus.nw  I might change that to a scheme variable which
> you could change mostly on the fly.  I don't use the right mouse button
> menu much nowadays.  I perfer having the mousepan attached to that button,
> but unfortunately the menu was written before the mousepan, and I won't
> change the default.

Actually, one can use the keyboard shortcuts to dive into hierarchy
the same way as the right mouse button.  I am just used to doing it
with the right mouse button after having used Viewdraw for several
years. . . . .  But anyway, having the ability to dive into a
schematic or symbol via a pop-up meny invoked by the right mouse
button is very intuitive, and is accordingly a good thing, IMHO.    

Stuart