Re: gEDA-user: Multiple user libraries, parametrized components

["a r" <nbs.public@xxxxxxxxx>]

On Jan 19, 2008 7:53 PM, a r <nbs.public@xxxxxxxxx> wrote:
> On Jan 19, 2008 6:35 PM, John Griessen <john@xxxxxxxxxxxxxx> wrote:
> > a r wrote:
> > >> [jg] Library search path is a good concept for dealing with this.
> > >> Proven in chip design setting to be low enough hassle, flexible.
> > >
> > > Can you tell me how to set this path? That's what I'm looking for.
> > >
> >
> > [jg]It's just a concept so far -- not implemented.
>
> OK.
>
> > >>> Also, is it possible to refer to an actual cell name / library name
> > >>> etc. from the symbol?
> > >> [jg]Sounds too complex, labor intensive.
> > >
> > > How component name conflicts are resolved then?
> >
> > First one found wins.  If it was done yet.
>
> One very typical use pattern is to copy whole library to another place
> (for backup, for reference for later changes). Without an explicit
> cell/library names this is asking for trouble.
>
> I understand that using explicit library names can be labour intensive
> (e.g. every time a cell/library is copied to another location its
> references need to be updated). How about a configuration editor then?
> A tool that can override library search order, or even permit explicit
> setting "this time take this cell from lib_B not from lib_A".

After thinking about this concept a bit more I came up with some more ideas:

1. Library search path, or rather some kind of a project file that
contains a list of libraries, is a good idea.
2. Component browser (and file open dialog) should display all defined
libraries and all cells from these libraries should be accessible.
3. Cell name conflicts should be resolved by:
a) searching in the current library (i.e. one that contains a
referencing/parent cell),
b) if (a) fails, by searching in other libraries in the order defined
in the library search path or a project configuration file.

Why not a simple library search path and why not explicit library names?
It is very often necessary to compare/edit same cell from two
different libraries (say, a different version, or a library that
someone else was working on). However, I agree that using explicit
library names in 90% of use patters would be an overkill and would
require additional tools/procedures (like renaming instances in copied
or moved library).

Also, I would rather avoid using separate scripts that make links in
the project area. This is going to be hard to maintain and does not
really do anything that couldn't have been done by a modification of
gschem/gnetlist itself.

-r.


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