Re: gEDA-user: Life and death for gEDA: portability...

[Bill Cox <bill@viasic.com>]

Sorry about ruffling feathers.  I've been working too hard.  Anyway, I 
want to help gEDA, not hurt it.  One way I might be able to help in the 
short term:

Gnetlist has lots of good netlisters, but for now, gnetlist is weak on 
recognizing some schematic features like instance arrays, hierarchy, and 
busses.  Gnetman is doing pretty well with these features, but doen't 
have many back-end netlists (just the one I'm using for my IC designs).

We could offer all the gEDA users out there a bit more advanced 
netlisting capability with the path:

   schematic files -> gnetman -> flat SPICE netlist -> gnetlist -> 
existing netlisters.

All we need is SPICE reader in gnetlist, which shouldn't be that hard. 
Or, I could write out whatever format is easy for gnetlist to read. 
Alternatively, if I generate some dummy schematics so no changes are 
needed in gnetlist.  That would have the advantage that it makes a good 
start to the schematic generator I wanted to write anyway.  The only 
disadvantage is that it would run slower and generate some extra files, 
but that shouldn't be so bad.

What do you think?

As for porting, guile has proven too hard to support for me...  So far, 
all users I know of who've attempted to use the guile based gnetman have 
asked me for help because of problems related to guile (which they 
percieve as gnetman problems).  So far, no users have asked for help 
about problems related to TCL, and I've even been able to create a 
native Windows port.  With limited time for supporting users, I have to 
go with TCL for now.  Also, it was easier to port to TCL from guile than 
from the old guile API to the new one!  When guile gets to the point 
that I can integrate it as easily as TCL, I'll do it.

So, here's a wish list for binaries that I could use right away.  If I 
can get these, I think I will be able to get our company to switch over 
to gEDA (a small victory: our company has only 7 people!).  Basically, 
here are the linux commands that need to work on either red-hat 8.0 or 
red-hat 9.0.

-- rpm -i geda_rh8_<date>.rpm
-- rpm -i geda_rh9_<date>.rpm

The RPMs should include all the other stuff, like symbols, docs, 
examples, and local copies of any .so file needed.  Shared libraries are 
great for saving memory, but let's not worry too much about disk space.

Also, if anyone has the binary executables of gEDA for Windows, that 
would be great.  We need it!  Running on top of cygwin is ok.  It would 
be nice if the geda packages included gwave.

If anyone builds such RPMs, our little company can help test them.

Bill