"Jim Thompson" wrote in
message ...
On Sat, 22 Sep 2007 10:11:58 -0700, "Mike" wrote:

(2) What are you trying to accomplish by parsing?

All I'm interested in is parsing. The program parses out all the primitives,
subcircuits, instances, and so on, and generates lists and hash-tables of
everything. After that, it's a snap to do things like find a net and report
on everything it attaches to. That can be real convenient if you're trying
to do something like estimate how much driving strength you need for a net.

There are other uses as well. For example, digital synthesizers put buffers
in signal lines. When you're troubleshooting a problem, it's hard to know
whether the signal reaching the output buffer is the right one. Once you
have a parser, you can easily trace back through buffers to find the signal
name at the driver.

I haven't done anything with bipolar in quite a while, but we used to do
lots of stuff with multi-level CML, and the standard problem we'd run into
was connecting the wrong levels together. SPICE wasn't much help with those
mistakes - the circuits would work find in simulation, then fail in silicon.

So, that's what I want to do with the parser.

The only thing I want to do with other netlists is read and parse, to check
the parser with as many different 'standard' SPICE formats as I can. There
are lots of variations, and every schematic / layout / synthesis tool seems
to produce some variation that take special code to parse correctly.

-- Mike --