[EMAIL PROTECTED] (Stuart Brorson) writes:
> Before I stick the patch in CVS I thought I would ask:
> What is the goal of this change? It looks like a
> condensation/clarification of existing code, but does it also change
> any behaviors of gnet-PCB?
i will have to ask you to kindly not simply stick the patch in CVS.
never do that (especially w/ my code)! instead, here is what you can do
to answer your questions:
- characterize the change
- what procedures are involved?
- have the calling sequences (arglists) of these procs changed?
- if so, have all callers been updated as necessary?
- are there any side-effects?
- have the side-effects changed?
- try out the change
- does the proc:
(define (try nets)
(equal? (OLD-IMPL nets) (NEW-IMPL nets)))
always return true?
- what value(s) of `nets' would cause `try' to return false?
i hope as a result of these experimentations you will agree w/ my
calling the patch "iso-functional". (this way, i can use that term in
the future and we will have enhanced shared-context. :-)
> Also, does ice-9 come with all guile installations? I did some quick
> Googling, and it looks like it does, but before we commit this change,
> I would like to make sure that it doesn't introduce a new dependency.
> OTOH, since ice-9 looks like it has a lot of useful stuff, perhaps
> making it available for other netlisters would represent a useful
> addition.
many modules named (ice-9 FOO) are usually installed w/ guile. the
particular module (ice-9 format) is quite stable, as far as i can tell.
it certainly makes the code more concise once you learn some of the its
idioms[0]. (one could even define netlisting as an exercise in finding
the right `format' string. ;-)
thi
[0] here is a link from http://www.glug.org/docbits/
(see also section) you might find interesting:
http://www.supelec.fr/docs/cltl/clm/node200.html#SECTION002633000000000000000
