Re: gEDA: Icarus Verilog progress

[Stephen Williams <steve-at-NOSPAMicarus_dot_com>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Tom Hawkins wrote:
| Stephen Williams wrote:
|
|>
|> Or "This is my excuse for being unresponsive to all those
|> recent bug reports.":-)
|>
|>
|> The main trunk of CVS is starting to show the sense of the major
|> changes I've been working on. I've got assignments and behavioral
|> code back to functional, and the ivl_target API changes for signal
|> types seems to be working. You can start to see the new style that
|> is coming of these changes. These are the sorts of changes that
|> effect the people working on ivl_target code generators.
|
|
| You're making this too easy for us.  I'll hardly have to make any
| changes to the FNF generator.  :-)

Not as easy as you think, maybe. ivl_signal_t objects now always
have exactly one pin, and the ivl_signal_pin methods are gone. That
is a change with significant implications. Also, the various LPM
devices that used to have lots of data and datab pins now have
one each of data and datab ports. It's really amazing how much
cleaner that makes a lot of things:-)

The good news is that these changes will probably allow you to
remove more code then you add. At least that's what I'm finding
in the tgt-vvp directry.

| So bit selection is performed by IVL_LPM_PART.  Does ivl_lpm_base point
| to the MSB or the LSB of the new vector?

Uh, LSB. Internally, Icarus Verilog always treats vectors as
little-endian.

| Also, is concatenation in the mix yet?  I didn't see it listed as a new
| LPM.  Or is it handled elsewhere?

Not yet. You seem to have correctly surmized that these changes
imply the new need for a concatenation operator node. I haven't
got there yet, although l-value concatenations are happening
already. In fact, I've been spending a lot of time recently on
assignments and their l-values. mostly behavioral stuff, actually.

As I'm going, I'm also improving and filling out ivl_target
documentation in the ivl_target.h header file. That's going to
grow to be very much more huge comments then code, but that will
be *the* documentation for the ivl_target API.

Also, the tgt-stub target is going to grow into an internal
consistency checker, so it too will encode more knowledge of
how the ivl_target API works.
- --
Steve Williams                "The woods are lovely, dark and deep.
steve at icarus.com           But I have promises to keep,
http://www.icarus.com         and lines to code before I sleep,
http://www.picturel.com       And lines to code before I sleep."
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFBwmiErPt1Sc2b3ikRAkQaAKCa9Ls1POEcj1RpOulJNcyxaya/4gCfWBpo
XsolHFA666ttCBYHBZYWppg=
=9NKL
-----END PGP SIGNATURE-----