Re: Let's stop feeding the NVidia cuckoo
* Justin Pryzby (email@example.com) [050225 22:35]:
> On Fri, Feb 25, 2005 at 04:23:07PM -0500, David Nusinow wrote:
> > I've just taken a quick (~10min) look through it. It's definitely
> > readable, and makes sense for the most part as far as I could see.
> > It's got comments and is fairly cleanly written. The caveat is that
> > there are a lot of magic numbers scattered about the code. Some are
> > commented (such as specific chip ID's) and others are not, since
> > they're things like bitmasks. I quickly looked through the ati
> > driver code as well, and while it seems to have significantly less
> > of these, those that are there are nicely commented telling you
> > where they came from. No such niceties in the nv code. Still,
> > nothing that would make me call it obfuscated.
> > I'll see about taking a closer look at parts to see if it actually
> > makes sense, but so far it looks fine to me. As it is, I don't see
> > any difference between this and any other vendor not releasing
> > hardware specs and yet a Free driver exists. Not a good thing, but
> > not non-free either.
> Well put. I think it is arguably not "source code", however, if the
> source we are seeing is the result of some sed-like script which
> converts a sort of custom #defined MAGIC_NUMBERs to id numbers, and
> then removes the #definitions.
Is there some proof that the files are created that way, or is this just
PGP 1024/89FB5CE5 DC F1 85 6D A6 45 9C 0F 3B BE F1 D0 C5 D1 D9 0C