> > (there was a buffer cache problem). Right now, I think a better way to
> 
> Why not pass off the source and see if somebody wants to debug it? I know
> cache problems can be weird but it aught to be solvable. :)

Two things. One, the code's sitting in my private directory on puma.
:-) Second, I really thing hfsfs (Paul Hargrove's one) would be
better as it wupports writing.

> > get hfs support would be Paul Hargrove's hfsfs module, which unfortunatly
> > is GPL'd. I've had lunch a few times w/ Paul, and he'd rather not remove
> > the GPL.
> 
> So what's wrong with GNU Public License? Ok, I know NetBSD has a more open
> code license than GNU but would it be so terrible to have some GPL'd software
> in the source tree?

I'm not really sure. I'm not as strong-willed about the GPL questions
as others.

> > The problem is that he received non-trivial help from a MacOS
> > simulator company (he got to read their source), and they asked him
> > to keep it GPL'd. There's a one-word reason: Microsoft. :-)
> 
> Kind of a weird reason. I'm sure Microsoft could get a couple of guys together
> and write an HFS package if they wanted to. A friend of mine wrote the
> HFS code they use for stuff by DataVis and that's a much smaller company :)

I thik the point is that Macrosloth should have to do it. Plus, as I
understand it, the colating sequence (the part given by the company to
Paul) is NOT what you'd expect. For instance, Some punctuation symbol
comes between 'A' and 'B'. It's THAT goofy stuff, which took
quite a while to find, which the company wants to make Macrosloth
sweat over. :-)

> Are the h* utilities truly free? Couldn't we build a file system off of them
> or is that what you tried to do already?

Paul and I just tried macfs, and looked at hfsfs. I think the H* utilities
are GPL'd.

Take care,

Bill