On Thu, Jul 06, 2000 at 02:49:00AM +0200, Claus-Justus Heine wrote:
> > Reading is (compared to writing) easy. Writing is a different story; I
> > somehow have to get bad segment information while it's writing (seems
> > like a patch to ftape may be necessary, I haven't really looked into
> > it yet).
>
> Well, that is not really needed but might be convenient. You have
> access to the header segments via the MTIOCRDFTSEG ioctls. Afterwards
> you can parse the data and get the bad sector information.
Ah - I wasn't aware that bad segment information was in the header
segments; I'll be reading up the specs on that - thanks.
BTW, when using MTIOCWRFTSEG, do I have to leave gaps in the buffer
which correspond to bad bits, and when using MTIOCRDFTSEG, do I get
regions in the buffer which correspond to bad bits, and therefore
undefined? It just occurred to be that this might be one of the things
that is going wrong ATM.
> > My plans currently only include reading. So far the decompression has
> > been a right pain; my implementation has been unreliable at best. I'm
>
> Well, the compression algorithm is NOT in the public domain. You might
> get problems when releasing your program to the public.
I assume you mean patents? Do you happen know which countries these
apply to, or who owns them? I live in the UK, which AFAIK doesn't
support software patents (although the EU may soon :-(
Of course, I don't like the idea of running the risk of being sued the
moment I set foot in the US.
Robie.
--
Robie Basak <[EMAIL PROTECTED]>
PGP signature
