🚀 go-pugleaf

In article: <923779875.1797.0.nnrp-06.d4e48b20@news.demon.co.uk>  
alisdair@nospam.demon.co.uk (Alisdair Wren) writes:
> 
> Roger@ttelmah.demon.co.uk uttered thus:
> >In article: <923681977snz129.os2.7@sunshine.tm>  {R}@sunshine.tm 
> >(Richard Ashton) writes:
> >> 2b) 2GB? There is a specific maximum DOS partition size caused 
by ?
> >FAT16. Uses a _signed_ integer, to address the allocation units. 
> >Allows 32768 units, each up to 65536 bytes in size. This gives 
2GB. 
> >In fact NT can use an unsigned integer to give 4GB, but DOS never 
did 
> >this (a pity).
> 
> That's not quite correct.  64KB clusters are illegal in normal DOS 
> partitions, only NT can handle 64KB clusters.  Using what you call 
a 
> "signed integer" would be pretty absurd [clearly, you're never 
going to 
> end up in a negatively addressed section of the hard disc].
> 
> Both DOS and NT can address 65535 clusters [2 byte value].  DOS can 
> handle 32KB clusters, NT can handle 64KB clusters.
Yes. The 15bit value is on the size, not the number, got them 
reversed. However it _is_ defined in the 'C' includes for DOS 
development by MS, as a _signed_ integer, which is why only 15bits of 
the value can be used. The same include files for NT, define the 
value as unsigned. I agree, the 'idea' of having a 'sign' on a size 
does seem rather strange, but there are several other similar 
examples in MS stuff...

Best Wishes

-- 
--------------------------------------------------------------------
| EMail roger@ttelmah.demon.co.uk  http://www.ttelmah.demon.co.uk/ |
| A beard! A beard! cried Fly Nicholas.'By God, that's a good one!'|
| (Chaucer)                                                        |