Andy Roberts wrote in a message to George White:



 AR>> That is 23 more than A-Z.  But I already know by default OS/2
 AR>> auto-assigned drive designators crap out at Z.  So how can we get
 AR>> OS/2 to auto-assign such a partition?

 GW> You've gone the wrong route. OS/2 can only support drives up to
 GW> drive number 31,

 AR> Humm... That's more than I have ever been able to get it to
 AR> auto-assign. 

 GW> the HPFS file system reserves the first 5 bits of the 32 bit sector
 GW> number to identify the drive, the remaining 27 bits identify the sector
 GW> within the drive, giving the max HPFS volume size of 64 Gig.

 GW> I would expect the drive letter sequence to
 GW> be:-'X','Y','Z','[','\',']','^','_'.

 AR> Perhaps the clue lies in the IFS being assigned.  In my case
 AR> the drives beyond Z were always a CDROM or CDR.  In that
 AR> case they simply failed. 

That last bit triggered a memory here,  about when I was installing Connect on 
_this_ box (which has a whole lot of smaller partitions on it).  At one point
I'd reached the stage of booting to OS/2,  and when I did it burped and
refused to continue.  I realized in there somewhere that my last HD partition
had become drive Z:,  and that the system wasn't seeing the cdrom at all.

Now,  one of the reasons for me wanting to run this stuff in the first place
was that HPFS would allow me to get away from the cluster size nonsense and
use fewer partitions anyhow,  but this has me wondering if there was perhaps
some way of me reaching the cdrom under that setup...

---