From: A06012XT@vm.univie.ac.at Subject: Re: Linux rookie questions Date: Thu, 08 Apr 1993 11:48:26 MEZ
In article <C4y7Gn.HyE@news.cso.uiuc.edu>
jaswal@mrcnext.cso.uiuc.edu (Vijendra Jaswal) writes:
>I'm REALLY NOT sure, but I think that Stacker does its compression
>trick by creating a number of partitions. One partition is used
>for directory information and tables of some sort while the
>other partition holds the compressed data. The user sees no
>difference but Stacker does all the translation at a low level,
>so you see only one logical drive.
I'm REALLY NOT sure too, if my version is better, but it seems more logical to
me:
STACKER does not generate partitions. ||| (This is proven, because my linux
fdisk doesn't report any mysterious partitions.)
I think STACKER works by playing around with the CDS.
--- This is only for anybody interested in MSDOS
The CDS is the structure telling DOS what logical drives exist.
(There are entries also for net-drives (except NOVELL, it's got another
trick)
So if you use SUBST or JOIN (or LAN-Manager) this program change the CDS.
There is good book about this subject "Undocumented DOS". If we think about
following situation:
/dev/hda1 -> primary DOS partition
Then we have got a C: drive without Stacker. After stacking with got D: and C:
C: is the compressed drive, and D: the uncompressed one. There is program in
your CONFIG.SYS (The 2nd line specific to Stacker) which invokes a program that
swaps C: and D:, because D: is generated as the native Stacker drive. This is
inconvient to users, so Stacker swaps the CDS-entries for C: and D:
I didn't verify this completly, but there is strong evidence for this scenario.
Andreas Kostyrka (A06012XT@HELIOS.UNIVIE.AC.AT)