🚀 go-pugleaf

In article <adtDFBIvC.5p9@netcom.com> adt@netcom.com (Anthony D. Tribelli) writes:
> Jerry Shekhel (jerry@cybercom.net) wrote:
> : Dmitry V. Irtegov (fat@cnit.nsk.su) wrote:
> : : I participated in development of small application on Win16.
> : : Win16 uses handles too - I wonder, why: 286 segments have
>                  			      ^^^ typo #1 ?
No, se below.
> : : some form of VM. Officially handles are needed to maintain
> : : compatibility with 8086 real-mode, but Win3.1 doesn't support
> : : real mode at all...

> : Right.  In Windows 3.0 and above, unlocking GlobalAlloc()'ed blocks isn't
>                      ^^^ type #2 ?
No, se below.

> : necessary, because the OS can move them around even if they are locked...

> Was "286" a typo? I don't think Intel supported VM until the 386.

Intel supported VM since 286 (TWO-Eight-Six). Period. 286 had segmented VM, 
which has some disadvantages in comparison with paged VM, as implemented
in 386 and above, but it definitely is a VM. Protected mode of 286 offered 
16Mb of physical memory and 512Mb of virtual one, though 286 hasn't good 
"virtual 86" mode; so protected mode OS couldn't run real mode programs 
without physical switching to real mode. It was a main problem of OS/2 1.3 
and Xenix286. So most 286 were used to run in real-mode only, as fast
replacement for 8086.

> Was "Windows 3.0" a typo? I don't think you could:
> lock and forget about unlocking until you are ready to free
> until Windows 3.1, when running in enhanced mode (not real or standard). I
> don't remember 3.0 having an enhanced mode (standard + VM ?). 

Standard mode of 3.0 and 3.1 works in 286 protected mode. Technically it 
is posible to implement "lock and do not unlock" in 286 protected
mode; only real mode have problem with it. It was done in OS/2 1.x 
and Xenix286. I wonder why it wasn't done in Windows.

Cheers,
	Fat Brother.

Fatal Error: Couldn't enter Windows, use Doors.