10 May
2024
10 May
'24
9:12 p.m.
From: Fred Cisin via cctalk
[mailto:cctalk@classiccmp.org]
Sent: Friday, May 10, 2024 4:36 PM
To: Stuff Received via cctalk <cctalk(a)classiccmp.org>
Cc: Fred Cisin <cisin(a)xenosoft.com>
Subject: [cctalk] Re: DOS p-System Pascal: (Was: Saga of CP/M)
Sorry Stuff, ain't so.
Bob Wallace wrote the Microsoft Pascal compiler, while he was at
Microsoft. He was their tenth employee. He told me that their runtime
library (which he didn't write) is buggy and slow.
So slow that it made benchmarks with their Fortran compiler (which also
used the buggy and slow runtime library), perform SLOWER than interpreted
BASIC.
But, it certainly did NOT do anything to the disk; certainly no messing
with the FAT.
If you had FAT corruption issues, perhaps you had SMARTDRV enabled with
write cacheing (which did occasionally mess up the FAT).
--
Grumpy Ol' Fred cisin(a)xenosoft.com
I wrote SW with MS Pascal for an accelerometer experiment on the Space Shuttle using DOS
on an 8086 Multibus system (OARE). MS Pascal supported a modifier for procedures named
"Interrupt" that was supposed to be used for Interrupt Service Routines, problem
was they forgot to push and pop the register set before and after entry. It didn't
work very well. IIRC, we had to create an assembly macro at entry and exit to do the
register saves and restores.
DOS, of course, is/was not reentrant, so no DOS services could be used in the interrupt
code. The SW worked well, but was on Columbia when it was lost on reentry. The flight
spares for that program ended up going on the ISS, with a different system I ported to a
PC-104 80486 system in C under VxWorks (MAMS) that had been in use for 20 years until it
was brought back down for refurbishment.
Gary