[Coco] Y2K fixes ever developed or posted?

[Gene Heskett]

On Sunday 14 December 2003 09:26, David wrote:
>On Sat, Dec 13, 2003 at 11:01:09PM -0600, Dave Kelly wrote:
>> ----- Original Message -----
>> From: "John Guin" <johnguin at hotmail.com>
>>
>> > Does anyone know if any programs were ever patched, and if there
>> > was an archive made anywhere?
>>
>> No one that I know of. Br Jeremy was going to post the patch.
>> Don't know if he did. However DeskMate is a horse of different
>> color. All its modules are named something different than OS9. The
>> same technique should work for both.
>>
>> Use your favorite binary editor (dEd) and edit OS9Boot.  Look for
>> the string "19??" ( without the quotes), change to "20??" (without
>> quotes).
>
>Several of the programs used a routine that did sort of a digital
>countdown.  They stored $30 (ascii "0") and then subtracted 10 from
> the year, each time adding 1 to the stored value until a negative #
> was reached. This was the 10's.  10 was added back to the
> date-value and the remainder, of course, would be the one's place. 
> I have a few programs that would print the date for this year
> something like "*3", or some such.
>
>> On second thought, it is in the date module which is in the boot
>> module. So save the date module, edit that and rebuild your boot
>> module. This is the safer way and IIRC the CRC module will not
>> work on a single module merged in with other modules.
>> On the MM1, OS9 has several modules that need patching.
>
>There was (still is, I think) a patch for RBF and dir on Daniel
> Simon's website.  All I did was apply these patches and I modified
> my clock module, but now I am not sure that the clock module even
> needed patching.  There's a "setimex" program on RTSI that you can
> use in place of "setime".  It will do a correct setime.
>
>RE fixing the clock module, both Gene Heskett and I patched our
> clock modules, to do a conversion before storing/retrieving data
> from the RTC chip, but I notice that there's no such patch in the
> os9 project's clock and it works.  Apparently the clock chip can
> store values >= 100.
>
Of the ones I played with, the dallas semi and the mm6242, neither of 
these clocks can handle the rollover at 99>100.  The dallas semi used 
in the B&B XT-RTC did take the rollover, but then failed on the next 
minute IIRC, apparently stopping completely.  The mm6242 apparently 
threw away any effect of the rollover and continued on its merry way 
thinking it was now the year 1900.  I don't recall if I ever 
published the dallas semi clock in any event as it has other, far 
more pressing problems that would prevent its use by me in any design 
I would ever have contemplated.

The seriel shiftng in of the 32 bit access code string took so long, 
with the interrupts locked out, that you could not download anything 
even at a paltry 300 baud without causeing all sorts of grief for the 
zmodem protocol.  The original B&B code got its minute updates from 
the chip on the minute, making it plumb useless for online work. I 
extended the counters in the clock module eventually and put it down 
to boot and once a day, at midnight IIRC, which made it usable 
online.

The disto clock I did got its time on the second, but its access was 
fast enough to not be a problem.  I also repaired the access protocol 
of that chip when it was found to keep very good time turned off, but 
lost several seconds per hour when the coco was running.  That was a 
read the chip protocol error, as they were using a reset to the top 
of the second command to stop the counters long enough to get a 
stable reading.  There was another command that froze the ripples 
without actually resetting the clock, using that stopped the time 
loss when the coco was powered up.

Actually bus glitches at power up occasionally fubared it, so it had 
to be corrected like a forgetfull child from time to time.

>> Microware did not fix the Y2K problem in OS9 until version 3.1 or
>> 3.2.
>
>They issued an upgrade for 2.4 - called it 2.9 which included Y2K
> fixes for several utilities plus a few bugfixes.  However, the
> price was a bit prohibitive.  IIRC, they were asking something in
> the way of $2,000.  I didn't think I needed it _that_ badly.

Sales would have had to been miniscule, or undetectable in accounting 
I'd think.  Sheesh.

-- 
Cheers, Gene
AMD K6-III at 500mhz 320M
Athlon1600XP at 1400mhz  512M
99.22% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attornies please note, additions to this message
by Gene Heskett are:
Copyright 2003 by Maurice Eugene Heskett, all rights reserved.