[Coco] Issues with NitrOS9 build

Tormod Volden lists.tormod at gmail.com
Sat Jan 25 16:04:53 EST 2014


On Sat, Jan 25, 2014 at 9:28 PM, Gene Heskett wrote:
> On Saturday 25 January 2014 15:06:25 Tormod Volden did opine:
>> Robert, if you mean Gene's and your stuff that we are massaging behind
>> the scenes, the make-48-TPI-disk-readonly-on-96-TPI-drive patch has
>> not been pushed to the repo yet because it hasn't been tested. Some of
>> the other (unrelated) changes have been pushed in separate commits. I
>> am /not/ just copying a "new rb1773" wholesale. We already found bugs
>> and unconfirmed changes.
>
> The original patch we sent, has been tested, I have at least 1 of every
> style of floppy ever used on a coco except the original Texas Peripheral
> junk yard dogs.  Those I bin on sight, maybe using the motors for something
> else.  The "bugs" were old ones that existed for yonks. The miss-matched
> spelling goes back at least to v3.2.6.  The format of the downloads changed
> some, and my copy of 3.2.0 is .dsk's only requiring I get DW to work again
> before I can really check them.  I suspect the coco os9's have either been
> using write precomp permanently disabled or permanently enabled for at
> least 20 years.

The bugs I had in mind were the omission of ",u". I guess the testing
was successful even if that was wrong. OK, I am willing to commit that
thing, I just had understood that you were going to test it once you
had your coco running again.

The fixes for the precomp misspelling and more have already been
pushed to the repo. Yes, precomp is disabled by default anyway.

The remaining part of that patch cluster (and what I meant by
unconfirmed changes) is the delay time in FDCDelay. The comments says
minimum 58 us but the current and suggested values make for much more.
I don't know how often this routine is called and if the extra delay
causes any performance drop.

Tormod



More information about the Coco mailing list