Re: [cc65] dio_query_sectcount() on Atari ?

Hi,

>> > I think is query the percom block, and if it's not present, just
>> > return 0. So, it's
>> > up to the calling application knowing the best method to access the
>> disk.

>> How much code would be necessary to additionally detect DOS 2.5 720/1040
>> disks?

> It can be done (it involves looking for the VTOC in the upper sectors and
> then the one in the lower sectors),

I see.

>  but that ignores all the "raw" formatted
> disks that a programmer may want to query the sector size on. I don't think
> we should make assumptions that all DIO access is going to be on DOS
> formatted disks.

The approach only makes sense if it can detect with a let's say 99%
confidence that it doesn't mistake a "raw" formatted disk for a DOS
2.5 disk.

> What's wrong with the solution of querying for the percom block, and if that
> doesn't exist, fall back on the brute force binary read method?  That will
> always return you a max sector size for any disk.  I don't like the solution
> of returning zero personally... that's so counter-intuitive, more so than
> reading the disk to determine the max sector (imo).

Thanks for bringing this up. There is probably a misunderstanding:

My question above regarding DOS 2.5 disk detection is only relevant
for the scenario of looking at percom and returning 0 in case of
failure. In that scenario I'd ask for an additional DOS 2.5 disk
detection.

I personally don't argue against the approach of looking at percom and
do a binary size search in case of failure. And in that case an
additional DOS 2.5 disk detection is most likely not reasonable
because of
- code size overhead
- risk of mistaking a "raw" formatted disk as DOS 2.5 disk

If I understood things right than especially hard disk drives are
supposed to support percom so a binary size search is supposed to only
be necessary on comparably small floppy disks (?) Maybe one could
invest a little code into specific support for "classic" floppy drives
by checking sector 1040 and 720 for failure first before starting the
actual binary search (?)

Regards,
Oliver
----------------------------------------------------------------------
To unsubscribe from the list send mail to majordomo@musoftware.de with
the string "unsubscribe cc65" in the body(!) of the mail.