[comp.sys.apple] ProSel 16 SCSI warning

TMPLee@DOCKMASTER.ARPA (01/16/90)

The most recent version of ProSel that I have (5/89 -- I know its not
THE most recent) came with the following warning:

"It seems that the optimizer and perhaps the volume repair in ProSel-16
will not work with the new Apple SCSI card or Apple SCSI driver.
Considerable circumstantial evidence points to the SCSI driver as the
probable culprit.  Regretfully, this will probably not be fixed in the
near future unless Apple will release the source to the SCSI driver for
examination; an unlikely event.  This does not affect the 8-bit programs
Beach Comber 32 and Mr.  Fixit, which work fine on all HDs."

Can anyone shed any more light on this?  Since it MAY have come with an
earlier version (the one I have did come with something called a SCSI
patch) is it still a problem?  Does it relate to the recent 5.0 and
5.0.2 >64K/cross boundary problem that has been discussed?

TMPLee@Dockmaster.ncsc.mil

(Yes, the version I have is legit -- I just forget which one the warning
applied to; I don't have a more recent one because I get mine direct,
having no other reason to pay for a compuserve or other account on the
services he seems to post updates on.)

mattd@Apple.COM (Matt Deatherage) (01/18/90)

In article <900115194155.699954@DOCKMASTER.ARPA> TMPLee@DOCKMASTER.ARPA writes:
>The most recent version of ProSel that I have (5/89 -- I know its not
>THE most recent) came with the following warning:
>
[warning about SCSI and concerns deleted]

This was a problem in the 4.0 SCSI Driver that was fixed by a patch by Glen.
Apple didn't fix it because the 5.0 SCSI Manager was about 2 weeks away from
announcement at that point.  The Manager was totally redone and did not
suffer from this particular bug?

So wait, I can see you folks panicking now - "But if there's problems with
the 4.0 SCSI Driver and the 5.0 SCSI Manager, what am I supposed to use?"

If you ask me, the 5.0 SCSI problem is nothing to lose sleep over; at least
not too much sleep.  (Forgive that semicolon.)  The bug occurs when data over
3 banks is copied - in other words, you start somewhere in bank A, continue
through all of bank B and end in bank C.  This could be as little data copied
as 64K + 2 bytes, but will always happen with 128K + 1 byte or more of data.
When this happens, the Manager messes up the last 512 bytes in the block.

Only the Finder (to my knowledge) is copying data in such huge chunks that this
could even be noticed as a problem.  I personally have few qualms about using
the 5.0 SCSI Manager and I recommend you do the same.

-- 
============================================================================
Matt Deatherage, Apple Computer, Inc. | "The opinions represented here are
Developer Technical Support, Apple II |  not necessarily those of Apple
Group.  Personal mail only, please.   |  Computer, Inc.  Remember that."
============================================================================