ephraim@think.COM (ephraim vishniac) (02/02/88)
Replies to two items from the Delphi Mac Digest, volume 4, number 2: VASMUG asks: > 3. Does anyone know what happened to ZAP Computer products? > They made and marketed ZAP harddrives last summer. We > would like to buy some more. Number now disconnected! Thanks for all > your help! > Fred ZAP sold out their products to some other company - I don't know who. I was under the impression that they never did sell any disks under their own name, since they only advertised in Icon Review :-). At one time, I was negotiating with ZAP for them to use my SCSI software. Could you do me a favor? Do a "Get Info" on one of your ZAP drives, or check the title screen of the disk formatter, and tell us all what it says? If the Get Info says "Where: Ephraim's SCSI driver" or if my name is in the copyright notice, it's time to call my lawyer... TONYN asks: > Subject: Disk Initialization of foreign drives > I am the author of RamStart. I am trying to make RamStart's RAM > disk compatible with the Disk Initialization package so, eg, Finder > Special menu Erase Disk will work. What does my driver need to do > to make Pack 2 happy? Where does Apple document it? Also, it might > be an advantage to my users to be able to make (at will) either an > HFS or MFS RAM disk. How do I control this, when a RAM disk can be > nearly any size at all? The most popular bug in device drivers is that they don't update dCtlPosition in the DCE. This only affects calls to the driver when relative positioning is used. The file manager always uses absolute positioning, but _DIzero uses relative positioning to sequentially erase the sectors of the volume directory on MFS disks. I don't know whether it uses relative positioning when initializing HFS disks, but it seems likely. The need to update dCtlPosition is documented in Inside Mac (in the Device Manager chapter, or perhaps the Disk Driver chapter), but it's just one sentence buried in a long paragraph. BTW, every ramdisk I've ever seen for the Mac (except the one I wrote for MassTech Development Labs, R.I.P.) has this bug. It took me a couple of weeks of experimentation before I hit on the idea that the problem was with a particular positioning mode. The only way to force a large ramdisk to be MFS is to build your own volume header. The sample ramdisk that was (is?) part of the Aztec C distribution did this. It's strictly against the rules, of course, so caveat emptor. I don't know any way to force a small volume to be HFS. Ephraim Vishniac ephraim@think.com Thinking Machines Corporation / 245 First Street / Cambridge, MA 02142-1214