home *** CD-ROM | disk | FTP | other *** search
Text File | 1991-03-06 | 1.6 KB | 41 lines | [TEXT/GEOL] |
- Item 8727995 13-Sept-90 00:13PDT
-
- From: MOOF Rollin, Keith A
-
- To: MACAPP.TECH$ MacApp Technical
-
- Sub: RE>DoInitUMemory bug with
-
- Attn: DacEasy, Wilma Blair,PRT
- SentBy: Keith Rollin
- Reply to: RE>DoInitUMemory bug with 2.0p
-
- Les,
-
- I wouln't consider this a “DoInitUMemory bug”. I think that what your ISAM
- vender is doing is pretty weird. DoInitUMemory is doing the correct thing in
- trying to account for that bogus CODE segment; the problem is that the CODE
- segment is too high up.
-
- Ideally, I think that the ISAM routines you add to your program should simply
- be a library you link in along with all of the MacApp and MPW libraries. If,
- for some reason, this post patching is necessary, you should assign a name to
- that CODE segment, and then have your ISAM loading routine should look for its
- extra CODE segment by that name. That way, the ISAM routines are not dependent
- on any specific number, and should work with that segment being numbered 82.
-
- If you can't get this developer in Podunk, Alaska, to fix his routines, then
- by all means, revert to the way DoInitUMemory used to work. Frankly, though,
- I'm in a bit of shock. I simply can't believe that there was someone out there
- RELYING on this bug...
-
- Sigh...I wonder if MacApp should be more defensive about this. It's possible,
- for example, to implement an array of two field records - one word to hold a
- CODE resource number, and one to hold its size/residency/etc. That way, the
- array wouldn't be so sparse and large in odd cases like this.
-
- - Keith Rollin
- - Apple Developer Technical Support
-
-
-