home *** CD-ROM | disk | FTP | other *** search
Text File | 1990-06-01 | 1.3 KB | 39 lines | [TEXT/GEOL] |
- Item 9770627 29-May-90 17:19PDT
-
- From: CDA0004 VAR Shana Corp, Don Murphy,IVR
-
- To: MACAPP.TECH$ MacApp Technical
-
- Sub: Potential Failure Problem
-
- Attn: MacApp.Tech$
- SentBy: Wayne Malkin
- Date 5/29/90
- Subject Potential Failure Problem
- From Wayne Malkin
- To MacApp.Tech$
-
- Memo Subject:Potential Failure Problem
-
- For anyone out there like me who tends to ignore the good advice of the people
- that designed and wrote MacApp, here is something to watch for:
-
- If you go ahead and CatchFailures with a global failure handler (even though
- Larry Rosenstein said it wasn't designed for that), make sure the failure
- handler is in a different segment from the CatchFailures call (or in a
- resident segment).
-
- If you aren't careful, you could find your failure handler getting moved when
- its segment gets unloaded. Since CatchFailures saves the address of the
- failure handler procedure, this is a BAD THING. Putting the failure handler in
- a different segment fixes it, since the CatchFailures call will pass the jump
- table address. Putting it in a resident segment is ok, since then it won't
- move on you. There is also a compiler flag {$B+} you can use to force
- CatchFailures to use the jump table address, even if the failure handler is in
- the same segment.
-
- -- Wayne Malkin
-
-
-
-