NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/comp/sys/mac/programm/20161

Internet Message Format | 1992-12-22 | 1.7 KB

Path: sparky!uunet!olivea!spool.mu.edu!agate!boulder!ucsu!rintintin.Colorado.EDU!steelem
From: [e-mail] (STEELE MARK A)
Newsgroups: comp.sys.mac.programmer
Subject: THINK C bug??
Message-ID: <[e-mail]>
Date: 22 Dec 92 00:12:48 GMT
Sender: [e-mail] (USENET News System)
Distribution: na
Organization: University of Colorado, Boulder
Lines: 21
Nntp-Posting-Host: rintintin.colorado.edu

A few days ago a thread appeared questioning a bug in Think C on a Quadra.
I replied to the author that I hadn't yet experienced this problem. I have
yet to see any more posts about this, however this past weekend I turned
on the profiler and have stdout redirected to a file so I could review it.
Here is what I encountered: The file that was open before running the program
to be profiled would become corrupt. A #define at the beginning of the file
suddenly became #ddddd folloed by some garbage characters. the last line
of the file became duplicated, since it contains the closing bracket, this
was an error. I clean this file up and reran the program, the debugger would
hang at this point. I then rebuilt the project by removing all objects and
discovered that some headers would be too big for the current segment.
These headers are supplied by symantec so I would think that they are ok
and the project has just been corrupted.
Has anybody else with a Quadra and Think C 5.0.4 ran into this problem?
Remember that I have only had this problem when the profiler has been turned
on and I run with the debugger. The problem occurs after running the app a
few times. I have run a memcheck program that checks out, I have also run
Norton Disk Doctor on the HD to look for problems, none found.

-Mark
[e-mail]