[comp.graphics] Borland C++ malloc problem with DKBTrace

craig@hpctdkz.HP.COM (/craig lindley) (05/01/91)

Has anyone yet figured out the problem the Borland C++ compiler is having
with the DKBTrace 2.11 code ? Even with the new version of DKBTrace the
problem remains the same. The code compiles correctly but hangs on execution.
It gets hung up in the malloc routine somewhere. The malloc code is in a
tight loop waiting for something to happen which obviously never does. I'm
using the IDE with the large memory model. I've tried changing the mallocs
to farmalloc's to no avail; same thing happens. I know that there is plenty
of malloc'able memory available when the problem presents itself. I also
believe that no requests for blocks larger than 64K are being requested. Any
other ideas on what to try. The code compiles and runs with Microsoft C and
unix C, Amiga C compilers also seem to accept it. Thanks in advance for 
your help.
Craig A. Lindley

Lee Sailer <UH2@psuvm.psu.edu> (05/01/91)

AS just a wild guess, declare the pointers receiving the farmalloc()
pointers to be huge, and cast the farmalloc() results to huge.

                                                              lee