::>  HB> Is there an FCBS bug with Borland C++ 4.5?

::> Not that I know of; and before you ask, this isn't the FCB bug in OS/2 War
::> that people have encountered.  If you had suffered from that bug, your wou
::> have seen a kernel TRAP and a dirty shutdown.

It wasn't an FCBS problem per se, that was just a side effect...

::>  HB> but when I run the resulting EXE, it traps with a SYS0035 error [...]

::> Usually an error caused by a DOS program performing an FCB file I/O call w
::> be returned to the program in the AX register.  It's odd that this raises
::> hard error and invokes the hard error handler for the VDM.

::> However, one thing that is important is to _use the debugger_ and _find ou
::> what the program is doing_.  It's no good saying "my program has an error
::> somewhere".  Get out the Borland debugger, and find out where
the error is

The problem was that when I compiled the DOS version, I was running out of 
Stack space. The OS/2 compile I had set up a DEF file to increase the stack 
space to 30k. I upped the stack size (and simplifying quite a bit of the 
code and making more liberal use of the heap) and all the strange errors 
disappeared :)

::> Without more information, I can only make one guess, and that guess is tha
::> have passed an incorrectly sized or corrupted buffer to a Borland C++ for
::> library function that makes FCB calls (such as _dos_findfirst).

Actually, all the file IO was done with if/ofstreams and overloaded
> 
operators... I've been trying to stick completely with C++ iostreams with 
this program. It's been good practice :)

Thanks!

#  Herbert Bushong    harchon{at}centuryinter.net       [TEAM OS/2]    
-  Blackbeard's BBS   Intelec: 239:600/0            
+  Fido: 1:19/19      http://www.win.net/eunicecity/stltcc/hbush/

---
   RM 1.31 2508   Cat - small animal when defurred, resembles Chinese food.