Hi All,

The 32-bit Windows port of timEd sort of works, but I'm getting
unexplained segfaults at certain stages during execution that seem to be
caused by code that was executed during startup.  Possibly some sort of
stack corruption.  There is some code in config.c that looks like it
could be the culprit, eg. there are calls to function pointers where a
different number of parameters are sent to the function depending on
what is in the config file.  Really dodgy stuff.  The compiler is
flagging these as "call to function with no prototype".  But I don't
want to rewrite all that code just to find out the bug isn't there. But
if I just comment out that code, timEd won't read the config file and I
can't test it.

I'm thinking maybe I should switch to MSVC to see if I can replicate the
bug, or maybe narrow it down.

My time is a bit limited, so if there are any Windows coders reading
this who'd like to have a go at debugging the fault, grab the free
version of the Windows Borland C++ compiler [1] and do a checkout of the
timEd [2] and XMSGAPI [3] CVS tree, then run "make -f
makefile.win32.borland" to build it.  You might want to download the
free version of Turbo Debugger also, then add "-v -y" to the COPTS part
of the makefile, then you can step through the timEd source code during
execution.

[1] http://www.borland.com/products/downloads/download_cbuilder.html

[2] cvs -r -z3 -d:pserver:anoncvs{at}blizzard.dnsalias.org:/cvsroot co timed

[3] cvs -r -z3 -d:pserver:anoncvs{at}blizzard.dnsalias.org:/cvsroot co xmsgapi

Regards
Andrew

-- mail{at}ozzmosis.com

--- timEd/FreeBSD 1.11.b1