:>How did you get the Warp toolkit to work with the Watcom Compiler ?

#define __IBMC__
#define __IBMCPP__

The IBM toolkit essentially has a bunch of defines for CSET/VAC++ and a bunch
of defaults for anything else (that usually don't work :-) - Watcom
understands/translates _Optlink, _System, etc. so just "tell" the toolkit
you're using IBM's compiler. There were still a couple of things that broke
with 10.0 doing this, but 11.0 worked perfectly through two betas and the GA
with the DevCon toolkit (and yes, you can tell the install prog to NOT install
the old toolkit, and it'll find the docs, etc and shadow them into the Watcom
folders now). 

Since nobody has mentioned it: The PowerPuke^H^H^H^HSoft folks may be keeping
"cool extras" from being added to the OS/2 side, but the compiler folks
haven't ignored us in the least. The code generated is noticeably tighter and
faster than 10.x (and makes VAC++ 3.x EXE's look positively Microsoftish
(read: bloated and slow)... yes, I own and like VAC++) - as far as I'm
concerned, they've taken what was already one of the worlds best compilers and
improved it by an order of magnitude - no small feat - and it is, after all, a
COMPILER, not a visual builder or what-not. So, for all that it lacks frills
for OS/2 (and hey, anybody in their right mind doing serious Win* work is
going to shitcan that BlueSky junk bundled with Watcom and buy a real tool
anyway :-), it is IMHO the best damn compiler for OS/2 period (at least until
VAC++ 4 anyway). Upgrades are $129 US... not cheap, but far from prohibitive. 

You might also like to know that they have DLL'd runtimes now also (with no
bogus requirement to rename the DLL I might add :).

Lastly, lest you think the compiler team doesn't care; they were extremely
responsive to myself and at least 2 others I know who participated in the beta
(running OS/2) - no second-class citizen treatment at all. Microsoft... err..
PowerSoft may want us OS/2 people to die and go away, but that's far from the
impression I got from the compiler team.