BO>  >I was under the impression (maybe the wrong impression) that when you
BO>  >included OS/2.h it included all the .h files in the os/2 h directory
BO>  >that your program need to compile.
BO>
BO> Hi, Carl.
BO> Right before the "#include " statement in
your source file,
BO> you can put different #define directives to control which headers
BO> are included, depending on what services you program will use.
BO>
BO> #define INCL_BASE
BO> #include 
BO>  .
BO>  .
BO>  .

Can I comment on this? Well, you've touched a nerve, so I'm going vent
my spleen anyway.

Basing all headers off  and then including the relavent parts
via macros is one of the most horrible things IBM has done for (to?)
developers. (Actually, I'm pleased with most of their other efforts...)
I've been writing some C++ wrappers for system services and using one
header file for each class. The dependancies between the _otherwise_
_unrelated_ class headers is horrible, since the first one included
gets 'what it wants' out of OS2.H and the rest of the class headers
that include OS2.H get what the first header specified. So otherwise
unrelated classes clash horribly at compile time whewn they're put
together. Strike a blow against partitioning. Oh, and if you include
, it will also include  without otherwise defining any
macros.

I used to dislike the appearance of line after line of system headers,
but now that I see the alternative, They don't look so bad.

 þ KWQ/2 1.2g NR þ Look Ma, I'm a *Beta Tester*: KWQ Mail/2 for OS/2 v2.X!


--- WM v2.04/92-0551