Vitus Jensen wrote in a message to Mike Bilow:

 VJ> {at}KEY-WORDS: internet hasch drogen porno Israel Kohl IRA RAF

Trying to attract the attention of international intelligence agencies?

 VJ> For your information: I moved to MS-C 7.0 in 1993 for ADDs. 
 VJ> I got sick of recoding drivers to prevent MS-C 6.00A to spit
 VJ> out "internal error". I would like to move to WatCom
 VJ> (because it is supported), but to write a new dhcalls.lib
 VJ> plus changes in driver code is not an option.

I use the stock DHCALLS.LIB from the DDK with Watcom.  You need to write a
custom ASM stub at the head of your project to account for the differences
in segment alignment, and you can choose to rebuild DHCALLS.LIB from the
DDK source using the Watcom compiler.  MS C settles for DWORD segment
alignment, but Watcom wants PARA segment alignment.

The Bollo freeware library also provides the necessary modules for using
Watcom with device drivers, although he does not specifically support
BASEDEVs or MCA, and this could be a problem -- certainly it would be for
an ADD.

Other than doing odd things such as overrunning the parser stack, I have
never had MS C 6 spit out an "internal error."

 VJ> There are differences between MASM 5.1 and 6, but the change
 VJ> is no real problem. 

I like Watcom a heck of a lot better.  My preferred configuration for OS/2
device drivers is the Watcom compiler with the Microsoft/IBM linker.

 VJ> I agree that moving to a single assembler would benefit the
 VJ> DDK, but until now IBM has only increased the count of tools
 VJ> (MASM 5.1, MASM 6, MS-C 6, CL386, CSet/2, Borland C++ and
 VJ> now ALP).

I agree with you, but there is no tools czar at IBM to fix this.
 
-- Mike


---