JDBP>> Indeed, I explicitly made the point that the OS/2 kernel itself 
 JDBP>> makes *no* distinction whatsoever between "text mode" and 
 JDBP>> "graphical" processes.  This was the whole point that I was making. 

 MR> Part of the priority table is something called "window boost", which
 MR> near as I can tell, serves no purpose other than to give a priority
 MR> boost to a graphical program.
 MR> [...]
 MR> So, in that limited respect, the kernel may be considered "aware" of a
 MR> difference between text-mode and graphical programs, though it's an
 MR> awareness that can be removed with a hex editor.

Actually, that doesn't mean that the kernel is aware of the difference between 
textual and graphical programs.  The "foreground priority boost" was in the
very first 16-bit version of OS/2, well before Presentation Manager was
introduced and thus well before there even were such things as graphical OS/2
programs.

This boost gives an increase in priority to "the foreground process".  But the 
kernel does not determine what "the foreground process" actually is.  It
leaves that decision up to others.  All that the kernel knows is that one
process is singled out as being the process that is "in the foreground" and is 
as a result to be given preferential treatment when it comes to doling out CPU 
time.

I suspect that it is the session manager that determines what the foreground
process is.  The 16-bit OS/2 CP reference describes the foreground process as
the process that "has the input focus", so I suspect some collusion by the KBD 
subsystem in the making of this decision as well.  Unfortunately, that is
pretty much all that is said anywhere on the subject.

 ¯ JdeBP ®

--- FleetStreet 1.22 NR
114/477
143/1