🚀 go-pugleaf

This has been an interesting discussion to watch.  It was my impression from
the work I have done with several packages that use DIAG58 to handle full-
screen I/O that anything outside of PA1 is reflected directly to the program,
and PA1 can be handled by setting a bit in the CCW.  In fact, there is one
package I work with directly in which this was precisely the problem; if they
had control of the screen and the user (inadvertently?) pressed PA2, the
user was immediately sent to the CMS subset, which made me particularly IRATE.
I complained to the software vendor, who sent me a zap (which, being the
hacker I am directed me to the area where all this stuff was done...) and I
found that they merely checked to see whether PA2 was pressed, and issued the
SUBSET command to send the user to the CMS subset.  Bypassing this check
allowed the software to treat the condition as if the user had pressed the
CLEAR key, which is effectively a no-op in this situation.  Is there something
else I am missing here????


R;