Odd behavior for GETQUI-related operations

[The_Doc_Man]

I'm going to toss in that extra F$GETQUI( "CANCEL_OPERATION" ) since the users will be captive to a script that after defining some other stuff, runs the code in question.  I can sneak in the lexical in the pre-amble to running the actual program.  Thanks for your research, Hein.  I was beginning to doubt my own sanity.  (Of course, the guys around me don't doubt that I've used up my sanity quota long ago...)  As wordarounds go, an extra lexical is not a high price to pay.

[The_Doc_Man]

I tried to post code snippets but it blew up on me.  I'm having some trouble with this interface and because it is on a government system, I cannot chose an alternate browser without first purchasing Archimedes's Lever (the lever from an Archimedes comment that given a big enough lever, you could move Heaven, Hell, and Earth).

Environment:  OpenVMS 8.4 on rx2800 i2, fully patched through 2015Q3, the last "big" patching being, I believe, VMS84I_SYS v 6.0 and a LIBOTS patch that was also recent.  Some of the patches of category 3 are not installed because we don't run the particular program.   We have the newest TCPIP (5.7-13ECO5) and the TELNET PAT.  We have recent LMF, LDAP, SSL, FIBRE_SCSI, UPDATE 10.0, etc.