Siblings of #5119: BLL cell idioms (NULL compare, read-command SET(cell) end-to-end) #5197

Siblings of #5119 (class: OS/VS COBOL BLL cell idioms lowered onto the pointer lane, register X56). Found in the review of PR #5196, which modelled chained storage areas and SET(cell) on the read commands. The members PR #5196 doesn't cover:

• A cell compared with ZERO / NULL (IF USERPTR = ZERO, IF NEXTAREA = ZERO on a chain's end): still refused, because the det translator refuses NULL in any condition (det/expr.py:207). This is a different mechanism from the lowering. IBM's p. 22 example tests a data field instead, but real chain walkers test the pointer. Refused by name today.
• End-to-end proof for READ / READ UPDATE / READNEXT / READPREV / RECEIVE SET(cell): these are translate-level plus CICS SET(ptr) pointer-lane siblings: RECEIVE, RETRIEVE, READQ TS, GET CONTAINER, READ/READNEXT/READPREV (Part of #5042) #5175's step lists only. A proven case needs a KSDS file (CSD + IDCAMS definitions, generated entity).

Already tracked elsewhere: LOAD PROGRAM SET (#5042), ADDRESS CWA/TWA/TCTUA/CSA into a cell (#5079), scanner era facts (#5117), estate-crucible S4-S6 (#5118).

Siblings: #5119