equivalence: cases for the CBSA programs made whole (Part of #4270)

docs/language_status/register/C12.md

@@ -5,7 +5,7 @@ status: "DIFFERS (the numbers) / ASSUMED (the interface)"
 title: "FUNCTION RANDOM — the numbers DIFFER from z/OS, the interface ASSUMED; refused where IBM does not allow the seed"
 area: "compiler"
 summary: "FUNCTION RANDOM: the oracle's generator (glibc via GnuCOBOL), not IBM's unpublished one; a seed IBM does not allow refused"
-reached: "translated, no proof yet (CBSA CRDTAGY1-5, INQCUST; GenApp LGICVS01)"
+reached: "yes (CBSA CRDTAGY1-5, INQCUST: cbsa-crdtagy1..5, cbsa-inqcust, seeds 0 and a stated `\"taskn\"`); translated, no proof yet (GenApp LGICVS01)"
 ---
 - **IBM** (Enterprise COBOL 6.4 Language Reference, RANDOM,
   https://www.ibm.com/docs/en/cobol-zos/6.4.0?topic=functions-random): argument-1, if given, "must be zero or a

docs/language_status/register/L3.md

@@ -5,11 +5,16 @@ status: "ASSUMED"
 title: "WORKING-STORAGE with no VALUE clause — ASSUMED"
 area: "LE"
 summary: "WORKING-STORAGE with no VALUE clause: GnuCOBOL's spaces vs LE's STORAGE option on z/OS"
-reached: "yes (CardDemo READACCT OUTFILE, 2 bytes)"
+reached: "yes (CardDemo READACCT OUTFILE, 2 bytes; CBSA CRDTAGY1-5 container-short, WS-CONT-IN past a short container)"
 ---
 - **What.** GnuCOBOL initialises such an alphanumeric item to spaces. On z/OS its first contents follow the LE
   runtime option STORAGE (commonly binary zeros, or whatever the storage held): the program's source does not say.
 - **Where it shows.** CBACT01C moves the 10-byte `CODATECN-0UT-DATE` to `OUT-ACCT-REISSUE-DATE`, but COBDATFT (A1)
   writes only its first 8 bytes and `CODATECN-REC` has no VALUE clause. The last 2 bytes of each OUTFILE record's
   reissue date are GnuCOBOL's spaces, and the Java side writes the same.
-- **Assumed** to be spaces. A z/OS run with STORAGE(00) would give X'0000' there.
+- **Where it shows (CBSA, #4270).** CRDTAGY1-5's `WS-CONT-IN` (261 bytes, no VALUE clause) is filled by GET CONTAINER
+  INTO; a container shorter than 261 bytes (scenario `container-short`) leaves the rest as it started, and PUT
+  CONTAINER writes all 261 bytes back: GnuCOBOL's spaces in the alphanumeric items and its zeros in the unsigned
+  numeric DISPLAY items (date of birth, review date), the same bytes on the Java side.
+- **Assumed** to be spaces (zeros for a numeric DISPLAY item, as GnuCOBOL initialises it). A z/OS run with
+  STORAGE(00) would give X'0000' there.

docs/language_status/register/Q3.md

@@ -11,6 +11,13 @@ reached: "CICS: yes (CBSA XFRFUN); batch: yes (synthetic UOWDEMO, `tests/equival
   the task ends, rolled back with the task's recoverable files by SYNCPOINT ROLLBACK or an abend (X3). The transaction
   is the harness's (`equivalence_cics.py`, a `TransactionTemplate` around the task); a deployment must give each task
   the same unit of work. Reached: CBSA XFRFUN's four ROLLBACK paths are proven.
+- **CICS, SQL after a SYNCPOINT ROLLBACK (#4270).** IBM: SYNCPOINT ROLLBACK ends the unit of work and the task's next
+  recoverable change starts a new one, committed when the task ends. The Java side rolls the task's Db2 connection back
+  at the SYNCPOINT ROLLBACK and goes on in the same transaction, so the SQL that follows is committed with the task, as
+  the COBOL side's Db2 commits it. Reached: CBSA INQACCCU's cursor failing inside CREACC's task (its SYNCPOINT
+  ROLLBACK, then CREACC's CONTROL updates and INSERTs: cbsa-creacc, cbsa-bnk1cac). A recoverable FILE changed after a
+  SYNCPOINT ROLLBACK is still undone with the task on the Java side (the transaction is marked rollback-only): no case
+  reaches one, and it would show as a difference.
 - **CICS, EXEC SQL COMMIT / ROLLBACK.** IBM: both are invalid under CICS -- Db2 for z/OS answers -925 / -926 (SQLSTATE
   2D521; Db2 for z/OS Codes) and the task's unit of work is CICS's (SYNCPOINT). The det translator REFUSES a CICS
   program's ROLLBACK by name. A CICS program's COMMIT stays a reset of the SQLCA on the Java side and a real commit on

docs/language_status/register/X24.md

@@ -5,7 +5,7 @@ status: "ASSUMED"
 title: "A task started with a channel; the containers it leaves compared — ASSUMED (#4270)"
 area: "CICS"
 summary: "A task started with a channel (a scenario's `channel`: what a RUN TRANSID CHANNEL parent or a LINK CHANNEL caller passed), made its first program's current channel; the containers left on that channel compared at the task's end (dropped on an abend), byte for byte; RUN TRANSID children not run"
-reached: "yes (async credit-card CRDTCHK, CSSTATS2, CSSTATUS, GETADDR, GETNAME)"
+reached: "yes (async credit-card CRDTCHK, CSSTATS2, CSSTATUS, GETADDR, GETNAME; CBSA CRDTAGY1-5, an explicit CHANNEL(CIPCREDCHANN))"
 ---
 - **What IBM documents** (CICS TS 6.x, "Scope of a channel"; RUN TRANSID CHANNEL: "the name of the channel that is
   to be passed to the child task"; FETCH CHILD CHANNEL: "the channel returned by the child task"; LINK CHANNEL: the

docs/language_status/register/X26.md

@@ -5,7 +5,7 @@ status: "ASSUMED (REFUSED where IBM is silent)"
 title: "BIF DEEDIT, INQUIRE / SET TERMINAL UCTRANST, DFHVALUE — ASSUMED, REFUSED where IBM is silent (#4415 slice 1)"
 area: "CICS"
 summary: "BIF DEEDIT FIELD [LENGTH] edited in place in the region's EBCDIC page (CCSID 037); INQUIRE / SET TERMINAL UCTRANST as the CVDAs UCTRAN 450 / NOUCTRAN 451 / TRANIDONLY 452, the terminal's value stated by the run (the case TYPETERM's UCTRAN), a terminal RECEIVE after a SET refused; DFHVALUE(name) in the stub's programs is IBM's CVDA number; a DEEDIT field with no digit left, beyond 7-bit ASCII or past LENGTH, and the UCTRANST of a terminal other than the task's, refused"
-reached: "yes (cics-crucible hc-deedit-uctranst, unreleased)"
+reached: "yes (cics-crucible hc-deedit-uctranst, unreleased; CBSA BNK1DCS's UCTRANST store / restore and the BNK1 screens' DEEDIT: cbsa-bnk1dcs, cbsa-bnk1cac ...)"
 ---
 - **BIF DEEDIT** (CICS TS 6.x, EXEC CICS BIF DEEDIT, https://www.ibm.com/docs/en/cics-ts/6.x?topic=summary-bif-deedit).
   "alphabetic and special characters are removed from an EBCDIC data field", "the remaining digits right-aligned and

docs/language_status/register/X27.md

@@ -5,7 +5,7 @@ status: "ASSUMED (the STARTCODE; REFUSED where IBM is silent)"
 title: "RETURN ... IMMEDIATE, LINK ... SYNCONRETURN — ASSUMED, REFUSED where IBM is silent (#4270 slice)"
 area: "CICS"
 summary: "RETURN TRANSID ... IMMEDIATE: the task of TRANSID attached at once with the COMMAREA, ahead of any terminal input and any START request, the terminal's next operator step left alone; its EIBAID is not stated by IBM (the crucible runner gives none and a case never reads it), its STARTCODE TD; INVREQ RESP2 1 (no terminal), INVREQ RESP2 2 (below the highest level), LENGERR RESP2 11 return to the program; LINK ... SYNCONRETURN accepted and ignored (IBM: \"ignored if the link is local\"); IMMEDIATE without TRANSID, and both INVREQs at once (no terminal below level 1), refused"
-reached: "yes (cics-crucible pc-return-immediate, unreleased)"
+reached: "yes (cics-crucible pc-return-immediate, unreleased; CBSA BNKMENU and the BNK1 screens' PF3, the task's terminal a case's `\"termid\"`: cbsa-bnkmenu, cbsa-bnk1cac ...)"
 ---
 - **RETURN IMMEDIATE** (CICS TS 6.x, EXEC CICS RETURN, https://www.ibm.com/docs/en/cics-ts/6.x?topic=summary-return).
   IMMEDIATE "ensures that the transaction specified in the TRANSID option is attached as the next transaction regardless

docs/language_status/register/X29.md

@@ -5,7 +5,7 @@ status: "ASSUMED (REFUSED where IBM is silent)"
 title: "INQUIRE ASSOCIATION, DELETEQ TS, QUERY COUNTER, RECEIVE MAP TERMINAL — ASSUMED, REFUSED where IBM is silent (#4415 slice 2)"
 area: "CICS"
 summary: "INQUIRE ASSOCIATION(EIBTASKN) ODAPPLID / ODUSERID / ODFACILNAME / ODNETWORKID / ODFACILTYPE: the task's own origin data, the five values the run states (a case's `origin`, for a task terminal input started; unstated, refused), the facility type as IBM's CVDA, a command with no origin option refused (INVREQ RESP2 2 is ambiguous); DELETEQ TS: the whole queue and its READQ NEXT position, QIDERR, INVREQ for a name of binary zeros, RESP2 0; QUERY COUNTER (COUNTER / POOL / VALUE): the value left unchanged, INVREQ RESP2 201 for a counter that is not there, a value beyond a fullword, MINIMUM / MAXIMUM / NOSUSPEND, a task number other than EIBTASKN and the origin of a task not started by terminal input refused; RECEIVE MAP ... TERMINAL: the task's terminal, as every RECEIVE MAP"
-reached: "INQUIRE ASSOCIATION, DELETEQ TS, RECEIVE MAP TERMINAL: yes (cics-crucible hc-inquire-deleteq, unreleased); QUERY COUNTER: no (the reference region has no named counters; unit tests on both runtimes)"
+reached: "INQUIRE ASSOCIATION, DELETEQ TS, RECEIVE MAP TERMINAL: yes (cics-crucible hc-inquire-deleteq, unreleased; INQUIRE ASSOCIATION and RECEIVE MAP TERMINAL also CBSA BNK1CRA / BNK1DCS: cbsa-bnk1cra, cbsa-bnk1dcs, a case's `\"origin\"`); QUERY COUNTER: no (the reference region has no named counters; unit tests on both runtimes)"
 ---
 - **INQUIRE ASSOCIATION** (CICS TS 6.x, CICS SPI command INQUIRE ASSOCIATION,
   https://www.ibm.com/docs/en/cics-ts/6.x?topic=commands-inquire-association). ASSOCIATION "Specifies the 4-byte number of the

tests/cobol_mainframe/test_equivalence_cics.py

@@ -486,6 +486,35 @@ def test_events_are_compared_field_by_field():
     assert ec.compare_events(cobol, java[:1])["diffs"][-1] == {"event": 2, "cobol": "RETURN", "java": None}
 
 
+def test_send_control_is_an_event_both_sides_compare_by_options_and_cursor():
+    """#4270 (CBSA's CLEAR key): the stub's SEND-CONTROL line is an event like CicsTask's, its options and CURSOR
+    compared -- a port that left it out, or sent other options, differs."""
+    res = {"events": ["SEND-CONTROL pgm=BNKMENU cursor=-1 opts=ERASE FREEKB", "RETURN pgm=BNKMENU level=1 transid= len=0"],
+           "screens": [], "text": [], "return": {"transid": "", "commarea": None}}  # fmt: skip
+    cobol = ec.cobol_events(res)
+    assert cobol[0] == {"event": "SEND-CONTROL", "options": ["ERASE", "FREEKB"], "cursor": None}
+    java = [{"event": "SEND-CONTROL", "options": ["FREEKB", "ERASE"], "cursor": None},
+            {"event": "RETURN", "transid": None, "commarea": None}]  # fmt: skip
+    assert ec.compare_events(cobol, java)["equal"] == 2
+    assert ec.compare_events(cobol, [{**java[0], "options": ["ERASE"]}, java[1]])["equal"] == 1
+    assert ec.compare_events(cobol, java[1:])["equal"] == 0
+
+
+def test_a_tasks_terminal_and_origin_are_stated_facts_or_none():
+    """#4270: "termid" / "uctranst" / "origin" -- the scenario's, else the case's; unstated None (the commands that
+    read them are then refused by both runtimes); a malformed one refused by name."""
+    case = {"termid": "T001", "uctranst": "NOUCTRAN"}
+    assert ec.terminal_facts(case, {"name": "s"}) == {"termid": "T001", "uctranst": "NOUCTRAN", "origin": None}
+    assert ec.terminal_facts({}, {"name": "s"}) == {"termid": None, "uctranst": None, "origin": None}
+    origin = {"applid": "CBSAREGN", "userid": "CICSUSER", "facilname": "T001", "networkid": "NET1",
+              "faciltype": "TERMINAL"}  # fmt: skip
+    assert ec.terminal_facts(case, {"name": "s", "origin": origin})["origin"] == origin
+    for bad in ({"termid": "T 01"}, {"termid": "TOOLONG"}, {"termid": "T001", "uctranst": "YES"},
+                {"uctranst": "UCTRAN"}, {"origin": {**origin, "userid": ""}}, {"origin": {"applid": "A"}}):  # fmt: skip
+        with pytest.raises(ec.Unsupported):
+            ec.terminal_facts({}, {"name": "s", **bad})
+
+
 def test_low_values_and_spaces_in_a_commarea_are_different_values():
     """#4635: a field RECEIVE MAP left LOW-VALUES and one the port left spaces are different bytes to the next program
     (COSGN00C: the RETURN COMMAREA's CDEMO-USER-ID); trailing spaces alone are still not data."""