Repository navigation
det: EXEC DLI SCHD PSB / TERM under CICS, DIBSTAT supplied (Part of #5278) - #5383
Conversation
…ed (Part of #5278) Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
|
No description provided. |
|
No description provided. |
|
No description provided. |
|
PR #5383 @ 420b9c5: 1 failed check(s) == Small audits (failure) -- real |
|
|
||
| def command(gen: Any, text: str, ind: str) -> list[str]: | ||
| """One EXEC DLI command of a CICS program: SCHD PSB(...) [NODHABEND] and TERM, DIBSTAT answered.""" | ||
| from gitgalaxy.tools.cobol_to_java.det.gen import Untranslatable, jstr |
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
|
No description provided. |
|
No description provided. |
|
PR #5383 @ 1f338b8: 1 failed check(s) == ground-truth part 1/2 (failure) -- real |
…java run (Part of #5278) Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
|
No description provided. |
|
No description provided. |
Part of #5278 (slice 1: SCHD / TERM / DIBSTAT, as the brief's split says).
What
det/dli.py+gen.py+program.py: under CICS,EXEC DLI SCHD PSB(name | 'lit' | (area)) [NODHABEND]andEXEC DLI TERMtranslate toDetDli.schedule/terminate; the answer is moved toDIBSTAT(the program gets a minimal DIB item,FILLER+DIBSTAT).DetDli.java:schedule(blank;TCif already scheduled with NODHABEND; refuses without it) andterminate(refuses with nothing scheduled).buildPcbs()extracted fromenter().I5.mdupdated.Class table
6 hand-written files (exception: the Java runtime twin, the translator trio, the test and the register page are one mechanism).
Evidence level
Translated only (no harness case: the oracle has no EXEC DLI). Unit translation tests, a JDK-compiled runtime test of schedule/terminate/TC. The PSB name is not checked against psb.cfg (it carries no PSB name); the
loaded/scheduledflags are per JVM.Gates
pr_gates --fast6/6;--ratchetsports, port-surface, estate, fact-crosscheck, completeness, ground-truth, time pass; ruff/mypy audits no new findings. Golden masters not touched (no parsing change). Note: local ruff 0.16.0 vs CI 0.16.10.Remaining slices
Oracle side (ggdli.c / equivalence_cics: EXEC DLI, DIB, PCB-by-number); segment commands GU/REPL/ISRT with SEGMENT/WHERE/FROM/SEGLENGTH; C17 INITIALIZE REPLACING; KEYLENGTH holes; COPAUA0C equivalence case (DB + MQ scenarios); register/ledger/det-sweep pins.
Sibling gaps
MQINQ/MQSET (no burned program uses them; frontier #4892); batch MQ (frontier); COPAUS0C / COPAUS1C / CBPAUP0C (same EXEC DLI mechanism, burned CardDemo); DBUNLDGS (GSAM PCBs, different mechanism).
Lessons for agents
.pyfile; build it from parts.Cross-repo
None.
🤖 Generated with Claude Code