[squid-protocol]

Checklist: partial [c:5ddf2663] -- the java-ported and java-facade cells of gt-web-client now pass with the crucible's equal log (det port committed); the bare java cell is the generated skeleton and fails for every case by design, and baseline.json is unchanged because the default runner reads ports/, not det-ports/ (same as pc-return-channel, #6330).

Part of #4767. Burned estate: cics-crucible gt-web-client (program GTWCLI, the zECS WEB OPEN / CONVERSE / CLOSE shape).

What

• tests/cics_crucible/det-ports/gt-web-client/GTWCLI/overlay/service/GtwcliService.java: det translation, 15/15 statements, 0 holes (cics_case_ports.py, pin v0.10.0). Not hand-edited.
• cics_crucible_det.py: registers the case in PROGRAMS so ports_dir lays the runtime beside it.
• test_web_client_det_port_4767.py: structure tests plus an e2e (EQUIVALENCE_E2E=1) that runs ok and notfound on java-ported and java-facade.

Evidence

• cics_crucible.py --cases gt-web-client --sides cobol-stub java-ported java-facade java --ports <det>: cobol-stub 2/2, java-ported 2/2, java-facade 2/2 pass; java (skeleton) 0/2, as in baseline.
• e2e test passes (22 s); pr_gates --fast 7/7, --ratchets 9/9 (ports, port-surface, fact-crosscheck, ground-truth, det-mutation included). No engine change, so no golden bless.

Class / Sibling gaps

Class: crucible gt-web-* cases and their java-ported / java-facade cells. Covered here: gt-web-client (GTWCLI). Not covered: gt-web-parse-url (GTWPARS; its cobol-stub parse/lengerr cells also fail on WRITEQ-TS, so it needs the stub fix first); gt-web-server-receive and gt-web-server-send (cobol-stub unsupported: translator refuses WEB RECEIVE/SEND CHARACTERSET and HOSTCODEPAGE, WEB READ HTTPHEADER; owner ruling on those pending per #4767); the java skeleton cells (by design failing for every case with a port).

🤖 Generated with Claude Code

[squid-protocol]

No description provided.

[squid-protocol]

No description provided.