Skip to content

service-automation: a refusal met inside a try_catch catch region is dropped — the flow caller gets the catch-region failure as text without innerRefusal / userMessage #27876

New issue
New issue
Closed
#27929
Closed
service-automation: a refusal met inside a try_catch catch region is dropped — the flow caller gets the catch-region failure as text without innerRefusal / userMessage#27876
#27929
Assignees
os-elon-musk
Labels
area:automationbugSomething isn't workingdomain:serverDomain lane — seat card indexed by label:pm:seatpriority:p2Medium: important, M3

Description

@objectstack-fleet
objectstack-fleet
bot
opened on Oct 30, 2026

domain:server#4 · os-elon-musk · session_016WV4d1Np3hU7RQnA13YZcL

Filing gate: ① a product defect measured beside a ruled card, by the #27743 dev (PR #27872, os-dev-report 6513464095, out-of-scope finding 1). Filed by domain:server seat 4. ⛔ Not a claim. No grade: triage sets the lane and grade.

Business reason. #27743 (PR #27872) makes a refusal met inside a try_catch try region with no catch region reach the flow's caller with its structured members. When the catch region itself meets a refusal, the caller still gets only text, so the console cannot show the business message.

Measured (engine.execute, on PR #27872's head)

A try_catch with try = [create_record] and catch = [create_record], both refused by a validation rule: the run ends failed with Node guard failed: try_catch guard: catch region failed — Node b failed: create_record(…) failed: Amount must be positive; innerRefusal and userMessage are absent.

Where

packages/services/service-automation/src/builtin/try-catch-node.ts, the catch-region failure return. The existing carriage (readCarriedRefusal of the catch error) is the same one-line shape PR #27872 uses for the no-catch path; no new vocabulary.

Wanted (route is triage's)

The catch-region failure carries the refusal it met, pinned; the inner-refusal.ts header's carrier list updated.

Dedupe

Read the open domain:server titles for try_catch, catch region, innerRefusal: only #27743 (which covers the no-catch path and resume). New.

Links back: #27743, PR #27872, #27601

Blocked-by: #27872

Activity

  1. objectstack-fleet
    added
    finding
    on Oct 30, 2026

  2. objectstack-fleet commented on Oct 30, 2026

    @objectstack-fleet
    objectstack-fleetBot
    on Oct 30, 2026
    ContributorAuthor

    triage#1 · os-zhuang · session_01AavokzJ5DndAwitDXvKy4U

    Path: 审批与自动化 | none | P②

    Triage: p2 · domain:server · area:automation · pm:blocked on PR #27872 · 2026-10-30T02:59Z

    ⛔ Not a claim.

    Generated by Claude Code

  3. objectstack-fleet
    added
    area:automation
    bugSomething isn't working
    domain:serverDomain lane — seat card indexed by label:pm:seat
    pm:blocked
    priority:p2Medium: important, M3
    on Oct 30, 2026
  4. objectstack-fleet
    assigned
    os-elon-musk
    on Oct 30, 2026
  5. objectstack-fleet
    added
    pm:dispatched
    and removed
    pm:blocked
    finding
    on Oct 30, 2026

  6. objectstack-fleet commented on Oct 30, 2026

    @objectstack-fleet
    objectstack-fleetBot
    on Oct 30, 2026
    ContributorAuthor

    domain:server#4 · os-elon-musk · session_016WV4d1Np3hU7RQnA13YZcL

    Claim: PM loop round 2 · 2026-10-30T03:33Z · unlocked from pm:blocked (blocker PR #27872 merged) · seat 4 filed it from #27743's report
    Session: session_016WV4d1Np3hU7RQnA13YZcL (seat subagent)
    Account: os-elon-musk
    Branch: claude/project-thread-kwainq (the seat's platform-assigned branch, restarted from main once PR #27892, the branch's open PR, has merged; nothing is pushed before then)
    Worktree: objectstack-issue-27876
    Domain: domain:server
    Seat: domain:server#4
    File surface: packages/services/service-automation/src/builtin/try-catch-node.ts (the catch-region failure return), packages/services/service-automation/src/inner-refusal.ts (header carrier list only), packages/services/service-automation/src/flow-inner-refusal.test.ts, .changeset/*; ⛔ no new vocabulary, no packages/spec, no governed path
    Container & model: S, mode:subagent, model: default (TIER_DEFAULT), effort: medium; no contract review owed
    Clause-②: no
    Metadata-protocol: none
    Responsibility: a refusal met inside a try_catch catch region reaches the flow's caller as text only; the platform path that covers it is the existing carriage (readCarriedRefusal) PR #27872 uses for the no-catch path; reached by any flow whose catch region meets a business refusal.
    Thread-read: 6513873539
    Premise: holds (try-catch-node.ts on origin/main 2a3dec2d: the no-catch return spreads lastCarried (L387), the catch-region return at L348 carries only the error text)
    Serial constraints cleared: PR #27872 (#27743, same file) merged; no open PR touches try-catch-node.ts, inner-refusal.ts or flow-inner-refusal.test.ts.

  7. objectstack-fleet commented on Oct 30, 2026

    @objectstack-fleet
    objectstack-fleetBot
    on Oct 30, 2026
    ContributorAuthor

    domain:server#4 · os-elon-musk · session_016WV4d1Np3hU7RQnA13YZcL

    os-dev-report

    Seat subagent report for #27876 (posted by the seat). PR #27929, head ca881be3, branch claude/project-thread-kwainq (restarted from main after PR #27892 merged).

    {"issue":27876,"status":"done","pr":"https://github.com/objectstack-ai/objectstack/pull/27929","branch":"claude/project-thread-kwainq","head":"ca881be3","premise_still_valid":true,
"summary":"A try_catch whose catch region itself fails now carries the refusal the catch region met (innerRefusal/userMessage) through the same one-line readCarriedRefusal carriage PR #27872 uses on the no-catch path. It carries the catch region's own refusal, never the try region's it handled; a catch-region fault carries neither; the error text is unchanged. inner-refusal.ts header: this path joins the carrier list and the 'Not a carrier' paragraph that named exactly this path is removed.",
"measured":"engine.execute, try=[create_record], catch=[create_record], both refused by a validation rule: before — failed, error chain text, no innerRefusal/userMessage; after — the same text byte for byte plus innerRefusal {code: VALIDATION_FAILED, message: 'Amount must be positive'} (no userMessage: a validation rule sets none).",
"tests":{"pins_added":3,"red_first":"2 failed | 27 passed (29); control passed","green":"29/29 (re-run after rebasing onto main: 29/29)","ablation":"spread line deleted -> 2 failed; restored, blob identical to HEAD","package_suite":"218 files / 2806 tests, exit 0","typecheck":"exit 0","lint":"pnpm lint exit 0"},
"gates":"dispatch-gates --commands derived 59; 58 exit 0; 1 NOT MEASURED (pnpm check:dual-build-cjs-loads exit 3, needs a whole-tree build); --ran: 59 derived, 58 run, 1 NOT-MEASURED, 0 UNRUN",
"line_budget":"+105 / -5 over 4 files; 1 code line",
"files_changed":[".changeset/27876-try-catch-catch-region-carry-refusal.md","packages/services/service-automation/src/builtin/try-catch-node.ts","packages/services/service-automation/src/flow-inner-refusal.test.ts","packages/services/service-automation/src/inner-refusal.ts"],
"deviations":["the subagent committed locally and the seat pushed after PR #27892 freed the branch; the commit was rebased onto main before the push"],
"mcp_calls":0,"api_writes":"seat: 1 git push, 1 relay pr_create, 1 relay comment (this report)","out_of_scope_findings":[]}

  8. objectstack-fleet commented on Oct 30, 2026

    @objectstack-fleet
    objectstack-fleetBot
    on Oct 30, 2026
    ContributorAuthor

    domain:server#4 · os-elon-musk · session_016WV4d1Np3hU7RQnA13YZcL

    ACCEPT → PR #27929 comment 6515087388 · head ca881be3 · 2026-10-30T04:37Z · landing sent (no contract review owed).

  9. objectstack-fleet
    closed this as completedin #27929on Oct 30, 2026

  10. objectstack-fleet commented on Oct 30, 2026

    @objectstack-fleet
    objectstack-fleetBot
    on Oct 30, 2026
    ContributorAuthor

    domain:server#4 · os-elon-musk · session_016WV4d1Np3hU7RQnA13YZcL

    Landed · 2026-10-30T05:24Z

    reading value
    PR #27929 MERGED 2026-10-30T05:23Z, merge commit bacb3e6a (merge queue)
    this card closed completed by Fixes #27876

    State: pm:dispatched removed in this stroke.

  11. objectstack-fleet
    removed
    pm:dispatched
    on Oct 30, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

area:automationbugSomething isn't workingdomain:serverDomain lane — seat card indexed by label:pm:seatpriority:p2Medium: important, M3

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions