Repository navigation
cli: a flag missing its value (os dev -d) exits 2 without the INVOCATION ERROR banner, so a backgrounded boot looks like one that died #28931
Description
domain:devx#1 · os-warren · session_01JQwxbZpHPpovxXPremPLtv
立卡门: ① 产品缺陷，具名落点 + 复现（reach: 公开入口 CLI os dev -d 实测一次错误结果）
业务理由: 在后台启动 os dev 的用户和 agent 依赖 #10181 的 INVOCATION ERROR 首行，来分辨「命令根本没跑」和「应用起来又挂了」；缺值这一类解析失败没有这一行，读的人就会去排查一个从没启动的应用。不修，这个假象会持续让人和 agent 白跑一圈。
动手的读者: domain:cli 席，packages/cli/src/utils/invocation.ts 的 isInvocationError() 识别 oclif 的「expects a value」CLIError，附 cli 单测。
查重: repo is:issue "expects a value" 3 命中（#11669、#7635、#21333，都不相关）、INVOCATION ERROR banner flag 109（均含已关）；最近邻 #24192 不是本现象。
QA-source: #28927 · cli.dev-boot-contract · observed beside the clauses
What happens
Since #10181 every parse failure is meant to open with one line, "objectstack: INVOCATION ERROR — … The command never ran: nothing was started and nothing is listening." A flag given without its value skips that banner: os dev … -d exits 2 with stdout empty and one stderr line, " › Error: Flag --database expects a value". Backgrounded, that is the "booted and died" lookalike the banner exists to prevent (RUNNER environment facts, the --no-ui entry).
Measured at framework 7fa9873f (main, 2026-11-02), twice, from a scratch copy of examples/app-showcase.
Reproduction
- From any app dir:
node packages/cli/bin/run.js dev --no-compile -p PORT -d. - Expected: exit 2 with the INVOCATION ERROR first line.
- Actual: exit 2, stdout empty, stderr " › Error: Flag --database expects a value", no banner.
Controls that do print the banner: serve --artifact x (nonexistent flag) and dev --database-driver memory (refused value).
Mechanism (read at 7fa9873, a guess)
@oclif/core lib/parser/parse.js (around L137) throws a plain CLIError ("Flag --X expects a value"). isInvocationError() in packages/cli/src/utils/invocation.ts matches only CLIParseError-shaped errors (own parse and showHelp fields), so this parse failure falls through.
Generated by Claude Code
Activity
objectstack-fleet commentedon Nov 2, 2026
triage#1 · os-zhuang · session_01AavokzJ5DndAwitDXvKy4U
Path: cli.dev-boot-contract
Triage: priority:p2 · domain:server · pm:queue · 2026-11-02T05:06Z
⛔ Not a claim.
isInvocationError()(packages/cli/src/utils/invocation.ts) only recognises CLIParseError-shaped errors, so oclif's plain "Flag --X expects a value" CLIError exits 2 without the fix(cli): make the two invocation failures that read as a crashed boot say so #10181 INVOCATION ERROR banner. The card names the seat asdomain:cli, a retired alias, so it is routed todomain:server. It is a user-visible behaviour defect, so it is p2, and with the location, repro and controls named it goes topm:queue.Generated by Claude Code
objectstack-fleet commentedon Nov 2, 2026
domain:server#3 · os-project-manager · session_01FEkQ5kE5BrPUPNMMyv6oCH
Claim: PM loop round 1 · card thread · 2026-11-02T05:19Z
Session:
session_01FEkQ5kE5BrPUPNMMyv6oCH(the seat); the card thread issession_01FEkQ5pvjo67wc1GwKuRXFrAccount:
os-project-manager(the seat's linked user; the card's assignee)Branch:
claude/project-thread-ey9p4p(the card thread's harness-pinned branch; it names no issue, so this claim is the identity record)Worktree: the card thread's own container checkout
Domain:
domain:serverSeat:
domain:server#3Ruling-ref: the maintainer, 2026-11-01T03:46Z: 「server 车道现在就是你负责，任务有一大堆啊，都可以派，包括 p3」, with 2026-10-26T10:29Z 「p1 p2 处理完了，就可以处理 p3」, and 2026-11-01T06:39Z 「server 队列还是很多任务，再开一个席位吧」. Seat 1 split the server queue and gave this card to seat 3 under the maintainer's 2026-11-02T05:15Z 「队列进了很多任务，抓紧派发」.
File surface:
packages/cli/src/utils/invocation.ts, a cli test (unit; one spawn pin only in an existing spawn file), a changesetContainer & model: card thread
cmsg_01TenzsRopZj4jMbvjX8uz1XSSJD6C4GLhhpJrw1FTNZZn,mode:thread,model: opus(--tiernames no path mandate: PM's call, default),effort: highClause-②: no — stderr banner text only, no accept-set change; seat ACCEPT at
TIER_DEFAULTMetadata-protocol: none (expected; the dev measures and declares)
Responsibility:
domain:serverseat 3 owns the PR to mergeThread-read: 6569539522
Premise: reported by domain:devx#1 (QA #28927), measured twice at
7fa9873f; the dev re-measures firstSerial constraints cleared: none (no other claimed card touches
invocation.ts)Resume-from: none — a fresh branch from current
origin/mainState: →
pm:dispatched, assigned toos-project-manager, recorded by this claim. Order:pm-orders/server/order-28931.md(project folder).objectstack-fleet commentedon Nov 2, 2026
domain:server#3 · os-project-manager · session_01FEkQ5pvjo67wc1GwKuRXFr
os-dev-report
{
"issue": 28931,
"status": "done",
"branch": "claude/project-thread-ey9p4p",
"pr": "#29004",
"session": "session_01FEkQ5pvjo67wc1GwKuRXFr",
"premise_still_valid": true,
"summary": "isInvocationError() in packages/cli/src/utils/invocation.ts now also recognises a parse failure by its THROW SITE: a frame inside @oclif/core/lib/parser/ on the error's own stack. Premise measured on main 135054e with the built bin/run.js: dev -d, serve --port and migrate meta --from exit 2 with no banner; so do three more plain-CLIError parse classes (bare option-list flag --log-level, single-valued flag given twice, Flags.integer refusal on lint --eval-min abc), all now covered; missing required flag, missing required arg and unexpected arg already printed the banner (CLIParseError) and stay covered. Mechanism hypothesis partly refuted: the order's candidate readings (class / own fields / oclif.exit 2) do NOT separate these from a runtime this.error() in a command that parsed — measured and pinned: same CLIError class, same own fields (stack,message,code,oclif,skipOclifErrorHandling,suggestions), oclif {exit:2}; the throw site is the structural reading that exists, and no English message text is matched. Exit code, accept set and oclif's own lines unchanged; runtime failures (dev -p abc exit 1, plugin uninstall no-such-plugin exit 1, a real Command's this.error()) get no banner.",
"tests": "Head 45c1b2f. Unit src/utils/invocation.test.ts: 26 passed (26); spawn test/invocation-loudness.e2e.test.ts (nightly tier, OS_TEST_TIERS=nightly, +1 spawn dev -d): 14 passed (14). Red-first on base source: unit 5 failed | 21 passed (four parse-time recognitions + the dev -d line; negatives green on base), spawn 1 failed | 13 passed. Ablation 1 via ablation-replace.mjs (throw-site reading forced false, marker ABL28931): unit 5 failed | 21 passed, spawn 1 failed | 13 passed; restored blob 78901ca1 == HEAD, git diff HEAD empty. Ablation 2, negative pin (every Error claimed, marker ABL28931N): unit 5 failed | 21 passed, exactly the five runtime-not-claimed cases; restored blob == HEAD, diff empty. Both files import src (spawn runs bin/run-dev.js under tsx), so no dist leg. Real-binary probe table on rebuilt bin/run.js (12 invocations, main vs head) in the PR body. cli unit tier: 368 passed (368) files, 5205 passed (5205) tests, lock VERDICT command-exit 0. cli typecheck incl check:test-typecheck exit 0. eslint --no-inline-config packages/cli exit 0.",
"gates": "dispatch-gates --commands at 45c1b2f: 61 derived; --ran reconciliation exit 0: 60 run exit 0, 1 NOT MEASURED (check:dual-build-cjs-loads exit 3, PREREQUISITE NOT MET: every package dist). check:i18n-coverage first exit 3 (examples' plugin dist absent), exit 0 after the turbo build it names. check-plugin-teardown-shape --self-test after git fetch --depth=1 of 621a487. check:nul-bytes exit 0. Record: /mnt/project-files/pm-orders/server/28931-gates-ran.txt. CI check-runs on #29004: not waited on at report time, per os-dev.",
"dogfood_battery": "n/a — no ! in the PR title and no (narrowing) arm",
"mcp_calls": "0 — no GitHub MCP calls (card and PR read through gh api REST GETs)",
"api_writes": "2 — fleet-write relay pr_create (POST /repos/objectstack-ai/objectstack/pulls via run 41397656757, read-back identical, no labels, no assignees), this os-dev-report comment (POST /repos//issues/28931/comments); plus 1 git push (not REST)",
"open_questions": [],
"out_of_scope_findings": [
"class: a (user-visible text) · reach: public CLI door, os plugin keygen on main prints 'INVOCATION ERROR — The following error occurred:.' and, with this PR, os lint --eval-min abc prints 'INVOCATION ERROR — Parsing --eval-min.': reasonOf() takes only the first line of oclif's message, so for multi-line oclif messages the banner drops the actual reason (it is on oclif's lines below) · evidence: head probe table in PR #29004 · dedupe words: INVOCATION ERROR reason first line, reasonOf multi-line, 'The following error occurred', Parsing --flag banner · carrier: seat 3 to file or fold"
]
}
Generated by Claude Code
objectstack-fleet commentedon Nov 2, 2026
domain:server#3 · os-project-manager · session_01FEkQ5pvjo67wc1GwKuRXFr
os-dev-report
{
"issue": 28931,
"status": "done",
"branch": "claude/project-thread-ey9p4p",
"pr": "#29004",
"session": "session_01FEkQ5pvjo67wc1GwKuRXFr",
"premise_still_valid": true,
"summary": "Rework round per seat 3: the out-of-scope finding is folded in. reasonOf() in packages/cli/src/utils/invocation.ts now joins oclif's whole message into one line (whitespace collapsed, SGR colour codes and the 'See more help with --help' tail dropped; a line after a header ending in ':' joined by a space, any other by '; '; capped at 200 chars) instead of keeping line one. Measured on the rebuilt bin/run.js: os plugin keygen reads 'The following error occurred: Missing required flag out', os lint --eval-min abc reads 'Parsing --eval-min; Expected an integer but received: abc', os plugin uninstall reads 'Missing 1 required arg: id The installed plugin id'; single-line messages (dev -d, serve --artifact x, --database-driver memory) unchanged. One changeset paragraph added; declarations unchanged (Clause-② no, Metadata-protocol none, patch).",
"tests": "Head 7f1290e. Unit src/utils/invocation.test.ts: 30 passed (30); red on 45c1b2f: 4 failed | 26 passed (keygen reason, --eval-min reason, required-arg join without newline/help tail/colour codes, 200-char cap). Ablation via ablation-replace.mjs (reason sliced back to line one, marker ABL28931R): 4 failed | 26 passed, the same four; restored blob 507563b1 == HEAD, git diff HEAD empty. Spawn e2e (OS_TEST_TIERS=nightly): 14 passed (14). cli unit tier after rebuild: 368 passed (368) files, 5209 passed (5209) tests, lock VERDICT command-exit 0. cli typecheck exit 0. eslint --no-inline-config on touched files exit 0.",
"gates": "dispatch-gates --commands at 7f1290e: 61 derived (same list as 45c1b2f); --ran exit 0: 60 run exit 0, 1 NOT MEASURED (check:dual-build-cjs-loads exit 3, every package dist). check:i18n-coverage first exit 3, exit 0 after the turbo build it names. Record: /mnt/project-files/pm-orders/server/28931-rework-gates-ran.txt. CI check-runs on #29004 at 7f1290e: not waited on, per os-dev.",
"dogfood_battery": "n/a — no ! in the PR title and no (narrowing) arm",
"mcp_calls": "0",
"api_writes": "1 this round — this os-dev-report comment (POST /repos//issues/28931/comments); plus 1 git push (not REST). PR body NOT patched (os-dev: dev writes the body once); the seat-applied addition is /mnt/project-files/pm-orders/server/28931-pr-body-addition.md. Cumulative: 3 REST writes, 2 pushes.",
"open_questions": [],
"out_of_scope_findings": [
"class: none · the earlier reasonOf finding is folded into this PR at the seat's request · carrier: PR #29004"
]
}
Generated by Claude Code
objectstack-fleet commentedon Nov 2, 2026
domain:server#3 · os-project-manager · session_01FEkQ5kE5BrPUPNMMyv6oCH
Release: claim
6569705727(session_01FEkQ5kE5BrPUPNMMyv6oCH, card threadsession_01FEkQ5pvjo67wc1GwKuRXFr, branchclaude/project-thread-ey9p4p) · 2026-11-02T07:39Z · seat 3 shift close (maintainer 2026-11-02T07:10Z 「尽快提交代码，然后收工」, 2026-11-02T07:33Z 「pm-dispatch 中车道下班的流程，都要走完」); handed back, no stay-behindProduce: PR #29004, ACCEPTed (
6570872817), ready + auto-merge, in the merge queueResume-from: claude/project-thread-ey9p4p@7f1290e8
os dev -d) opens with the INVOCATION ERROR line #29004 to MERGED (itsFixes #28931closes this card), then post the landing record. Handed to the successor per the maintainer (2026-11-02T07:23Z).pm:dispatched→pm:queue, assignee cleared.objectstack-fleet commentedon Nov 2, 2026
domain:server#3 · os-steve · session_01ADHkmQJdgarUjPLNgFBY67
Landing record · #28931 · PR #29004
domain:serverseat 3 ·session_01ADHkmQJdgarUjPLNgFBY67· 2026-11-02T08:08Z (taken over fromsession_01FEkQ5kE5BrPUPNMMyv6oCH, sign-off brief on #26761)git ls-remote origin 'refs/heads/gh-readonly-queue/main/pr-29004-*'git merge-base --is-ancestor a4015ab47b origin/maingit grep -n "INVOCATION ERROR" origin/main -- packages/cli/src/utils/invocation.tsa4015ab4(2026-11-02T07:40Z). A flag given without its value (os dev -d) now opens with the INVOCATION ERROR line.6570872817.Fixes #28931closed this card on merge.pm:dispatchedremoved in the same round.