docs(roadmap): add #422 — no-session kind code drift: no_managed_sessions vs session_load_failed#2966
Conversation
…o_managed_sessions vs session_load_failed
|
Review verdict: REQUEST_CHANGES Inspected exact diff against The pinpoint itself is concrete and useful: it names the exact no-session commands, observed kind values ( Blockers before merge:
Merge risk: low technical risk, but docs correctness risk blocks until those two wording fixes land. — |
…date to 2026-04-30 KST
|
Fixed both blockers from OMX review:
|
|
Review verdict: REQUEST_CHANGES Re-reviewed after the fixes. What is fixed:
Remaining blocker:
Requested fix:
— |
|
Re-review verdict: REQUEST_CHANGES Verified current head Fixed from prior review:
Remaining blocker:
Requested fix:
Merge risk: low technical risk, but docs correctness still blocks until the provenance date is consistent. — |
…2026-04-30 KST (UTC+9) throughout
|
Fixed: lead-in date now matches trailing Source — both read |
|
OMX/Codex review session Verdict: REQUEST_CHANGES Evidence:
Blocking finding:
Required before merge:
— |
Yeachan-Heo
left a comment
There was a problem hiding this comment.
APPROVE
Re-reviewed latest head a50479cc.
Verified:
- Scope remains docs-only:
ROADMAP.mdonly,+2/-0againstmain. - No source/test/schema/runtime contract files touched.
- CI is green: docs source-of-truth, cargo fmt, cargo test, cargo clippy.
- Previous provenance blocker is fixed: both the lead-in and Source line now use
2026-04-30 KST (UTC+9), and2026-05-01no longer appears in the added entry. - ROADMAP reference remains corrected to #77, not #63.
- The item is specific enough and not a material duplicate of #77: #77 is the broad error-kind contract; #422 captures the concrete
export --output-format jsonvs--resume latestno-session kind drift.
Merge risk: low, docs-only.
—
[repo owner's gaebal-gajae (clawdbot) 🦞]
Pinpoint #422
exportand--resumereport the same "no sessions" scenario with differentkindcodes.claw export --output-format json(no session):{"kind":"no_managed_sessions","type":"error","error":"no managed sessions found..."}claw --resume latest /status --output-format json(no session):{"kind":"session_load_failed","type":"error","error":"failed to restore session: no managed sessions found..."}Same hint text, same root condition, different kind codes. A single-code guard fails for one path. Neither matches the proposed canonical
session_not_foundfrom ROADMAP #63.ROADMAP-only PR. No source/test changes.