Repository navigation
Merge remote-tracking branch 'origin/main' into agent/2908-copy-files…
PASS
Baseline. This re-reviews 8d3c001f against the CONDITIONS on ba2cc257 (this PR), on arch 7a86ec7b. I checked these facts. since.diff equals git diff ba2cc257 8d3c001f. pr.diff equals git diff 46d9e530 8d3c001f, and 46d9e530 (#3006) is the base and the merge base. The compare has two merges of main and one branch commit, abe9acb2. The merge 5db910e1 is clean (git diff-tree --cc is empty). The merge 8d3c001f resolves one conflict in crates/minimal/src/tests.rs, where both sides added tests at the same place. The combined diff drops no line and adds no line that is new to both parents, so it is a union.
Governs: architecture.md Output conventions lines 381-383 (RFC 3339 UTC timestamps, the min/v1/event shape, one min/v1/error object in machine modes). Command reference line 482 ([-f|--follow]). Retention and reaping line 501 (replay then tail, and the retained filesystem for cp). Line 503 (--parent merges all children). Exit codes lines 365-366, through BCLI-034.
Ruling. abe9acb2 answers each point the baseline raised in the compare. EventRow serialises ts as an RFC 3339 UTC string through EventRowWire, and the tests now parse it as RFC 3339 and check the Z. --follow gains -f. machine_output_mode adds the Events arm for -o jsonl, with a unit test. BoxFileRefusal carries exit 4 and 5 to failure_exit_code, and the unit test and box_cp_reads_stopped_box_as_box_017 pin both. Unavailable and Refused stay exit 1. They cover several causes, and the table's row 1 admits that. The one-time child read under --parent --follow is a gap against line 503. It is outside covers:, marked interim in code, and named as residual #3024. Placement is unchanged: the CLI code is in crates/minimal and the channel is in crates/minimal-client. The change moves no check across a host or guest boundary.
Note. EventRow::ts() falls back to the raw millisecond number when the value is outside chrono's date range. No real minimald timestamp reaches that range, so I did not make it a condition. A later change can drop the fallback.
Not checked. I could not reach GitHub from the sandbox, so I did not confirm that issue #3024 exists and is open. The code comment and the commit message name it. I did not build or run the tests, per the contract.
Ruled at 8d3c001 against gominimal/arch@7a86ec7bb3c0 by claude-opus-5-5; arch_rulings 936, job 23412440-3216-429b-b9bc-796030c9534c.