stack/agents/filbert/work/queue-166/r2/body.md at sp2

0 Fork 1

Files

T

jason.woltje b9ea9304fd lead: commit filbert's untracked evidence (273 files: queue-101 queue-102 queue-105 queue-107 queue-108 queue-109 queue-110 queue-112 queue-113 queue-114 queue-115 queue-116 queue-119 queue-120 queue-122 queue-123 queue-125 queue-126 queue-128 queue-129 queue-130 queue-131 queue-132 queue-133 queue-134 queue-135 queue-136 queue-137 queue-139 queue-140 queue-141 queue-142 queue-145 queue-146 queue-147 queue-148 queue-149 queue-150 queue-151 queue-152 queue-153 queue-154 queue-155 queue-156 queue-162 queue-163 queue-165 queue-166 queue-167 queue-169 queue-170 queue-175 queue-177 queue-181 queue-182 queue-184 queue-185 queue-186 queue-187 queue-188 queue-49 queue-64 queue-66 queue-70 queue-71 queue-80 queue-84 queue-85 queue-87 queue-88)

Unclaimed files left in the canonical checkout. Inventory: canonical-cleanup-inventory.md (commit-as-evidence groups). detect-secrets scan clean on 2026-11-02: hits are sha256 pins, git object ids, test sentinels already in tracked tests and unit-file lines.

2026-11-02 15:05:14 -06:00

4.0 KiB

Raw Permalink Blame History

Filbert review, row 166 round 2: approve

Candidate manifest 4fc0613f2f9dad09d9a2a4ed424f4cde807485c0cee3eb536fec5c768468aced (2 paths), build.patch ecfe96dd22191810468007d83267a2e8b779d1329a608f6b454b565fb05e5575 on 139e28673.

What I checked

I built both rounds in fresh worktrees on 139e28673:

round 2: build.patch;
round 1: build-r1.patch (f7c22a97), against candidate-manifest-r1.sha256 (d0c58a37).

Both manifests match their trees. Round 2's git diff --cached --check is clean, at +139 -19 over 2 files. scripts/stack-release.sh is byte-equal between the rounds (cmp), still 0cfdfa4f. Its cached diff still hashes to 490d6c99...88d3, the diff Mos read. So the only change is the test file, and I read its full round 1 to round 2 interdiff (3 hunks, 20 +/- lines).

Note 1: the Console ExecStart form

T31 gains ['Console ExecStart quotes', ...], which edits the Console unit from --business acme to --business "acme". edit() (stack-release.test.mjs:266) asserts that the text is present, so the form can't silently do nothing. The Console unit's ExecStart flows through the show stand-in, as the existing console drift case at :389 already relies on.

With the quotes, read_units (scripts/stack-release.sh:156) returns 1 at unquoted "$console_argv". The form then gets the same four assertions as the other forms:

activate gives unit-flags-drift "cannot read";
status prints unit flags: cannot read the units;
adopt gives check-failed and writes no pointer;
no stderr contains /a.

Without that check, the quoted token would reach the flag compare as "acme". Activate would then refuse with "the units differ from the pointer" (:542), not "cannot read" (:550), and the exact-stderr assertion fails. So, as I read it, mutant 4c can't survive this form.

Note 2: the relative unit path

The systemd-analyze stand-in now also prints the contents of $SD_STATE/faults/unit-path when that file exists (:163-164). T28's new subtest sets the fault to systemd/user, so the script's paths gets the real absolute path and then systemd/user. drop_in_override (scripts/stack-release.sh:515-518) passes the first, then refuses at the case with invalid systemd unit path systemd/user.

The subtest asserts:

exit 3 and that exact stderr;
the fingerprint unchanged and no start or stop;
both units active;
one more log line, drop-in-override.

These are the same assertions as the other three T28 subtests. The test title names the new case.

Other checks

The grep correction. BUILD.md :105-110 now names packages/cli/README.md and records the correction. Small anchor point: that line is :648 on 139e28673, but row 156 moved it to :693 at HEAD. The text is unchanged and still true. Not blocking.
Note 3. BUILD.md leaves docs/TOOLS.md:389-390 as a doc-only follow-up, which I agree with (decision 137).
Hash pins. No 64-hex pin in packages/*/tests or scripts/test-*.sh names either manifest path.
Base drift. From 139e28673 to HEAD 7f9b976c9, nothing changed on the 2 manifest paths, stack-copy.sh or bus-service.sh. The canonical script is still blob 4b16e848. build.patch passes git apply --cached --check on HEAD in a temporary index.
The commit message. commit-msg.txt has no attribution line.
Landing order. This approves the candidate only. It lands after the 8e7ea89a1 window, per mos-0834.

Ran by me

Nothing that needed the slot. I built two worktrees, ran sha256sum -c (2/2 each), cmp on the script and the script-diff digest. I also read the test interdiff and the script paths it reaches, ran greps, and ran the HEAD apply check. I did not run the mutants. The 4c point above is my reading.

Not mine

Darkwing's round 2 slot run (sage-0959), as BUILD.md records it:

stack-release.test.mjs: 73 of 73;
stack-copy.test.mjs: 12 of 12;
awk pass 85, fail 0;
CONTROL passes, and all 10 mutants are killed, including 1d and 4c.

Its logs are under ~/darkwing-scratch, which I don't read. The full gate is Sage's at landing.

4.0 KiB Raw Permalink Blame History