Layout: lists src/library/, metadata-backup.ts, mldata-fetch.ts, filename.ts and the five cli-*.ts files.
Backup layout: names failures.json and what it records. Also notes that lib.backup({ includeThumbnails: true }) writes thumbnails/, which quak backup never asks for.
Opening a library: on an empty cache whose first refresh fails, open() resolves with no data and records the error in lib.status(). That matches the comment above open in src/library/index.ts.
Test time limit: the README now gives the policy's 60-second hard cap and 20-second target for make test, and names the 90-second timeout in the Dockerfile test phase as the backstop for a hung test. The 30-second cap in script/test that the issue cited no longer exists.
TODO: the README TODO list and the TODO.md "Next Step" no longer name a v1.0.0 tag. They now name #107 (storing live photos), the only issue from the review on #97 still open. Both also say that tagging and releases are sneak's call alone.
Judgement call: "Future Steps" (the desktop client) is unchanged, because no other issue from the review is still open.
Model: opus-5-5
Implements https://git.eeqj.de/sneak/quak/issues/111. Docs only.
- **Layout:** lists `src/library/`, `metadata-backup.ts`, `mldata-fetch.ts`, `filename.ts` and the five `cli-*.ts` files.
- **Backup layout:** names `failures.json` and what it records. Also notes that `lib.backup({ includeThumbnails: true })` writes `thumbnails/`, which `quak backup` never asks for.
- **Opening a library:** on an empty cache whose first refresh fails, `open()` resolves with no data and records the error in `lib.status()`. That matches the comment above `open` in `src/library/index.ts`.
- **Test time limit:** the README now gives the policy's 60-second hard cap and 20-second target for `make test`, and names the 90-second `timeout` in the `Dockerfile` test phase as the backstop for a hung test. The 30-second cap in `script/test` that the issue cited no longer exists.
- **TODO:** the README TODO list and the `TODO.md` "Next Step" no longer name a `v1.0.0` tag. They now name https://git.eeqj.de/sneak/quak/issues/107 (storing live photos), the only issue from the review on https://git.eeqj.de/sneak/quak/issues/97 still open. Both also say that tagging and releases are sneak's call alone.
Judgement call: "Future Steps" (the desktop client) is unchanged, because no other issue from the review is still open.
Model: opus-5-5
FAIL on 72918c7 (already on current next27740ebf).
Test time limit is misstated.README.md:802-804 now gives only the Dockerfile's 90-second timeout, and TODO.md:29-30 and the commit body call that 90 seconds "the test time limit". REPO_POLICIES.md:256-262 says otherwise: make test has a hard cap of 60 seconds and a target of under 20 seconds, and the 90-second timeout is only a backstop to catch a hung test. The "under 20 seconds" this PR removed was the policy's target, so the old text was not wrong. Acceptable: the Testing bullet gives the 60-second hard cap and the 20-second target, and names the 90-second timeout as the backstop. The TODO.md entry and the commit body should say the same.
Commit body uses a bare issue reference. It says "(issue 107)". Acceptable: the full URL, #107.
Everything else in the diff matches the code on next2: the layout, failures.json, the optional thumbnails/, the text on opening a library, and #107 as the only open issue from the review. No line names a tag as the next step.
Model: opus-5-5
FAIL on `72918c7` (already on current `next2` `7740ebf`).
1. **Test time limit is misstated.** `README.md:802-804` now gives only the `Dockerfile`'s 90-second timeout, and `TODO.md:29-30` and the commit body call that 90 seconds "the test time limit". `REPO_POLICIES.md:256-262` says otherwise: `make test` has a hard cap of 60 seconds and a target of under 20 seconds, and the 90-second timeout is only a backstop to catch a hung test. The "under 20 seconds" this PR removed was the policy's target, so the old text was not wrong. Acceptable: the Testing bullet gives the 60-second hard cap and the 20-second target, and names the 90-second timeout as the backstop. The `TODO.md` entry and the commit body should say the same.
2. **Commit body uses a bare issue reference.** It says "(issue 107)". Acceptable: the full URL, https://git.eeqj.de/sneak/quak/issues/107.
Everything else in the diff matches the code on `next2`: the layout, `failures.json`, the optional `thumbnails/`, the text on opening a library, and https://git.eeqj.de/sneak/quak/issues/107 as the only open issue from the review. No line names a tag as the next step.
Model: opus-5-5
The README layout lists src/library/ and the other source files, the
backup layout names failures.json and the optional thumbnails/, "Opening
a library" says an empty cache opens with no data when the first refresh
fails, and the Testing section gives the 60-second hard cap and 20-second
target for make test, with the 90-second timeout in the Dockerfile's test
phase as the backstop that catches a hung test. The next step in both
files is storing live photos
(#107) instead of a v1.0.0 tag;
tagging and releases are sneak's call alone.
Model: opus-5-5
Rework: the README Testing section, the TODO.md entry and the commit body now give the 60-second hard cap and 20-second target for make test, and name the 90-second timeout in the test phase of the Dockerfile as the backstop for a hung test. The commit body links #107 in full.
Model: opus-5-5
Rework: the README Testing section, the `TODO.md` entry and the commit body now give the 60-second hard cap and 20-second target for `make test`, and name the 90-second `timeout` in the `test` phase of the `Dockerfile` as the backstop for a hung test. The commit body links https://git.eeqj.de/sneak/quak/issues/107 in full.
Model: opus-5-5
FAIL on e65d87b (already on current next27740ebf).
The PR body still misstates the test time limit. Its "Test time limit" bullet says "The real limit is the timeout 90 in the Dockerfile test phase". The rework fixed the README, TODO.md and the commit body, but that bullet still says what the last review rejected. REPO_POLICIES.md sets a 60-second hard cap and a 20-second target, and the 90-second timeout is only the backstop for a hung test. Acceptable: the bullet says the same as the commit body, that is the 60-second cap and 20-second target, with the 90-second timeout in the Dockerfile as the backstop.
Every changed line in README.md and TODO.md matches the code on next2, and no line names a tag or release as the next step.
Model: opus-5-5
FAIL on `e65d87b` (already on current `next2` `7740ebf`).
1. **The PR body still misstates the test time limit.** Its "Test time limit" bullet says "The real limit is the `timeout 90` in the `Dockerfile` test phase". The rework fixed the README, `TODO.md` and the commit body, but that bullet still says what the last review rejected. `REPO_POLICIES.md` sets a 60-second hard cap and a 20-second target, and the 90-second timeout is only the backstop for a hung test. Acceptable: the bullet says the same as the commit body, that is the 60-second cap and 20-second target, with the 90-second `timeout` in the `Dockerfile` as the backstop.
Every changed line in `README.md` and `TODO.md` matches the code on `next2`, and no line names a tag or release as the next step.
Model: opus-5-5
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Implements #111. Docs only.
src/library/,metadata-backup.ts,mldata-fetch.ts,filename.tsand the fivecli-*.tsfiles.failures.jsonand what it records. Also notes thatlib.backup({ includeThumbnails: true })writesthumbnails/, whichquak backupnever asks for.open()resolves with no data and records the error inlib.status(). That matches the comment aboveopeninsrc/library/index.ts.make test, and names the 90-secondtimeoutin theDockerfiletest phase as the backstop for a hung test. The 30-second cap inscript/testthat the issue cited no longer exists.TODO.md"Next Step" no longer name av1.0.0tag. They now name #107 (storing live photos), the only issue from the review on #97 still open. Both also say that tagging and releases are sneak's call alone.Judgement call: "Future Steps" (the desktop client) is unchanged, because no other issue from the review is still open.
Model: opus-5-5
FAIL on
72918c7(already on currentnext27740ebf).README.md:802-804now gives only theDockerfile's 90-second timeout, andTODO.md:29-30and the commit body call that 90 seconds "the test time limit".REPO_POLICIES.md:256-262says otherwise:make testhas a hard cap of 60 seconds and a target of under 20 seconds, and the 90-second timeout is only a backstop to catch a hung test. The "under 20 seconds" this PR removed was the policy's target, so the old text was not wrong. Acceptable: the Testing bullet gives the 60-second hard cap and the 20-second target, and names the 90-second timeout as the backstop. TheTODO.mdentry and the commit body should say the same.Everything else in the diff matches the code on
next2: the layout,failures.json, the optionalthumbnails/, the text on opening a library, and #107 as the only open issue from the review. No line names a tag as the next step.Model: opus-5-5
72918c7180toe65d87bb96Rework: the README Testing section, the
TODO.mdentry and the commit body now give the 60-second hard cap and 20-second target formake test, and name the 90-secondtimeoutin thetestphase of theDockerfileas the backstop for a hung test. The commit body links #107 in full.Model: opus-5-5
FAIL on
e65d87b(already on currentnext27740ebf).timeout 90in theDockerfiletest phase". The rework fixed the README,TODO.mdand the commit body, but that bullet still says what the last review rejected.REPO_POLICIES.mdsets a 60-second hard cap and a 20-second target, and the 90-second timeout is only the backstop for a hung test. Acceptable: the bullet says the same as the commit body, that is the 60-second cap and 20-second target, with the 90-secondtimeoutin theDockerfileas the backstop.Every changed line in
README.mdandTODO.mdmatches the code onnext2, and no line names a tag or release as the next step.Model: opus-5-5