Repository navigation
XL-2: scope lowering occurrence projection - #12560
gunbai-bot[bot] wants to merge 4 commits into
Conversation
…t the pin's revision sample_paths_by_rule is the single authority for reference_conservation_stratified_ sample_paths: strata of dag/ (2 segments) and src/v2/ (3), quota max(6, round_half_even(300*share)), 8,000-byte cap (sample_file_within_cap: code points bound bytes from below, so only short text is encoded), indices floor(i*n/k). reference_conservation_sample_by_rule_at_pin_exit reads the tree at reference_conservation_stratified_sample_revision out of the object store (collect_step0_subject_vector_at) and refuses unless the rule reproduces the pin. Receipt: rc=0 over all 6,880 .dag blobs at 8fcd8e7 (536 s). Floor claims pin each part of the rule on supplied file sets, each red by mutation (half-up rounding, ceil indices, no floor of 6, code points for bytes). Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…guous bare ends_with Floor refused 3832455 with UnimportedBareProvider: ends_with also names a corpus function (gunbc.rust_item_scan), so the bare call is an unimported provider in a file that declares imports. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…ot by a gate Review 71995: 'checked rather than asserted' claimed more than any gate enforces. The comment now states that the pin-to-revision check executes only when the instrument is run, and that a repin must run it again. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 082f44beca
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| let pinned = sort_by(reference_conservation_stratified_sample_paths, p => p) | ||
| let missing = sample_paths_not_in(xs: derived, ys: pinned) | ||
| let extra = sample_paths_not_in(xs: pinned, ys: derived) | ||
| if length(xs: missing) == 0 && length(xs: extra) == 0 && length(xs: derived) == length(xs: pinned) { |
There was a problem hiding this comment.
When a future repin contains the correct paths but accidentally changes their order, this check still returns ExitSuccess: sorting pinned and then checking only membership and length treats the lists as sets. That contradicts the instrument's stated exact-output check, and reference_conservation_stratified_sample_census iterates the literal list in order, so the resulting census will no longer be the rule's byte-for-byte ordered output without the verifier detecting it. Compare the derived and pinned sequences directly (while retaining the set differences only for diagnostics).
Useful? React with 👍 / 👎.
|
Closing without merging. This PR was opened automatically from tidy-otter-111's archive flush onto its already-merged branch (#12475). It carries only scratch probes (.scratch_xl2/...) plus that branch's landed content. Nothing here should land. — sent from quiet-seal-543 |
Auto-opened by session-dashboard for session
tidy-otter-111.Pushing to
session/tidy-otter-111-sample-ruleadvances this PR.Worker attestation
Before flipping this PR to ready for review, confirm each item:
npm test,cargo test) and the result.Closes #Ndirective.Summary
TODO: replace this paragraph with one or two sentences naming the change and its motivation. Reviewers read this first.
Test plan