Repository navigation
coverage: judge every receipt against the parser and fail on new contradictions (D141) - #71
Merged
Merged
Conversation
…radictions (D141) The snapshot's case lists only reach receipts through catalogue forms, so a receipt on a retired or uncited case id never counted. The check now judges every current receipt by its statement (SQL, setup and session), with SQL from the citing case or from the hash-checked inputs of the run reports, and records receipt_contradictions and unjudged_receipts in the snapshot. diff fails on a contradiction the base did not have; documented (D113) ones go on their own non-failing line (D135), and existing ones are listed without failing. On catalogue 8c9639e and evidence a82d976, 4203044 -> cbe394e reports 53 new contradictions, including all 27 the D118 audit of #68 found, plus the 68 documented flips on their own line.
Contributor
Author
|
Landed with #72 via D94: main = abd037f, which contains this branch's head ff4f6d0. Gate exit 0 on abd037f; tree a1628e77 on both main and the head. The gate now fails on any receipt that the head newly contradicts compared with main. Regression check: against 4203044→cbe394ec it reports 53 new contradictions, including all 27 in the auditor's list, and 68 documented flips on their own line. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
D141 (lane B). Lands through the D94 gate by fast-forward; this repo has no GitHub CI.
Why #68's regressions passed the gate.
make check-mergerunssyntax-coverage diff coverage/syntax/baseline.jsonon main's tree. That tree carries only the first 10 receipt imports, and the baseline (catalogue 443acff) has no parsed/refused case lists, so it can never report a parsed case newly refused. Separately,snapshotonly reaches receipts through catalogue forms → obligations → cases, so receipts on retired or uncited case ids never counted, even on full evidence.Change. The check judges every current receipt by statement (SQL, setup, session), with SQL from the citing case or from the hash-checked
inputsof the run reports. It recordsreceipt_contradictionsandunjudged_receiptsin the snapshot.difffails on contradictions the base did not have.Regression check (catalogue 8c9639e, evidence a82d976), D141 binaries built at 4203044 and at cbe394e:
diffexits 1 with 53 new receipt contradictions. They include all 27 in the auditor's D118 list for #68, plus 26 more real ones: 8 catalogue accept cases now rejected (ALTER DYNAMIC TABLE ×7, CREATE EXPERIMENT) and 18 on uncited case ids (10 CREATE PROCEDURE backtick ARTIFACT_REPOSITORY 1003s now accepted, 4 CREATE EXTERNAL TABLE and 4 old hand rows now rejected). The 68 documented flips are on their own line, matching the auditor's 68. 30 receipts have no SQL in the tree and are listed as unjudged.Tests: the coverage crate's tests, including new fixture tests for an uncited receipt judged from report SQL, an unjudged one, and the diff lines. A mutation that drops the report-SQL lookup fails the new test.