Repository navigation
Parse dollar symbols when a matching currency is configured - #446
Open
ryanduguid wants to merge 1 commit into
Open
ryanduguid wants to merge 1 commit into
ryanduguid wants to merge 1 commit into
Conversation
ryanduguid
added a commit
to ryanduguid/release-policy
that referenced
this pull request
Oct 14, 2026
## Summary Moves the central PR-review caller from reviewer `6b20437` (3 October) to current main `1cd5539`, and updates its baseline digest. At `6b20437` the reviewer caps output at 12,288 tokens. MiMo reached that cap (`finish_reason: length`) on the largest chunk of JavaMoney/jsr354-ri#446, so the review failed with `incomplete_completion` on 3 October (run 37084986655) and again today (run 38495499693). The current policy allows 32,768 output tokens and includes #139 to #141 (reviewer output qualification, generation metadata, Python 3.14). ```diff - uses: 6b20437 + uses: 1cd5539 ``` The `central_public_report_v1` inputs and the `OPENROUTER_API_KEY` secret are unchanged in `pr-review.yml` between the two commits. `pr-review-consumer.yml` keeps its pin; #142 is changing that caller. ## Evidence - **Before:** run 38495499693's receipts show GLM 5.3 finished both chunks (`stop`), and MiMo v2.6 Pro stopped at 12,288 completion tokens (`length`, `completion_contract`). - **After:** `tests/test_repository_baseline.sh` passes with the new digest, and actionlint is clean. ## Merge danger **Door:** two-way. **Blast radius:** advisory reviews The review is advisory and runs only on Ryan's manual dispatch. The max review spend stays at US$3. ## Unverified The paid review at the new pin runs after merge. Co-authored-by: Ryan Duguid <ryanduguid@users.noreply.github.com>
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.
Summary
A symbol format configured with a dollar currency currently rejects
$1000.00with$ is not a unique currency symbol.Use that suppliedCurrencyUnitwhen the input matches its full symbol in the format's locale, and consume only that symbol. For example, a US format configured with USD can now parse the amount as USD 1,000.Fixes #423. The change covers the existing dollar-rejection branch. Dollar input without a matching configured unit still fails; the configured unit remains a fallback for other parsed currencies.
Evidence
$US, adjacent amount text, small and negative amounts, mismatching hints and ambiguity controls.mvn -B -DskipTests -Dmaven.test.skip clean packagebuilds pass, using isolated settings and the public 1.5-SNAPSHOT parent installed locally.CurrencyToken.parse.testFailureIgnore.git diff --checkpasses. No dependencies or POM files change.Risk
Previously rejected dollar inputs become valid when the configured unit's locale symbol matches. The tests require the supplied unit's identity and full symbol consumption, and retain rejection when a hint is absent or does not match.
Unverified
A fully passing reactor build, Java 21.0.6 (hosted CI's version), Windows, other JDKs and upstream CI are unverified. The Coveralls upload needs credentials and was skipped; local tests and JaCoCo ran.
This change is