test: the reading pass over test/o* - #740
Merged
Merged
Conversation
3 files read; changes in 2. oromo is clean — its 322 lines carry a lot of
"no longer" phrasing, but every instance describes a RULE-ORDERING consequence
rather than the file's own history ("step 5 consumes the meridiem, so a
following enclitic no longer has digits in front of it"), which is the porting
fact.
A DANGLING FRAGMENT, eighth variant: occitan's "thin single source." after
asturian, bavarian, haitian, icelandic, kabuverdianu, luxembourgish and
maltese. Restated with the part that actually constrains the reader: 675
headwords, SMALL and DIALECT-MIXED — which is the ceiling on what any referee
can say about a PAN-DIALECTAL orthography, where one spelling has
dialect-specific readings by design.
Referee percentages out of occitan and odia.
Clean on the other detectors: no mixed-script runs, no missing-space typos, no
attesting tallies.
227 files / 3131 tests unchanged, typecheck clean. Comment-only.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012Lc3WnUgogC7okV7n53vjr
…gate main was broken: test/manifest-script.test.ts imports MANIFESTLESS_SCRIPTS and the table was not there, so two tests failed on the tip. MY MISTAKE, and worth recording because the shape is generic. To prove the new coverage gate was not vacuous I deleted the `lo` row, confirmed it failed, then reverted with `git checkout -- src/core/scripts.ts`. But the table was still UNCOMMITTED at that point, so the checkout did not undo my one-line break — it undid the whole addition. I then committed the test that imports it, and merged. ⚠ THE VERIFY-BY-BREAKING TECHNIQUE IS RIGHT; DOING IT ON UNCOMMITTED WORK IS NOT. `git checkout --` reverts to HEAD, not to "before my last edit", so it is only safe once the thing being broken is committed. Commit first, break, revert, then continue. It also slipped through because I tail-ed the verification output and read a timing line rather than a pass/fail line — the run after the revert showed only "Duration 461ms". A check whose result I did not actually look at. 227 files / 3131 tests pass, typecheck clean. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_012Lc3WnUgogC7okV7n53vjr
…rong I justified the table with "a manifest exists to hold DATA, and a file whose only content is its own script name is a worse record". That inverts what manifests are for: a manifest is a language's encyclopedic entry — phonology, orthography, the concerns a reader needs before touching the engine — and the data being IN it is what keeps the entry honest, not the reason it exists. So the 26 single-.ts engines are missing a RECORD, not a field. Their documentation is written — 8–20 line .ts headers, longer than the prose in several manifests that do exist — it is just in the wrong file, with ~800–3,300 chars of tables inline beside the code. The table stays for now because it is what makes coverage complete and the union assertable, but it is interim for that third case and says so. Tracked in #741. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_012Lc3WnUgogC7okV7n53vjr
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
1 participant
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.
3 files read, changes in 2.
oromois clean. Its 322 lines use "no longer" a lot, but every instance describes a rule-ordering consequence rather than the file's own history — "step 5 consumes the meridiem and emits a half-day WORD, so a following enclitic no longer has digits in front of it" — which is exactly the porting fact a reader needs. Different thing from theThis expectation previously pinned…pattern I've been cutting.A dangling fragment, eighth variant
occitan's
thin single source.— after asturian, bavarian, haitian, icelandic, kabuverdianu, luxembourgish and maltese. Eight files, eight slightly different truncations of one sentence.Restated with the part that actually constrains a reader: 675 headwords, small and dialect-mixed — which is the ceiling on what any referee can say about a pan-dialectal orthography, where one spelling has dialect-specific readings by design. A low folded score there is a property of the writing system, not a defect.
Also
Referee percentages out of
occitanandodia.Clean on the other detectors: no mixed-script runs, no missing-space typos, no attesting tallies.
227 files / 3131 tests unchanged, typecheck clean. Comment-only.
🤖 Generated with Claude Code
https://claude.ai/code/session_012Lc3WnUgogC7okV7n53vjr