Dictionary curation 2025 03 06

[hippietrail]

Issues

N/A

Description

• Add words to dictionary
• Adjust affix annotations
• Move placenames into the new canonical case system
• Add "VS Code" to the canonical case system
• Sort parts of the canonical case JSON file that I modified

Demo

N/A

How Has This Been Tested?

Tested on place names that previously didn't work or needed special rules.

Checklist

• I have performed a self-review of my own code

[elijah-potter]

These hyphenation rules can be made into phrase_corrections entries.

[hippietrail]

Cool. Will look into that today then....

As it turns out, unless I'm missing something, phrase_corrections will only add the hyphens, it won't change the case. So it can correct Guinea Bissau but will incorrectly "correct" guinea bissau to guinea-bissau, Guinea bissau to Guinea-bissau, etc.

If you go with the quick fix for those then the canonical case linter will kick in, but this might not be what we're really aiming for.

I'll move the one case of each that can be fixed, but maybe the best solution is to change "canonical case" into "canonicalize" and it should handle at least case and hyphenation, but ideally also closed vs open vs hyphenated compounds and hyphenation and missing/extra apostrophes, missing/extra periods, standardizing abbreviated vs full spelling, for things like "Saint" vs "St."

[elijah-potter]

They can be made into separate rules. An entry for proper noun capitalization and an entry for hyphenation. In other words: just do the quick fix.

I may start looking into how we can collapse multiple Lints into one soon, since we're starting to hit the critical mass you mentioned a while ago.

[hippietrail]

They can be made into separate rules. An entry for proper noun capitalization and an entry for hyphenation. In other words: just do the quick fix.

I may start looking into how we can collapse multiple Lints into one soon, since we're starting to hit the critical mass you mentioned a while ago.

I'm not sure how to proceed with this? Is it possible to at least merge in the dictionary.dict curation part as it's going to get more and more out-of-sync with later curation PRs.

[hippietrail]

Oops I accidentally pushed 3d0c3c3 here instead of my new curation branch. I tried to revert it with git reset --hard HEAD~1 and git push -f origin dictionary-curation-2025-03-06 which I thought worked, but apparently it didn't. I don't want to fumble around and make it even worse \-:

[elijah-potter]

I'm a little confused to your meaning. Are you wanting to only submit the dictionary.dict changes?

[hippietrail]

I'm a little confused to your meaning. Are you wanting to only submit the dictionary.dict changes?

I think one problem is I keep forgetting which file can fix which things because it's not in the name of the file, the name of the linter, or the comment at the top of the file.
By this I mean which file has lints that can add hyphens to phrases vs which has lints that can correct the case of hyphenated phrases.

I'm going to try to narrow this down, include tests, and include comments. But I might miss something. I'll push my best effort and then merge everything that looks right to you.

Maybe we could have a file somewhere just for testing place names and/or proper nouns that can include tests for cases, hyphenation, apostrophes, accents, etc, without having to know which source file is responsible.

[hippietrail]

Is the precommit really still failing the PHP test? I can't grok why? It says it gives correct diagnostics but then there's an error.

Failures:
1) Languages > gives correct diagnostics for PHP files
  Message:
    Error: Expected 1 diagnostics, got 0.
  Stack:
    	at compareActualVsExpectedDiagnostics (/home/runner/work/harper/harper/packages/vscode-plugin/build/tests/suite/helper.js:38:15)
    	at UserContext.<anonymous> (/home/runner/work/harper/harper/packages/vscode-plugin/build/tests/suite/languages.test.js:50:61)

30 specs, 1 failure
Finished in 8.873 seconds
Error: Tests failed
	at Object.run (/home/runner/work/harper/harper/packages/vscode-plugin/build/tests/suite/index.js:19:15)
	at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
Extension host test runner error Error: Tests failed
	at Object.run (/home/runner/work/harper/harper/packages/vscode-plugin/build/tests/suite/index.js:19:15)
	at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
Asking native host service to exit with code 1.
[main 2025-03-21T15:32:42.370Z] Extension host with pid 11581 exited with code: 0, signal: unknown.
Exit code:   1
Failed to run tests TestRunFailedError: Test run failed with code 1
    at ChildProcess.onProcessClosed (/home/runner/work/harper/harper/node_modules/.pnpm/@vscode+test-electron@2.4.1/node_modules/@vscode/test-electron/out/runTest.js:110:24)
    at ChildProcess.emit (node:events:518:28)
    at ChildProcess._handle.onexit (node:internal/child_process:293:12) {
  code: 1,
  signal: undefined
}
 ELIFECYCLE  Test failed. See above for more details.
error: Recipe `test-vscode` failed with exit code 1
Error: Process completed with exit code 1.

---

After the last merge with master this is no longer happening. 🎉