Automattic / Automattic/harper
A bunch of false positives
- Dominant language
- Rust
- Stars
- 15.4k
- Forks
- 627
- Avg merge
- 1d 15h
- Merged PRs (30d)
- 106
Description
1. `The projects that want to take it seriously are the best.`: "Did you mean 'wants'" → no, "projects" is plural (#1302)
2. `This is nothing new.`: "Did you mean 'knew'" → obviously not (#1303)
3. `A printable format and layout.`: "This word might be a phrasal verb rather than a compound noun." → I think this is perfectly fine here though (#1304)
4. ~~`That's a very stupid idea.`: "The possessive noun implies ownership of the closed compound noun 'Avery'" → no idea what's going on here~~
5. ~~`The root directory of a repository.`: "The word 'of' is not needed here" → I think it is...~~
6. ~~`The inverse of a function.`: "The word 'of' is not needed here" → Here too...~~
7. `A 2.5" hard drive.`: "This quote has not termination" → Doesn't like quotes used as symbol for inches
8. `Plot the data on the x-axis and y-axis.`: "Did you mean to spell 'y' this way?" → Seems to support x-axis but not y-axis
9. ~~`It is composed of a chain of words.`: "The word 'of' is not needed here" → a 3rd case of this error~~
10. `A series of tests that cover all possible cases.`: "Did you mean the closed compound noun 'coverall'?" → no
11. `OpenGL and OpenPGP on OpenBSD.`: "Did you mean to spell 'OpenBSD' this way?" → apparently it knows a bunch of Open* projects, but obvious ones are left out. I don't know what the policy overall with regards to the incomplete dictionary, for example how about "CD" and "UTF-8" which aren't recognized either?
12. `Which call(s) exactly?`: "Did you mean to spell 's' this way?" → classic case of optional plural not recognized
13. `The story of how and why things came to this point.`: "Insert 'to' after 'how'" → not appropriate here
Version: Harper 0.36.0
Edit: updated to Harper 0.37.0
Edit: for `13.`, another test case: "How everything is tied together."
Contributor guide
Research direction
Reproduce the reported examples against Harper 0.37.0 and identify which diagnostics produce each false positive. Review the affected grammar rules and their existing tests, if present; done means the listed valid sentences no longer trigger incorrect warnings while genuine errors remain detected.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- devtools, testing-qa
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100