How code is analyzed
The spell checker is designed to report genuine spelling errors in source code while keeping the number of false reports low.
Identifiers are split into words
Compound identifiers are separated before each part is checked:
| Identifier | Checked as |
|---|---|
camelCase | camel case |
HTMLInput | html input |
snake_case_words | snake case words |
camel2snake | camel snake |
All-capital words
Trailing s, ing, ies, es, and ed are retained with the preceding word, so
CURLs is checked as curls and CURLedRequest as curled request.
What is not checked
- Words of three characters or fewer.
- All symbols and punctuation.
- URLs, hex values, escape characters, and Base64 blocks longer than 40 characters.
- Email addresses.
Known limitations
- Checking is case insensitive:
englishis not reported in place of English. - The dictionaries contain omissions, and occasionally errors.
- Compound word support is set per file, not per region.
The full list of predefined include and exclude patterns is documented in In-document settings.