[Bug] Case helpers (strCamelCase, strKebabCase, strSnakeCase, strCapitalizeWords): case not normalised, inconsistent word boundaries, non-ASCII letters dropped #634

Version: @nevware21/ts-utils 0.16.0 (0.15.0 gives the same output for the inputs checked)
Environment: Node v22.23.4, the package's dist/es5/mod build

Summary

All four helpers are built on two regular expressions:

• _convertCase: /((_|\W)+(\w){0,1}|([a-z])([A-Z]))/g
• strLetterCase: /(_|\b)\w/g

Without the u flag, \w, \W and \b cover ASCII only, and the conversion only rewrites the character after a boundary. That causes the five behaviours below. Sections 1, 3 and 4 look like bugs; section 2 is a word-boundary rule worth deciding; section 5 is a question. Every "Actual" value comes from running 0.16.0. The "Suggested" column is a proposal, not a claim about intended behaviour.

1. Bug: strCamelCase does not normalise the case inside a word

Only the first character is lower-cased (/^\w/) and each later word's first letter upper-cased. The rest of each word keeps its case, so all-caps input loses its word boundaries entirely (hELLOWORLD).

2. Word boundaries inside a word are found inconsistently (kebab and snake case)

• ([a-z])([A-Z]) splits after a lone lower-case letter (h + ELLo).
• It never splits an acronym from the word after it (ABCDef, XMLHttp).
• It never splits a digit from an upper-case letter (2Baz).
• strCamelCase on the same hELLo input does not split there at all, so camel case and kebab case disagree about where the words are.

3. Bug: letters outside ASCII are treated as separators and dropped

é, è, ü and ß match \W, so camel, kebab and snake case replace them with the separator and lose them. \b sees a word boundary after them, so capitalise-words capitalises the next letter (CrèMe). Unicode property escapes ([\p{L}\p{N}] with the u flag) would keep them. They need ES2018, which may not fit the ES5 build, so a character-class alternative or an option may be needed there.

4. Bug: leading and trailing separators become a dangling separator

Only whitespace is trimmed before conversion, so any other leading or trailing punctuation becomes a separator at the edge. Camel case drops it, so camel case and kebab/snake case disagree here too.

5. Question: should strCapitalizeWords keep camel-case word boundaries?

It lower-cases everything before capitalising, so camel-case input becomes a single word. This may be intended; if so, it would help to document it.

Repro

const { strCamelCase, strKebabCase, strSnakeCase, strCapitalizeWords } = require("@nevware21/ts-utils");

console.log(strCamelCase("hELLo dARKness-my old_friend")); // hELLoDARKnessMyOldFriend
console.log(strCamelCase("HELLO WORLD"));                  // hELLOWORLD
console.log(strKebabCase("hELLo dARKness-my old_friend")); // h-ello-darkness-my-old-friend
console.log(strKebabCase("XMLHttpRequest"));               // xmlhttp-request
console.log(strKebabCase("fooBar2Baz"));                   // foo-bar2baz
console.log(strCamelCase("café crème"));                   // cafCrMe
console.log(strCapitalizeWords("café crème"));             // Café CrèMe
console.log(strKebabCase("__init__"));                     // -init-
console.log(strSnakeCase("O'Brien & Sons, Ltd."));         // o_brien_sons_ltd_
console.log(strCapitalizeWords("fooBarBaz"));              // Foobarbaz

Compatibility

Any of these fixes changes output that existing callers may rely on, including code that ports these helpers and must match them output for output. A changelog entry, or an option on each helper that selects the new rules, would let callers move deliberately.