Discussion
Papyrix includes a support library (integrated into its fancier text layout) for auto-hyphenating words, finding all the place might be hyphenated. So how does it work?
There's a public API which reformats the results from its auto-hyphenation & exposing the language-setter.
That setter invokes which helper which strips the region off the language-code & converts ISO639-2 into into ISO639-1 codes via lookuptable, to get the appropriate hyphenation data via another lookuptable.
1/?
In the happy path for each char it checks whether we're at the end of the word or at an explicit hyphen & if either of those are the case it hands that segment of text off to the language-specific hyphenator reformatting results.
There's a sourcecode file aiding detecting/processing different charsets, some of which I've mentioned. And I've already discussed the language registry.
Through a layer of indirection Papyrix uses Liang's algorithm, which looks a bit involved so... Tomorrow!
3/3Fin!
Meanwhile to compute break offsets Papyrix validates it has a word to process, reformats the text (via a helper in another file) into UTF-32 annotated with UTF-8 offsets, parses off surrounding punctuation & footnotes, locates various forms of hyphens & contiguous latin characters into the tentative results, if data's available iterates over the codepoints sorting & deduplicating results, otherwise defers to a method on the hyphenator or breaks on all chars in centre of the word.
2/?