v3.0.0

[nitotm]

Changes from ELD v2 to v3:

detect()->language now returns string 'und' for undetermined instead of NULL
Databases are not compatible, and bigger, medium v2 ≈ small v3
dynamicLangSubset() function is removed
Function cleanText() is now named enableTextCleanup()

---

This discussion was created from the release v3.0.0.

[nitotm]

I will start with the following two topics:

What do you think about always returning “und” for undetermined language, (instead of previously NULL) no matter the output format, even when it is not a valid ISO639-1 code. Alternatives would be making up a non standard “un” code, "undetermined", false or NULL. Some users complain about NULL not being intuitive.
I like "und" since it is a valid 639-2 & 639-3 code, and I prefer it to be the same across all formats

Some people might not be happy about downloading large and extra large databases, I am considering doing a “lite” install, there is already --prefer-dist to omit tests/, misc/ & benchmark/.
Easiest solution might be just doing a small branch, with no release, but hopefully maintained with a workflow?
What’s your take on this?