Add text primitives in DAML 1.dev for experiments.

The idea is to add LF primitives for:

• toUpper/toLower
• replicate/fill
• break/breakOn/stripInfix… and other missing List counterparts
• containsOnlyAlphabet: Text -> Text -> Bool (useful for isDigital, isAlpha, isWhitespace, isAscii and similar functions)
• split/join

cc @cocreature @jberthold-da

[ghost]

@cocreature pointed out that take and drop aren’t primitive, so I’m adding those.

[cocreature]

I took a brief look at the unicode standard, in particular, section 4 in https://www.unicode.org/versions/Unicode12.0.0/ch04.pdf paragraph “Case Mapping”. It looks like a nice option might be to use the data from UnicodeData.txt which is not dependent on weird things but not use SpecialCasing.txt because nobody likes special cases and that is context dependent.

However, I have no clue how easy it is to do that on the JVM or anywhere else so while this seems easy to specify and less restrictive than nly allowing ASCII, I’m not sure it is actually a reasonable option.

I don’t really like A2 and it seems to go against the spirit of those primitives which are all about performance and I agree that A3 is not a very nice specification even though it probably does what you want™ in basically all cases. So I’m leaning towards A1, what do you think @remyhaemmerle-da? @filmackay Would ASCII be sufficient for your usecases?