In scan for TimeUnit fragment allow two unicode representations of mu

[jlapeyre]

Previously only one unicode character for mu was recognized as the literal for microsecond. Now two are recognized. The new one appears frequently in Python. It also is used in qss-qasm.

Add tests for the three ways to specify microsecond.

Closes #422

Summary

Recognize two different characters representing mu in the Lexer. Write them as unicode literals in order to remove ambiguity, since the glyphs are typically indistinguishable.

[jakelishman]

Please can this come with an update to the text of the spec to make the explicit Unicode points clear as well?

[k4rtik]

According to Wikipedia, there seems to be a difference between \mu and \micro in Unicode. The lexer already supports us as an alternative for writing microseconds (if I am reading it correctly).

Another, perhaps worse, solution to this issue is to explicitly state that the character is \micro, i.e., U+00B5. Though this PR's solution does not seem problematic as the conflict between the two characters is going to be restricted to time unit parsing.

Further, the article for micro- states:

Micro (Greek letter μ (U+03BC) or the legacy symbol µ (U+00B5))

[jakelishman]

From a pragmatic perspective, it'd be good to make sure that people who use TeX-to-unicode plugins for their text editors have an easy way to insert the correct symbol, since I'm guessing (I don't use one myself) that's the intended usecase for this over just using "us". I would imagine that \mu will produce a mu-like letter, and \micro isn't a standard TeX macro, even though the micro codepoint is probably semantically more appropriate.

[jlapeyre]

This should not be merged yet. See #440.

(Is there a don't merge label? I am unable to add labels)

[blakejohnson]

To be revisited after #477.

[blakejohnson]

Superseded by #477 .