[DAPHNE-#659] Scientific notation for float literals and digit separators

[AlexRTer]

This PR closes #659. It adds scientific notation for float literals as well as digit separators for both float and integer literals to improve readability.

Features:

• Scientific (E) notation that is case insensitive (accepts e and E)
  • e.g. 1.2e3 = 1200, 1E-3 = 0.001, both of type f64
• Digit separators ' and _ for float and integer literals
  • 1e6 = 1'000'000 = 1_000_000

Changes:

• Added scientific notation and digit separators to DaphneDSLGrammar.g4
• Updated marked daphne scripts with scientific notation
• Added a few test cases for literals
• Update to the LanguageRef doc

[AlexRTer]

Thank you for your feedback, I will go ahead and implement some better testing as well.

[AlexRTer]

I changed the grammar slightly to disallow digit separators right behind the decimal point.

The tests now include some negative tests, i.e.

• infff, 1., 01.0, 10.0_, 1.'0

and cover (corresponding file index in brackets)

• [1] special values (e.g. nan, inf)
• [2] zeros
• [3,4] simple floats with varying length (e.g. 12.34)
• [5,6] floats with digit separators (e.g. 1_000.0)
• [7,8] scientific notation without decimal points (e.g. 1e6)
• [9,10] scientific notation with decimal points (e.g. 0.000000e-6)
• [11] scientific notation with digit separators (e.g. 1.0_0e6)

For all tested ordinary values there is a test without signs, with a negative sign, as float type and mixed (negative and float type). The tests also cover small and capital e for scientific notation as well as +, - signs.

I believe a test to verify that e.g. 1.0f is of float type is still missing in Daphne. Maybe a built-in typeof function would be a good addition to DaphneDSL.

[pdamme]

By the way, there is an open issue on introducing a typeof() built-in function in DaphneDSL: #357.