scheme.py: wrong match of ASCII digits

[locsite]

In all regular expressions defined by lex/scheme.py, the sequence "\d" is erroneously used for matching an ASCII digit. By default, in Python 3 that sequence matches too much: all Unicode digits. For example, it matches the full-width digit "４" (U+FF14), which is not a legal ASCII digit, as required by the lilypond syntax.

This generates wrong colorization patterns in "Frescobaldi" when a source file contains Unicode digits outside the ASCII set.

The "\d" sequence should be replaced with the safer "[0-9]" sequence in all regular expressions where an ASCII digit is intended.