Mostly so that the values have readable toString representations.

The openjdk images are no longer updated, aside from early-access
builds. See:

* https://hub.docker.com/_/openjdk/
* https://github.com/docker-library/openjdk/issues/505

Fixes platform-dependent behavior when using non-ASCII characters in
grammar source files.

Was added in 2012, not clear why this is needed.
d160106c

If anything, it seems to be *faster* with the default initial
capacities, but the difference is not very significant.

And also expand string concatenations in arguments of
StringBuilder.append, to avoid javac creating unnecessary intermediate
StringBuilders.

StringBuffer is synchronized for thread safety, but Lexer is already not
thread-safe anyway, so this is unnecessary overhead. The other uses of
StringBuffer were local in a single method and so can also be safely
replaced.