Implement guessing and resolution between TypeScript and TypoScript.

These values would trigger problems when guessing content without a filename,
and really aren't needed. They've been removed, allowing the standard behavior
to take place.

On top of this, the regex for TypeScript has been slightly altered to include
a whitespace match after "export" to be within the group for the "export"
text.

The new TypoScript lexer was taking precedence over TypeScript, which share
the same file extension. This meant that any previously-working TypeScript
source files would be misidentified and highlighted incorrectly. This was due
to the lack of a guesser for TypeScript, and TypoScript having a naturally
higher priority.

This change balances things a bit between the lexers. TypoScript now has an
explicitly-higher priority than TypeScript, both for the file extension and
defaults when guessing content, as it has less we can rely on in terms of
content. TypeScript has a lower priority, but has an analyse_text() method
that checks for common symbols in code, giving it a better chance of matching
for actual TypeScript source files.

TypeScript also gains an additional file extension, *.tsx. Ideally, this would
have its own lexer that understood inline JSX syntax, but giving general
TypeScript syntax highlighting for these files is a step in the right
direction.

update LassoLexer

Add support for string literals to TypeScriptLexer.

javascript.py edited online with Bitbucket

This introduces support for string literals to the TypeScript lexer. The rules
are consistent with those in JavascriptLexer, helping to ensure the same
behavior and capabilities between the two.