Kotlin allows three keywords to be used inside a list of generic
parameters: `in`, `out` and `reified`. In addition, it supports the use
of `:`. This commit adds support for these elements to the Kotlin lexer.

The list of keywords that is recognised in the current LLVM lexer is
not as complete as in some other syntax highlighting libraries (such as 
Pygments). This commit adds a Rake task that checks a source file in
the public LLVM repo and generates a keyword file that is loaded on
request when the LLVM lexer is used.

Rouge currently loads various components of the library using `Kernel#load`. In certain conditions, this can cause issues where
`load` is redefined at the global level. To ensure maximum
compatibility with other libraries, this commit replaces all calls to
`load` with qualified calls to `Kernel::load`. It also updates one call
to `Object.send` with a call to `Object::send` for consistency.

Version 3.18.0 of Rouge included a new mechanism for handling strings
in the Python lexer. One of the consequences of that change was that
raw strings would break if they included 'invalid' escape sequences.
This is a mistake as raw strings do not have 'invalid' escape sequences.

This commit fixes this error by changing the approach that the Python
lexer takes to escape sequences more generally. Rather than dividing
recognised and unrecognised escape sequences into 'valid' and 'invalid',
it simply treats unrecognised escape sequences as ordinary strings.

The result of this is that all escape sequences in raw strings produce
`Str` tokens. In other types of strings, recognised escape sequences
produce `Str::Escape` tokens and unrecognised escape sequences produce
`Str` tokens.

This commit adds the `*.cljs` fille glob to the JavaScript lexer.
Node.js treats files ending in the `.cjs` extension as CommonJS modules.

Co-authored-by: Michael Camilleri <mike@inqk.net>

Python supports the use of Unicode in identifiers. This commit uses
POSIX bracket expressions that match against Unicode characters rather
than the more common character classes that only match ASCII characters.

Co-authored-by: Michael Camilleri <mike@inqk.net>

The use of a ternary in code added to the Python lexer as part of the
#1508 pull request results in a RuboCop warning concerning parentheses
and grouped expressions. This commit fixes that warning.

This commit improves the handling of numbers in the Kotlin lexer.
Specifically, it:

1. adds support for underscore separators;
2. adds support for binary literals;
3. removes invalid support for the lowercase "l" suffix; and
4. adds support for unsigned literals.

An empty prefix (`:` as distinct from `<prefix>:`) is permitted in
Terse RDF Triple Language. However, the current lexer requires at least
one character before the `:`. This commit removes that requirement.

The case-sensitive keyword `a` is a valid keyword in SPARQL. At
present, the SPARQL lexer does not recognise it as such. This commit
fixes that error (the rule is similar to the one used in the Turtle lexer).

A number of the Rake tasks that generate lists of built-in keywords
for particular languages have bugs. These range from actual errors (as
in the case of the Lua task) to deprecation warnings (in the case of
tasks using `Kernel#open` to open URIs).

This commit fixes these bugs and includes freshly generated lists of
the built-in keywords that have changed.