A single line if-statement in the LiveScript lexer causes an error when
YARD runs over the project. Replacing it with a ternary statement fixes
the problem.

The `@` symbol tells the command interpreter not to show any prompts.
This commit special cases it for `rem` so that using `@` also indicates
a comment.

This commit adds a lexer for BrightScript.

Co-authored-by: Dom Rowan <rowand@amazon.com>
Co-authored-by: Michael Camilleri <mike@inqk.net>

This commit adds a lexer for SSH configuration files.

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

The `&` selector does not match correctly if it is combined with certain
characters. This commit fixes the problem by adding a rule matching `&`
to the `:content_common` state.

The change in #1495 to the way that the `template` keyword caused
template parameter packs not to be highlighted correctly. This commit
fixes that bug while also introducing a more robust set of rules for
highlighting identifiers succeeding the `template` keyword.

#1551)

The ` character escapes characters in PowerShell statements. This
commit adds support for it to the `:parameters` state.

In addition to adding the `*.Dockerfile` file glob, #1059 added a file
glob for `Dockerfile_*` to address #1070. However, as noted by #1539,
this causes issues for files that begin with `dockerfile_` but that are
not Dockerfiles.

Adding the `Dockerfile_*` file glob was an error and this commit
removes it. It is a non-standard name and is not recognised by other
syntax highlighting libraries (e.g. Pygments, Chroma, Linguist). The
correct fix for the issue that precipitated #1070 is to use a
`.gitattributes` file.

The current Jinja lexer ends a `raw` or `verbatim` block when the lexer
encounters either `endraw` or `endverbatim`. This behaviour is
incorrect. The block should only end with a matching tag. This commit
fixes that behaviour.

Most of Rouge's lexers use rules that only match ASCII characters. This
is often not strictly correct as many languages support the use of
non-ASCII characters in their identifiers.

This commit adds support for non-ASCII characters to the CSS, HTML,
JavaScript, Julia, XML and YAML lexers. The regular expressions used
are more permissive than they should be if they were to be completely
correct but this is intentional. Ease of maintenance has been
prioritised over syntactic correctness.

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

This commit adds a lexer for Janet.

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

In single- and double-quoted strings, the Perl lexer currently
tokenises `\` as an error if it is not followed by a character in part
of a recognised escape sequence. This is a bug. Perl accepts the use of
`\` in single- and double-quoted strings even if it is not part of a
valid escape sequence.

This commit permits the use of `\` in single-quoted and double-quoted
strings as well as increasing the range of escape sequences that are
recognised in double-quoted strings.

The Rego lexer matches certain identifiers using `\b`. This will match
identifiers that include `_`. This commit replaces these rules with a
more robust generic regular expression. In addition, it changes the
token used for variable names to `Name`.

The `?` symbol is an alias for the `Where-Object` in PowerShell. This
commit adds support for it to Rouge.

The `=>` operator (sometimes called the 'fat comma') is an operator in
Perl that causes the identifier to the immediate left of the fat comma
to be treated as if it were a string. This commit adds support for it
using a simple lookahead heuristic.