TokenTypes documentation contains double encoded symbols

[pbludov]

This file
http://checkstyle.sourceforge.net/apidocs/com/puppycrawl/tools/checkstyle/api/TokenTypes.html
http://checkstyle.sourceforge.net/apidocs/com/puppycrawl/tools/checkstyle/api/JavadocTokenTypes.html
is hard to read. Many symbols are displayed in encoded form.

Some examples:

Modifier and Type	Field and Description
static int	LNOT The ! (logical complement) operator.
static int	LE The <= (less than or equal) operator.
static int	BSR The >>> (unsigned shift right) operator.
static int	BSR_ASSIGN The >>>= (unsigned right shift assignment) operator.

It looks like encoded symbols were not decoded when <code> </ code> was replaced with {@ code}. It's a bug in the Idea actually. But the docs need to be fixed anyway.

regression caused by 7078670 in scope of #4805

see limitations of @code tag at - #5425

[rnveach]

I agree with this. Users shouldn't need to know HTML encoding to understand the document.
Nothing matching regular expression \&\w\w; should be displayed.
I can't imagine we would display something like this to user on purpose.

[romani]

fix is merged.