ambiguity, complication in the lexer ?

[SVGAnimate]

Description

Hello,

In the file /Zend/zend_language_scanner.l

Isn't the following code redundant?

<ST_IN_SCRIPTING>"&"{OPTIONAL_WHITESPACE_OR_COMMENTS}("$"|"...") {
	yyless(1);
	RETURN_TOKEN(T_AMPERSAND_FOLLOWED_BY_VAR_OR_VARARG);
}

<ST_IN_SCRIPTING>"&" {
	RETURN_TOKEN(T_AMPERSAND_NOT_FOLLOWED_BY_VAR_OR_VARARG);
}

If so, this complicates the reading of rules parser (/Zend/zend_language_parser.y)

ampersand: ...
is_reference: ...
...

why not just use the token T_AMPERSAND_NOT_FOLLOWED_BY_VAR_OR_VARARG (T_AMPERSAND) ?

Thank you for your explanations

PHP Version

PHP 8

Operating System

No response

[iluuu1994]

See f9c78be It's needed to work around a parser ambiguity with & in intersection types and by-reference parameters.

Please ask these questions on a more appropriate channel. E.g. in #php-internals on https://phpc.chat/.