Re: [STANDARDS] Questions re keywords

On Fri, Dec 18, 2015 at 9:36 PM, Rex Jaeschke <[email protected]> wrote:

> As part of my work at Facebook in updating the formal spec for V7, I was
> working on the new intrinsic, assert, which led me to the following
> questions:
>
>
> 1.       Should `assert` be added to the list of keywords?
>

Nope, assert is an ordinary function with some extra handling.


> 2.       If so, which other intrinsic names should be listed as keywords?
> (Right now, only `echo`, `list`, and `print` are listed.)
>
> Stas pointed me to
> https://github.com/php/php-src/blob/master/Zend/zend_language_parser.y#L100
> where I see
>
> %token T_ARRAY           "array (T_ARRAY)"
> %token T_ECHO       "echo (T_ECHO)"
> %token T_EMPTY      "empty (T_EMPTY)"
> %token T_EVAL         "eval (T_EVAL)"
> %token T_EXIT      "exit (T_EXIT)"
> %token T_ISSET      "isset (T_ISSET)"
> %token T_LIST            "list (T_LIST)"
> %token T_PRINT        "print (T_PRINT)"
> %token T_UNSET      "unset (T_UNSET)"
>
> From that it seems to me they should all be listed as keywords.
>

Yes, all of these are keywords.

This list might be helpful to compare against, though I didn't check if
it's up to date: http://php.net/manual/de/reserved.keywords.php


> 3.       If `exit` is a keyword, shouldn't `die` also be one?
>

Yes, "die" is also a keyword. The list in the parser doesn't include it
because "exit" and "die" are merged into the T_EXIT token.


> 4.       Based on the following extract from the Zend parser grammar
> file,  %token T_YIELD_FROM   "yield from (T_YIELD_FROM)", it seems to me
> that `yield from` is treated as a keyword containing whitespace, correct?
> While unusual, this not new to me. (When I produced the Ecma spec for "C++
> with extensions for .NET", we added a number of such names in order to
> avoid reserving new keywords.)


The reason is the same here. We didn't want to add "from" as a separate
reserved keyword.


> In that spec we wrote such keywords as xxxx$xxxx, where $ was actually
> some odd Unicode character that in Word rendered as ░. Then we said, "The
> symbol ░ is used in the grammar to signify that white-space appears within
> the keyword. Any white space that appears in the program text after
> translation phase 1 is permitted in the position signified by the ░
> symbol.  It is unspecified whether white space generated by comments,
> documentation comments, and macro invocations is permitted in the position
> signified by the ░ symbol. Following translation phase 4, a keyword with ░
> will be a single token." It seems to me that we need to say just what kind
> of whitespace can separate the two parts, and whether /*...*/ is allowed as
> a separator.
>

In PHP the allowed whitespace matches the white-space-character production
and comments are not allowed.

Nikita