Hi,

This proposal is for the often called line like this:
$var = isset($_GET['var']) ? $_GET['var'] : '';

Only a shorter and imho a cleaner solution to get the same:
$var = varset($_GET['var']);

It should be called var_set() - better on name space pollution.

oke

However there is a slight issue with this approach. If notices are
turned on
this code will generate a notice while i think it should not do that. But
perhaps this approach is "to short".
A slightly different implementation (and longer) prevents the notice:

If is is a language element (like isset()) then you can avoid this problem.

Could you explain that a bit more?

I do find a lot of code, in simple scripts, that does just that.

It might be nice to extend it such that if the 1st argument is a list then
the
first in the list which is set is returned, eg:

   $var = var_set(($_GET['var'], $_POST['var']), 'default');

I might be missing the point here, but the way i understand it that can give
unexpected results.. since it returns the first element from an array in
your suggestion but that's not what you want to do.

--
Alain Williams
Linux/GNU Consultant - Mail systems, Web sites, Networking, Programmer, IT
Lecturer.
+44 (0) 787 668 0256 http://www.phcomp.co.uk/
Parliament Hill Computers Ltd. Registration Information:
http://www.phcomp.co.uk/contact.php
#include <std_disclaimer.h

It might be nice to extend it such that if the 1st argument is a list then the
first in the list which is set is returned, eg:

$var = var_set(($_GET['var'], $_POST['var']), 'default');

If that is the usage, I would suggest coalesce() to be consistent with
the same concept in SQL. And you would not need a list as the first
argument. Just pass multiple arguments and return the first set value.

$var = var_set($_GET['var'], $_POST['var'], 'default');

http://en.wikipedia.org/wiki/Null_%28SQL%29#COALESCE

Brian.

On Wed, Apr 20, 2011 at 6:00 PM, Jonathan Bond-Caron jbondc@openmv.comwrote:

function varset($arr, $key, $default = '') { return (isset($arr[$key])
? $arr[$key] : $default); }

where the call would be:
$var = varset($_GET, 'var');

I personally like both ways...
My proposal is to make this function a core php function in PHP 5.4.
The added benifit is obvious. People can, with this, use a way shorter
notation to validate if a given array element exists. Right now that
needs to be done with a ternary, filter_var or some other method
(there are quite a few ways to check for existence).

I tried to look in the PHP source to see if i could make a patch to
add this but i couldn't find the function that defines the isset
function (i wanted to base it on that). So a pointer to the right
location would be nice (along with docs that tell me what i all need
to do to implement a new php function).

https://svn.php.net/viewvc/php/php-src/trunk/Zend/zend_language_parser.y?rev
ision=306938&view=markup

Look for "isset_variables:", then zend_do_isset_or_isempty

isset() lives in the parser and requires some advanced knowledge of the
opcodes (personally I'm not there yet)

Oh boy, i never ever did anything in the core PHP coding so i'm certainly
not likely to be able to understand all of that. (yet)

So, what do you think of this?

I like the idea, it could also be called vardef() or var_default()

@ the rest.
The list idea is nice, but i don't really see the added value for it.. Lets
keep it simple, oke :)
As for that other thread: "[PHP-DEV] Implicit isset/isempty check on
short-ternary operator"
I don't really know much of it, but does that mean that my suggestion is
rejected even before i made an RFC for it?

Regards,
Mark

Hi,
This proposal is for the often called line like this:
$var = isset($_GET['var']) ? $_GET['var'] : '';
Only a shorter and imho a cleaner solution to get the same:
$var = varset($_GET['var']);

The implementation for this in PHP code is this:

Arg 1 = the variable to check for existence

Arg 2 = the default return value which is an empty string by default

function varset($var, $default = '')
{
return (isset($var) ? $var : $default);
}

I proposed something similar over 5 years ago.  It ain't gonna happen because PHP language can't support it.  The Zend engine needs to be rewritten to remove the warnings and that's not something they are volunteering to do no matter how much people want it.

 http://markmail.org/message/yl26ebzcix35wtke

My proposal was called "filled" since it was the opposite of "empty" which already existed.  I extended the function to return the first non-empty value or null if all values evaluated as empty.

You could use the function like this:

 $x = filled($_GET['x'], $obj->something, $default, 'default');

It would return the first argument where !empty($arg) evaluates as TRUE.

There would need to be a companion function to check 'isset' opposite as you have proposed.

Like I said, though, I don't think this can be done in the language because I think they ran out of OPCODES and would have to tear apart the whole PHP engine to support such a feature.

-- Dante

--

Hi,

well, i did propose 2 possible ways although the second one is a
little longer but still clean and readable.
And the second one is certainly possible to implement!

Btw. i did look at the suggested core php code parts and even though i
can do nearly everything in php.. i have a hard time understanding how
the inner php parsing things actually work. There is no way i'm able
to make a patch for php.

Regards,
Mark

My proposal was called "filled" since it was the opposite of "empty"
which already existed. I extended the function to return the first
non-empty value or null if all values evaluated as empty.

You could use the function like this:

$x = filled($_GET['x'], $obj->something, $default, 'default');

Like I said, though, I don't think this can be done in the language
because I think they ran out of OPCODES and would have to tear apart
the whole PHP engine to support such a feature.

That's not the reason, there's 127 / 256 opcodes.

So far "filled", "varset", "coallesce", "ifsetor" or any operator would likely have to manipulate opcodes.

http://php.net/~helly/ze2-ifsetor-20040901.diff.txt

From what I understand, zeev and andi are strongly opposed to adding a new opcode. I'm sure they have their reasons.

Maybe there's another approach where ~ E_NONE could be passed to zend_error

ZEND_API void zend_error(int type, const char format, ...) / {{{ */
{
if(type & E_NONE)
return;
}

Then modify to zend_vm_gen.php to pass E_NONE if within 'ifsetor'.

Either way, there's a solution for it, I think the debate is over how it's implemented.

-- Worth reading

https://wiki.php.net/rfc/ifsetor#rejected_features

-- Why?

I think it simply boils down to this:

• PHP developers want a function for it.

Derick

I proposed something similar over 5 years ago. It ain't gonna happen
because PHP language can't support it.

It supports it. Several functions allow you to pass in variables that
are not set and don't throw an error. Not sure what you are talking about.

Brian.