On Fri, 16 Jul 2004, Sterling Hughes wrote:
not only that but those people who want this performance boost can use
apc. i'll give george a patch that solves this if no one else steps
up.
Do that before we revert it then, ok? :)
Derick
--
PHP Internals - PHP Runtime Development Mailing
i'll do it sometime, but no, this patch should be reverted (.) the
performance increase is neglible - its a *bad* optimization.
-sterling
On Mon, 19 Jul 2004 22:59:14 +0200 (CEST), Derick Rethans
[EMAIL PROTECTED] wrote:
On Fri, 16 Jul 2004, Sterling Hughes wrote:
not only that but those
woops, discussion should be on [EMAIL PROTECTED]
-- Forwarded message --
From: Sterling Hughes [EMAIL PROTECTED]
Date: Fri, 16 Jul 2004 00:11:53 -0700
Subject: Re: [ZEND-ENGINE-CVS] cvs: ZendEngine2 /
zend_language_parser.y zend_language_scanner.l
To: Marcus Boerger [EMAIL
Excuse my ignorance, but why does this breaks compatibility?
Sterling Hughes wrote:
woops, discussion should be on [EMAIL PROTECTED]
-- Forwarded message --
From: Sterling Hughes [EMAIL PROTECTED]
Date: Fri, 16 Jul 2004 00:11:53 -0700
Subject: Re: [ZEND-ENGINE-CVS] cvs:
?php
function null() {
}
null();
?
?php
class connected {
function true() {
}
}
if (connected::true()) {
echo bar;
}
?
On Fri, 16 Jul 2004 09:26:33 +0200, dharana [EMAIL PROTECTED] wrote:
Excuse my ignorance, but why does this breaks compatibility?
Sterling Hughes wrote:
Well,
the new word clone made the same :
class a{
function clone() {}
}
is not working anymore :)
andrey
Sterling Hughes wrote:
?php
function null() {
}
null();
?
?php
class connected {
function true() {
}
}
if (connected::true()) {
echo bar;
}
?
On Fri, 16 Jul 2004 09:26:33 +0200, dharana
I don't think it's critical to include this patch, but I do think it'd be a
good thing.
Do you really think it'll break BC for many applications? How many people
have functions that use null(), false(), true()?
Andi
At 08:17 AM 7/16/2004 -0700, Sterling Hughes wrote:
oh, i didn't notice it at
I would have been +0 for 5, I'm -1 for 5.1. FETCH_CONSTANT is not a
terribly expensive operation (as opposed to registering constants with
define()), and something like this is *easy* to optimize in a compiler
cache/optimizer. End of the day, we shouldn't break BC in any way
during a point
Nah, I dread the INI word. It makes applications less portable.
I prefer rolling back than doing this.
Andi
At 10:09 AM 7/16/2004 -0700, Sara Golemon wrote:
Andi Gutmans [EMAIL PROTECTED] wrote in message
news:[EMAIL PROTECTED]
I don't think it's critical to include this patch, but I do think
not only that but those people who want this performance boost can use
apc. i'll give george a patch that solves this if no one else steps
up.
-sterling
On Fri, 16 Jul 2004 10:20:08 -0700, Andi Gutmans [EMAIL PROTECTED] wrote:
Nah, I dread the INI word. It makes applications less portable.
I
10 matches
Mail list logo