Hi Dmitry Stogov, you wrote:
> Hi Michael,
>
> It is not a bug. You understood proper. Internal classes are stored in
> regular heap and they don't use e[memory] function family.
> This is done, because memory allocated by emalloc is freed after each
> request, but internal classes should still ali
Hi, I wrote:
> Isn't the behaviour of ReflectionClass then bogus, because
> it doesn't care about ZEND_INTERNAL_CLASSes?
And...
shouldn't there be a cleanup for static members of
ZEND_INTERNAL_CLASSes to reset the contents to the
declared values?
Also zval_internal_dtor() implies that one canno
Sebastian Mendel wrote:
how about if i need a name for object not loaded?
Class::getName( $object_id );
Weird, i thought you must know the classname for a static
call :)
--
PHP Internals - PHP Runtime Development Mailing List
To unsubscribe, visit: http://www.php.net/unsub.php
patrick müller wrote:
> Sebastian Mendel wrote:
>
>> how about if i need a name for object not loaded?
>>
>> Class::getName( $object_id );
>>
>
> Weird, i thought you must know the classname for a static
> call :)
not the classname, the objectname f.e. the name for the object of type
customer wit
Hello Michael,
sure but there's a difference Reflection whatever is only run during
request time, what you want is running outside and require malloc rather
than emalloc.
Thursday, July 21, 2005, 4:30:59 PM, you wrote:
> I wrote:
>> I have some issues with static members of internal classes.
If you've spotted a behaviour change, open a new bug with a short
self-contained reproducing script. I suspect some kind of engine bug
here, because what you've described doesn't make sense :)
That warning about the length comes from this code in the fread function:
len = Z_LVAL_PP(arg2);
has anyone figure it out if it was rejected or just forgotten ?
and if it was rejected, what was the reason ?
the only valid reason I seen so far was that you can do this using
something else (perl, or an apache module, or even something more
exotic). but, this also apply to php itself ... I can w
Hi Marcus Boerger, you wrote:
> Hello Michael,
>
> sure but there's a difference Reflection whatever is only run during
> request time, what you want is running outside and require malloc rather
> than emalloc.
Thanks, I already adjusted my code according to that.
But guess, I've still not solv
working on apc
--
PHP Internals - PHP Runtime Development Mailing List
To unsubscribe, visit: http://www.php.net/unsub.php
Hi,
I'm trying to compile the latest PHP without compiled-in extensions so I could
loaded them later from php.ini. Unfortunately I've found that standard
depends on spl (spl_ce_Countable) and date. At least the second is very small
library, but the first depends on libxml and other extensions,
Just noticed that something changed in HEAD since last week.
We have some code that does this:
class foo {
function foo() {
$this['data'] = array();
}
}
last week that ran fine, this week it gives me an error:
'Cannot use object of type foo as an array'.
I agree that the code above look
Wez Furlong wrote:
If you've spotted a behaviour change, open a new bug with a short
self-contained reproducing script. I suspect some kind of engine bug
here, because what you've described doesn't make sense :)
That warning about the length comes from this code in the fread function:
len
Hi Piotr,
On Friday 22 July 2005 15:15, Piotr Roszatycki wrote:
> I'm trying to compile the latest PHP without compiled-in extensions so I
> could loaded them later from php.ini. Unfortunately I've found that
> standard depends on spl (spl_ce_Countable) and date. At least the second is
> very smal
Hi Wez,
I didn't look on any related code but I guess the reason might be related to
the fix of #33710 (ArrayAccess objects doen't initialize $this)
johannes
P.S. sorry if this comes more than once - I failed to use my mailer :-)
On Friday 22 July 2005 17:44, Wez Furlong wrote:
> Just noticed
On Friday 22 of July 2005 19:56, Johannes Schlueter wrote:
> A build with spl statically compiled in is for me just a few k bigger than
> a build without spl. But building spl shared wouldn't work well for two
> reasons: It deals with quite engine-near stuff (like overriding the []
> syntax or __au
15 matches
Mail list logo