The patch is fine.
I applied it.
Dmitry.
> -Original Message-
> From: Ilia Alshanetsky [mailto:[EMAIL PROTECTED] On Behalf Of
> Ilia Alshanetsky
> Sent: Tuesday, November 07, 2006 11:35 PM
> To: Matt Wilmas
> Cc: internals@lists.php.net; Dmitry Stogov
> Subject
Looks like a good optimization to me.
On 5-Nov-06, at 7:30 AM, Matt Wilmas wrote:
Hi Nuno,
Late reply, but I'm glad the idea was able to be used. :-)
To the other hash people: While checking how Ilia made implode()
faster in
5.2, I think I found another optimization that can be done. In
Hi Nuno,
Late reply, but I'm glad the idea was able to be used. :-)
To the other hash people: While checking how Ilia made implode() faster in
5.2, I think I found another optimization that can be done. In
zend_hash_copy/merge, it seems the zend_hash_quick_* functions can be used
instead for as
Aaahh nice catch ;)
From a quick lookup in PHP sources I also found these functions were the
same optimization can be applied:
zend_default_exception_new_ex() - trivial
convert_to_array() - not as easy as others
zend_object_std_init() - possibly
clone_wrapper_hash() - trivial (php-src/main/stre
Hi all,
I noticed awhile ago how most every use of zend[_u]_hash_init has nSize as
0. Of course it isn't always known how many elements will be added to the
array (nTableSize), but there are places where an accurate value could be
used instead of getting the minimum default (8). For anyone who d
