|
|
Posted by Andy Hassall on 10/27/29 11:51
On Wed, 28 Jun 2006 18:00:49 +0200, "B.r.K.o.N.j.A" <thebrkonja@inet.hr> wrote:
>example code
>
>function myfunc1()
>{
> return myfunc2();
>}
>
>function myfunc2()
>{
> return myfunc3();
>}
>
>function myfunc3()
>{
> return very_large_string; // let's say that very large string is 30kb
>of data
>}
>
>
>print myfunc1();
>
>The question is, would these 30Kb of data be copied into each function
>wasting resources or only reference to these data would be propagated to
>myfunc1? PHP version is 5.0.x
PHP uses copy-on-write for the contents of variables. So, there'd only be one
copy of the large string in memory in the above.
Whilst "=" assigns by copy, it's my understanding that it initially just sets
up references. But when it comes to writing to the new variable, it only then
does the copy, giving the same semantics as copy-by-value but giving savings
where the copies are actually just read-only.
$a = 'long string'; // one instance of long string
$b = $a; // still one instance - $b and $a refer to same data
$b .= 'some more'; // only at this point is the data copied, so that $b
// can be modified
More information here: http://www.zend.com/zend/art/ref-count.php
In the case of functions, there's a further hint in the manual that multiple
copies may be avoided automatically:
http://www.php.net/manual/en/language.references.return.php
"Do not use return-by-reference to increase performance, the engine is smart
enough to optimize this on its own."
--
Andy Hassall :: andy@andyh.co.uk :: http://www.andyh.co.uk
http://www.andyhsoftware.co.uk/space :: disk and FTP usage analysis tool
Navigation:
[Reply to this message]
|