count optimization #37
Conversation
samsonasik
commented
Jul 8, 2018
samsonasik
commented
- Is this related to quality assurance?
There was a problem hiding this comment.
Hi @samsonasik
thanks a lot for your contribution. I just added a minor change - note. But nothing critical, I could do this on merge. If you are able to change this, I'd appreciate.
| @@ -134,7 +134,7 @@ private function buildParametersCode(string $type) | |||
| $tab = str_repeat(' ', $intention); | |||
| $code = ''; | |||
|
|
|||
| if (count($withOptions)) { | |||
There was a problem hiding this comment.
This should be imported with use function for performance. You may actively do this, but don't worry, I can do this during the pull as well.
There was a problem hiding this comment.
truthy check is faster than count($var)
There was a problem hiding this comment.
Shouldn't make a difference with the current opcode optimisations in the language, but yes, a direct truth check is faster indeed.
There was a problem hiding this comment.
Misread of the diff on my side never mind.
There was a problem hiding this comment.
@Ocramius VLD opcodes for 7.1 show that count() actually does function call, initializing function context is the most expensive op here:
line #* E I O op fetch ext return operands
-------------------------------------------------------------------------------------
if ($var) :
5 6 > JMPZ !0, ->8
6 7 > ECHO 1
if (count($var)) :
9 8 > INIT_FCALL 'count'
9 SEND_VAR !0
10 DO_ICALL $4
11 > JMPZ $4, ->13
10 12 > ECHO 2
if (! empty($var)) :
13 13 > ISSET_ISEMPTY_VAR 293601280 ~5 !0
14 BOOL_NOT ~6 ~5
15 > JMPZ ~6, ->17
14 16 > ECHO 3
Not that it really matters, it is optimized in 7.2
|
@samsonasik one note though: you keep sending these against |
|
@Ocramius rebased against develop and set target to develop |
|
@Ocramius right, thanks. |
