Re: [ZEND-ENGINE-CVS] cvs: ZendEngine2 / zend_compile.c zend_compile.h zend_globals.h zend_language_parser.y /tests break_label01.phpt
break_label02.phpt break_label03.phpt break_label04.phpt break_label05.phpt break_label06.phpt break_l

From: Date: Fri, 03 Mar 2006 21:23:11 +0000
Subject: Re: [ZEND-ENGINE-CVS] cvs: ZendEngine2 / zend_compile.c zend_compile.h zend_globals.h zend_language_parser.y /tests break_label01.phpt
break_label02.phpt break_label03.phpt break_label04.phpt break_label05.phpt break_label06.phpt break_l
References: 1 2 3 4 5  Groups: php.zend-engine.cvs 
Request: Send a blank email to [email protected] to get a copy of this message 
At 01:16 PM 3/3/2006, Marcus Boerger wrote:

Hello Andi,

Friday, March 3, 2006, 10:09:07 PM, you wrote:

At 10:29 AM 3/3/2006, Rasmus Lerdorf wrote:

Marcus Boerger wrote:

Hello Dmitry,
 the code is the opposite of our beloved KISS approach. When i read:
  Label: while(1) { break Label; }
i clearly expect an endless loop, guess why?


I'd have to agree with Marcus on this one.  We are making one of the
first forward compatibility breaks in the procedural language for this?


Sorry, I don't get it. Why do we expect this to be an endless loop?
break Label just says what loop to break out of. Or am I
misunderstanding the concern?

A label is commonly used for a 'goto' point. Now, think again without the
stuff you had in mind while brewing the patch but instead what you are
used to from any other language that offers labels.


Uhm, OK I thought again. You are getting confused with goto. This isn't goto but saying what loop you want to break/continue. You are marking the loop with a label and naming it, instead of counting. I don't see what other kind of semantics you'd expect. If you'd change it to jump then it's goto.

If you find it too confusing, then maybe this is a good reason to roll back and not support labeled breaks/continues. I wasn't too enthusiastic to begin with but did it out of peer pressure :)

Andi

Thread (28 messages)

« previous php.zend-engine.cvs (#4649) next »