Re: Why PHP sucks - farce or is there a bit of truth?

On 10/17/2013 8:50 PM, Tim-Hinnerk Heuer wrote:

http://php.net/manual/en/language.operators.comparison.php#example-122 Found this in relation, so this behaviour is documented. Tim-Hinnerk Heuer Twitter: @geekdenz Blog: http://www.thheuer.com On 18 October 2013 13:46, Tim-Hinnerk Heuer <th.heuer@gmail.com> wrote:

Hi, I've been a PHP programmer for several years now and have a bit of a love-hate relationship with it. It's great for doing something quickly, especially web stuff, but recently I have heard people moaning about PHP a lot and did some research and found this: http://webonastick.com/php.html One thing I had to get my head around is this: The ternary operator <?php

     $foo = 1;
     print(($foo == 1) ? "uno" : ($foo === 2) ? "dos" : "tres");
     print("\n");

outputs

dos

because the operator is left-to-right associative instead of right-to-left as in other languages. I was thinking there must be a reason for this. Speed? Is it faster to evaluate/implement all operators as left-to-right? I noticed that the above could easily be fixed by saying: <?php

     $foo = 1;
     print(($foo == 1) ? "uno" : (($foo === 2) ? "dos" : "tres"));
     print("\n");

outputs

uno

Was this a deliberate design decision or is it a flaky implementation of the ternary operator? Tim-Hinnerk Heuer Twitter: @geekdenz Blog: http://www.thheuer.com

Having endured this entire thread (from T-to-B) I have to add my $.02.

I recently experienced the rather unexpected outcome of a ternary within a ternary. Rather than consider the order of evaluation being L-to-R or R-to-L, I merely diagnosed it as a simple bug in PHP - which BTW I love, despite it's less than consistent syntax. Reading all the talk of order of evaluation, I have to say in my 40+ years of programming in a variety of languages, I can't recall ever having to consider the topic. When I wrote code I always expected and relied upon the order being the way I wrote it. And - I usually added parentheses to help clarify in my mind what I was doing with the statement as I wrote it, although sometimes I knew they were necessary to achieve the analysis I had in mind. So - as expected, the ternary operator does work from L-to-R except when joined with a second (or third?) ternary operator. Why that is I can't figure out other than to assume it's an un-resolved bug, that occurs so seldom it's probably never been issued. Oh, well.....