Word of warning for anyone about to use new versions of PHP

[Sgarissta]

Dude, I just said it, PHP is not targeting the same target as Python/Java/.NET. PHP is a shared-nothing procedural easy-to-deploy language that is dependent on Apache.

Cool, glad that's what YOU are using it for. Zend on the other hand is attempting to position PHP in the enterprise, hence their moves at getting support from IBM and Oracle, two BIG names in the "enterprise".

Wherever Apache is, so is PHP. PHP, for its small to medium who-gives-a-damn-if-not-scalable web scripts doesn't need threads. But Ruby is targeting the world of application servers and heavyweight web scripts. And it shows imaturity when a language feature (threads) is not supported in the VM.

First off, Ruby is just a language, not a web development movement. I believe you mean Rails. To top it off, Ruby is a scripting language, that set off to be a better, cleaner perl, not exactly what I'd call targeted towards the enterprise. Also, it doesn't have a VM (I assume you mean Virtual Machine) it has an interpreter, just like PHP. Which amazingly gives it the same "shared nothing" mentality that you mention for PHP. There is a move to get Ruby 2.0 based on an actual VM, but that's more for performance and platform reasons, that to get it to be head-to-head with Java et al.

Now if you can provide me some evidence of where Ruby (or even Rails, which is what I believe you to mean) is targetting the enterprise, as opposed to the same basic niche that PHP fills, by all means do.

God, please don't let MiiJaySung to talk about Ruby anymore, I feel like I can't take it

Please refrain yourself from talking about Ruby until you at least get a grasp at whether you're arguing about Rails (a web development framework) or Ruby (a scripting language).

[bonefry]

Cool, glad that's what YOU are using it for. Zend on the other hand is attempting to position PHP in the enterprise, hence their moves at getting support from IBM and Oracle, two BIG names in the "enterprise".

As pointed out allready, maybe that's the problem.

Also, it doesn't have a VM (I assume you mean Virtual Machine) it has an interpreter, just like PHP.

From Wikipedia: A virtual machine "is a piece of computer software that isolates the application being used by the user from the computer.....The application is run on the computer using an interpreter or Just In Time compilation.".

For a list of Ruby VMs as in Virtual Machines check out: http://rubygarden.org/ruby/ruby?acti...MachineOptions

Oh, and btw, to clear things out, I was talking about Ruby "the platform" which allready identifies itself with RoR as in Ruby on Rails on every forum discussions I have met so far.

You, my friend, are the one who does not know what you are talking about.

Please refrain yourself from talking about Ruby until you at least get a grasp at whether you're arguing about Rails (a web development framework) or Ruby (a scripting language).

OK, I won't. I admit I do not know what I am talking about. BTW: What is PHP ?

[MiiJaySung]

Oh seeing as everyone else has turned this into a "Why PHP isn't Ruby thread" (and as they say if you can't beat them, join them), I'll add my say again.

@bonefry

as for PHP vs Java. Well PHP won appeal origianlly because PHP <> Java. However, PHP seems to be making the mistake of following after Java (it's OO model gets closer and closer). I'm not saying PHP is the same as Java linguisticly, coz it's still got some distance to go.

There still might be more Java jobs, but at the same time, Java has started to loose it's market share to .NET. This is a trend foreseeable for the future as well.

I know many still see Ruby as a toy, and yes it might be, but its a damn good toy to play with. After all, the creator said was was created to be fun. However, there is support from big names, such as Fowler, and if this trend continues, Ruby will edge more into the main stream. Personally, I would be quite happy if it didn't end up in the mainstream, because at the moment the Ruby community is small, and focused and therefore there is decent quality projects coming out of Ruby, such as Damage Control, Rails etc. One can argue that Ruby isn't enterprise ready, but there are large organisations adopting it. You could say the exact same about Python, not many people are using it in the enterprise, but it is not an arguement *not* to use it. I doubt it was the way Google looked at things when they planned to adopt using Python.

No one build stuff with PHP GTK. I know few people use Ruby for GUI apps either. However ruby uses mature libraries for it and therefore there's nothing to agues against using it for business apps. If I was gonna do serious GUI coding, I would bother to learn C# .NET Forms, Obj C + Cocoa, or C++ (well I sort of know this) and Qt, depending on the platform. However Ruby is a well rounded language for doing most things. It's not the ideal language for GUI stuff, and the same might apply to web programming if you were to exclude Rails. However it's well rounded as a language (much like C++ is, but at a different level)

As for your arguement of target/purpose. Unlike other languahges which have commercial support and specific goals / roles, I don't think Ruby has any. It never claims to be for "enterprise" programming. It's neither geared for shell scripting, or GUI apps. Ruby was just designed to be a fun language to use, which could be productive at the same time as a happy side effect (if you enjoy something, you are more productive at it right?). This is pretty much what Matz, the creator says. Likewise PHP was never targeted for the enterprise market originally. Ruby doesn't have target markets because it's not commerially backed like other languages.

Anyway, that's my side of the arguement there.

[Sgarissta]

As pointed out allready, maybe that's the problem.


From Wikipedia: A virtual machine "is a piece of computer software that isolates the application being used by the user from the computer.....The application is run on the computer using an interpreter or Just In Time compilation.".

For a list of Ruby VMs as in Virtual Machines check out: http://rubygarden.org/ruby/ruby?acti...MachineOptions

Nice of you to make the distinction for Ruby but not PHP, but your choice of language is fine. Given this, both Ruby, and PHP run on VMs.

Oh, and btw, to clear things out, I was talking about Ruby "the platform" which allready identifies itself with RoR as in Ruby on Rails on every forum discussions I have met so far.

You, my friend, are the one who does not know what you are talking about.

Pardon me. I apologize. As a person who frequents both the Rails, and Ruby mailing lists, and is actively developing in Rails, I DO make the distinction between the language and the framework. You have also obviously not actually discussed Ruby or Rails with anyone having done substantial work in either, as they are often quick to make this distinction.

I do this just as you would make the distinction between your choice of framework be it Phrame, Mojavi, WACT or PHP-GTK and PHP the language.

I'm done here. It is obvious to me that you do not wish to have an actual discussion, but prefer arguing semantics. If you wish to have an actual discussion, other than arguing semantics, feel free to contact me via PM.

[bonefry]

Off Topic:

I finally learned how to write offtopics

There still might be more Java jobs, but at the same time, Java has started to loose it's market share to .NET. This is a trend foreseeable for the future as well.

Java is not loosing ground to .NET. .NET is just picking the broken pieces after older M$ technologies. .NET did had an inpact in 2002-2003 but it's over because all the marketing bullsh*t is now gone. Do not be afraid to learn Java just because someone who does not know what he is talking about tells you that Java is loosing ground. Just think about this: the innovation is never in M$'s camp. And here are some things to think about (although the arguments aren't 100% valid): http://www.manageability.org/managea...tterThanDotNet

OK, lets get back to PHP.

However, PHP seems to be making the mistake of following after Java (it's OO model gets closer and closer).

This is a misstatement, and I see many have the wrong impression. Just because the language got some new keywords that exist in Java like (interface, implements, instanceof) it does not mean it's object model is derived from Java. What about the magic trio (__set, __get and __call). Those certainlly get far away from the Java object model. In fact, now you can really say PHP is dynamic (if we talk about it in a Smaltalk way).

[bonefry]

I'm done here. It is obvious to me that you do not wish to have an actual discussion, but prefer arguing semantics

OK, I am sorry (allthough you attacked me first ). But my hole point against your choice of language and platform (I am not mentioning it again in here since it attracts other comments ) is that it is different from PHP and it cannot replace PHP in the near future. From your experience with it, why not give some guidelines on how to improve PHP.

[Eric.Coleman]

Hasn't this topic gotten a little off trac?

Edit:

trac... hrm, I work too much

[bonefry]

Hasn't this topic gotten a little off trac?

Edit:

trac... hrm, I work too much

Maybe we should start a new thread like "PHP is Evil and Zend coders are cowboys" then we can really party without getting offtopic

[MiiJaySung]

Installing PHP 4.4 isn't as easy as thought

The code is failing in PHP 4.4, it's returning:
Notice: Only variable references should be returned by reference in D:\webroot_apache\flaw0001.php on line 3

<?php
error_reporting(E_ALL);
function &a() { return array(); }
function b(&$data) { /* ... */ }
b(a()); 
?>

That ain't too bad in PHP4

the trouble is in PHP 5.1.0b2, it's a fatal error. That is a little extreme. This is what annoys me about Zend is these stupid descisions. Fair enough I can understand a notice / warning, as it does at least allow code to continue working and therefore give you a migration path.

It's not the first time the Zend time have done this sort of glorified **** up in order to "fix" problems. Look at Magic Quotes. That was Zend's dumb "solution" to SQL injections. Anyone with half a brain cell will realiase this "solution" not only creates more problems than you started with, but it doesn't even solve the original problem anyway, because the quote encoding is not ANSI compatible anyway.

It's not just the stupid "solutions" Zend come up with, it's the stdpid amount of time and lack of attention they pay to engine level bugs, such as this famous PHP4 bug.. http://bugs.php.net/bug.php?id=8130 and this well known one that Marcus reported... http://bugs.php.net/bug.php?id=31449. It took years before 8130 was properly resolved. They still haven't fixed Marcus' bug yet and according to http://phplens.com/phpeverywhere/?q=node/view/205 they've know about the reference mem leak issue since last year.

[MiiJaySung]

Maybe we should start a new thread like "PHP is Evil and Zend coders are cowboys" then we can really party without getting offtopic

Nah, seeing as everyone is gouing on about Ruby in this forum, lets rename the forum (again) to "Why PHP isn't Ruby". That way other threads would not be seen as off topic

[bonefry]

Nah, seeing as everyone is gouing on about Ruby in this forum, lets rename the forum (again) to "Why PHP isn't Ruby". That way other threads would not be seen as off topic

Hmm, seeing your other posts, actually you were the one that brought it up and advocated it more
Maybe it should be:
"PHP rides the wild horse and Ruby is no cowboy"

[jayboots]

It is technically correct not to return a reference to anything but a variable in the Zend engine (even returning one from a static is questionable). Changing this behaviour is the right thing to do -- at least in the 5.x branch.

On the other hand, the engine has been working in a certain way for 4 years. Did the PHP Group really have to fix this issue (assuming it is fixed properly) in the 4.x branch? If so, couldn't they have at least co-ordinated this with the full 5.1 release and then told people very clearly about it?

The worst part is that the 4.4 release announcement doesn't even mention that upgrading could cause problems and break existing code. Instead, they make it sound like it is an innocuous bug fix release. The kicker, of course, is that reading about it on internals, at least a few people have stated that it is the script writers fault that their code is broken -- after all they misused a construct and they had a whole month to test the change. Har har har! The PHP Group had 4 freaking years to deal with this. To blame users for a bug in the engine and not warn the userbase of consequences is not acceptable.

To those blaming Zend -- isn't this really a PHP Group maintenance/release cycle issue?

[Ezku]

What about the magic trio (__set, __get and __call). Those certainlly get far away from the Java object model. In fact, now you can really say PHP is dynamic (if we talk about it in a Smaltalk way).

What I'd really like is C-style setter-getter methods that in PHP could double as means to enforce type. Nowadays you'll have to hack around that.

[{RainmakeR}]

Just FYI, it appears that phpBB 2.0.16 is also afflicted by this change in the latest PHP release (for example, you receive a blank page upon submitting a post, and the error log reports a fatal error of "Only variable references should be returned by reference").

[{RainmakeR}]

The more I think about it, the more I think this variable reference change is completely stupid. I must admit, I haven't been doing much OO work in PHP4 for a while, but 4.4.0 and above will now break a hell of a lot of my PHP4 code, because I use references heavily to make sure copies of objects are not passed around. This change to the scripting engine means that code such as this:

function &getDefaultModule()
{
   foreach ($this->modules as $module)
   {
      if ($module->isDefault())
      {
         return $module;
      }
   }
} 

...will create an error because if no 'module' object is found, it tries to return NULL. Does that mean this function will have to be re-written as:

function &getDefaultModule()
{
   $module = NULL;

   foreach ($this->modules as $module)
   {
      if ($module->isDefault())
      {
         return $module;
      }
   }

   return $module;
} 

That's just dumb.

[Super Phil]

http://bugs.php.net/bug.php?id=33643 has been reclassified as an assigned bug.

[bonefry]

Anyway, the reference "bug" shouldn't have been handled like that. No other language has this behaviour. If it was a memory coruption issue, maybe they should have found a real sollution and fix it without breaking the syntax. They choose the easy way.

[DougBTX]

Its only for personal home pages ye know.

[MiiJaySung]

http://bugs.php.net/bug.php?id=33643 has been reclassified as an assigned bug.

Yea, it's taken a lot to make them realise it was a bug. As I said in another post, they first marked it as bogus, then a documentation problm because they didn't want to face up to the fact it was a bug.

They have now finally changed the bug back to what I marked it as, a scripting engine bug!

This is the attitute problem that Zend I am fed up with. I know that to some degree people are wrong because they shouldn't have abused references, but like wise, there is no mention in the documentation stating not to use references in the way they have disallowed. Also, if they have let people use references like this for almost 4 years, they can not expect people to suddenly change. They need to realise this as to be phased out. I think it was wrong in the 5.1 branch to classify this as a fatal error. It should be a notice (or warning at the worst) to allow people to phase this out.

I wouldn't be so cheesed off if Zend marked this in the documentation (even though it does sort of break backward compatiblity still). It's the problem that loads of people suddenly are realising that LOADS of code is breaking.

This is certainly not an isolated case either. I reported this to Zend a while back. http://bugs.php.net/bug.php?id=32745 Again, they closed this bug as bogus with no changes. I've reopened it as really it's a documentation problem, as I haven't seen it in the docs.

[MiiJaySung]

Its only for personal home pages ye know.

ROFL. I wonder how many people are prepared to shell out loads of money for Zend encoder and the Zend IDE for their "personal home pages"

[pachanga]

MiiJaySung thanks for your bug reports to Zend! It's funny(actually very sad) to read how less experienced PHP developers close your bugs(or mark them bogus) while more experienced ones do re-open them...

[Dr Livingston]

Before I go again, I think you are all on some kind of 'Ruby Junket' which may be fascinating but where is your loyalty huh?

Meaning of Junket for those not in the know, http://dictionary.reference.com/search?q=junket

[R. U. Serious]

but where is your loyalty [to PHP] huh?

I know you said this half-jokingly, but it's a good oportunity to sneak in a reference to this timeless article:
"Why I Hate Advocacy"
http://www.perl.com/pub/a/2000/12/advocacy.html

Yes, it's about perl, but it's not really about perl. It's about people and what the author calls "tribalism" and how it's really bad, mmkay.

[DougBTX]

Thanks for the link Serious

[Luke Redpath]

Nah, seeing as everyone is gouing on about Ruby in this forum, lets rename the forum (again) to "Why PHP isn't Ruby". That way other threads would not be seen as off topic

Why not just rename it to "Application Design" full stop? I never worked out why PHP had a dedicated advanced forum on here, and no other language did - its not like application design is exclusive to PHP, and most discussions on here often branch off into discussions on other languages and platforms, such as Java and .NET, and recently, Ruby/Rails.