Comments on: Objective-C 3.0 http://www.losingfight.com/blog/2006/10/25/objective-c-30/ smarter than your average squirrel, on most non-acorn related topics Mon, 08 Sep 2008 06:59:23 +0000 http://wordpress.org/?v=2.5 By: tyler http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-24911 tyler Sat, 20 Oct 2007 04:45:19 +0000 http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-24911 http://developer.apple.com/documentation/Cocoa/Reference/ObjCRuntimeRef/Reference/reference.html#//apple_ref/c/func/objc_addClass note comment in the sample code: // Allocate empty method lists. // We can add methods later. // http://developer.apple.com/documentation/Cocoa/Reference/ObjCRuntimeRef/Reference/reference.html#//apple_ref/c/func/objc_addClass

note comment in the sample code:

// Allocate empty method lists.
// We can add methods later.
//

]]> By: Mark Aufflick http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-14003 Mark Aufflick Wed, 25 Apr 2007 13:37:30 +0000 http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-14003 Closures would be hard in Objective-C because it's still, at its core, C. C uses a stack for holding the contexts of the chain of function calls and so has no way to 'hang on to' the context for a closure after the parent function has returned. There is an excellent message on comp.lang.objective-c that explains this better: http://groups.google.com/group/comp.lang.objective-c/msg/37b994ca69cde086 Certainly not impossible, but a whole lot of changing at the core of the language would be required. Closures would be hard in Objective-C because it’s still, at its core, C. C uses a stack for holding the contexts of the chain of function calls and so has no way to ‘hang on to’ the context for a closure after the parent function has returned.

There is an excellent message on comp.lang.objective-c that explains this better:

http://groups.google.com/group/comp.lang.objective-c/msg/37b994ca69cde086

Certainly not impossible, but a whole lot of changing at the core of the language would be required.

]]> By: Andy http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-668 Andy Thu, 26 Oct 2006 17:45:08 +0000 http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-668 Hey Scott, That's great info, I didn't know you could actually add a single method dynamically. It does futher my belief that what I've proposed is feasible. I was worried about my hand-waving when it came to calling anonymous functions. However, it's still not as cool as closures. :-) Hey Scott,

That’s great info, I didn’t know you could actually add a single method dynamically. It does futher my belief that what I’ve proposed is feasible. I was worried about my hand-waving when it came to calling anonymous functions.

However, it’s still not as cool as closures. :-)

]]> By: Andy http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-667 Andy Thu, 26 Oct 2006 17:41:25 +0000 http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-667 Hey Joachim, After writing this article I was convinced that closures are completely doable in Objective-C, albeit with the couple of gotchas that I outlined. I even started wondering if I should download the source to gcc and start tinkering. :-) It would be interesting if someone smarter than I could comment on if I totally missed the boat, and suggested something that's not implementable. Hey Joachim,

After writing this article I was convinced that closures are completely doable in Objective-C, albeit with the couple of gotchas that I outlined. I even started wondering if I should download the source to gcc and start tinkering. :-)

It would be interesting if someone smarter than I could comment on if I totally missed the boat, and suggested something that’s not implementable.

]]> By: Andy http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-666 Andy Thu, 26 Oct 2006 17:37:12 +0000 http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-666 Hey Keith, You're right about fast array enumeration being in Objective-C 2.0. My choice of examples for closures probably wasn't the best one. Fortunately there are lots of better examples of the power of closures out on the web. Hey Keith,

You’re right about fast array enumeration being in Objective-C 2.0. My choice of examples for closures probably wasn’t the best one. Fortunately there are lots of better examples of the power of closures out on the web.

]]> By: Keith Duncan http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-664 Keith Duncan Thu, 26 Oct 2006 11:57:19 +0000 http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-664 Yeah I think I missed the point of what you were saying when I read it this morning, I was mulling it over in my head durning my driving lesson (the instructor was less than pleased with my concentration) and now see what your getting at - a driving lesson evey day for a week can really tire you out! Yeah I think I missed the point of what you were saying when I read it this morning, I was mulling it over in my head durning my driving lesson (the instructor was less than pleased with my concentration) and now see what your getting at - a driving lesson evey day for a week can really tire you out!

]]> By: Joachim Bengtsson http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-663 Joachim Bengtsson Thu, 26 Oct 2006 09:12:41 +0000 http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-663 Closures in a compiled language would be beyond awesome (although I'd guess it's already been done by some language long forgotten...). And with a real syntax, yes please! Not all the mess with "Define a function in the global scope, get a function pointer from that, and inject that into the runtime"... However, perhaps all that's needed is a preprocessor... ;) Next summer project here I come! (It's funny that Objc started as a preprocessor...) Although that wouldn't preserve locals, hmm... Closures in a compiled language would be beyond awesome (although I’d guess it’s already been done by some language long forgotten…). And with a real syntax, yes please! Not all the mess with “Define a function in the global scope, get a function pointer from that, and inject that into the runtime”…

However, perhaps all that’s needed is a preprocessor… ;) Next summer project here I come! (It’s funny that Objc started as a preprocessor…) Although that wouldn’t preserve locals, hmm…

]]> By: Scott Stevenson http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-662 Scott Stevenson Thu, 26 Oct 2006 08:58:17 +0000 http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-662 The future is here now. Categories can load at runtime, and you can also <a href="http://theocacao.com/document.page/327" rel="nofollow">add methods at runtime</a>. You do have to get your hands messy, though. Great post, though. The future is here now. Categories can load at runtime, and you can also add methods at runtime. You do have to get your hands messy, though.

Great post, though.

]]> By: Keith Duncan http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-661 Keith Duncan Thu, 26 Oct 2006 08:03:43 +0000 http://www.losingfight.com/blog/2006/10/25/objective-c-30/#comment-661 Just a little note on the closures you mention, if you read the web documentation on Obj-C 2.0 it features fast enumeration - faster than an NSEnumberator. So unless im missing something this already covered this in version two by using for(id someObject in array) NSLog(@"%@", [someObject description]); Just a little note on the closures you mention, if you read the web documentation on Obj-C 2.0 it features fast enumeration - faster than an NSEnumberator. So unless im missing something this already covered this in version two by using for(id someObject in array) NSLog(@”%@”, [someObject description]);

]]>