LLVM 3.1 released

http://llvm.org/releases/3.1/docs/ReleaseNotes.html

by octopus • 14 years ago

73 13 14 years ago

llvm.org

bobbypage• 14 years ago

Sweet, especially the Objective-C Literals: http://clang.llvm.org/docs/ObjectiveCLiterals.html

lukeman• 14 years ago

Yeah, I've been waiting for this release. Just in time for WWDC, too.

DeepDuh• 14 years ago

Sweet indeed. Certainly makes Obj C a little bit more concise.

misnome• 14 years ago

Yes, I am ridiculously excited about this, especially the ability to create single-item NSArray's without a tonne of messy typing - this is probably the largest cleaning effect it will have.

Any idea how long it has historically taken apple to move the newer Clang releases?

Someone2• 14 years ago

I don't know, but I guess that the compiler itself has been seriously tested on Mac OS X, and WWDC is June 11-15. Esepcially if they want/dare to use this for compiling Mountain Lion, I expect there to be some kind of release around that time.

Rickasaurus• 14 years ago

Oh cool! I didn't know Julia was LLVM backed! Now if only it had a good IDE package.

jvoorhis• 14 years ago

Looking forward to playing with autovectorization in ruby-llvm. https://github.com/jvoorhis/ruby-llvm

program• 14 years ago

Collections subscripting operator [] and boxed C strings are life-savers. Instantiating obj-c object is always a pain cause the syntax is very verbose, expecially for NSString:

  NSString *s = [NSString stringWithCString:"hello" encoding:NSUTF8StringEncoding];
  NSString *s = @("hello");

it's 11 characters vs 67. Now what I really, really miss is the implementation of the NSString concatenation operator. stringWithFormat isn't always the answer nor appendString. Please forgive me if you find that my idea is silly:

  // String concatenation
  NSString *s = @+(@"hello ", @"world");

Or maybe some concise dictionary operators.

  NSDictionary *d1 = @{ @"first":@"Hello", @("second"):@"World" };
  NSDictionary *d2 = @{ @"third":@"Greetings" };
  
  NSDictionary *cd = @|(d1, d2); // Dictionary merge
  NSDictionary *id = @&(d1, d2); // Dictionary intersect
  NSDictionary *dd = @^(d1, d2); // Dictionary difference

Someone2• 14 years ago

I think that would be inconsistent. All existing @<whatever> fragments are evaluated at compile-time. Your @+, @|, @&, @^ feel more appropriate as proper functions (the compiler could still know of them, so it could optimize calls away)

Also: dictionary merge/intersect/difference aren't easily defined. For example, what should

  @&( @{ @"a":@"aye", @"b":@"bee" }, @{ @"a":@"ai ai", @"Bee":@"bee" })

return?

sjwright• 14 years ago

I don't like it, mainly because it spreads the at-sign too far, and the result is a cornucopia of dense, unreadable at-signs. When you want to do thinks like dictionary operations, I'd prefer to have the intent clearly spelled out in the code.

luriel• 14 years ago

A cool, seems to include the changes that were needed by llgo: https://github.com/axw/llgo

seunosewa• 14 years ago

Is there any successful language/runtime that runs primarily on LLVM?

cristiantincu• 14 years ago

What do you mean by “primarily”? Do you mean “specifically”? “Succesful” how?

http://en.wikipedia.org/wiki/Llvm

LLVM 3.1 released

Related Stories