switch to room list switch to menu My folders
Go to page: First ... 18 19 20 21 [22] 23 24 25 26 ... Last
[#] Fri Jan 23 2009 03:22:03 EST from dothebart @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

why use a garbage collection if you have to learn how to operate it as well as malloc/free?

and once you get into live/heavy load conditions the garbage collection can give you much of vodoo and fancy stuff.

theres valgrind and profilers for c/c++ to get it done cheap and easy. thats why I prefer it.

well... actualy I prefer plain old C.

[#] Fri Jan 23 2009 07:42:49 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

you're really most of the way there. So what if it isn't enforced by

the compiler?

well, if you're going to go through all the effort anyway, you can save yourself a lot of grief by having the compiler not forget to call your destructors.

[#] Fri Jan 23 2009 22:48:59 EST from IGnatius T Foobar @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

theres valgrind and profilers for c/c++ to get it done cheap and easy. thats

why I prefer it.

Now thee's an interesting thought. A think layer on top of a C compiler that comes with a built-in replacement for malloc(). The compiler adds some code here and there to keep track of pointers, and when the last pointer to an object is dereferenced, *poof* it gets freed by the garbage collector.

I wonder if it's even possible.

[#] Fri Jan 23 2009 22:54:35 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

They have that in C++, it's called smart pointers.
You remember me bitching about smart pointers, don't you?

You can't do it in C unless you explicitly call something to deactivate a pointer at which point you might as well call free() or dereferencecount()

That's where the magic of destructors come from.

I have two words that neatly explain why you don't want to do that:
Weak pointers.

[#] Fri Jan 23 2009 22:59:15 EST from IGnatius T Foobar @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Yeah, I think I'll stick with valgrind :)

[#] Fri Jan 23 2009 23:22:15 EST from Peter Pulse @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Destructors in C++ (timely ones, not ones that happen whenever..) were a good idea. The whole idea of local objects being just like any other stack variable.
It actually made programming easier... Files that close themselves, etc.
Some of the other stuff was (probably still is) dumb. I am sure ls will tell me why this would never work.. but for example instead of requiring all type safety to be assured at compile time, it could have been done at link time for things that aren't known at compile time. So then you could have things like method pointers without knowing the class, and yet it would all be checked ahead of time, thereby making everyone happy.

[#] Sat Jan 24 2009 09:17:23 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

well... depends on your goal.
I'm a big speed freak, I want to do as little after compile time as possible.
I ever give you my rant on the itanium and what a fucking genius idea it was?
It takes all the thinking of piplining and shit out of the processor and puts it in the compiler. Fucking brilliant.
We need more of that.

[#] Sat Jan 24 2009 09:19:26 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

I aggree that link or run time binding/type checking is functional and cool. I also see that what I do is a dying art and I should just shut the fuck up already since I've already sold out and prefer java t o C++ at this point (only because there's no point in writing fast software anymore, and eclipse is just so much fun)
But there's a little part of me that cringes still when I think about calling a function and having the os load something from disk and bind it because it hasn't been loaded yet. Sigh.

[#] Sat Jan 24 2009 12:11:25 EST from Peter Pulse @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Not sure if that's in response to me.. but I'm not saying move it to runtime.
I'm just saying move some stuff from cc to ld.

[#] Sat Jan 24 2009 14:17:48 EST from fleeb @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

C++ doesn't perform type-checking during linking because the design of the language doesn't really account for how the binary information is formatted (from what I recall reading on this topic).

I think they didn't want to impose binary standards on the language because the designers of the language wanted it to be useful on as many different platforms as possible, and some platforms do not have 8-bit bytes (for example).  In fact, there's a lot of weird things about C++ because it doesn't want to be limited to a particular platform.

[#] Sat Jan 24 2009 14:58:51 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Not sure if that's in response to me.. but I'm not saying move it to
runtime. I'm just saying move some stuff from cc to ld.

Oh, I thought you mean run time linking.

[#] Sat Jan 24 2009 14:59:14 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Not sure if that's in response to me.. but I'm not saying move it to
runtime. I'm just saying move some stuff from cc to ld.

Which would be interesting because it would obviate the need for headers.

[#] Mon Jan 26 2009 12:23:39 EST from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Ok, so you miss some extra goodies like inheritance and overloading,

but you've still got objects, you still get the clean, organized code

that results from using objects, and you didn't have to learn a
byzantine language like C++ to make it happen.

If that's all you wanted to do, you can accomplish it better in C++ but without using C++'s misfeatures. After all it's a superset of C.

[#] Mon Jan 26 2009 12:28:37 EST from fleeb @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Which is kind of what I've been saying.

You don't *have* to use the full language.

[#] Mon Jan 26 2009 12:30:01 EST from fleeb @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

I just thought of a kind of funny (but potentially arrogant) comparison.

If you use some of the more arcane features of the C++ language without really understanding the language, youwind up looking like that guy on In Living Color who tried to use big English words, but clearly didn't know what they meant.

[#] Sun Feb 01 2009 14:45:28 EST from dothebart @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

ajax... without xml an json instead... yes. but... ajax without javascript?????


15. ZK - Simply Ajax and Mobile
ZK is Ajax Java framework without JavaScript. With direct RIA, 200+ Ajax
components and markup languages, developing Ajax/RIA as simple as desktop
apps and HTML/XUL pages. Support JSF/JSP/JavaEE/Hibernate/.., and Ajax
script in Java/Ruby/Groovy/Python/..

[#] Mon Feb 02 2009 09:10:41 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

I don't get it.
Or have we come completely full circle and they do without the A in ajax.

[#] Mon Feb 02 2009 10:22:47 EST from IGnatius T Foobar @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

I think they mean that *you* don't have to write any JavaScript because they've done it all for you; you just write a bunch of declarations and they generate the code (or more likely, link up a bunch of pre-written code).

Declarative "languages" are nothing unique; the terrorists of Redmond have been trying to force XAML on us for a while now.

[#] Mon Feb 02 2009 22:29:29 EST from Animal @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]


Go to page: First ... 18 19 20 21 [22] 23 24 25 26 ... Last