Quantcast
[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

If you can see this message, the SSL certificate expiration has been fixed.
Become a Patron!

/g/ - Technology


View post   

[ Toggle deleted replies ]
File: 139 KB, 706x674, gccegg-deprecated.png [View same] [iqdb] [saucenao] [google] [report]
27021261 No.27021261 [Reply] [Original] [archived.moe] [rbt]

It has come to my attention that many /g/entlemen are unaware or refuse to believe that GCC is defective by design and crippled for ideological reasons.
Straight from the horse's mouth:

http://gcc.gnu.org/wiki/GCC_Plugins

This intentional crippling is one of the reason GCC is obsolete and deprecated.

>> No.27021382

isnt it the only thing that supports c++11 though?

>> No.27021487

>>27021382
From this page clang and gcc seem to be close in their support of c++0x

http://wiki.apache.org/stdcxx/C++0xCompilerSupport

>> No.27021520

>Not using LLVM

>> No.27021525

>>27021261
1. It just werks [tm]
2. Propose better solutions

>> No.27021582

>>27021525
The Glasgow Haskell Compiler (GHC) has a working LLVM backend that achieves a 30% speed-up of the compiled code when compared to native code compiling via GCC.

>> No.27021613

>>27021582
For C please.

>> No.27021629

>caring about a compiler

>> No.27021631

>>27021261
Oh, I didn't realize BSD devs were so prominent on this board.

>> No.27021661

>willingly using GPL software when better free software is available
LLVM Master race.

>> No.27021666

>>27021613

Clang is an "LLVM native" C/C++/Objective-C compiler, which aims to deliver amazingly fast compiles (e.g. about 3x faster than GCC when compiling Objective-C code in a debug configuration), extremely useful error and warning messages and to provide a platform for building great source level tools. The Clang Static Analyzer is a tool that automatically finds bugs in your code, and is a great example of the sort of tool that can be built using the Clang frontend as a library to parse C/C++ code.

dragonegg integrates the LLVM optimizers and code generator with the GCC 4.5 parsers. This allows LLVM to compile Ada, Fortran, and other languages supported by the GCC compiler frontends, and access to C features not supported by Clang (such as OpenMP).

The LLVM Core libraries provide a modern source- and target-independent optimizer, along with code generation support for many popular CPUs (as well as some less common ones!) These libraries are built around a well specified code representation known as the LLVM intermediate representation ("LLVM IR"). The LLVM Core libraries are well documented, and it is particularly easy to invent your own language (or port an existing compiler) to use LLVM as an optimizer and code generator.

>> No.27021695

> clang
> liking apple shit
> liking taking it up the ass

>> No.27021713

>>27021661
>implying GPL isn't free

>> No.27021719

>>27021695
Not everything Apple is involved in is shit you massive freetard.

>> No.27021724

gcc is the only compler that has a --vroomvroom option

>> No.27021734

>>27021666
Problems:
1. Clang generates slower code than GCC.
2. A lot of software uses GCC "extensions". Linux is AFAIK one of them. Does clang compile the latest kernel?

>> No.27021769

People still use GCC?
B- but, Clang exists...

>> No.27021771

>>27021734
clang supports GNU extensions as of GCC version 4.2

GCC is now 4.7, so no it does not support everything.

>> No.27021788

>>27021719
You don't have to defend your life choices from me, I am not homophobic.

>> No.27021827

>>27021734
change takes time...

>> No.27021854

>>27021827
> clang takes time
Not as much time as GCC

>> No.27022022

I read that article and thought fine, GCC has some downfalls.

Then I read people suggesting Clang and laughed my ass off.

>> No.27022116

RMS and the FSF are right up there with MS and other proprietary software vendors when it comes to lock-in.
The non-standard GNU extensions in gcc and every unix utility makes it a pain in the ass to port code to systems not based on GNU.
I, for one, can't wait until they either support the standards or become irrelevant because of the shenanigans.

>> No.27022218
File: 55 KB, 492x600, autismalasfuck.jpg [View same] [iqdb] [saucenao] [google] [report]
27022218

>Gentoo Portage uses GCC LLVM ICC GHC Path64 etc.

>> No.27022244

>mfw in the short time LLVM has been around it's seen more support than GCC
>LLVM confirmed for better compiler backend
>BSDL confirmed for more developer friendly than GPL

>> No.27022291

>>27021487
Based on that page it would seem Clang supports more of it than GCC does.

I know it's an ongoing effort, but Clang and LLVM have come very far in the last year or two. The package represents a very good compiler toolchain, arguably better than GCC.

>> No.27022300

>>27022244
LLVM isn't BSD licensed.

>> No.27022346

oh look its this butthurt thread again

>> No.27022385

>>27022346
Not butthurt, just raising awareness.

>> No.27022425

>>27022385
go away bono

>> No.27022601

>>27022116
>RMS and the FSF are right up there with MS and other proprietary software vendors when it comes to lock-in.
HAHAHAHAHAHAHA! Pic related!

>The non-standard GNU extensions in gcc and every unix utility makes it a pain in the ass to port code to systems not based on GNU.
Bull. If you want to stick to a standard, do so and pass GCC the correct parameter and it too will stick to them.

>I, for one, can't wait until they either support the standards or become irrelevant because of the shenanigans.
No one is pulling any shenanigans. You're just so full of hate that you can't help but make up bullshit.

>> No.27022621
File: 57 KB, 375x500, hahahahahahaha.jpg [View same] [iqdb] [saucenao] [google] [report]
27022621

>>27022601
Pic. Stupid captcha.

>> No.27022647

>>27022116
>RMS and the FSF are right up there with MS and other proprietary software vendors when it comes to lock-in.

Being forced to use a free license is not "lock-in." It actually prevents lock-in.

>> No.27022709

>>27022647
It's a kind of lock-in in itself. GNUshit restricts your freedom to use the code in any way you wish, which is why actual free licenses like BSD's three clause is better if you care about the end user.

>> No.27022726

>>27022709
But that's wrong. GPL gives you a license to use the code for whatever you want. It's a FREE SOFTWARE license.

>> No.27022768

>>27022647
>GNUshit restricts your freedom to use the code in any way you wish
No it doesn't. It simply prevents you from restricting others. You can use it however the fuck you want, but when you go to give it to others, something you can't do by default, then its terms come into play.

>actual free licenses like BSD's three clause is better if you care about the end user.
But what about people who get my software from others? Why shouldn't they get the same freedom that the person who got the software from me did?

>> No.27022814

>>27022726
What if 'whatever I want' means linking to your GPLshit in my programming project at work, which is unfortunately non-free software?

>>27022768
>No it doesn't. It simply prevents you from restricting others.
You're delusional if you don't realize that's a restriction. The best you could do is argue that it's a necessary restriction.

>But what about people who get my software from others? Why shouldn't they get the same freedom that the person who got the software from me did?
Your original code is always available.

>> No.27022874

>>27022814
You want to restrict your users and take away their freedom, therefore you don't get to use GPL software. I see no problem with this.

>You're delusional if you don't realize that's a restriction.

The GPL restricts you from exerting power over your users. Power is not freedom. So yes it is a restriction, but a necessary one to preserve your freedom.

>> No.27022951

>>27022814
>What if 'whatever I want' means linking to your GPLshit in my programming project at work, which is unfortunately non-free software?

So long as you don't redistribute it, fine. The GPL comes into play -only- regarding redistribution. But you'd have to use logic and read the damn thing rather than lash out with stupidity.

>You're delusional if you don't realize that's a restriction.

It's a minor one, frankly. Or, as it were, that's the price I ask for granting you permission to redistribute.

>Your original code is always available.

But whatever they added is not. And if there are people dependent on those addons, then you've used my work, added a pittance of your own and lord it over your customers. If I use the GPL to license my software, it's explicitly to ensure that you don't reserve all the benefits for yourself.

Typical BSD fanwank bullshit. Sadly, this BS is not restricted to /g/.

>> No.27023016

>>27022874
Nobody cares about your redefinitions. rms zombies mindlessly roam where they're not wanted moaning about `muh freedoms' when it's not important at all.

The code that you originally wrote will always be free if you so choose, so the users are protected: they'll always be able to download your Free software. You're not entitled to the code other people write, and I'm tired of this crap viral license forcing people to give up their freedom to use whatever license they want when they're using code from so-called `free software'.

>>27022951
>So long as you don't redistribute it, fine. The GPL comes into play -only- regarding redistribution. But you'd have to use logic and read the damn thing rather than lash out with stupidity.
If you weren't a moron you'd be able to figure out that I was talking about having to have it be distributable.

>It's a minor one, frankly. Or, as it were, that's the price I ask for granting you permission to redistribute.
The price of my freedom, which your type never seems to give up on whining about.

>>27022951
>But whatever they added is not.
You are not entitled to their code, especially not when you pretend your license is actually free.

>> No.27023054

>>27023016
>You're not entitled to the code other people write
How ironic considering your entire complaint is that you can't take someone else's free code, add features to it, and then extort users by making the whole thing non-free.

>> No.27023068

>>27023016
Neru pls go

>> No.27023095

>>27023054
>How ironic considering your entire complaint is that you can't take someone else's free code, add features to it, and then extort users by making the whole thing non-free.
It is /not/ free code.

Anyone who wants to use free code is entitled to use it. That's what it means to the world by distributing free software: you're entitling people to it.

>> No.27023119

>>27023095
Yes, we've already been over this. GPL code can be used by anyone at any time. It is free software. What you aren't allowed to do is distribute it without source.

>> No.27023169

>>27023119
Code that can't be distributed unless I let you coerce me into doing something with my source that I don't intend to isn't free.

>> No.27023186

Why does every gcc thread degenerates into GPL vs BSD discussion?

>> No.27023213

>>27023169
You want to keep your source from your users. There is no practical reason for you to do this other than to restrict and control them. The ability to do that is not freedom, it's power. This has already been explained numerous times.

>> No.27023252
File: 298 KB, 1200x1500, 1339018676475.jpg [View same] [iqdb] [saucenao] [google] [report]
27023252

>>27023186
Because it's literally a farce of a license that pretends it's free and even gets the GNU GPL users to go around saying that it's free, but then they're reluctant to admit that there are actually retarded arbitrary restrictions on how I can and can not distribute the code and that I shouldn't care that I have to pay a price in my freedom to use their `free' shit.

It's like a false advertiser who'll make you go out of your way to do dumb shit and in the end you don't get what you were promised at all, and they're ashamed to admit it.

>> No.27023261

>>27023186
These GCC hate threads are in essence GPL hate threads and the only people who truly hate the GPL are BSD fanbois that can't are pissed at the limited amount of non-GPLed code available to build off of.

>> No.27023290

>>27023261
Ignore that extra can't. Editing the post went wrong

>> No.27023301

>>27023252
You are free to redistribute the code so long as you don't do it in a manner that is restricting to your end users. How is this a problem?

>> No.27023404

>>27023301
You're not free to redistribute the code with non-free licenses that restrict everyone. How is this a problem?

What's the answer to this question, in one word?

Freedom. Same to yours.

You've ended up as a policeman on the other side of the line, and you're no better.
My code is better free to use and distribute in whatever you can come up with, even if someone else disagrees with that usage. I want a free license that can help me do that, and it will never be the GPL.

>> No.27023462

>>27023252

GPL code *is* free. In much the same way that a free society implements "arbitrary" laws to limit what people can and cannot do, the GPL places basic restrictions on what one is able to do with with GPL licensed code.

As someone else has repeatedly pointed out, you are completely free to do what you wish with GPL code. But should you distribute the result, there is, like society, certain rules you have to meet to ensure a net gain in freedom for the community as a whole. It is a basic conflict between negative liberty and positive liberty that ardent fans of permissive licenses seem to get stuck up on so often.

The basis of your argument is dubious at best, and at worst irrelevant. It just hinges on your interpretation of 'freedom' as nearly doing as you wish, when in reality that is not the only type of freedom, nor is it the one that will ultimately create the biggest gain in freedom on a wider scale.

Your constant inclination to throw petty insults, and use idiotic pictures to support you argument don't do you much good.

>> No.27023506

>>27023404
No, again that isn't freedom. It's power. I really don't see why this is so hard for you to understand when it's been said so many times in so many different ways.

>> No.27023574

>>27023506
Because surprisingly, there are people in the world who have different opinions than us. It's not up to me to preach to them about how I think they ought to use or distribute their software.

>> No.27023590

>>27023186
Because trolls.

>> No.27023591

>>27021261
Every now and then you try to make this thread, with almost the same words

The general consensus is always the same: that OP's mind is deprecated.
The scary thing is that there could be *BSD users out there that really believe OP's bullshits; but there are so many inane gents in the world - no one can take care of them all

>> No.27023670

>>27023574
Ok, that's nice, but that isn't an opinion. That's actually the definition of the word. If you desire power over your users then so be it, but don't act like that is freedom.

>> No.27023716

>>27023670
I didn't say that part was the opinion.

>> No.27023754

>>27023670
That argument would only hold up if you were being forced to use the GPL. Don't like the GPL, then don't use the GPL and don't reuse any code that is licensed under the GPL

>> No.27023756

The problem with the GPL is not that it forces you to release other peoples' code under it, the problem is that it takes away your freedom and your control over your own code. GPL2 was acceptable because it didn't do this too much, but GPL3 is entirely focused on taking away your freedom and forcing communistic rules on people and their code. This difference between the versions is exactly why Torvalds won't use the GPL3.

>> No.27023789

>>27023670
No one has freedom if no one respects other's freedom.

>> No.27023831

>>27023756
>communistic
Stopped reading there. Fucking kids.

>> No.27023849

>>27023756
>The problem with the GPL is not that it forces you to release other peoples' code under it, the problem is that it takes away your freedom and your control over your own code.

Then don't put it under the GPL. And don't modify GPL projects. Write your own and stop bitching.

>GPL2 was acceptable because it didn't do this too much, but GPL3 is entirely focused on taking away your freedom and forcing communistic rules on people and their code.
>communistic
If you even knew what the fuck that meant. Anyway the GPLv3 is about keeping asshole hardware vendors from abusing their customers by using GPLv2 software, releasing the sources, but making it impossible to replace the software with rebuilt or modified versions.

It's a direct reaction to the system-level DRM that's being rammed down our throats.

>This difference between the versions is exactly why Torvalds won't use the GPL3.
Torvalds couldn't use it even if he wanted to.

>> No.27023851

>>27023831
>replying to copypasta
typical GPL babby

>> No.27023860

>>27023756
>GPL
>communistic

Nope. If anything it's the only capitalistic license because it allows free use of code for any purpose while preventing anti-competitive patent abuse and hardware lockout.

>> No.27023876

>>27023016
What? No-one's forcing you to use GPL code in your own creations.

>> No.27023912

>>27023756

This is what fucktards don't understand. GPL places restrictions on the next developer, how the fuck is that freedom. It infects whoever wants to use it. If they really cared about freedom they'd use cddl or some shit.

That way their shit stays free and my shit is released how I want it.

>> No.27023915

>>27023876

"No-one's forcing you" is not a valid argument against anything. The basis of the argument is that you WANT to use the thing in question.

I'm a GNUfag btw, just pointing out how that argument is stupid and you should refrain from using it in the future

>> No.27023929

>>27023876
Which part of that post says “I'm being forced to use GPL code”?

More like it's forcing me not to because it's restrictions don't always play nice in the real world.

>> No.27023939

>>27023912
But it's not your shit. It's their shit with your modifications bolted onto it.

>> No.27023948

>>27023860
>he doesn't know what capitalism nor communism are!

>> No.27023955

>>27023213
But i own the code, it's my freedom to keep it.

>> No.27023965

>>27023939

Exactly! Their shit stays free like they want with cddl.
My shit is distributed how I want.
Freedom FOR BOTH OF US.

>> No.27023979

>>27023929
Use GPL code or don't. Please don't act like a loathsome cunt, however. Your blustery opinion doesn't carry as much weight as you think it does.

>> No.27023980

>>27023849
>Write your own and stop bitching
It seems like he would. Also, your the one who's bitching.

>Anyway the GPLv3 is about keeping asshole hardware vendors from abusing their customers
Stop listening to Stallman and actually read the damn thing.

>It's a direct reaction to the system-level DRM that's being rammed down our throats.
[citation needed]
By who, wtf are you even talking about.

>Torvalds couldn't use it even if he wanted to
stopped reading there.
""The FSF doesn't like that Linux in particular turned the GPLv2 into something pragmatic. The GPLv3 is designed to take the FSF back to its original 'good old days,' when 'Free Software' was a war, and RMS (Stallman) was its proselytizing general. But the fact is, it's not a war, and peaceful and happy co-existence is actually much preferable to moral jihads." " ~Linus Torvalds on the GPLv3

>> No.27023996

The GPL is the botnet of licenses. It infects every project that remotely links to a GPL projects and forces it to join a botnet of GPL software.

>> No.27024006

>>27023965
But the way you want to distribute shit is morally reprehensible and wrong because you want to restrict your users.

>> No.27024017

Can someone explain to me how the fuck anyone is going to prove i used GPL code if i release a closed source binary only?

how is the GPL even enforceable?

>> No.27024046

>>27024006
Thanks for the lecture. Also eating meat is morally reprehensible because you're contributing to enslaving and murdering millions of non-human animals and also because I say so.

>> No.27024057

>>27023980

>stopped reading there.

As would everybody else; it's the end of the post.

Besides, your quote proves nothing. It's true that the Linux kernel could not move to GPLv3 even if Torvalds wanted to.

>> No.27024058

>>27024017
Hackers can reverse engineer your binary or look for behavior peculiar to a GPL implementation.

>> No.27024085

>>27024046
Cool non-sequitur bro. We are talking about humans.

>> No.27024104

>>27024017
It inserts certain strings into the binary saying it was compiled by GCC.

>> No.27024105
File: 40 KB, 494x367, bsd-vs-gnu-hammer.jpg [View same] [iqdb] [saucenao] [google] [report]
27024105

>>27024006
>morally wrong
Pillaging, rape, and murder are morally wrong. Deciding not to give things away to ungrateful bitches is not. You must still be in high school.

>> No.27024131

>>27023915
Am I misreading your post? It sounds like you are saying others should use a specific license because it makes your life easier.

>> No.27024150

>>27023980
>Also, your the one who's bitching.

Err, no. I'm pointing out that he's whining about others not using his preferred license. Or it's copypasta.

>Stop listening to Stallman and actually read the damn thing.

I've read it. The most prominent bits are the patent clauses and the anti-tivoization clause.

>By who, wtf are you even talking about.
Proof that you're totally ignorant of what you speak.

>irrelevant quote
That doesn't negate what I said. Dumb fuck.

>> No.27024166

>>27024104
you're assuming he's using gcc. and that's irrelevant anyway.

>>27024017
Analysis of the binaries would show segments of assembly that are identical.

>> No.27024190

>>27024105
Hah, what a stupid image. If only it weren't a completely stupid analogy.

>> No.27024232

>>27024150
> >By who, wtf are you even talking about.
>Proof that you're totally ignorant of what you speak.
Typical GPL fanboy, can't provide examples, only logical fallacies and ad hominem attacks.

>> No.27024255

>>27024232
fuck you

>> No.27024287

>>27024131
No. People should use the GPL because it makes everyone's lives easier.

>>27024105
>Deciding not to give things away to ungrateful bitches is not.
Ok, I don't see what that has to do with the GPL. The GPL gives freedom to all users, which they all deserve.

>> No.27024342

>>27024232
>Typical GPL fanboy, can't provide examples, only logical fallacies and ad hominem attacks.

Oh -please-. They wouldn't have added the anti-tivoization clause if attacking lock down wasn't a goal. TiVO was just the first, best example of the dishonest use of GPLv2 software. Since then lots of Linux using companies lock hardware down, like Motorola and many other handset/tablet vendors.

>> No.27024347

>>27024166
My bad, I read GCC.

>> No.27024363
File: 53 KB, 230x230, 1345355175037.jpg [View same] [iqdb] [saucenao] [google] [report]
27024363

>>27024255
>>27024255
>>27024255
>>27024255
>>27024255
>>27024255
сheсk em

>> No.27024366

>>27024104
>>27024166
>>27024017
only if they know which binaries to compare

>> No.27024409

>>27024366
There's no instant way to know. If you think someone has stolen your code, you check all of their binaries to the assembly output (via compiler) [or by dumping the assembly of your binary] of your code.

>> No.27024480

Oh boy, the same GNU vs BSD thread with the same GNU vs BSD arguments and the same GNU vs BSD quotes.

>> No.27024536
File: 52 KB, 604x453, 1342415316574.jpg [View same] [iqdb] [saucenao] [google] [report]
27024536

>>27024342
>tivoization
wow, you can wiki

>Motorola and many other handset/tablet vendors
you really don't know the difference between proprietary drivers and DRM do you?

>> No.27024644 [DELETED] 
File: 62 KB, 350x331, 1345354915730.jpg [View same] [iqdb] [saucenao] [google] [report]
27024644

>>27024366
>>27024366
>>27024366
>>27024366
>>27024366
>>27024366
>>27024366
сheсk em

>> No.27024673 [DELETED] 
File: 167 KB, 500x574, 1345354632655.jpg [View same] [iqdb] [saucenao] [google] [report]
27024673

>>27024644
>>27024644
>>27024644
>>27024644
>>27024644

>> No.27024749

LLVM/Clang on FreeBSD 9.0 master race reporting in.

>> No.27024764

>>27024536
You're an idiot, aren't you?

>wow, you can wiki
Err, what?

>you really don't know the difference between proprietary drivers and DRM do you?

This has nothing to do with drivers. This is about the fact that many device ship with locked bootloaders that don't allow you to replace the kernel and support platform-level DRM.

>> No.27024893

>>27024764
>You're an idiot, aren't you?
typical GPL babbbby

>locked bootloaders
That was never the problem with getting your phone to acknowledge your freedom and run your Android firmware of choice. Most of them were easily hackable via shitty coding and security flaws. And now, due to consumer demand, none of the phone makers are doing it anymore. The current problem is drivers, specifically those related to graphics, cameras, and 4G.

Do you even have an Android phone?

>> No.27024984
File: 371 KB, 800x450, 1332748305651.png [View same] [iqdb] [saucenao] [google] [report]
27024984

>>27023756
>>27023849
>>27023980
>>27024150
>>27024232
>>27024342
>>27024536
>>27024764
>>27024893
samefag alert

>> No.27025142
File: 223 KB, 1024x768, kagamisip.jpg [View same] [iqdb] [saucenao] [google] [report]
27025142

>> No.27025235
File: 219 KB, 536x599, just-linux.jpg [View same] [iqdb] [saucenao] [google] [report]
27025235

>>27021261
>guaranteed replies

>> No.27025280

>>27025235
There is no Linux, only GNU.

>> No.27025453

>>27024342
baka

http://www.youtube.com/watch?v=bw58LZTuZjA

captcha: eeekee employed

>> No.27025504

Why all the gcc hate? It's a good enough compiler for most cases, gets the job done and it's free. What else do the haters want?

>> No.27025523

>>27025235
why is he so much cooler than rms

>> No.27025735

>>27025523
Because he's not a fat stinking toejam-eating hypocrite like rms. Also, he uses a MBA.

>> No.27025745

One thing that most people don't understand is that the GPL is "restrictive" like that because of the existence of proprietary code.

You wouldn't have any issue with the GPL or any related licenses if proprietary code didn't exist. Stallman, through the GPL, wants to make sure that all the code is free and is kept free, this has nothing to do with whatever vision of "freedom" you think. Freedom is not "do what the fuck you want, I don't care", because doing so would mean that you'd be able to fuck over other people (see: proprietary code) because "you're free to be a jerk". Freedom is related to code, it's all about free code (see the 4 rules).

Now, if proprietary code were "illegal" and shipping software required also source code and GPL-compatible licensing, then we wouldn't need any GPL license anymore, because everything would be free and available for everyone.

However we live in a world where corporations and proprietary developers are always looking for a way to screw over other companies/the users/other developers, stealing their code, adapting it and then re-selling it, hiding backdoors, hiding viruses/malware/botnet/whatever in order to damage the end user.
Running programs that don't let you access the source code is equivalent to eating food where the recipe is unknown. Who knows what kind of poison you'd be ingesting? As far as I know (at least where I come from, the EU) selling food without a list of the ingredients is illegal, why would it be different for software?

If you don't use GPL you're indirectly supporting proprietary software because you give a chance to other developers/companies to "steal" (it's not literally steal, since your "freedom" license allows it) your code and release it as proprietary, which in turn helps spreading more non-free proprietary software.

>> No.27025779

>>27025504
If you're a normal coder you should try clang, the error reporting will blow you away.

If you like to do source code analysis, clang is good for that too as it's essentially a library in and of itself.

>> No.27025795

Ooh, ooh, this looks like a good time to open the letter I just got from the FSF.

>> No.27025818

>>27025745
>if proprietary code were "illegal" and shipping software required also source code and GPL-compatible licensing, then we wouldn't need any GPL license anymore, because everything would be free and available for everyone
how is the GPL implemented in law not communism?

>> No.27025889
File: 1.15 MB, 1920x1440, IMG_9288.jpg [View same] [iqdb] [saucenao] [google] [report]
27025889

>>27025795

>> No.27025897
File: 2.84 MB, 235x180, carlin.gif [View same] [iqdb] [saucenao] [google] [report]
27025897

>>27025795
What does it say?
Tell us, tell us!

>> No.27025916

>>27025745
And your telling me Mr Stallman can stop Apple or Microsoft if they reuse GPL code as proprietary with a virus in it? Riiight

BSD 4 lyfe

>> No.27025919
File: 1.10 MB, 1920x1440, IMG_9290.jpg [View same] [iqdb] [saucenao] [google] [report]
27025919

>>27025889

>> No.27025946
File: 1.13 MB, 1920x1440, IMG_9293.jpg [View same] [iqdb] [saucenao] [google] [report]
27025946

>>27025919

>> No.27025967
File: 1.59 MB, 2272x1704, IMG_9291.jpg [View same] [iqdb] [saucenao] [google] [report]
27025967

>>27025946
thank you based gnu/stallman

>> No.27025998
File: 147 KB, 646x960, ISHYGDDT2012.jpg [View same] [iqdb] [saucenao] [google] [report]
27025998

>>27025967
>2012
>giving money to stallman
ISHYGDDT

>> No.27026006

>>27025998
gnu with it

>> No.27026017

>>27025967
>free
>have to pay

>> No.27026021

I can definitely sympathize with those who use the GPL, especially for large and substantial software projects that would be worth a lot if it were proprietary. Who the fuck in their right mind would want to help the majority of corporations out there, like Apple, which snaps up all the free resources they can, gives practically nothing of value back, and goes around patenting the fuck out of all their 'innovations'?

So either you fuck over the indie dev to prevent large corporations from cashing in on your hard work and using that money to further constrict everyone's freedom

or

you help the indie dev out at the cost of helping out people/entities that abuse foss at the same time

>> No.27026042

>>27026021
Apple only gave back obfuscated code to khtml and only made webkit free because they had to.
>this is what happens when you don't use gpl

>> No.27026062
File: 87 KB, 384x400, reagan-signed-christmas-smokes.jpg [View same] [iqdb] [saucenao] [google] [report]
27026062

>>27026021
>majority of corporations out there, like Apple
>gives practically nothing of value back
Apple and Google practically bankroll llvm/clang development. What the fuck are you smoking?

>> No.27026084

>>27026062
don't forget the money they put into webkit

>> No.27026110

>>27026062
>>27026084
>give tons of money into free as in bsd projects
Of course, so they can take all the code and make it proprietary for their own exploits later.

>> No.27026140

>>27026084
you might want to read on the history of khtml and webkit

>> No.27026154

>>27026110
why would they even make it proprietary? what benefit does that actually give them?

inb4 expecting gnutards to think beyond rms preaching

>> No.27026162

>>27025795
I already got mine a week ago you pleb

>> No.27026193

BSD: All freedom to the developer, including but not limited to development for evil.
GPL: All freedom to the end user, including but not limited to use for evil.

>> No.27026196

>>27026154
I don't know, why don't you ask Apple why OS X is proprietary?

>> No.27026208

>>27026140
from wikipedia:
"Developers: Apple, KDE, Nokia, Google, RIM, Palm, Samsung, others"

if apple and google hadn't forked khtml, we wouldn't have modern safari, chrome, chromium, ios, android, <insert modern software using web 2.0 backend features>

>> No.27026233

>>27026196
>complaining that OS X is proprietary
doesn't /g/ constantly argue that OS X sucks, why would you care?

>> No.27026243

>>27026154
> why would they even make it proprietary?
Apple bankrolled clang so that they could among other things provide proprietary closed source extensions in their closed source proprietary development tool.

>> No.27026261

>>27026243
>so that they could among other things provide proprietary closed source extensions in their closed source proprietary development tool
but they've been doing that for years with gcc and xcode

>> No.27026285

>>27026261
If you could prove that then I'm sure the FSF would be very interested.

>> No.27026290

>>27025818
Stop equating the GPL to communism. All open source software is doing is showing you the "how".

If you went to the store and bought a bike, you can see exactly "how" it works. It's open source.

If you went to the store and bought a bike, but the bike was a black box that you could sit on and ride around, that's closed source. What happens when a tire goes flat? Does it even have tires? What about gears? A chain?

The problem is that the software industry uses a nonsensical business model, and they use closed source software to PUSH this model because it maximizes profit, and minimizes their own responsibility towards a software product.

Selling software like its fruit at the market doesn't make sense because software is fucking intangible. Software shouldn't be sold 'per copy', because software doesn't have a manufacturing cost -- if you have one copy, you have a million. To push this shit system we get license keys, activation, DRM up the ass.

>> No.27026367

>>27026285
lrn2google faggot
http://opensource.apple.com/release/developer-tools-43/

>> No.27026402

>>27026285
>>27026367
[X] TOLDASAURUS REX
[X] Cash4told.com
[X] No country for told men
[X] Knights of the TOLD Republic
[X] TOLDSpice
[x] The Elder TOLDS IV: Oblivious
[x] Command & Conquer: TOLDberian Sun
[x] GuiTold Hero: World TOLD
[x] CounTOLD Strike
[x] Unreal TOLDament
[X] Half Life 2: Episode TOLD
[x] World of Warcraft: CaTOLDclysm
[X] Roller Coaster TOLDcoon
[x] BattleTOLDS
[x] S.T.A.L.K.E.R.: ShaTOLD of Chernobyl
[X] TOLDasauraus Rex 2: Electric Toldaloo
[x] TOLD of Duty 4: Modern TOLDfare
[x] TOLDmember
[X] Pokemon TOLD and Silver
[x] The Legend of Eldorado : The Lost City of TOLD
[x] X-men: The first TOLD
[x] TOLD Fortress Classic
[x] Toldman: Arkham TOLD
[X] The Good, The Bad, and The TOLD
[x] Super Mario SunTOLD
[x] Legend of Zelda: TOLDacarnia of Time
[X] TOLDstone creamery
[x] StoneTOLD Steve Austin
[x] Mario Golf: TOLDstool Tour
[X] Super TOLD Boy
[x] TOLDboy
[x] TOLD mountain
[x] October TOLD
[X] Left 4 TOLD
[X] BatTOLDfield: Bad Company 2
[x] TOLDman Sachs
[x] Avatold: The Last Airbender
[X] Dragon Ball Z TOLDkaichi Budokai
[x] TOLDcraft II: TOLDs of TOLDberty
[x] Leo TOLDstoy
[x] Metal Gear TOLDid 3: Snake Eater
[x] J.R.R TOLDkien's Lord of the TOLD
[x] TOLDfinger
[x] Man with the TOLDen Gun
[x] TOLDenEye
[X] LitTOLD Big Planet
[x] Rome: TOLDAL War
[x] Gran TOLDrismo 5
[x] TOLD Fortress 2
[x] Castlevania: RonTOLD of Blood
[x] Guilty Gear XX Accent TOLD
[x] Cyndaquil, Chicorita, and Toldodile
[x] was foretold
[x] Tolden Sun: Dark Dawn
[x] Tic-Tac-TOLD
[X] BioTOLD 2
[X] Toldbound
[x] iceTOLD
[x] TOLD of the Rings

>> No.27026406

>>27026062
What they're doing is cheaper and more effective than developing the software in house. Apple, even Google aren't fucking philanthropists. If it was more financially sound to develop anything in a proprietary manner, that's exactly the way it would be done.

They fund llvm. They get professional devs to develop a compiler they can do whatever they want with (make money, impose more restrictions on competition, , eat up smaller companies, proliferate software patents). They also get the input and contributions of hundreds if not thousands of people fixing the software they're funding, make it more robust for FREE.

>> No.27026437

>>27026406
Isn't this the very reason why open source is better?

>> No.27026509
File: 700 B, 56x56, Spr_1b_129.png [View same] [iqdb] [saucenao] [google] [report]
27026509

>>27026406
You do realize that you can compile proprietary code using GCC and that compiling with GCC doesn't make your code GPL licensed?

>> No.27026511

Does anyone remember that terrible article that was posted somewhere that stated how open source anything was communism and in inferior to capitalism? I wish to read it again for laughs.

>> No.27026522

>>27026406
Breaking news: businesses only do things that will make them money.

>make it more robust for FREE
Free for EVERYONE, not just them. They've helped create a toolset that everyone can do whatever they want with. I don't see the issue.

>> No.27026541

>>27026511
Open source is not communism. The GPLv3 is communism.

Also, I'm looking for that one too.

>> No.27026581

>>27026522
Exactly. And it's all due to the BSD license.

>> No.27026694

>In this thread
GPL is communism
BSD is THE BEST
GNU babbys get TOLD
Clang is compiler of the year
General trolling of Stallmanu's sheepy sheepies
Taking anti-GPL copypasta seriously and getting buttraged

>> No.27026703

>>27026511
>>27026541
I don't know if it's the same one I'm thinking of, but there is a famous post from adequacy.org that is pretty much exactly what you've said

>> No.27026709

>>27026522
I'm just trying to argue that they're not "giving back". Any notion of "giving back" that you get from their funding of llvm is a side effect of just doing what's more financially viable. The notion of 'giving back' is more like going out of your way to do something philanthropic.

Its great that everyone can use the result of something that's been funded by Apple and Goog and whoever else, but its just a side effect. Not an act of good faith. You just need to look at Apple's history with kde and webkit or OpenStreetMaps (not technically code, but a similar idea applies) to see what kind of company they are.

>> No.27026757

>>27026290
Okay, then make all software lose demand, and now we have no more software.

Oh the FSF will make some you say? They only exists BECAUSE there is proprietary software.

Your bike analogy is flawed. Not everybody has the metal to make the bike and sell it. Everybody with a computer as the materials needed to make the software, thus software loses its demand, and no one will sell it anymore. It's more complicated to compare software to a real world market than that.

>> No.27026765

>>27026367
Those are not proprietary extensions, you can see the licenses there you moron.

>> No.27026778

>>27026709
Open source as a moral issue will never get off the ground and is rooted in fantasy.

On the other hand, open source being better, more secure, and more efficient than everything else is a fact that needs to be spread.

>> No.27026826

>>27026757
>Okay, then make all software lose demand, and now we have no more software.
That is demonstrably false.
It's even false in the proprietary world - most software is custom/in-house software, meaning most programmer don't actually work on software be sold, they work more like contractors hired to add a custom feature or maintain a project.

>They only exists BECAUSE there is proprietary software.
No they don't, in the beginning of computing practically all software was free, selling software as a product was something that came later.

>> No.27026830

>>27026290
um, I can read x86 and arm assembly so I can understand a binary just as well as someone who can take apart a bike and look at the pieces if they know how bikes are made. I understand assembly languages because (well i read up on them) I understand how compilers work.

Your bike analogy is flawed because the bike is like a binary, a finished product not meant to be understood except by those who make them. The bike's blueprints/parts list/manufacturing details are akin to source code.

>> No.27026859

>>27026765
That was my fucking point.

>> No.27026892

>>27026826
>>They only exists BECAUSE there is proprietary software.
>No they don't, in the beginning of computing practically all software was free, selling software as a product was something that came later.
The FSF and GNU came about b/c AT&T stopped sharing the source code to UNIX. baka

>> No.27026893

>>27026830
>the bike is like a binary, a finished product not meant to be understood except by those who make them.
No, not really, plenty of things 'users' can modify and change on their bikes, and they are allowed to sell their modified bikes.

>The bike's blueprints/parts list/manufacturing details are akin to source code.
The analogy fails because there are no physical copiers, and there's a very real practical limit on how much you can change physical objects.
Blueprints would not be analogous to source code because source code is actually a direct 1:1 correspondence to machine code, whereas a blueprint is not.

>> No.27026904

>>27026709
I completely agree. The only reason Apple put any money into llvm was so they could continue locking their hardware. But GPLv3 caused the change. If not Apple then someone else would have. The fact remains that GPLv3 is completely incompatible with commercial software.

>>27026830
No, the process used to manufacture that bike is akin to source code. The difference is that, with source code, any average Joe can go on the internet and compile a program, but, in the case of the bike, tools machinery, and skills are required to manufacture it.

>> No.27026923

>>27026859
You responded to a post that was asking for source on Apple making proprietary gcc extensions and thus violating the license, yet you linked to a page that shows them writing free extensions.
So what was your point then?

>> No.27026940

>>27026893
>source code is actually a direct 1:1 correspondence to machine code

If this were true, it wouldn't matter what compiler you used. Just sayin'.

>> No.27026952

>>27026893
people do modify proprietary binaries and firmware. it's hard but there are people that can and do do it. Which is similar to myself, as I know about assembly and compilers but I don't know shit about bikes.

but source code isn't a 1:1 with assembly. if you understood compilers you'd know this.

>> No.27026959

>>27026940
Not him, but you're completely missing the point.

>> No.27026963

>>27026703
Sadly that's not the one. It was the kind of article you could actually believe a concern and misguided patriotic parent could have written when learning that people volunteering time to create software is an actual thing that happens.

I remember it stating something along the lines of how open source software must be bad, because if the people writing were capable of writing good software, they would have a paying job instead.

>> No.27026967

>>27026892
>The FSF and GNU came about b/c AT&T stopped sharing the source code to UNIX.
Exactly, so if proprietary software never came about, free software would still exist, since that was the natural state of software being made - getting the source code was pretty much as natural as an instruction manual to a physical machine.

>> No.27026985

>>27026923
I was saying that they did nothing illegal. They do the same legal bundling of open source llvm/clang with xcode as they do with gcc and xcode.

>> No.27027020

>>27026940
Just because there are multiple different mappings doesn't mean there isn't 1:1 mappings, just mean there are multiple sets.
The point is that when you have the source code, you can always get a functionally identical program by using a compiler, and if you use the same compiler you even get the exact same program.
This is not the case in physical manufacturing, there the final products is always different, there are no two identical bikes in the world, it's highly dependent on the person doing the assembling, it depends on the materials, etc.

>> No.27027030

>>27026967
But proprietary software _did_ come about. And the issue was never even that, it was just that AT&T didn't care that people modified the code. It was like a midway anti-GPL interception where people had been modifying the code freely and then the original author told them that they had no control over it anymore.

>> No.27027032
File: 70 KB, 840x768, 840px-2010_Richard_Stallman_(Copyright_vs_Community)_at_CRI_of_Paris_014.jpg [View same] [iqdb] [saucenao] [google] [report]
27027032

Isn't it ironic that the proprietary software developers call us communists? We are the ones who have provided for a free market, where they allow only monopoly. … if the users chooses this proprietary software package, he then falls into this monopoly for support … the only way to escape from monopoly is to escape from proprietary software, and that is what the free software movement is all about. We want you to escape and our work is to help you escape. We hope you will escape to the free world. The free world is the new continent in cyberspace that we have built so we can live here in freedom. It's impossible to live in freedom in the old world of cyberspace, where every program has its feudal lord that bullies and mistreats the users. So, to live in freedom we have to build a new continent. Because this is a virtual continent, it has room for everyone, and there are no immigration restrictions. And because there were never indigenous peoples in cyberspace, there is also no issue of taking away their land. So everyone is welcome in the free world, come to the free world, live with us in freedom. The free software movement aims for the liberation of cyberspace and everyone in it.

>> No.27027058

>>27027030
I was only disproving the statement that without proprietary software there would be no software.

>> No.27027070

>>27027032
that's an ok copypasta, a little rough around the edges. 6/10
work on it a little and it'll be good

>> No.27027097

Oh look, another paid Wintroll circlejerk complaining about the best compiler suite on the planet that is even used all the fucking time to compile Windows programs.

What hypocritical, ignorant, FUD-spewing wastes of human life.

>> No.27027121

>>27027058
I mostly agree, it is just that the software was never really "free" to begin with. I would argue instead that without actively enforced copyright protection of proprietary software there would be no free software licenses.

>> No.27027134

>>27027097
> best compiler suite on the planet
[citation needed]

>> No.27027147

>>27027097
are you talking about mingw?

>> No.27027161

>>27027134
Aww, how cute, the paid wintroll is trying to fight back. It's really pathetic.

>> No.27027181

>>27027097
Where is the FUD in OP's post?
He linked to an article from the gcc website that supports his position that gcc is being held back because of the fear of proprietary extensions to gcc.

>> No.27027184

>>27027121
>I would argue instead that without actively enforced copyright protection of proprietary software there would be no free software licenses.
Yes, no doubt about that, the LICENSES are indeed a direct result of proprietary software existing - one of GPLs purposes is after all to combat proprietary software - but the concept of free software would and did exist before and without proprietary software.

>> No.27027216

>>27027181
>proprietary extensions in free software
>holding back the best compiler suite that doesn't need any proprietary extensions
Wintroll logic.

>> No.27027230
File: 78 KB, 600x900, 1338892423616.jpg [View same] [iqdb] [saucenao] [google] [report]
27027230

>>27027181
>implying it'd be better with proprietary software
0/10 didn't even troll me

>> No.27027245

>>27027184
The concept of sharing your source has existed yes, but "free software" has other connotations to it that require "free software licenses" due to the Berne convention (copyright by default unless explicitly noted). And back then the free/open sharing of source code was more akin to the public domain.

>> No.27027250

>>27027181
> that supports his position that gcc is being held back
Allowing proprietary plugins would be a step backwards, not progress.

>> No.27027268

>>27023016
>Doesn't use proper quotes
>Hates freedom
Hello there, Mr. von Raadt, how's your pathetic excuse for draft dodging working out?

>> No.27027269

>>27027216
>>27027230
Do you even know how to read?
The same thing that prevents proprietary extensions and plugins also prevents free extensions and plugins. The project as a whole is being held back.

>> No.27027286

>>27021734
BSD will be using clang instead of gcc because of GPL3

>> No.27027292

>>27027269
>free software program preventing free software extensions
>apparently wintrolls don't know about my patch
Even more wintroll logic.

>> No.27027297

>>27027250
That's not what we're arguing about. We're arguing that trying to hurt everyone to hurt proprietary plugin development is bad for the good people who want to code open source plugins.

>> No.27027298

>>27027286
And the best software will on forwards to GPLv3, the same thing happened with GPLv1.

>> No.27027316

>>27027298
no it won't
linux won't move to GPL3 too

>> No.27027334

>>27027245
>but "free software" has other connotations to it that require "free software licenses" due to the Berne convention (copyright by default unless explicitly noted)
The only connotations it requires are those setup by FSF, namely to use the program however you want, to be able to share it with whoever you want, and to modify and share modified copies with whoever you want.
All those were commonplace principles in the beginning of software, it was the de-facto standard, the software was nothing more than a way to control a machine, and no one had ever thought of legally preventing someone from using his machine however he wanted.
Software was like cooking recipes - not being able to tell your friend about a new dish would be absurd.

>> No.27027336

>>27027286
BSD can suck my dick. Talk about shit that can be made non-free easily and completely prevent ANY patches and addons.

>> No.27027341

>>27021734
That's a misnomer based on like clang 2.3 (current is 3.1). Recent builds are varied in favor both depending on what the project being compiled is. GCC tends to only beat Clang when OpenMP is involved because Clang doesn't support OpenMP.

>> No.27027344

>>27027250
The point is that the design decision to prevent proprietary extensions make it hard to make free extensions as well. The lack of modularity prevents experimentation with new parsers, optimization techniques, integration with IDEs, etc. The defective design is holding back the whole project as a whole.
Further proof of OP's point is that only know, under pressure from the competition from llvm is gcc trying the correct those design mistakes.

>> No.27027376

>>27027292
>Freetard doesn't know how hard is to get patches accepted in gcc if they implement big changes in any subsystem because the design is not modular.

>> No.27027379

>>27027297
It's not being hurt because allowing a plugin infrastructure like clang would allow proprietary plugins, thus it would be a net loss.
It's better to have no plugins at all, than to have propriety + free plugins.

>> No.27027391

>>27027344
The lack of modularity doesn't do shit. People patch GCC all the time to add new shit and change things, and pretty much every single IDE in GNU/Linux integrates with GCC beautifully.

There is NO challenge making GCC extendable via standard free software methods. It's only asspained Wintrolls trying to subvert GCC that are mad that they can't add their proprietary shit into GCC.

>> No.27027403

>>27027334
Stallman does not have the right to define what freedom is and if you think so you're a fucking pawn.

>> No.27027405
File: 83 KB, 676x496, Theotrain-oncoming.jpg [View same] [iqdb] [saucenao] [google] [report]
27027405

>BSD and "true freedom licenses in a nutshell"
Everything should be truely like free as in true freedom man! Roll another dooby.
>What about microsoft dude, i mean they make the nonfree software and stuff
MAN WHO CARES as long as software IS TRUE FREEDOM it doesn't matter
>Why make it free if it's just going end up proprietary any ways
MUH
MUH
TRUE FREEDOM
AND THE FREE COPY IS STILL THERE
>It's still there but now your software has been further developer as proprietary and now the freedom of others is being denied and there are less users of the free version than proprietary.
DOESN'T MATTER OUR VERSION IS STILL TROO FREEDOM
>W-why do you care about freedom if you don't care about people making the free code proprietary?
MUH TROO FREEODM
FUCK GPL

>> No.27027411

>>27027344
>The lack of modularity prevents experimentation with new parsers, optimization techniques, integration with IDEs, etc. The defective design is holding back the whole project as a whole.
Except it has no problem in those areas, gcc is doing fine.

>> No.27027429

>>27023574
They still don't have to start thread every other day yelling about how our shit sucks.

GNU's restrictions are: The only restriction that you can allow, is this, but you cannot do anything to restrict this.

If this is too restrictive, use more permissively licensed stuff.

If there isn't anything out there, then either roll your own, or hire someone to do so. But don't come complaining on anonymous imageboards about it.

>> No.27027434

>>27027403
Seems like he did it much better than anyone else did because everyone keeps crying over it.

>> No.27027438

>>27027379
If it was better to have no plugins, there wouldn't be tons of people who want a plugin api to make open source plugins and the maintainers of GCC wouldn't have even acknowledged creating one.

>> No.27027463
File: 14 KB, 307x400, CheckmateD.jpg [View same] [iqdb] [saucenao] [google] [report]
27027463

The only "freedom" that the GPL denies is the "freedom" to deny other people freedom and rule over them
Anybody that disagrees with the GPL wants to deny other people freedom and rule over them.

>> No.27027474

>>27027403
He hasn't defined 'freedom' either.
He and I are telling you want is meant by the term free software, you yourself said that 'free software' had connotations, and I assumed you were referring to the de-facto definition of free software - that's how language works after all, meaning is determined by consensus.

You are free to make up your own definition if you want, but calling it 'free software' is likely to cause confusion, so you would have to be explicit or make up a new word.

>> No.27027475

>>27023574
Then don't use that software.

Subverting free software with proprietary stuff is forcing proprietary software on free software.

Let's just face it, this thread is yet another paid MS circlejerk and the sooner we get together and tell moot to block MS IP addresses, the better /g/ will be.

>> No.27027490

>>27027391
>pretty much every single IDE in GNU/Linux integrates with GCC beautifully
false. nothing integrates w/ GCC because it is a locked down blob that can't be used for analysis. That is why clang came about. google clangcomplete.

>> No.27027504

>>27027391
gcc barely integrates with anything. You can't get the intermediate steps and work on them or just use the parser or many other things you can with llvm.

>> No.27027507

>>27027438
There wasn't a demand before, but there is now, and it is being met, so what's the problem?

>> No.27027538

>>27027490
>nothing integrates w/ GCC because it is a locked down blob that can't be used for analysis
Holy fuck, only paid MS trolls can be this fucking delusional and dare to tell such lies.

Everyone should just report this and other obvious paid MS troll threads. I mean, no one can really be this stupid and daring.

>> No.27027546 [DELETED] 
File: 532 KB, 631x477, 1345297388703.png [View same] [iqdb] [saucenao] [google] [report]
27027546

>>27027463
This.

>> No.27027547

>>27027507
It's being met by clang/llvm which is gaining popularity because of it. That's why the GCC people are even thinking about doing it now.

>> No.27027560

>>27027475
>paid MS
Paranoid much? /g/ is below MS' notice.

>> No.27027576

>>27027463
b-b-but it's the opposite of that. The GPL denies other peoples' freedom to do what they want with their code.

>> No.27027580

GCC-
Garbage
Controlled
Compiler

Why aren't you using Clang?

>> No.27027582

>>27027403
it doesn't matter what definition you wanna use, you completly missed the rest of the post - which contained the actual point.
'Free software' - whatever definition you wanna use - was there in the beginning, before proprietary software, ending the ridiculous notion that 'waah without proprietary software there would be no software!!!'

>> No.27027587

>>27027560
>522th most visited site in the US, one of the biggest if not the biggest tech forum is below microsoft's notice
alright buddy

>> No.27027594

>>27027547
>ms viral marketing clang/llvm to remove its hated GCC

>> No.27027602

>>27027547
and clang itself was only created because of gcc, we can play this chicken-and-egg game all day long.

>> No.27027637

>>27027580
Clang-
Compiler
Looking
Autistic
N
Gay

>> No.27027640
File: 707 KB, 600x795, 1345104757064.jpg [View same] [iqdb] [saucenao] [google] [report]
27027640

>>27027576
>The GPL denies other peoples' freedom to do what they want with their code.
That's a power, not a freedom.
>their code.
If they're being forced to use GPL, it's not their code. It's the result of the collaboration of many people.
>The GPL denies other peoples' freedom
The GPL allows you to do anything with GPL code apart from deny other people freedom and rule over them.

>> No.27027644

>>27027580
>Why aren't you using Clang?
produces shit machine code, doesn't support any platforms.

>> No.27027660

>>27027602
llvm was created as a platform to experiment with compilation techniques. Gcc didn't fulfill that role because it was crippled by design.

>> No.27027674

>>27027403
>namely to use the program however you want, to be able to share it with whoever you want, and to modify and share modified copies with whoever you want
BSD does all of that. It's the GPL that restricts peoples' freedom to take control of their own code. If you don't want anyone to use your code for any purpose (including commercial proprietary software) you don't want them to have control over their own code (that isn't yours). The GPL is not freedom, it is anti-copyright which is the same, but opposite, of normal copyright in several of the freedoms it takes away from people.

>> No.27027685

/g/:
Go to /q/.

Start a new thread asking the investigation of /g/ being overrun by posters paid by MS to troll. Explain that most of the shitposting done on /g/ is because of obvious viral marketing done by MS that has no basis in any sort of reality.

Ask that we get a moderator able to remove said threads and a sticky explaining that we should report all threads with a blatant pro-MS troll slant as they are viral marketing threads done by MS.

Do not respond to any of these threads. Simply hide them and move on.

MAKE /g/ A BETTER PLACE: REMOVE MICROSOFT.

>> No.27027700

>>27027334
>namely to use the program however you want, to be able to share it with whoever you want, and to modify and share modified copies with whoever you want
BSD does all of that. It's the GPL that restricts peoples' freedom to take control of their own code. If you don't want anyone to use your code for any purpose (including commercial proprietary software) you don't want them to have control over their own code (that isn't yours). The GPL is not freedom, it is anti-copyright which is the same, but opposite, of normal copyright in several of the freedoms it takes away from people.

>> No.27027710 [DELETED] 
File: 41 KB, 200x200, astroturf.jpg [View same] [iqdb] [saucenao] [google] [report]
27027710

>>27027640
>It's the GPL that restricts peoples' freedom to take control of their own code.
If they're being forced to use GPL, it's not their code. It's the result of the collaboration of 2+ people.
>The GPL is not freedom
The GPL allows you to do anything with GPL code apart from deny other people freedom and rule over them.
>commercial
Commercial does not mean nonfree.

>> No.27027718

>>27027674
Piss off, paid MS troll trying to destroy GPL. Your FUD and lies are laughably stupid and not worth even debating since anyone with any intelligence knows you are horribly, blatantly wrong.

Go back to Ballmer and tell him your push to discredit all things GNU/Linux has failed AGAIN. Tell him he should be fixing his Win8tanic shit OS rather than wasting time telling lies about its eventual and inevitable replacement.

>> No.27027720

A question to the gcc supporters:
Do you deny that gcc is crippled by design or do you claim that such crippling is a necessary evil to prevent proprietary software vendors taking advantage of the work done with gcc?

>> No.27027721
File: 177 KB, 378x500, 1344070103620.jpg [View same] [iqdb] [saucenao] [google] [report]
27027721

>>27027700
>It's the GPL that restricts peoples' freedom to take control of their own code.
If they're being forced to use GPL, it's not their code. It's the result of the collaboration of 2+ people.
>The GPL is not freedom
The GPL allows you to do anything with GPL code apart from deny other people freedom and rule over them.
>commercial
Commercial does not mean nonfree.

>> No.27027726

>>27027674
>It's the GPL that restricts peoples' freedom to take control of their own code
No license can ever control any of YOUR code.
GPL only applies to redistribution of yours AND someone else's code.
If you redistribute only your piece of code, GPL (or any license for that matter) has nothing to say about it.

>> No.27027739

>>27027644
>doesn't support any platforms
Supports *BSD, Linux, and OSX

Are you seriously debating in a best free software license thread while using a non-free OS like windows?

>> No.27027741

>>27027720
>Do you deny that gcc is crippled by design
Lovely flawed logic tactic, paid Wintroll.

>> No.27027757

>>27027720
Loaded question, logical fallacy.

>> No.27027758
File: 279 KB, 900x900, 1326901959200.png [View same] [iqdb] [saucenao] [google] [report]
27027758

>>27027376
>I'mma do big changes in a project that has millions of users and is used EVERYWHERE.
>Why is it so hard to get patches accepted?

Do you even listen to yourself?

>> No.27027774

>>27027720
False dichotomy.
I deny it's even crippled at all.

>> No.27027776

I like how people think that this is a pro-MS thread when clang doesn't even run on windows.

>> No.27027788
File: 35 KB, 322x504, 1342164418826.jpg [View same] [iqdb] [saucenao] [google] [report]
27027788

>>27027721
why does colour make that food look so good?

>> No.27027809

>>27027758
And yet, if gcc were modular your patches were more likely to be accepted because they don't mess with the whole thing. Or even if they are not accepted in the main branch you can maintain them separately with ease because the whole system is modular.

>> No.27027842

>>27027685
You can ban MS addresses all you want but you'll achieve nothing because OP is not a paid shill.

>> No.27027853

>>27027739
clang only targets x86 and ARM plus partial support for PPC, SPARC and MSP430.

This is not even a complete list of all platforms gcc supports:
http://gcc.gnu.org/install/specific.html

>> No.27027858

>>27027726
Wrong, once your code is in GPL it can never be un-GPL'd. Even if you rewrote all of the other pieces of a GPL'd program. Yes you could take all of your code and put it into a proprietary version, but you still have to share the GPL version.

>> No.27027866

>>27027809
But then again, most people on the wiki are in favour of doing so.
It's going to happen, if nothing else then because new compilers force them to.

My point was merely that if you have anything that's huge and widely used, then obviously, changing substantial parts of it carries a lot of risk.

>> No.27027872

>>27027858
If they're being forced to use GPL, it's not their code. It's the result of the collaboration of 2+ people.

>> No.27027892

>>27027853
clang is the frontend compiler. llvm makes the machine code. llvm only makes machine code for those platforms because most of the ones in the gcc list are fucking dead.

llvm/clang is recent, gcc is ancient. it makes sense that gcc would support dead architectures that clang/llvm don't. also platform != architecture.

Go learn how compilers work.

>> No.27027937

>>27027872
You do realize that rewriting a program based on knowing what it is supposed to do is at grayest, reverse engineering and that by the end all of the code is the result of the collaboration of 1 person with him/herself. Retyping code =/= rewriting a program piece by piece because the original code sucks.

>> No.27027940

>>27026402
>>27026367
Holy shit you fucking retard.

What you linked to was open source free software extensions to GCC, Apple offered this functionality optionally for use in their closed source proprietary development tools.

Are you too fucking dumb to know what open source/closed source even is?

>> No.27027954

>>27027858
>once your code is in GPL it can never be un-GPL'd.
Yes it can, you are the author of the code, you can always do whatever you want unless you've explicitly transferred ownership to someone else - which GPL does NOT do.
You can't _revoke_ it of course - but then again you can't do that with any free license, including BSD - but you are always allowed to re-license, make your next version proprietary, or whatever you want.

>but you still have to share the GPL version.
You - as the author - don't HAVE to do anything, the only time you have to share it is if you've incorporated OTHER peoples code, and then it's not solely your code anymore.

>> No.27027963

>>27027866
That's the whole point: many people have wanted plugins for gcc for ages but they were prevented from working on them because of fear of proprietary plugins. It took llvm becoming a serious contender to make gcc's people to change their mind. Gcc was being held back because of ideological reasons.

>> No.27027967

>>27027940
Why do you even need to ask?

He posted a TOLD list. That should tell you everything.

>> No.27027972

>>27027937
You do realize that they used GPL and that's the end of the story. You're lucky you're being granted the freedom to redistribute and see the code. Stop complaining about GPL and go complain about proprietary software, they're the real problem.

>> No.27027994

>>27027954
sorry but you're wrong, go read the GPL.

>> No.27028014

> can't into openmp
> can't into most architectures for computing nodes
> produces slow as shit binaries
clang is beyond bad, it's totally irrelevant.

That's why the fanboys are extremely fervent, they're like Haskell users, infinitely annoying and have no results to back their shit talk up with.

>> No.27028017

>>27027963
Then you should have said that.
What you said instead, was kind of ridiculous.

>> No.27028022

>>27027940
see >>27026985

>> No.27028027

>>27027994
You go read the GPL you fucking retard.

You can re-license, dual-, triple-, quadruple-license your code if you're the owner.

>> No.27028054
File: 404 KB, 446x400, Girls-1.png [View same] [iqdb] [saucenao] [google] [report]
27028054

>>27028014
> openmp

>> No.27028058

>>27028022
See >>27026285, >>27026261 and >>27026243 you stupid piece of shit.

Note how you have posted no evidence that Apple has been violating the GPL and providing closed source proprietary extensions to GCC.

But you're probably too fucking dumb to even understand what I'm talking about.

>> No.27028069

>>27028054
> I don't know what openmp is
> I'm just going to laugh at the industry standard for high performance numerical computing

>> No.27028072

So how long until Clang can compile from and for every platform and produce the fastest binaries? It's almost like a fucking Intel compiler without the speed at this point seeing as how it is developed by Apple.

>> No.27028073

>>27027892
>llvm makes the machine code.
llvm doesn't compile C code.

>llvm only makes machine code for those platforms because most of the ones in the gcc list are fucking dead.
Then you're clearly clueless if you think x86 and arm are the only computers out there.
AVR, MIPS, Sparc, POWER (PPC != POWER) are used a lot.

>llvm/clang is recent
Hence my statement about it being shit.

>it makes sense that gcc would support dead architectures that clang/llvm don't. also platform
They are not dead architectures.

>platform != architecture.
They can be used interchangeably when talking about compiler targets, supporting a platform implies supporting an architecture, a certain computer running a certain architecture is called a platform, supporting that platform thus means supporting that architecture.

>Go learn how compilers work.
Should take your own advice.

>> No.27028080

>>27028027
Let me repeat this. You can do whatever you want with code you yourself wrote. However, once any section of it is GPL'd, THAT GPL'd code cannot be un-GPL'd. You can however still release the code as proprietary (or under any license you want).

In addition the GPL specifically states that you have to make the GPL'd code publicly and easily available. Even if you maintain a closed-source version.

>> No.27028096

>>27028080
> However, once any section of it is GPL'd, THAT GPL'd code cannot be un-GPL'd.
Unless you know, you decide to actually un-GPL it.

How the fuck do you think dual-licensing even works? You give people a choice, you either choose this license or the other, if they don't chose the GPL one they don't have to give two shits about it.

>> No.27028101

>>27028014
>Haskell
>Statically typed language
>Made by people who think that mathematical notation is neat.
>But still have different paren-rules because you type fewer characters.

Yeah, this sounds about right. clang-complete is still neat though.

>> No.27028103

>>27027994
Go look at MySQL or QT you moron, they have dual licensed their code with one being (L)GPL, Stallman was even the one who recommended MySQL be dual licensed.

>> No.27028106
File: 37 KB, 337x368, laughing girls.jpg [View same] [iqdb] [saucenao] [google] [report]
27028106

>>27028069
> can't defend shittiest parallelism framework on the planet
> resorts to age-old "you don't know what it is" troll

>> No.27028132

>>27028073
>llvm doesn't compile C code
frontend compilers take the c code and translate it into an intermediate representation. idiot.

>AVR, MIPS, Sparc, POWER (PPC != POWER) are used a lot.
it targets mips and sparc http://llvm.org/docs/doxygen/html/dir_4dd1d6cde120bbd790a89b34785ed945.html

>> No.27028142

>>27028106
Oh yes, I'm sure whatever shit you use is better than what has been carefully developed over several years by the leading players in the industry collaborating with top academics.

Go on, make fun of yourself some more.

>> No.27028145

>>27028080
>However, once any section of it is GPL'd, THAT GPL'd code cannot be un-GPL'd.
But that is true of ANY license.
Once you've released something under a license you of course can't suddenly come and revoke that, that would be a contract violation.

>In addition the GPL specifically states that you have to make the GPL'd code publicly and easily available.
Only if you're distributing binaries and only if you've incorporated other peoples code - if it's 100% your code, you can do whatever you want

>> No.27028159

>>27028103
They have a code submission agreement that gives copyright control to them. That's how most orgs do it.

>> No.27028188

What have we here? Is this how you greet me in the morning?

One Apple fan samefagging about clang and retards actually responding to him?

>> No.27028189

>>27028145
>Only if you're distributing binaries and only if you've incorporated other peoples code - if it's 100% your code, you can do whatever you want
wrong.

>> No.27028198
File: 57 KB, 552x752, whatthehell.jpg [View same] [iqdb] [saucenao] [google] [report]
27028198

>mfw this troll thread is the mist commented in /g/.

>> No.27028201

>>27027963
>>27028017
Oh fuck. I thought you were someone else.
That was retarded of me, sorry.

>> No.27028203

>>27028132
>frontend compilers take the c code and translate it into an intermediate representation. idiot.
There are several steps to translating C to machie code.
You cannot use clang to target anything other than x86 and ARM (+ partial ppc and sparc support)

>it targets mips and sparc http://llvm.org/docs/doxyge /html/dir_4dd1d6cde120bbd790a89b34785ed945.html
Not fully supported, it's the reason FreeBSD 10+ will still use gcc on their sparc ports.

>> No.27028208

>>27028189
>wrong.
Nope, read the license.

>> No.27028209

>>27028188
I think he is a BSDfag.

>> No.27028211

>>27028189
> wrong.
Nope.

You can even make proprietary extensions to free software as long as you do not distribute it to anyone else.

If you had to make your code freely available to everyone who asked then how on earth would you still have the capability to charge money for it? (You know, like Stallman did with Emacs.)

>> No.27028242

>>27028159
>They have a code submission agreement that gives copyright control to them.
Exactly, that means they are the sole authors of the code - meaning they can't choose to license it however they want.
If you make a project, release it as GPL, never accept any patches unless they hand over ownership to you, you can re-license it, make it propriety, or whatever you want.

>> No.27028244

>>27028203
It means they're working on it. So don't assume it will only be used for x86 and arm.

>> No.27028248

>>27028242
they CAN choose to license it however they want*

typo.

>> No.27028261

>>27028242
>Exactly, that means they are the sole authors of the code - meaning they can't choose to license it however they want
don't you mean they CAN choose to license it however?

>> No.27028265

>>27028244
>It means they're working on it.
In other words - it's currently shit.
You don't get points for trying, call me when clang can actually outperform gcc and supports the relevant platforms.

>> No.27028275

>>27028261
yes, it was a typo.

>> No.27028286

>>27028265
OK. And you call me when GCC isn't a bloated piece of shit that doesn't allow source code analysis for security purposes like finding and preventing exploitable bugs.

>> No.27028302
File: 31 KB, 335x198, hsface.png [View same] [iqdb] [saucenao] [google] [report]
27028302

>In this thread
People expecting new compilers to support all modern architectures

>> No.27028308

>>27028265
> You don't get points for trying, call me when clang can actually outperform gcc and supports the relevant platforms.
They won't ever do that of course.

I'm telling you these people are like Haskell fans, they go on about how everything is best done in Haskell, how Haskell theoretically can outperform everything and how much better it is than everything else. Yet there's not a single piece of good software out there written in Haskell and in benchmarks it lags way behind.

There is absolutely no point in arguing with clang fans, until their compiler is usable on multiple architectures and produces binaries that are worth running it will remain a laughable toy compiler, no amount of whining on the internet can change that.

>> No.27028330

>>27028308
apache httpd

>> No.27028335

>>27028286
> OK. And you call me when GCC isn't a bloated piece of shit that doesn't allow source code analysis for security purposes like finding and preventing exploitable bugs.
GCC is a compiler collection, compilers compile code.

If you want source code analysis maybe you should check out, oh I don't know, a source code analyzer?

>> No.27028363

>>27028330
Yes, what about it?

>> No.27028366

>>27028335
Compilers have all of that functionality, but gcc isn't modular and prevents such uses unlike clang.

see >>27027344
and >>27028017

>> No.27028380

>>27028366
Compilers compile code you stupid piece of shit.

How are you making the argument that they should also perform source code analysis while concurrently making an argument __against__ bloat?

You're fucking retarded.

>> No.27028385

>>27028308
It's more like Firefox vs. Chrome.

Firefox came in with a huge advantage, a committed dev community, and lots of user support. Unfortunately, their ancient code base, and questionable organizational goals hindered them. They failed to support companies interested in using their software (how damn difficult it was to administer Firefox). They supported everything under the sun, and people assumed that was they way things should be. But then Chrome came along, and it was just... less painful.

GCC is in a similar position. Clang is interesting as a compiler, but half of the exciting things going in with Clang come in the tooling world. Google is running MapReduces on fucking source code with clang, performing automatic refactoring, etc. Want to see something fucking amazing? Look at a project like AddressSanitizer. That shit would be suicide-inducing in GCC.

GCC has a good code base, and the devs have been pretty good about responding to the positive points of clang. It remains to be seen whether what they do is enough though.

>> No.27028391

>>27028286
Even clang static code analyzer is in it's infancy and thus shit.
There are way better tools out there, there was a comparison done on some major open source projects like Firefox, Apache, Linux, gcc (itself), and clang was way behind all other tools, can't remember the article, gonna try dig it up.

>> No.27028411

>>27028302
When you call other compilers deprecated I do.

>> No.27028413
File: 386 KB, 1024x768, successfultroll.jpg [View same] [iqdb] [saucenao] [google] [report]
27028413

>OP in his office.

>> No.27028416

>>27028385
This.
/thread

>> No.27028417

>>27028335
"Hey, we heard you have a program which can analyze source code, convert it into a well represented AST, and output machine code. Can we skip the last step and use it for program analysis"

NO, MUH FREEDOMS PREVENT THAT, GO WRITE YOUR OWN TOOL, WHY WOULD YOU WANT TO REUSE CODE.

Classic GCC fanboy.

>> No.27028425

>>27028302
>gcc is deprecated, use clang and llvm
>but clang and llvm doesn't support $architecture
>of course not its' still in its infancy.

for serious?

>> No.27028435

>>27028385
Except that Chrome can actually show webpages and actually does support a lot of extensions.

clang is just a toy compiler that's completely unsuitable for any large scale work.

It's more like Firefox vs Netscape 1.2

>> No.27028461

>>27028385
>GCC has a good code base, and the devs have been pretty good about responding to the positive points of clang. It remains to be seen whether what they do is enough though.
It's extremely rare that one piece of software surpasses another because of technical superiority. GCC has way too strong foothold, the only way clang can take a chunk of that is via aggressive marketing (ie, promote it via osx/ios development, whether clang is actually technically better is unfortunately kinda irrelevant in the software world.

>> No.27028462

>>27028417
You are saying that compilers should also be source code analyzers, yet you're somehow also saying that bloat is bad.

Why the fuck do you want to introduce bloat into compilers then?

>> No.27028484

In the HPC world there's a pretty famous joke, it goes like this.

'clang'

Everybody laughs every time.

>> No.27028516

>>27028462
Static code analysis is essentially the exact same task as compiling only you skip the 'output machine code' part, making a separate tool for that would require code and library duplication which would be bloat.

>> No.27028520

>>27028435
>Unusable for large scale work

Do... do you know what that word means?
Your adorable little Atmel projects are just downright sad compared to what Apple and Google are doing. Do you understand just how much Apple does with Clang? It's everywhere in their build systems.

But no, Clang is doomed, because who would use a compiler which doesn't support your sad little embedded platform. People had the same complaints about Chrome -- OMG, WHERE'S PRINT PREVIEW?

Guess what, people who do real work don't give a shit.

>> No.27028528

>not using the Plan 9 compilers
shigure daigi

>> No.27028531

Isn't GCC still faster at execution time?

>> No.27028535

>>27028516
> having less features introduces bloat
Are you sure you know what the word 'bloat' actually means?

And if you think code analysis is just building an AST you're so wrong it's not even funny.

>> No.27028557

>>27028461
To be honest, GCC's license is *really* hurting it in the corporate world. That was a huge factor in Apple's decision to back it, and it definitely was there for Google as well. But software definitely gets replaced because of technical superiority -- marketing isn't the last word of the world.

>> No.27028562

>>27028520
It can't even compile for Blue Gene/P, I'm sure whatever babby shit you're developing for the iOS is really important but clang is and continues to be a toy compiler.

>> No.27028576

>>27028535
Do you understand the concept of modularity? It really isn't that hard -- you have a frontend which can be adapted for use in a static analyzer, and a backend which can be adapted for all sorts of language agnostic optimization and instrumentation.

Oh wait, you're a GCC fanboy, of course you don't understand modularity.

>> No.27028587

>>27028531
No.

>> No.27028604

>>27028587
>>27028531
Good lie, of course it still is, at least according to every benchmark.

>> No.27028616

>>27028535
>Are you sure you know what the word 'bloat' actually means?
Yes, it means unnecessary features, extra space or time requirements or duplicated effort - which writing a separate tools would be.

>And if you think code analysis is just building an AST you're so wrong it's not even funny.
That's what I said you moron, it essentially leverages the same kind of infrastructure as the front-end of a compiler, so it makes complete sense to export that kind of functionality as an interface or put common compiler and analysis feature together in a library so both the code generator and the analyzer use it, which minimizes bloat and code duplication.

>> No.27028634

>>27028604
Not all.

>> No.27028645

tcc master race reporting in

gcc users jelly, clangfags mad.

>> No.27028656

>>27028634
Can you link to any one benchmark where GCC wasn't the definite winner?

>> No.27028679

>>27028656
http://www.phoronix.com/scan.php?page=news_item&px=MTA5Nzc

>> No.27028686

>>27028656
>suddenly the thread went silent

>> No.27028700

>>27028686
suuuuurrreeeee

>> No.27028716

>>27028679
Yes? Some of the tests has GCC about the same as clang and others have GCC way ahead of clang.

Overall the benchmark shows that GCC is way ahead of clang, as expected.

>> No.27028719
File: 178 KB, 800x810, 1345354857227.jpg [View same] [iqdb] [saucenao] [google] [report]
27028719

>>27028700
>>27028700
>>27028700
>>27028700
>>27028700
>>27028700
сheсk em

>> No.27028729

>>27028562
I'm not an expert in HPC, but targeting every backend ever just isn't a priority for Clang/LLVM right now. Checking the source does seem to indicate that there've been commits explicitly mentioning support for Blue Gene/P, but I'm not the one to ask about that.

>> No.27028735

>>27028716
Clearly you didn't read it right.

>> No.27028741

>>27028679
>>27028716
Not to mention that clang can't even compile openmp annotated code, so they can't even benchmark the heavily numerical stuff.

>> No.27028753
File: 59 KB, 582x582, curious-george.png [View same] [iqdb] [saucenao] [google] [report]
27028753

>>27028741
>openmp
is dis nigga srs?

>> No.27028769
File: 53 KB, 331x319, 1344645699084.jpg [View same] [iqdb] [saucenao] [google] [report]
27028769

>>27028741
>openmp

>> No.27028774
File: 76 KB, 519x534, 1332750573588.jpg [View same] [iqdb] [saucenao] [google] [report]
27028774

>>27028741
Call me when gcc gets halfway decent c++11 support.

>> No.27028814
File: 61 KB, 550x562, there-be-a-shit-storm-a-brewin.jpg [View same] [iqdb] [saucenao] [google] [report]
27028814

>>27028774
>c++11
inb4 ultra super mega shitstorm

>> No.27028828

>>27028484
Oh god why did I laugh

>> No.27028830

>>27028814
C++11 (stupid shit like UDLs excluded) is a godsend. C++03 had some fucked up shit in it.

>> No.27028834

>>27028774
>2012
>saging in an autosage thread
eye shiggy diggy

>> No.27028835

>>27028729
It's not solely about the backend, it's also about OpenMP, it's basically useless when it comes to HPC since the trend of high performance parallel machines of even less memory per node. It's way too expensive to create one MPI process per core on a BG/P (not to mention a BG/Q or any other top supercomputer for that matter).

Then there's that whole issue with it generating numerical code that are several orders of magnitude slower than that of other compilers like GCC and ICC.

>> No.27028847

>>27028769
>>27028753
Yes, MPI/OpenMP is the industry standard.

>> No.27028879
File: 36 KB, 300x250, allOver75_1.jpg [View same] [iqdb] [saucenao] [google] [report]
27028879

>>27028847
>industry standard
this non-standard shit
#include <omp.h>
#include <stdio.h>
#include <stdlib.h>

int main (int argc, char *argv[]) {
int th_id, nthreads;
#pragma omp parallel private(th_id)
{
th_id = omp_get_thread_num();
printf("Hello World from thread %d\n", th_id);
#pragma omp barrier
if ( th_id == 0 ) {
nthreads = omp_get_num_threads();
printf("There are %d threads\n",nthreads);
}
}
return EXIT_SUCCESS;
}

>> No.27028896

Recent benchmarks.
http://www.phoronix.com/scan.php?page=article&item=gcc48_llvm32_svn1&num=2

The differences are even bigger on AMD hardware.

>> No.27028902

>>27028879
Yes, you managed to copy paste something from Wikipedia, congratulations.

It's still the industry standard, together with MPI for distributed memory, have you found a way to quell your tears yet?

>> No.27028907
File: 2 KB, 254x192, winrar.png [View same] [iqdb] [saucenao] [google] [report]
27028907

>>27028879
*clap* *clap* *fap*

>> No.27028926

>>27028774
*calling*
http://gcc.gnu.org/projects/cxx0x.html

Only concurrency is missing, but that is a doubtful to be even worth it, practically everyone uses and will use 3rd party libraries for that anyway - there's just too much legacy code relying on existing code.

>> No.27028935

>>27028896
The tears of clang users are so delicious and salty.

> B-but it's better in theory
I'm telling you it's Haskell users all over again, they will never produce any real world results. You'll obviously get that one guy who looks at that benchmark, probably finds a single test (out of like 100) where clang is almost at par with GCC and then take it as evidence that clang produces better binaries, such is fanboyism.

>> No.27028941

>>27028835
Don't worry, I'm not a fanboy -- I'm well aware that GCC/ICC are a long way ahead in terms of intense numerical computation. Given that Clang really isn't that interested in trying to win that fight. Knowing they're the underdog, Clang/LLVM are trying to play up their strengths rather than fight a losing battle over their weaknesses.

So please, go ahead and use GCC/ICC to compile code for HPC systems. It's just not what the LLVM dev team is really interested in working on right now.

>> No.27028948

>>27028926
>Only concurrency
u best go get your eyes checked

>> No.27028983

>>27028948
No compiler has or ever will have 100% C++ support, the things missing in gcc are minor shit no one every uses.
clang doesn't support everything either - and it never will.

>> No.27028998

>>27028941
> my toy compiler can't do any heavy lifting
> actually doing calculations of worth is not something it's interested in doing
So what is it supposed to be used for then? Your new iPhone app or something worthless like that?

>> No.27029000

>In this thread
2 in this threads
clang makes slower binaries than gcc
clang is great for integration w/ shit
gcc is shit
clang has great error reporting
gcc is fucking up its plugin support on purpose
clang doesn't support backwards architectures from 30 years ago
>openmp

>for everything else, there's mastercard

>> No.27029017

>>27028983
clang is actually 100% conforming to C++03 and it's nearing 100% conformance to C++11 shows how much you know freetard.

>> No.27029027
File: 354 KB, 320x320, 1345354776321.gif [View same] [iqdb] [saucenao] [google] [report]
27029027

>>27029000
>>27029000
>>27029000
>>27029000
>>27029000
>>27029000
>>27029000
>>27029000
>>27029000
>>27029000
>>27029000
>>27029000
>>27029000
>>27029000
>>27029000
сheсk em

>> No.27029042

>>27028935
>Real world results
You... you don't understand benchmarks for recent hardware, do you?

God GCC fanboys are funny. Don't worry, you'll understand when you're older.

Oh and while I'm here, why the shit are people posting Phoronix links? I've seen them from both the clang and gcc people, so just to remind people -- if you think Phoronix is in any way legit... lol.

>> No.27029068

>>27028896
> Real C/S, more is better
> LLVM/Clang: 721
> GCC: 2102
> John The Ripper heavily benefits from GCC since there's proper OpenMP support there for multi-threading but OpenMP remains unavailable for LLVM/Clang.

Let's get this shit started
> clang

>> No.27029073

>>27029042
>if you think Phoronix is in any way legit... lol
what praytell on the entire internet is a source of unbiased benchmarks, kami-sama?

>> No.27029092

>>27029042
> You... you don't understand benchmarks for recent hardware, do you?
Yes I do, if you look at any of them, you'll see that GCC is far ahead.

> if you think Phoronix is in any way legit... lol.
What are you suggesting, that Stallman is paying them to trick with the benchmarks?

>> No.27029094

>>27029017
>clang is actually 100% conforming to C++03
Nope, doesn't support exported templates.

>and it's nearing 100% conformance to C++11
So is gcc.

http://clang.llvm.org/cxx_status.html

14 (gcc) vs 10 (clang) features missing, hardly a big difference, disregard concurrency (which no one will use) and you have 5 vs 4.

>> No.27029101

>>27029073
and silence has fallen yet again

>> No.27029115

>>27028896
> Smallpt v1.0
> Seconds, less is better
> LLVM/Clang: 205
> GCC: 70

> clang

>> No.27029122

>>27028998
http://www.youtube.com/watch?v=mVbDzTM21BQ
http://code.google.com/p/address-sanitizer/wiki/AddressSanitizer
http://code.google.com/p/thread-sanitizer/wiki/ThreadSanitizerAlgorithm

I could go on. It's *SO* much easier to add things to clang. Shit, AddressSanitizer has been responsible for finding some ridiculous number of the Chrome security bugs.

Of course, someone who jerks off over benchmarks wouldn't know anything about managing large code bases, or dynamic analysis tools.

>> No.27029139

>>27029094
>thinking exported templates was ever more than extension in c++03
"During the March 2010 ISO C++ standards meeting, the C++ standards committee voted to remove exported templates entirely from C++0x, but reserve the keyword for future use." -wikipedia

>> No.27029148

>>27028896
Wow holy shit, gcc is raping clang, there are only a couple where they are close and there are only 2 tests on all of those 4 pages where clang is a little bit better than gcc. Is clang really this bad?

>> No.27029153

>>27029094
Exported templates were removed in C++11. No one uses them, and they've made a conscious decision not to include them.

>> No.27029160

>>27029115
openmp again

>> No.27029173

>>27029139
They removed them from the NEW standard, they are still part of C++03 you idiot.
It's the same as C11 removing VLA (it's only an optional feature) whereas in C99 is a required feature.

>> No.27029187

>>27029122
> Of course, someone who jerks off over benchmarks wouldn't know anything about managing large code bases, or dynamic analysis tools.
Oh yes, I totally agree, benchmarks are totally worthless and doesn't actually teach you anything about the quality of a compiler.

Seriously, Apple fanboys are you really this retarded?
>benchmarks shows you something that you don't like to see
>whelp, must be best to ignore it completely then instead of facing reality

>> No.27029203

>>27029092
>What are you suggesting, that Stallman is paying them to trick with the benchmarks?
I was going to explain things like AVX2, new ISAs, and the process of bringing new features into a compiler, but then I read this.

Phoronix. Is. A. Joke. If you think Phoronix is anything resemble legit, there's no point in even discussing this with you. The compiler community literally jokes about the stupid shit they report. They've been actively trolled by devs, just for fun.

>> No.27029218

>>27029187
Ubuntu/Debian/Arch/FreeBSD user/system administrator reporting in.

I like clang and hate OSX. Go fuck yourself.

>> No.27029236

>>27029160
What about Timed HMMer Search v2.3.2
> Seconds, less is better
> LLVM/Clang: 32.57
> GCC: 26.31

Can't blame on openmp support there (lack of which is completely laughable and indicative of just another toy compiler by the way)
> clang

>> No.27029250

>>27029203
>Phoronix. Is. A. Joke
no sources

>The compiler community literally jokes about the stupid shit they report
>compiler community
suurree

post mailing list sources or GTFO

>> No.27029253

>>27029153
>Exported templates were removed in C++11.
Still need to support them if you wanna call yourself a 100% _C++03_ conformant compiler.

>> No.27029255

>>27029203
So do you have any proof, alternative websites, anything at all or are you just talking shit?

>> No.27029269

Man, next time an Apple fan starts talking about clang I'll just dig up a benchmark, these things are great.

>> No.27029270

I for one agree w/ the guy talking about phoronix being shit.

They don't even realise that clang svn builds have auto debugging built into the tree that slows it down.

>> No.27029272
File: 124 KB, 580x386, yawn2.jpg [View same] [iqdb] [saucenao] [google] [report]
27029272

>nonfree software
>nongpl software
>mfw

>> No.27029275

>>27029187
No, the Clang devs know that most numerical benchmarks favor GCC. It's just that no one really cares right now. Clang is getting so much cool shit added into it right now, and frankly, no ones really interested in getting every little micro optimization for every architecture added into the backend. Not to mention, of course, that Clang was a fucking grad students project in ~2005, so the fact it's even close to GCC is, well...

>> No.27029284

>>27029270
> They don't even realise that clang svn builds have auto debugging built into the tree that slows it down.
That would slow down the compiler binary but not the code it generated, confirmed for being dumb as shit.

>> No.27029314

>>27029275
I love how some of the Apple fans in this thread say stuff like "clang is way better than gcc at producing fast binaries" while others are like "yes the benchmarks all indicate that clang generate slow as shit binaries, but that's not important for the dev team".

Can you make up your mind Apple fans?

>> No.27029324

>>27029284
>being dumb as shit
you don't understand compilers do you?

>> No.27029339

>>27029314
And then other people talk about those benchmarks being unreliable.

>> No.27029351

>>27029324
Yes I do, very well.

Explain to me how building the compiler with debug information in the binary would slow down the generated binaries. I'm sure a compiler expert like you could tell me.

>> No.27029363

>>27029339
> And then other people talk about those benchmarks being unreliable.
And then they get asked for any source for their claims and can't provide any at all, so they get silenced.

>> No.27029364

>>27029314
Yeah bro, all clang fanboys are the same person, that guy is fucking dumb!

>gcc fanboys confirmed for incapable of critical thought.

In fact, I think most of the people in this thread are pants on head retarded... clang or LLVM

>> No.27029372

>>27029275
>so the fact it's even close to GCC is, well...
It's not anywhere close, doesn't support even a fraction of the targets and runs on hosts gcc does.

>> No.27029374

>>27029284
>>27029351
you may understand compilers, but obviously you can't detect obvious trolling.

>> No.27029390

GCC
> produces faster binaries
> doesn't allow companies to anus violate you

Clang
> produces slow as shit binaries
> allows companies to anus violate you

Since Clang seems to be linked with Apple the anus violation being regarded as a plus doesn't really surprise me.

>> No.27029413

>>27029363
Sorry, there are a lot of dumb people for me to respond to, and it's gonna be my bedtime soon :3.

Phoronix has a reputation for reporting on rumors and making ridiculous assumptions based on mailing list postings. Half their "news stories" about Clang/GCC come from skimming a mailing lost posting by some rando and then extrapolating to high hell. The results of their benchmarks (GCC produces faster binaries in a majority of cases) are accurate, but I have *VERY* little faith in their methodology.

>> No.27029418
File: 364 KB, 343x604, push-eax.png [View same] [iqdb] [saucenao] [google] [report]
27029418

>>27029364
>all clang fanboys are the same person
and i'm winning so hard it hurts

>> No.27029420

"Clang produces faster binaries than GCC!"
> posts benchmarks
> GCC absolutely rapes Clang
"Uhhh, the benchmarks are laughable and biased, Stallman is paying the authors of the benchmarks to twist the results in the favor of GCC."
> post some other benchmarks and post the source to your claims
"..."

Come on Apple fans, we're waiting.

>> No.27029425

Thread is over.
Has evolved into shitposting.

>> No.27029432

>>27029413
> The results of their benchmarks (GCC produces faster binaries in a majority of cases) are accurate, but I have *VERY* little faith in their methodology.
The methodology is to compile the code then run it.

If you don't want to face the truth (clang is a toy compiler) because it hurts your feelings then fine, but it doesn't make the truth any less true.

>> No.27029435

>>27029372
>GIVE ME MY TI9900, CONFIRMED FOR SHIT TIER COMPILER
Nobody cares

>> No.27029438

>>27029425
>Has evolved into shitposting
saging an autosage thread

>> No.27029472

>>27029420
>Come on Apple fans, we're waiting
all the shit benchmarks were in favor of gcc.

baka.

>> No.27029478
File: 490 KB, 449x401, Girls.png [View same] [iqdb] [saucenao] [google] [report]
27029478

>>27029432
>He thinks that benchmarking is "compiling and running"
Again -- clang devs know and don't care about tossing every optimization added to GCC since the beginning of time into the backend -- that's just not where the interesting work is happening right now. It's great if GCC wants to keep doing that, it's just not what the Clang devs are most interested in.

>> No.27029487

>>27029435
but my ti-89 got me through calc2

>> No.27029497

>>27029420
Anyone who says that clang produces faster binaries in general is an idiot. Phoronix remains a joke, but they've got the general trend correct (GCC produces faster binaries).

If someone disputes this, they're a clang fanboy.
If someone thinks this is the end all be proof that GCC is superior, they're a GCC fanboy.

Clear enough?

>> No.27029505

>>27029497
>got the general trend
benchmarks or GTFO

>> No.27029513

>>27029418
>shitposts
>LOL IM WINNING IM FINALLY COOL
... I'm happy for you?

>> No.27029533

>>27029497
Can you read? Or are you a clang fanboy? Not even the clang devs claim they're producing faster binaries in anything except a few special cases.

God I don't know why I bother replying to these threads anymore.

>> No.27029535

>>27029513
just look at the rage and self defecation everywhere, it's glorious

>> No.27029549
File: 40 KB, 349x642, hey-guys-look-how-retarded-i-am-jokes-on-them.jpg [View same] [iqdb] [saucenao] [google] [report]
27029549

>>27029535

>> No.27029568

Was OP's original claim (that gcc is crippled by design) disproved somewhere in the thread?

>> No.27029584

>In this thread
3 in this threads
clang devs care more about important shit than optimizations
gcc fanboys butthurt
no reliable benchmarks
>shitposting
some idiot pretending to have magnificent bastarded the thread
2 turtledoves
and a partrige in a pear tree

>> No.27029595

>>27029568
It's true to an extent. There's a really classy RMS email where he goes on about how modularizing the components of GCC would hurt the freedom of it, because it would make it easier for companies to hurt his freedoms.

Things have gotten better though, GCC isn't as scary as it used to be. But in terms of ease of extending, clang is a pretty clear winner.

>> No.27029606

>>27029568
No, you should read the whole thread.

>> No.27029645

>>27029595
>RMS
>classy
pick 1

>> No.27029692

and in a surprise knockout, the young and thin clang has overtaken the slow and bloated gcc in /g/'s license/compiler shitpost fest 2012

>> No.27029712
File: 85 KB, 544x400, 1305668923619.jpg [View same] [iqdb] [saucenao] [google] [report]
27029712

>>27029692
summer

>> No.27030444

bump

>>
Name (leave empty)
Comment (leave empty)
Name
E-mail
Subject
Comment
Password [?]Password used for file deletion.
Captcha
Action