/tech/ - Technology★

gcc

gcc

clang is only good if your boss is requiring you to cuck the developers.

I'm quite impressed with the turnaround gcc made. I thought the whole thing was doomed back during the egcs/pgcc days. I guess technically it actually was doomed as the egcs guys took over.

>>817480 (OP)

holyc

Don't be a nigger cattle, holyc is the only compiler that you ever need.

The best compiler will be the compiler which is the most capable. The dominant compiler will be the one that is most dominant.

compcert

gcc/clang will remain

>>817491

>compcert

<The performance of its generated code is often close to that of GCC (version 3) at optimization level -O1, and always better than that of GCC without optimizations

Why does this even exist.

>>817493

Did you even read the rest?

>>817495

I read it's formally verified by someone's cock to produce embarrassingly shitty output slower than a compiler from before many anons were born.

>>817498

>clang is open source

and gcc isn't?

>>817505

Are you literally retarded? gcc is Free Software.

Open Sores != Free Software

https://www.gnu.org/philosophy/open-source-misses-the-point.html

>>817497

I don't think you understand the number of P1/P2 bugs gcc has every release. Being able to produce good code and being sure of it is extra important.

>>817511

Clang and GCC are both open source as well as free software.

The page you linked says "The two terms describe almost the same category of software".

You're missing the point of the way open source misses the point. Educate yourself, for example by carefully reading the page you linked, because you're not providing any arguments against clang.

If you want a gnu.org page to link to to protest clang, try this one:

https://www.gnu.org/philosophy/pragmatic.html

Free software and copyleft are different things.

>best

(((intel))) compiler because it has the best (((x86))) optimization

>dominant

clang. Companies love cuck licensed open sores software and are afraid of free software.

>>817505

gcc is not open source

if I use even a small fragment of it in my private project I will be forced to release all my code on the same communism infected license

>>817519

Kill yourself, retard.

>Conclusion

>As the advocates of open source draw new users into our community, we free software activists must shoulder the task of bringing the issue of freedom to their attention. We have to say, “It's free software and it gives you freedom!”---more and louder than ever. Every time you say “free software” rather than “open source,” you help our cause.

>>817522 (checked)

not an argument

>>817532

>I will be forced to release all my code

Good. Stop being a CIAnigger.

Also, if this is shit that only you will be using, privately, you don't have to release the source.

>>817539

anyone who knows can force me to release it, so i can't tell anyone that i am using it and disconnect internet just in case

>>817531

Do any companies develop in-house additions to clang btw? If the snowflake license on it allows.

>>817532

>>817541

>anyone who knows can force me to release it

Only if they get a copy of your software. As long as you don't share the software itself nobody can demand anything from you.

The Open Watcom license does require you to publish your source code even for people who don't already have a copy of the software, which is why the FSF considers the Open Watcom license proprietary.

>>817544

I am still forced to do something instead of posibility to do it freely

>>817545

And as a producer with a goal to make money, I don't know why one would ever consider GNU. It also prohibits you from selling to any of the various walled gardens, which people clearly enjoy. (Apple, Google stores etc.)

>>817545

>why can't I enslave people and be enslaved?

Will they ever learn?

**Permissive = anarkiddies (goal: personal freedom)

Copyleft = fascism/natsoc (goal: group's freedom)**

>>817561

Good idea. We need race and ideology specific licences.

>>817566

they already exist https://www.rust-lang.org/en-US/conduct.html

>>817542

no, and even if they did they could do inhouse gcc extensions too without having to release them. companies are scared of the evil "viral" copyleft boogeyman.

>817480

>compilers

>Which compiler is the best?

gcc

>Which compiler will be the dominant one in the near future?

cucklang

Depends. Clang has some advantages (compiles faster, is much easier to extend, license). But so does gcc (works on anything, license). Which produces better code depends, I've seen huge differences between the two for both sides (usually, but not always, icc > gcc > clang). Very long-term clang has the advantage since their code is less fucked because it's more recent.

>>817532

>if I use even a small fragment of it in my private project I will be forced to release all my code

You only need to release the code to your users. Since the project is private you have few users and can add in the contract they can't develop a competing software.

>>817497

You don't always need maximum speed. Often you only need "does x before y seconds pass". Take the guidance software for a rocket for instance. You can afford wasting a few more watts to avoid some crazy bug.

>>817602

If I require reliability I'd already be solving the problem with three different builds of the program where the output is compared so I could detect and recover from program errors, tool errors, random errors, hardware faults, physical damage (combat aircraft), cosmic rays, etc.. A slow as shit "formally proved" language is of little value since I'm not going to be putting all my eggs in one basket.

>>817480 (OP)

Compiler for what?

>>817622

Webasm.

>>817622

of programming language

>>817511

What is this?

https://gcc.gnu.org/viewcvs

>>817498

>optimizes better than gcc

Except it doesn't, you lying fuck. The Clang fuckers kept crying about its lower compile times, and now it can't do that either.

>>817561

>portraying being a nazi faggot as anything to do with free software

Kindly hang yourself.

>>817567

>body size

>>817700

>inb4 phoronix 'benchmarks'

>>817480 (OP)

>Which compiler is the best?

The one you write yourself, nigger