/dpt/ - Daily Programming Thread

What are you working on, Sup Forums?

Old thread:

Other urls found in this thread:

blogs.msdn.microsoft.com/oldnewthing/20140411-00/?p=1273
en.wikipedia.org/wiki/Pooh-pooh
twitter.com/NSFWRedditVideo

Thread reminder your arguments should be based on solid logic built from reasonable premises.

Reminder that C++11 is more portable than C99.

>non-anime image
hi. please fix next thread.

>muh arguments, their so stronk!

Working on a project in Rust.

Reminder that 2+2=5

whatcha making lad?

my condolences

The best way to ensure your arguments are based on solid logic and your premises are reasonable is to argue in favor of Lisp and never argue against it.

Molymeme is a fag

Reminder that screeching your subjective opinions and calling it "an argument"/"a critique" doesn't convince anyone besides members of your own delusional cult.

fantastic post

I forget what you posted, and I'm too lazy to look, so I'm posting to be able to hover and see dat shit, so I can post a smug or helpful reply depending on what your post was

iow I post so I can post

>smug or helpful reply
crap, it's option #3: genuine question!
4chanx or summin?

>struct initialization
Compiler level stuff that's nearly trivial to implement. user was talking about "system libraries"

>VLAs
Again, not complicated at all to implement compared to ClusterPlus

GTK+ apps work on winDOS and they are written in C. I have no idea what they link against and how, but since they work there must be a way already. Checkmate.

Nothing wrong with raising the bar.

Is it theoretically possible to recursively inspect the #include "foo.h" statements starting from the main to figure out which files should be included in a C++ build? If so, why the fuck do we still have to list every file in the project in our CMakeLists.txt file like fucking cavemen?

>Mathematician, how much 2 + 2 equals to?
>*two full blackboards of proofs later* Central Theorem of Arithmetics: 2 + 2 = 4

>Physicist, how much 2 + 2 equals to?
>*weeks of experiments later* According to these experiments, 2 + 2 = 4 with a 0.005 absolute margin of error

>Computer scientist, how much 2 + 2 equals to?
>*after running several simulations* Well, according to the computer, 2 + 2 = 4.00001

>Accountant, how much 2 + 2 equals to?
>How much was it last year?

>Statistician, how much 2 + 2 equals to?
>How much would you like it to be?

If you use anything other than C89, then you are a fat cunt.

Coq agrees
Inductive nat : Type :=
| Zero
| Succ : nat -> nat.

Fixpoint plus (x : nat) (y : nat) : nat :=
match y with
| Zero => x
| Succ y => plus (Succ x) y
end.

Notation "0" := (Zero).
Notation "1" := (Succ 0).
Notation "2" := (Succ 1).
Notation "3" := (Succ 2).
Notation "5" := (Succ 3).
Notation "x + y" := (plus x y).

Check eq_refl : 5 = 2 + 2.

It's always fantastic when you point out the hypocrisy in your adversary, comrade.

But those are the best kinds of cunt. Not as good as slick cunnies, though.

Reminder that screeching their subjective opinions is how normal people interact, discuss issues, convince others and reach the agreement. Requiring that every argument should be a solid logical argument doesn't change anything, just moves the object of screeching from the argument itself to prepositions.

that would only work if every foo.h also has a corresponding foo.cpp. That would be a sensible assumption, but C++ devs don't accept it because of the rare one in a billion case where you might have to have multiple source files for one header/multiple headers for one source file and have it not be awful programming practice.

>Sure, you could write a 3rd party library that implements the missing stuff, but it's kind of telling that nobody has bothered, at least not in a free/open-source way.
How do they even build gimp then? Somebody bothered afaict. And even if msvcrt were to disappear it would not be very complicated to call C++ ABI functions under the hood, no?

If you use anything other than m68k assembly you're basically a retard

>The best way to ensure your arguments are based on solid logic and your premises are reasonable is to argue in favor of Lisp and never argue against it.
She says with a passive-aggressive tone, after openly admitting that she considers logical arguments unimportant, because what matters is that things "feel" true enough.

Only a Sith deals in absolutes...

Don't forget to use #![feature(nll))] .

>being this buttmad over a lisplet

By limiting themselves to C89 and the (extremely small) subset of supported C99/C11 library features. Jesus, user, this isn't rocket science.

>molymeme
>raising the bar

How come Microsoft hasn't updated it? How come nobody has made an open source version?
Is C11 literally that irrelevant? I would have thought that Microsoft would at least get an intern to throw something together.

appchan x, but i think it's been unmaintained for a while.

>Reminder that screeching their subjective opinions is how normal people interact
That's how normal people interact when discussing their favorite soup, and how "normal" low-IQ people interact when discussing the suitability of a tool for some practical purpose. Normal intelligent people can do better, but you wouldn't know.

>just moves the object of screeching from the argument itself to prepositions.
Only with retards like you whose prepositions have nothing to do with any practical considerations, and basically amount to "a good language has to have dependent types" or some other autism.

I'm writing a Reddit CLI client.
The goal is to make finding a post and copying the link to the clipboard as efficient as possible.
Rust is a pretty good language. Not perfect, but it fits all of my needs.
Is NLL already in stable?

Anything is raising the bar next to gook trash.

leave

Why? This is a programming thread on a technology board.

>That's how normal people interact when discussing their favorite soup, and how "normal" low-IQ people interact when discussing the suitability of a tool for some practical purpose.
Please stop screeching your subjective opinions.

wow, argumentz r stupid. there comrade i just crushed those right wingers amarite xddd

the absolute state of types

> "normal" low-IQ people
Wew, we have intellectual elite here, I wouldn't expect anything less from a lisper.
> prepositions have nothing to do with any practical considerations
What, do you have logical agument for this statement?
>"a good language has to have dependent types" or some other autism.
Is you calling this preposition "autism" a logical argument or you screeching your subjective opinion? Do you have logical arguments for all the propositions you hold/reject? Otherwise, why do you hold/reject them?

You're right actually. GlibC doesn't support WinDOS as a platform according to the docs. Weird that no one is guarding against that tho. Have CLang fags done anything?

>Is NLL already in stable?
No, just use nightly, it only breaks once a month or so.

That's not an opinion, user. It's just an observation. I never said there's anything wrong with being unintelligent, user. I think your existence is essential, because someone has to sweep the streets, clear toilets etc.

forgot cute pic

>How come Microsoft hasn't updated it?

According to Microsoft, they made the decision decades ago to focus on C++ and focusing on one thing means neglecting others. Why make a bunch of mediocre libraries and tooling for multiple languages when you can just support one low level language really well?

And there are private libraries. Intel offers one. It's expensive. As for open-source ones, who knows. It's boring and a lot of work and most C developers are happy in Unix land anyways. There's probably even less interest now since you can just run your C programs on the Windows linux subsystem in Win10.

It's true though. Just look at the evidence presented here.
If you want to write portable code, you write C89. If you want to write slightly less portable code in exchange for "nice" features, you write C++. If you want to jerk off and write code that will never be executed outside of your basement, you write C99, C11 or some nonstandard GNU dialect.

>claim to strive for "logical, factual arguments"
>make fallacious arguments instead
>accuse your opponent of being against "arguments, facts and logic!" when you get called out
Every time.

Threadly remember that this thread is only circlejerking and should be forbidden

nah, nigga's just wrong
read this, ya dumb fuck
blogs.msdn.microsoft.com/oldnewthing/20140411-00/?p=1273
(great blog btw, especially for baka moment generation)
>There's a DLL in the system directory called MSVCRT.DLL, and from its name, you might think that it is the Microsoft Visual C/C++ Run-Time library. That is a perfectly reasonable guess.
>But it would also be wrong.
it's (mostly) not microsoft's fault that windows devs/packagers are pants on hat retarded
OTOH it's their fault for letting the wrong thing be the easiest one, and letting it continue to work, but that's one con for all those sweet backward compatibility pros

>If you want to write portable code, you write Java

ftfy

Is the RLS still broken in like 30% of nightly builds?

But Guy Steele separated Church from state

>Reddit CLI client.
>The goal is to make finding a post and copying the link to the clipboard as efficient as possible.
nice!
post a link or a name we can google later
would be something I'd be interested in seeing the code for

I have no idea, I use racer with spacemacs, autocomplete, flymake and go-to-definition work fine.

>he thinks a series of conclusions are observations
You're really not too swift, are you?

Shame, I'm really looking forward to it.
I prefer to write my code for stable Rust and let Travis CI test beta and nightly. I had some problems with nightly in the past. I do use nightly, but only for rustfmt and clippy.

...

>If you want to write unoptimized, nonfunctional code, you write Java

ftfy

No idea what you're talking about, m8. msvcrt.dll is the C runtime. At one point in time they started versioning it because reasons which is probably what that article is referring to, but then they stopped because of changes to the way installations worked.

Mingw links C projects against msvcrt.dll. That's not hard to fact check.

>next thread
>doesn't make a new one but begs for another
fool

Disregard this thread, here's a real one:

>What, do you have logical agument for this statement?
Depends on how you took it.

>Is you calling this preposition "autism" a logical argument
Nope.

>Do you have logical arguments for all the propositions you hold/reject?
As typical for low-IQ people, you keep misusing "preposition"/"proposition" and I have no idea what you're trying to say. Are you talking about premises? If I told you "if you don't want to die, you should drink, because dehydration leads to death", would you ask me for the logic behind the premise that you dehydration can kill you, or the premise that you MAY prefer to not die from dehydration?

Please collect your thoughts and write them down next time.

>muh artisinal hand-crafted assembly got BTFO by optimizing compilers and JITs with million man hours in them, but I'm continuing to live in the 80s

(Checked)
I will eventually. The top priority is to deliver the program with the bare minimum of features. Otherwise, I'll lose interest in the project and work on something else.

...

>such and such people interact in such and such way
>that's a conclusion, not an observation
Simply subhuman.

Is it 2000 again?

>looking for interns

Is this a red flag?

en.wikipedia.org/wiki/Pooh-pooh

>Mingw links C projects against msvcrt.dll. That's not hard to fact check.
no argument, I had the mispleasure to experience it when I was forced to work in that frankenstein of an environment
pointing to mingw as an example of "proper" development is pants on hat retarded
niggas just be trying to make shit work, that's probably why they're doing the wrong thing

seriously, read Chen's blog, nigga knows all about dumbfuck devs abusing implementation-specific shit in Windows

>implying a shitpost deserves anything more than a shitpost

>top priority is to deliver the program with the bare minimum of features. Otherwise, I'll lose interest in the project and work on something else.
good priorities indeed
show us dat code tho, luv
wanna have more points of reference for when I reinvent that particular wheel in a more appropriate (for the task) language

Let me rephrase: why do you outright reject the premise "a good language has to have dependent types"? Do you have a solid logical argument against it? If so, on what premises is it based? If not, and you reject it just because you feel so, what's the point of building solid logic arguments on top of subjective premises?

>btfo by compiler

Only people who have never written assembly say stuff like this

>cnile doubles down on his delusions, news at 11
shocking

I'm sorry to hear you can't even beat a compiler. Maybe programming is not for you?

i wrote an unbounded spsc ringbuffer for soft real time whose rw operations only wait on lock when it expands

>why do you outright reject the premise "a good language has to have dependent types"?
Because it's arbitrary and doesn't appeal to anything besides one's presumed appreciation of dependent types.

>Do you have a solid logical argument against it?
What makes you think I need to "prove" something to dismiss it?

>what's the point of building solid logic arguments on top of subjective premises?
Not all subjective premises are equal.

>unbounded spsc ringbuffer

So a linked list? Or actual ringbuffer with mutex for resizing?

>swapping the value of their labour for money
>exploitation

It's either that or die in a ditch. Global poverty is decreasing as a direct result of capitalism.

Why not lock-free? Lock-free spsc ringbuffer should be a relatively easy task.

...

>wanna have more points of reference for when I reinvent that particular wheel in a more appropriate (for the task) language
You can always check out other projects in the meantime. That's how I got error handling working correctly.

he said it only locks when it resizes so supposedly in normal mode it is waitfree

dont really understand why you would want to resize a ringbuffer in the first place

oh no, a dpt shitposter told me I succ, he must be right! how will I ever recover?
get real, kid, there's a time and place for hand-written assembly, and in real-world non-autistic non-toy projects, that is "almost never, almost nowhere"

...

oh yeah, that's a given
just my lazy data/reference hoarder instincts firing

>Because it's arbitrary and doesn't appeal to anything besides one's presumed appreciation of dependent types.
So you reject it because you feel it's "arbitrary" and "doesn't appeal" to you? Nice solid logic here, m8.
>What makes you think I need to "prove" something to dismiss it?
Because you do, otherwise it's just a bunch of subjective ideas you subjectively liked enough to assume as premises.
>Not all subjective premises are equal.
I see, yours are better than mine because of reasons, and you can actually base your logical arguments on your subjective premises.

But to further our discussion, can you post a single premise you actually held in regards to programming languages? Like "a good language has to have homoiconicity" or something.

Dynamic linking is a mistake.

Dynamic typing is a mistake.

>having more than a snowball's chance in hell to fix critical vulns at the distro level is a mistake
absolute state of dpt posters

it's a dynamic array storage with a shared_mutex (aka "rw"lock). the shared lock is for reads, and writes within the current capacity, and exclusive lock to write with array growth.
it's unbounded

>dependency & security hell is better than having distro maintainers smart enough to update their shit

Critical vulns? Are you using dynamic typing again, user?

Dynamic programming is a mistake.

>I live in an utopia where all software can be rebuilt
wow, that sounds really neat, op!