/dpt/ - Daily Programming Thread

you guys worry about everybody being replaced by robots and advanced AI?

FIRST FOR TRAPS!

>>59078331
No, because according to the 60s we should already have been replaced by robots and live a happy and stress-free life thanks to robots doing anything for us.
50 years later and it still hasn't happened.

Ban C programming.


C is the reason why we still have shitty backdoors.
C is the reason why we cannot utilize multiple cores efficiently.
C is the reason why our programs continue to reveal security exploits after decades.
C holds back technology advancement to the 80's .
C is the reasons why Linux has no concurrency

File: CRUST.jpg (43KB, 366x281px) Image search: [Google]

43KB, 366x281px

itt C and RUST merge into CRUST the ultimate language to power the WORLD

File: lisp look out.png (32KB, 502x362px) Image search: [Google]

32KB, 502x362px

TemplateHaskell is ALMOST good

>>59078352
You can write C within Rust

>>59078361
i can also eat with a shovel

>>59078353
What does TemplateHaskell do?

>>59078369
But you were asking for a shovel

@59078349
>our programs
m8...

>>59078375
Quoting and splicing, for instance in >>59078353
I generate a list of 10 type declarations

data Test0
data Test1
..
data Test 9

and splice them in

>>59078378
i though we just exchange facts..

@59078349
Your fervent anti-C posting just makes me want to use it even more.

>>59078392
>>>/t/witter

#include <stdbool.h>

>>59078389
So after that I can use e.g. Test8 as normal type? That neat, but what are the real use cases?

>>59078422
Yes

>>59078424
one example is several libraries provide a function like

(Lens library)
$(mkLenses ''MyType)
(''Type quotes the name of Type)

This passes the type to the mkLenses function, which then generates functions and intsances based on its components (more powerful than GHC.Generics).

Unfortunately there are a lot of limitations, and the AST itself is quite bloated.

>>59078353
Ever since I tried Terra* I feel like the world needs more levels of metaprogramming for their types (at the cost of probably needing to ship LLVM to run the damn programs by the end!!!)

* http://terralang.org/
skip the first bit go to generative programming and enjoy the ride

>>59078422
why tho?

>>59078447
A lot of the time metaprogramming isn't needed, it's just a way to surpass limitations of the language.

>>59078347
but it is happening now for real

web 3.0 and advanced AI are making most people pretty much useless

>>59078534
>But some stupid boulevard papers and sites who only care for clickbait headings say so!

>>59078534
>advanced AI
The state of AI is nowhere near even slightly advanced. It's the same weak shit as always, and in almost any serious context numerical and statistical methods will kick the shit out of your bespoke RNN which falls into a worthless local maxima if you so much as include one slightly less than stellar piece of training data I am SICK of CIA Niggers pushing these dirty memes into our children

File: anim.png (163KB, 1131x835px) Image search: [Google]

163KB, 1131x835px

i added a neat tracklist view today

great start to a thursday

>>59078451
Why wouldn't you?
Define true and false yourself is stupid, and nobody wants to type _Bool.

>>59078715
>Define
Defining*

>>59078607
>The state of AI is nowhere near even slightly advanced.
already advanced enough to take over logistics

> CIA Niggers
why would they be interested in causing existential panic in society? you hate them just to hate, you can't even logically articulate it anymore you brainwashed libshit

>>59078715
just use 0 and 1

Is it possible to become a Hacker without python?
Also, what's a good language for triple A games like Call of Duty and stuff?

>>59078841
yes
javascript

>>59078841

Yes
C++

>>59078852
>>59078868
What do I have to learn for my first question then? I don't like Python and other similar stuff, I've tried it.
>javascript
Isn't that for scripting on the web and stuff? Or can it also be used for cool games?
>C++
Are most books on it fine?

>>59078765
>already advanced enough to take over logistics
You're saying that calculating a shortest path is something new? Or were you referring to moving packages though shelves/lanes that are standardized in a few shapes and weights in order for the robots to not crash and burn?

>>59078841
The CoD servers are actually written in Erlang. You should learn that.

>>59078922
autonomously driving trucks and ships on that shortest path is new and will fuck a lot of people

but that's just the beginning

>>59078891

>What do I have to learn for my first question then?
C, then JavaScript, and a reasonable scripting language. Then learn you some network protocols, computer architecture, and basics into software security. All the while, you should be programming for the fun of it. You don't have to create anything interesting, but you should at least be a tinkerer. Fuck with things. If nothing else, hacking is just making something do something it wasn't designed to do, for better or worse, and that's pretty fun. Take shit apart, fuck around.

>Isn't that for scripting on the web and stuff? Or can it also be used for cool games?
It's a general purpose programming language. You can do whatever you want with it. It's primarily used for browser side scripting and some web servers, however. These days, it's like a second franca lingua of programming, so you should know it in addition to C, the original franca lingua.

>Are most books on it fine?
If it's written by Bjarne himself, it's probably fine. I can make no judgments about other books.

>>59078965
None of that is done by AI though, they can only offer assistance (keeping the lane you're in, but only if on a highway, etc.) but are completely useless for replacing humans.

File: -5.jpg (9KB, 480x360px) Image search: [Google]

9KB, 480x360px

>>59078314
>*mouth frothing*
FIRST FOR... RUST!

RUST RUST RUST RUST RUST RUST RUST URS RUST RUST RUST RUST URS T UR SUT RUST URS RUST URST RUST RUST RUST RUST RUST RUST RUST RUST RUT RUST RUST RUST RUST RUST RUST RUST RUST RUST RUST RUST RUST RUST RUST RUST RUST RUST RURS TUR STUR STUR STUR RUST RUST RUST RUST RUST

>>59079036
You okay, anon?

>>59078988
it is all AI and it is replacing everything

>>59079069
You seem like one of those "business types" who always going on about "the cloud", "AI", and all of that sort of bullshit without actually understanding what any of it actually is.

>>59079077
no one knows what it is, it just works and getting more powerful

http://www.wired.co.uk/gallery/machine-learning-graphcore-pictures-inside-ai

>>59079102
>Reddit fomatting
>Goes on about AI without even knowing anything about it
You're an idiot, and need to go back.

>>59079117
>ignorant of recent progress made in AI field
>resorts to reddit-shaming on a blue board

they probably took your job already

>>59079117
What are you going to do about it?

>>59079169
fuck off back to your shitty "subreddit"

File: 1486388396743.jpg (45KB, 460x698px) Image search: [Google]

45KB, 460x698px

designing fault tolerant protocols is a fucking pain

>>59079102
>no one knows what it is, it just works and is getting more powerful
that's some mystic occultism right there

>>59079102
learn English before posting here, you fucking plebbitor

>>59079225
that's what advanced AI is like, you don't program it, it trains without your input

all the fast food workers like this fag are totally fucked >>59079239 >>59079217 >>59079117

redpill me on rust

>>59079292
You can't let it get too good, or it will replace humans

>>59079298
Don't fall into the trap. Avoid it like the fucking plague.

C and C++ are NEVER going to fucking die. Anyone who says otherwise is a brianlet and/or a webdev

>>59079314
I don't really see much keeping C++ alive.

Why does /g/ hate rust?

>>59079322
>what are video games?

>>59079314
>C and C++ are NEVER going to fucking die

Are you living in the 90's? They have been replaced by other languages. The remaining stubborn 5% will soon die out.

>>59079322
If there's no demand for C++ why the hell would anyone use Rust? Because it forces you to use smart pointers?

If you could replace C++ by being "C++ but not quite so ugly and with some options removed" it would have happened decades ago because everyone and their mother have already tried.

>>59079345
>They have been replaced by other languages
C is nowhere near being replaced.
People have been trying to displace C from its niche for decades, and nobody has even gotten close.

>>59079345
Does your name happen to Pajeet or some derivative thereof?

>>59079351
>not quite so ugly
That goes a long way. Developing in C++ objectively costs too much.

C is one of the best languages to learn if you want to become more than a CRUD developer. It really helps your understanding of computing as a whole.

>>59079385
Granted, but again, it's the first thing anyone ever thinks of and it's been tried a thousand times already. If it was gonna happen surely it would have happened already.

>>59079411
>>59079411
>If it was gonna happen surely it would have happened already.

While you've been in a coma Java took over almost entirely.

>>59078314
Working on a BeautifulSoup website parser and I'm stuck.

Basically if I have a HTML DOM like this:

<div class="container">
<span class="1">fizz</span>
<span class="2">buzz</span>
</div>

which repeats itself mulitple times and I want to print the content of each individual span, grouped by container, how would I do that?

>>59079452
Java isn't "C++, but..." Java won the industry over with its somewhat-exaggerated claim of "write once, run anywhere."

How do I make a 3d game? Ive got unity, I know c++ but I havent worked with any graphics libraries. Or is unity a bad idea?

>>59079411
It has. Still, the two languages have only been replaced in "mainstream" applications but are still going strong in their respective niches. And that *is* unlikely to change.

>>59079478
Doesn't Unity work with the graphics library for you? If it doesn't, what's the point?

I recommend starting with a Unity tutorial.

>someone figured out how to inject .so into other processes on linux
what a time to be alive, time to port some cheats

>>59079298
> https://www.rust-lang.org/en-US/community.html

How can I make this output 5 instead of Friday?

def day_num(d):
    days = ['Sunday', 'Monday', 'Tuesday', 'Wednesday',
    'Thursday', 'Friday', 'Saturday']
    return days[d]

print(day_num("Friday"))

What book is that

>>59079590
print(5)

>>59079590

len(days)

>>59079590
days.index("Friday")

>>59079545
>We are committed to providing a friendly, safe and welcoming environment for all, regardless of gender, sexual orientation, disability, ethnicity, religion, or similar personal characteristic.
What did they mean by this?
I mean, are they implying that you can't use some other languages if you're gay or something?
I'm confused.

File: mkv.webm (3MB, 1280x720px) Image search: [Google]

3MB, 1280x720px

I'm implementing the file detection part of a thumbnailing library and am having troubles distinguishing WebM and MKV. So WebM is basically a subset of MKV, any WebM is also an MKV and some MKV can be interpreted as WebM?

<- is actually an MKV

File: BlackMagic.png (50KB, 727x435px) Image search: [Google]

50KB, 727x435px

>>59078314
Working on some c++ black magic...

>>59079759
>2d game using some framework
>black magic
You sure you posted a right screenshot?

>>59078314
>>59078314
Just got through fixing 3000 lines of SQL procedures.

Now I have to use an internal software development tool that runs on Silverlight, and only works on Internet Explorer, to create an application using drag and drop.

...

Such is life on the Microsoft Stack.

>>59079733
WebM uses the exact same container format as MKV. WebM is essentially a subset of MKV that only allows certain video/audio codecs (intended for Web Streaming). I think it might impose other limitations, but either way it's a subset. So if you have and MKV file that contains a video encoded in the correct format (VP8/VP9), then it is a WebM.

>>59079759
That's way too many conditional branches to be honest famalam. I thought C++ was invented to prevent this.

File: 148590453.jpg (48KB, 1280x720px) Image search: [Google]

48KB, 1280x720px

holy fucking shit why are strings so hard in C?

I have been trying to extract links from html files and downloading them holy fuck the amount of shit I'm struggling with

strings are fucking painful

On that topic, what languages are good for this kind of stuff

>>59079810
Getting links from HTML files?

Easiest path would be to just use Python for this, since it has a library that does what you want to do already:

import httplib2
from BeautifulSoup import BeautifulSoup, SoupStrainer

http = httplib2.Http()
status, response = http.request('http://www.nytimes.com')

for link in BeautifulSoup(response, parseOnlyThese=SoupStrainer('a')):
    if link.has_attr('href'):
        print link['href']

Obviously, you can swap the hardcoded url with args[0] instead if you want.

These types of quick small scripts are where Python shines too bro.

File: itrue.png (3KB, 273x106px) Image search: [Google]

3KB, 273x106px

>>59079808
You are absolutly right :)... But i got 3 more hours to fix shitty code that has been 'taped with duck tape' for 5 years :| . The whole fucking engine is a piece of black magic. Here is one more masterpiece form same engine :D

>>59079810
Any language made after 1990. Or abandon the built-in strings and google a good string library for C.

Perl shines at parsing strings, but damn is it ugly as sin.

File: 14668678574668.jpg (45KB, 1280x714px) Image search: [Google]

45KB, 1280x714px

>>59079845
Second question:

Why do people mix languages for example c and python for software?

Aren't you throttling yourself by the slowest language?
Why not just use (for example) python for the whole thing

Edit source in C to make program utilize CPU (Windows, and Linux) resources up to a specific threshold, which is determined by total CPU load (like I want to use a quarter of free resources).

Help, what to read? Like, fundamental/essential reads besides K&R.

>>59079458
rtfm

Sup /g/, I want to read a stream of JSON from a socket in C#, encoded as utf8. How would you go about this? currently it's just a text stream, and I can insert delimiters between them on the sending side.

The JSON data will be on the form

{
  "electrodes": [1, 3, 5, 6],
  "stimFreq": [100.4, 145.6, 330.8, 54.6]
}
//whatever delimiter I want
{
  "electrodes": [1, 3, 5, 6],
  "stimFreq": [104.3, 135.6, 333.8, 78.5]
}

>>59079885
C[++] DLLs are often used for heavy-duty complicated tasks because it'll run them fast even if the GUI wrapper that you made in Python is inefficient. The Python code is "bottle-necking" only insomuch as the code is running in Python. Being wrapped in Python won't make the C code run any slower, it'll just take slower to get there.

>>59079885
faster development time
easier to find python programmers

>>59079885

1% of the code does 99% of the work, so it makes sense to only use C when you really need it.

>>59079885
Because the 0.1 seconds slower that small Python script is going to run doesn't matter.

What's going to take time is to get the site downloaded. The somewhat slower speed of the task isn't that relevant.

I mean, consider some inane task:
> FIND ALL THE LINKS ON OUR DOMAIN!
Yeah, you could write that software from scratch, or you could just use wget to grab all the pages, and feed them to a python script (like the one I linked) to get all the links, stuff them in a text file, do the sort | uniq -count trick, and you'd have a text file with all the links on our domain reachable from the root, with the number of times they popped up too.

Time to write that: 5 minutes.
Time to run it: 1-15 minutes depending on how big the site is, how fast the network is, etc.

Now, imagine sitting down and writing all that shit from scratch so that it ran slightly faster.

Would anyone really care that it runs twice as fast when you spent weeks building it? No.

>>59079926
Read the char-stream, tokenize until you have a complete JSON-object, and pass the whole object on down the pipeline?

>>59079298
It'll save the programming world... if anybody used it.

>>59079810
>why are simple arrays so hard
anon...

>>59079918
some epic hacker website

>>59079958

That's what I want to do, but I was hoping there was a convenient set of tools to do it so that I won't have to. like

Stream JSONStream = SomeTool.JSONFromSocket(socket, ... );

>>59079991
I don't know C#'s library ecosystem.
I know that you could use Jackson to do it in Java (you could even explicitly parse token by token and handle it from there if you WANTED to), but I won't be of much help here.

>>59079991
What language? Any modern one should have some type of JSON parser and it shouldn't be difficult to tweak it to read multiple Tokens

>>59079981
kys

>>59080053
Oh, sorry, I see you said C#.
http://www.newtonsoft.com/json/help/html/ReadMultipleContentWithJsonReader.htm

https://mitpress.mit.edu/sicp/full-text/book/book-Z-H-21.html#%_sec_3.2

How is this entire section not overcomplicated nonsense? The stack explanation could be done in 1 minute and didn't need all this.

>>59079339
>>59078891
>>59078841

C# is great for video games. I worked in unity 3d last winter break. Shit was so cash.

Can we make a list of extinct species? I'll start:
Neanderthal
Mammoth
Passenger Pigeon
Employed Haskell Programmer

>>59080072
Looks like what I was after, thanks

>>59080121
Funny posts on 4chan/g/

File: 1477912558539.png (325KB, 480x716px) Image search: [Google]

325KB, 480x716px

>>59080121
c programmers

>>59080124
>>>plebbit
>>>lamechan
>9fag
shoo shoo pleb

If you want to write a library that could be used from other languages what options do you really have?
C seems best but C as language isn't the nicest. You could go with C++ and extern "c" but sepples is cluster fuck that nobody wants to touch.
Some langs with gc can do it but it would require you to initialize the languages system/gc for it to work, I assume?

>>59080158
>Some langs with gc can do it but it would require you to initialize the languages system/gc for it to work, I assume?
Yes. Most languages are awful for interoperation, which is why people use C for that.

>>59080158
Really depends on what the purpose of your library is. If it's not perf critical you can make some epic webscale json api and not even bother with FFI

>>59080121
Haskell is just a Lisp in disguise, anyway.

File: dunce.png (64KB, 658x901px) Image search: [Google]

64KB, 658x901px

>>59080188
No

>>59080188
No, it's fundamentally different from Lisp in several ways.

It's approach to typing, restricting the programmer, multiparadigm vs PURITY FIRST, etc.

It's not at all the same.

>>59080182
Say you wanted to abstact opengl away a little or something like that. Heavy functions that you will call a lot.
Having another process for executing and using sockets for messaging is option but let's assume that the overhead from it is too much.

>>59080158
>what options do you really have?
just C

>>59080237
In that case I don't have anything useful to add, but I agree the socket thing would be a bad fit.

>>59080234
All easy to implement with macros.

File: 1470452005283.jpg (68KB, 500x498px) Image search: [Google]

68KB, 500x498px

Racket vs Haskell vs Scala
is like
C vs C++ vs Jewva

File: 1485273967694.jpg (190KB, 671x649px) Image search: [Google]

190KB, 671x649px

>compressing a 3gb plaintext database to 120mb

>>59080269
Type inference in lisp via compiler macro is not even funny as a joke.

And it wouldn't be all that funny either.

>>59080269
>easy

Yeah, both are turing complete..

Is there anyway to send data from phone to phone when it's outside wifi/Bluetooth range?

You can send text messages and make calls with just the phone number but is it possible to send data app to app or do I need to have my phone server to store and update IP data?

>>59080158
>>59080179
That said, as long as the language does not have a garbage collector, compiles natively and into the operating system's specific library format (.dll, .so) it doesn't really matter what language you use.

>>59079345
C/C++ will always be needed for programming on the bare metal. Rust is a nice idea but it can never do what C/C++ are doing, the safetry runtime is too heavy weight.

>>59080310
>do I need to have my phone server to store and update IP data
Yes.

>>59080318
Even if it has a garbage collector there's usually some language library that you have to include and initialize the runtime system before using any of your own library's functions. And when you're done you just call the function to de-initialize the runtime.
You'll have to live with the funny GC overhead and random pauses though.

There's a reason why pretty much every library in circulation has a C interface and pretty much every language brings a C FFI. It's the lowest, most common denominator which makes exchanging data between programming languages easy-ish.

>>59080281
That's a cute Moon :3

>>59080291
Lisps employ inference to begin with, so much of the work is already done. Static checks can be done with compilers like SBCL using DECLARE.

>>59080310
You could technically build a protocol on top of SMS.

>>59080373
That's what I'm thinking. There might be a way to have the app communicate via SMS without having the user interact with it

>interview question that expects tou to use some shitty library to get the best answer
ok fuck you

>>59080323
pretty much everything used to run on C so it's a huge demotion

> safetry runtime is too heavy weight.
hardware keeps improving though

>>59080745
>hardware keeps improving though
Not an excuse, Rust will never be appropriate for drivers, libraries, or microcontrollers.

I have a set [ A B C D ] and need to list all possible combinations for it:
A, AB, AC, AD, ABC ... ABCD, ... C, CD, ... D

Been searching a while, but only found algorithms for permutations of all elements (Heap's, SJT, etc.). I need something that includes singles, pairs, etc.
I heard the name for it before but I forgot.

>>59080144
>c programmers
non-forced memes are extinct too

>>59080745
>hardware keeps improving though
that's the marketing for javascript too

File: 1484462224863.jpg (111KB, 396x536px) Image search: [Google]

111KB, 396x536px

>>59080745
>hardware keeps improving though
Pacemakers still use 6502s.

>>59080777
But at some point, neither will C.
For a certain class of devices this date has already passed.

> muh speed!
If it's supposed to poll every 1 thousandth of a second, then you could write some drivers in fucking Python.

> It hasn't been done before!
Lisp Machines had their drivers written in lisp. You had drivers for loads of stuff there.

>>59080777
How so? People run even JS on microcontrollers, what makes Rust unsuitable?

>>59080792
That's called the power set. Just without the empty set.

>>59080824
>wanting medical equipment to have atmel microcontrollers and javascript
Enjoy killing people.

>>59080837
It wouldn't be a power set without the empty set.

>>59080883
>create power set
>remove empty set
wow so hard

>>59080697
Use TensorFlow to train an NN to compute FizzBuzz. You have 5 minutes and no internet access.

>>59080893
By definition, the power set includes the empty set. The set you describe is not the power set.

>>59080911
Have you considered that >>59080792 does not want the power set?

>>59080911
Does anyone have a filter for people being pedantic about things that weren't asked? That would be real helpful on this board.

>>59080837
>>59080883
>>59080893
>>59080911
Thanks, all of you,

>>59080821
It has nothing to do with speed. It's about response time guarantees. Only bare-metal languages will ever have these.

>>59080973
Not the guy you replied to but if you have a natively compiled language that supports functions being called by interrupts (i.e. doesn't optimize them away just because nothing in the code is actually calling them) then it can be used on bare-metal hardware.

>>59080933
>>59080939
Doesn't matter. They want the set difference between the power set and the null set. Saying something like, "that's called the power set," is wrong.

>>59080973
But that's wrong.
What prevents say, a language on the JVM or the CLR to have these?

You're not going to say something stupid like GC pauses right?

>>59081042
How are GC pauses "stupid"?

Would learning C now still be a good way to start and later use C++ or should i go for the Rust meme?

>>59081010
Oh I forgot: The language must of course be capable of writing arbitrary data to arbitrary memory locations as well as providing/supporting a way to place specific memory maps in specific parts of the binary.

>>59081042
We're not talking about "could theoretically run on a microcontroller", we're talking "people would actually want to use it on a microcontroller without being completely crippled".
Or what do you suggest happens when an interrupt fires during GC operations? Sucks for the data loss?

>>59081063
Go directly to C or C++. Just forget Rust is a thing. Nobody else will remember it in a year.

Please explain to me what a server is in computing?

What is apachi and what are "server-side languages"?

>>59081063
Learn C, then try Rust if you want, but avoid C++ at all costs.

>>59081063
Whatever you're going to do, start with C.

>>59081105
Servers serve data to a client program that usually requests that data.

>>59081063
use c++ or rust, preferably rust, but don't use plain c whatever you do

>>59081129
>use c++
Opinion discarded.

>>59081095
see >>59081138

>>59081116
Ok, so why do you need a server-side language instead of a regular programming language

>>59081163
I told the truth. You can't veto what I said by shouting lies.

>>59081173
They are not mutually exclusive.

>>59081184
Your post is wrong simply by virtue of mentioning the usage of C++ in a positive light.

>>59081056
You can avoid triggering GC.
You can have GC that doesn't pause execution (wait really? We have multi-core computers now?)
And if you can make guarantees about how much memory you will need, you can put bounds on how much execution time such a pause would eat up regardless.

IOW, it's a non-argument.

>>59081077
> It could theoretically work but it is currently not practical
Well then. If we know that any garbage collection will take at most time t, can we then make response time guarantees? What if we can guarantee that we will never GC at all?
What if the microcontroller has GC stuff built into the hardware?

Do you really think that if we were forced to write all microcontrollers in lisp since 1959, that we wouldn't have had microcontrollers?

No. We would have had generations of microcontrollers designed around a lisp-like language, and the hardware would have reflected that.

Instead, we have used C, and the hardware design have reflected that instead.

In this parallell universe I would have been saying that you could indeed write software for microcontrollers with a statically compiled low level language like C, and you would have complained about things like:
> SAFETY!
> Expressiveness!
> Toolkits will suck!
And so on.

>>59081218
>No. We would have had generations of microcontrollers designed around a lisp-like language, and the hardware would have reflected that.
We already had those. They were garbage. Nice try.

>>59081063

Start with C. Do not use Rust unless you're already able to write good programs in a large number of languages, its a meme for several reasons and learning a "new" language is never a good idea when you're new.

Depending on your purpose, you may find there's a lot of shit in C++ that you don't need/want to have anything to do with when you take the C-first approach. That's what happened for me - very little of what's commonly taught for C++ is applicable for the C-with-classes usage I and many other have used for their work successfully.

C is such a great foundation, and its a language with a lot of tenured "experts" with differing opinions that actually all have merit to them. Rust does not have this because its a language that has only been around for half a decade.

I recommend Zed Shaw's book for learning C, unless you can find a copy of Pointers on C somewhere, that book is great but old as fuck - they both have a similar style. Then take apart K&R and put it back together again as you wish.

>>59081228
>a language can be a "meme"
stopped reading right there. back to your subreddit.

>>59081228
>I recommend Zed Shaw's book for learning C
this is proof you don't know C

I've been offered a job where I'd be required to port a bunch of legacy VB6 code over to C#, how difficult a task would this be considering I've never touched VB6?

>>59081228
>Zed Shaw anything

Reminder that according to Zed Shaw Python is not turing complete because Python 3 is not fully backwards compatible with Python 2.

>>59081246

What's your opinion on the book, then?

>>59081218
>You can avoid triggering GC
So memory management without the ability the ability to actually manage memory? We'll call it SUPER JAVA FUCKYOU 9000.1 u12321321313.
What is the point of GC again?
>You can GC that doesn't pause execution
Some operations in the GC are atomic, but even then we're talking about microcontrollers retard.

All your arguments defeat the purpose of having a GC in the first place.

>>59081218
Why the hell do you space your posts like that, redditor?

>>59081277
>what is you opinion on a book written by someone that doesn't know what the fuck he's talking about?
take a guess

>>59081280
> YOU CANNOT DO THIS
> Yes you can, for example this way.
> Then what's the point?

Maybe there's more to a language than memory management.

>>59081223
Oh, so you're now admitting that you can use lisp with microcontrollers?

Your arguments are that it cannot be done, and yet you're saying that it could be done, but that the chips would suck.

Make up your mind broseph.

>>59081319
give some examples. propose your point with facts rather than conjecture

>>59081354
We already agreed you can do this, just that it's a waste of fucking time and energy hacking around to make something work less than the other options already did.

>>59081319

Explain to me why you think he "doesn't know what the fuck he's talking about?".

Also I'd love to read the book that got you started on C, seeing as you're experienced in the language enough to decide whether or not Shaw knows what he's talking about. Can you name it? Perhaps it would help the guy I was replying to see a different angle and would be an awful lot more constructive.

>>59081354
Stop fucking writing like that, plebbitor.

>>59081354
>Oh, so you're now admitting that you can use lisp with microcontrollers?
There used to be microarchitectures designed for Lisp execution. They were slow, useless wastes of electricity. Their manufacturers all went out of business for a reason. Good riddance.

Lisp is an inferior design for a language.

>>59081250

Not difficult, but absolutely unpleasant.

How much experience do you have with porting other people's code?

>>59081415
So you are saying that we have to use bare metal languages, and yet you acknowledge that there are examples to the contrary.

I don't care if Lisp microcontrollers were good or not. That's not the issue. The issue is that it's possible.

>>59081409

>haHAA im going to go to /dpt/ and reply with >reddit to everyone in the thread. EPIC TROLL. YES!

>>59081415
>Lisp is an inferior design for a language.
Your argument is one big emotionally charged non sequitur. "Hardware manufacturers go out of business, so a programming language is poorly designed!" Let the grown-ups do the arguing.

>>59081440
Pretty much none, only ever ported small segments of code from one language to another and it was never particularly difficult.

It is for a graduate position and I have been fully honest, so they're aware that this would be new for me and it isn't my main responsibility. Got a few weeks before I start, so was thinking of trying to gain some knowledge before getting thrown in the deep end.

>>59081454
Lisp machines worked by moving the complexity of the Lisp runtime out of the language and into the silicon. For microcontrollers, this isn't an option. They need to be as simple as possible.

>>59081372
>give some examples
>>59081386
>Explain to me why
here's a link where he makes a fool of himself by not understanding the concept of undefined behavior:
https://news.ycombinator.com/item?id=8833965
the code in question is from his """book""" where he thinks he can avoid invalid inputs by sprinkling more checks:
https://github.com/kveratis/lcthw/blob/master/Ex27/ex27.c
by his own admission, he doesn't "know" the languages he's writing books on, he doesn't use them anyway, he just learns enough to make a quick buck by """teaching""" others about them; kind of hard to trust that """teacher"""
here's his "sour grapes" whining about it:
https://zedshaw.com/2015/01/04/admitting-defeat-on-kr-in-lcthw/

>>59081095
>>59081106
>>59081108
Thanks,
>>59081386
Not the person you talking too but i'm reading modernC.pdf and the C book.

Is there a way to initilaize a table with only one dimension having values? Example:

string[,] cuck = new string[2,6] { {"This", "is", "a", "waste", "of","time"}, {"0",0",0",0",0",0",} };

This is a short example, but it gets tedious if you go anywhere above 20-30.

>>59081659
Write a macro to do it or a for-loop that initializes the matrix with 0s

File: rust1.png (153KB, 960x353px) Image search: [Google]

153KB, 960x353px

Hey /g/uys, have you started learning Rust yet?

>>59081747
Nope, I'm learning Lisp.

>>59081710
There's no other way?
Also a related question, if I have an array in a function, does it get overwritten with a new one, if I ask for one as a parameter?

Example, I want to calculate number of button presses, first dimension are buttons, second is numbers; if I ask for an array as a parameter, does it have to be the same size as the one in the function, or can it be different?

I want either the user to enter which buttons to track, otherwise it will pick default values. I have a hard time explaining, but, if user inputs an array for 'a', 'b' and 'c', but the array in the function goes to 'e', what happens? I want it to default to 'e'.

>>59081747
yes

There should be federally sanctioned suicide booths.

>>59081386
If you want to learn C

- get K&R and read through it once to understand basic things like pointers, realize that C declarations have completely changed sine this book was written.

- get the CS:APP book and read through it to see how C works at the assembly level. There's also vid lectures here: (click on old video) http://www.cs.cmu.edu/~213/schedule.html

- read tutorials on modern C afterwards
https://matt.sh/howto-c (everything in here explained in CS:APP book like Pointer printing macros)

- read a style guide, for a modern operating system (modern as in still developed) http://man.openbsd.org/style

If you do this you will have a solid introduction to compilers, x86 architecture, how exactly C abstracts memory and uses it, how undefined behavior can happen, ect. Now go look at a project like PostgreSQL https://github.com/postgres/postgres maybe even try to implement a feature or debug issues

>>59081747
>>59081770
All bindings for system APis are c or c++.

>>59081823
Which can be linked to

>>59081823
>He doesn't know what an FFI is
>He doesn't know Lisp compiles to assembly

File: emacs_gril.jpg (283KB, 960x1280px) Image search: [Google]

283KB, 960x1280px

Bored to death. What should I work on?

Most of us will have to sit in front of a glowing rectangle until we drop dead.

>>59081860
RedSeaFS Linux kernel driver

>>59081904
>will have to
>he chose programming as a profession
LMAO

>>59081904

Yeah but like...

...what else would we fucking do?

>>59081915
Yeah, I kind of regret it now, after 10 years
Wish I was a carpenter, able to build houses

That would be more useful after the apocalypse

Programming is a great hobby but a sad profession. Sitting too much is very unhealthy.

File: 1451995301143.jpg (141KB, 600x2158px) Image search: [Google]

141KB, 600x2158px

>>59081934
This, it's not like there's something else worth doing.

>>59081951
Churn out boiler plate code for 6hrs a day then go build a house in your spare time with all the money you have to afford tools to do such an endeavor.

File: 1486442355985.png (88KB, 630x630px) Image search: [Google]

88KB, 630x630px

>>59081967
That's an unusually muscular Pepe.

>>59081978
Working on it
Most employers don't want part time
Trying freelancing now, first project underway, fun so far

>>59081783
Please respond.

>>59082051
Please clap.

>spend two weeks trying to locate a bug in third-party code
>coworker comes along and hacks around the bug in 30 minutes

Feels awful.

If only I could find the actual underlying bug which still is in there, somewhere...

>>59081912

Has someone already started on it? I know there's been a lot of work on Shrine, but I don't know if anyone's porting stuff to Linux.

>>59082002
Good luck making more than $10 a week

>>59082051
Your question is worded poorly. Do you mean you have a statically allocated array in the function and pass one with the same name in? Post pseudo-code.

>>59081783
Seems you should have a function called 'default values' which has switch cases for key entries, falling through to default values, with loads of error checks.

>>59082179
With each switch case incrementing number_of_times_typed :=++ or w/e you're programming in. Post pseudo code

>>59082145

static void Something(string[,] arr)
{
arr[1,5]="0";
}

Main
{
string[,] yarr = new string[2,4];
yarr[1,4] ="2";
}

Will it go out of bounds? How do I overcome that, I want the array to be larger than [2,4], if the user makes it so small.

>>59082129
>Has someone already started on it?
Not that I know of.
Someone made a RedSeaFS userspace driver for BeOS, just for extra Internet hipster points
no FUSE or native Linux driver that I know of

>>59082215
That is a question that is kinda language dependent. If you called something in C/C++, unless you are passing a reference/pointer to the object (2d array in your case), the variable is actually a COPY of the argument in question, and will go out of scope when you return from the function (i.e. get deleted). If you passed by reference, you would be modifying the actual argument, and changes would be visible outside the function.

In something like Python this is done automatically.

In a more general case, your "Something" function should ALWAYS be checking what it was passed is valid (i.e. non-empty/correctly sized) before modification.

I implemented fast event objects using user-space queuing and eventfds.

http://pastebin.com/ddBjddG7

>>59082527
Why do all of your linked list nodes have a pointer to head? If you're going to have two pointers in every node, you may as well have one point to the next node and one to the previous.

>>59079314
>brianlet

Daily Reminder

OOP is far superior to functional programming in everyway

>>59080108
unity is written in c++, the c# is just scripting

In Java, how would you call a method which is in a class from another method which is in another class?

class Car {

public String someMeth() {
    return "hel";
}
}

class Truck {
    
    public String callMeth() {
        return this.someMeth();
    }
}

Why won't this work?

>>59082469
That's for C#.

>>59082609
nigga....

>>59082559

struct fdnotifier {
    int fd;
    _Atomic(struct fdnotifier *) next;
};

struct node {
    struct fdnotifier * trigger;
    _Atomic(struct node *) next;
};

They don't.

struct event {
    _Atomic(bool) triggered;
    _Atomic(struct node *) head;
    _Atomic(struct node *) tail;
};

is the root of the linked list and not the nodes themselves.

File: memearrowpython.png (389KB, 2560x2813px) Image search: [Google]

389KB, 2560x2813px

>>59082728

File: Screen Shot 2017-02-23 at 1.03.46 PM.png (14KB, 141x76px) Image search: [Google]

14KB, 141x76px

I noticed an extra newline but code worked and it surprised me.
In Python you can have arbitrary number of new lines after IF?

what's going wrong with this?

justin@debian:~$ ghci haskell/*.hs
GHCi, version 7.6.3: http://www.haskell.org/ghc/  :? for help
Loading package ghc-prim ... linking ... done.
Loading package integer-gmp ... linking ... done.
Loading package base ... linking ... done.
[1 of 2] Compiling Mergesort        ( haskell/Mergesort.hs, interpreted )
[2 of 2] Compiling Collatz          ( haskell/Collatz.hs, interpreted )
Ok, modules loaded: Collatz, Mergesort.
λ> mergesort $ collatz 5

<interactive>:2:1: Not in scope: `mergesort'
λ> :l haskell/Mergesort.hs
[1 of 1] Compiling Mergesort        ( haskell/Mergesort.hs, interpreted )
Ok, modules loaded: Mergesort.
λ> mergesort [9,7..1]
[1,3,5,7,9]
λ> 

I want to use multiple modules i've made myself simultaneously in ghci

>>59083119
>haskell
found your problem

>>59083119
>deficient module system
Use a better language.

>>59083119
import Mergesort
import Collatz

>>59083167

λ> import Mergesort

<no location info>:
    Could not find module `Mergesort'
    It is not a module in the current program, or in any known package.
λ> 

>>59082980
In hindsight this makes sense because you can add newlines to space stuff but it still looks weird to me.
On that note how does Python handle repeated newlines?
Looking at https://docs.python.org/2/reference/grammar.html
An IF must be followed by at least one stmt

stmt: simple_stmt | compound_stmt
simple_stmt: small_stmt (';' small_stmt)* [';'] NEWLINE
small_stmt: (expr_stmt | print_stmt  | del_stmt | pass_stmt | flow_stmt |
             import_stmt | global_stmt | exec_stmt | assert_stmt)

But I don't see how small_stmt can be emptystring.

>>59083214
:also [file]

>>59083273
A newline isn't an empty string.

Almost finished Think Python, got into those shitty OOP chapters.
I run this:

class Point(object):
    """
    attributes: coordinates x, y
    """

class Rectangle(object):
    """
    attributes: width, height, corner
    """

box = Rectangle()
box.width = 100.0
box.height = 200.0
box.corner = Point()
box.corner.x = 0.0
box.corner.y = 0.0

def find_center(rect):
    p = Point()
    p.x = rect.corner.x + rect.width/2.0
    p.y = rect.corner.y + rect.height/2.0
    return p
    
center = find_center(box)
print(center)

but get this

<__main__.Point object at 0x7f3cc716d4a8>

instead of the result. What do?

>>59083303

print(center.x)

>>59083277
it says it doesn't know the command. the only way i can get it to work it seems is to make a big file which imports all my other files

>code fails to compile with cryptic error messages because I accidentally wrote data in a type alias instead of type
motherfuck

>>59083328
Are you using cabal or stack?

>>59079036
SECOND FOR LINUX CAN'T IN TO CONCURRENCY. PORT LOONIX TO RUST. INSTALL RUSTOO

>>59083290
I know but we have

if_stmt: 'if' test ':' suite ('elif' test ':' suite)* ['else' ':' suite]
suite: simple_stmt | NEWLINE INDENT stmt+ DEDENT

So if_stmt must be followed by a stmt. A newline isn't compound_stmt so I look at simple_stmt. The simplest being small_stmt NEWLINE so to get just NEWLINE small_stmt should be empty string correct?

File: andthatnumberisgoingtogetfaster.png (352KB, 949x533px) Image search: [Google]

352KB, 949x533px

>>59079339
Jai will lead us into the future of video game development. and Jonathan Blow will be our shepherd.

>>59083324
God, why does this shit should work like this in new versions.
Fuck OOP.

>>59083355
i was kind of secretly hoping to bait someone into explaining to me how to use cabal

>>59082583
>C is far superior to procedural programming in everyway
what did he mean by this?

>>59083303
Because you return the Point() object, not its attributes, you can add a __repr__ method to the Point class if you want:

def __repr__(self):
    return '<Point({}, {})>'.format(self.x, self.y)

how's haskell for server side applications?

>>59083428
Use cabal to make a project

cabal init

>>59083369
does he have a lambda prompt while using windows?

>>59083508
does this publish this online? because the questions seem like it does. i just want to have these in a local library for personal computing

>>59083515
he's using a console emulator called cmder

>>59083547
no, it's just something you can do with it.

>>59083303
>>59083377
Do something like

class Point(object):
    """
    attributes: coordinates x, y
    """
    def __str__(self):
        defined = hasattr(self, "x") and hasattr(self, "y")
        return "({}, {})".format(self.x, self.y) if defined else "Undefined"

>>59083303
why no constructors?

Back end dev making my own site, can't style for shit. Anyone got good advice for more of a straight programmer on how to make things look decent?

>>59083515
so?

>>59083580
i did have the cabal file for Collatz.cabal and it still tells me it doesn't know what i'm talking about when i tell it to import Collatz. it's in a folder called Collatz with Collatz.hs and Collatz.cabal and a setup.hs file. i'd like to be able to just do import collatz anywhere on my system and have the code imported

>>59083697
Boostrap.

>>59083700
windows as an operating system isn't very compatible with the hacker ethos which a lambda prompt seems to imply you aspire to

>>59083746
does using windows at all (even though it's not my main system) now disqualify me from using a lambda prompt? but I have official permission

>>59083746
You don't seem to grasp just how pretentious Jonathan Blow is.

>>59083866
>pretentious
>using that meaningless word at all
mate...

>>59083833
i doubt you'd be so defensive if you really used linux primarily for your computing

>>59083921
pretentious - attempting to impress by affecting greater importance, talent, culture, etc., than is actually possessed.

what's wrong with OOP I don't get it?

>>59083970
most things shouldn't be modelled as objects

>>59083970
there's nothing inheritly wrong with it, ancient elitist programmers like to rage against it because they weren't brought up using it and bad programmers use it as a crutch to write shitty code

File: 2017-02-23-201614_805x365_scrot.png (60KB, 805x365px) Image search: [Google]

60KB, 805x365px

>>59083921
>assuming the word you don't know doesn't have a meaning
You should have checked my dubs instead

>>59083970
I don't mind it as long as it makes sense for what you're doing.
I dislike when it's forced into everything.
I had design patterns in java book and the lengths it went to do things that would be simple in other language.

>>59083991
when should you model things as object and when shouldn't you model them as objects?

This is what I don't understand

>>59083949
>lincux
why would I be using literal garbage as my primary system?

>>59083970
it requires way more code and encourages an enormous number of states

>>59084059
>when should you model things as object
very rarely if at all

File: Screenshot_2017-02-23_21-15-55.png (282KB, 1134x604px) Image search: [Google]

282KB, 1134x604px

Still working on thumbnailing library. Seems to be well, except for detection failing on some video container combinations and OGG with cover art is giving an FFMPEG error.

>>59084059
If it makes things easier to model as a object then model it as an object. Otherwise don't. Shitters tend to model really easy problems that can be solved with some simple procedural code as object hierachies aswell which ends up with them writing more code, so don't do that either

>>59084059
Generally anything that has to persist in a certain state until some event happens, like GUIs and game entities.

>>59084090
post that kumiko image

>>59084174
so basically if I am working with GUIS or video game programming?

What if I am working with Java though and everything is OOP

>>59084220
>What if I am working with Java though and everything is OOP
then you switch careers

In 8086 dos programs, can I make any assumptions about the register values on entry?
for example the AX register always appears to be FFFF on entry, and CX appears to be the size of my program. Are these values guaranteed to be the same every time?

File: no_sound.webm (248KB, 640x360px) Image search: [Google]

248KB, 640x360px

>>59084194
I think this one is downscaled and 2 times over lossy-encoded though. Don't have the original.

Explain Finite State Machines in one to two sentences.

>>59084311
Goto sentence 2.
Goto sentence 1.

>>59084311
It's a machine. A finite state machine to be more precise.

>>59084311
A finite state machine is just a monoid in the category of endofunctors. What's the probleⅿ?

>>59084220
don't work with java

>>59084311
>>59084322
>>59084333
checked mate

>>59084360
No it's not. This is blatantly false.

switch(currentState){
   case kys:
  {
   kys();
   }break;
   case dontkys:
   {
    dontkys();
   }breakM´;
}

>>59084420
I think he's doing a meme

>>59084420
You can implement an FSM as a monad, which makes it a monoid in the category of endofunctors

>>59081646
thanks anon. I was just trying to raise the bar on conversations by making people backup their claims. you're a big guy

>>59084507
>what's a FSM
>it's a Java class

>>59084458
>that bracketing
jesus christ man

>>59084507
Does this make it a monad by definition? I don't think so.

>>59084458
Disgusting.

switch currentState {
case kys:
    kys()
case dontkys:
    dontkys()
}

>>59084311
It's a special case of a turing machine without storage.

>>59084557
>>59084600

//bad
    switch (c)
    {
        case Black:
            int x = 1;
            break;
        case White:
            int x = 5; //redefinition, multiple init
            break;
    }

    //good
    switch (c)
    {
        case Black:
        {
            int x = 1;
        }break;
        case White:
        {
            int x = 5; //ok
        }break;
    }

you're both dummies

>>59079102
"Turns a machine learning framework such as Tensor flow, into a computational graph."

Dude my sides wired please stop this

How do I implement a stack with push, pop and an O(1) min function?

>>59084616
>needing explicit breaks
// bad

>>59084616
>Brackets on case statements
>break after closing bracket
I don't know what retard school you went to, but that is not industry standard anywhere

>>59084458
jesus christ ..

new thread
>>59084645
>>59084645
>>59084645

File: 13f.jpg (26KB, 600x449px) Image search: [Google]

26KB, 600x449px

>>59084616
>declaring variables inside a switch case

>>59084616
>>59084458

fun Black = 1
fun White = 5

>>59084621
>O(1) min function?
always keep track of the minimum

>>59083970
OOP is implemented in way is very slow to high performance application, on game development, don't want to pay price, mostly things don't had JAI, is avoid pay prices on performance,complexity or time compile.

In my code why is the do futex option faster?

http://pastebin.com/PqiNL5n4

Oh god. So many fucking pajeets

http://www.spinics.net/lists/linux-c-programming/

>Hi,
>
>is there size limit on the buffer content as i am using "cat" command
>( so ,not allocating any buffer in userspace, so not doing any
>copy_to_user from kernel; simply doing sprintf(buffer, %s , data )
>from kernel for a big data )
>
>-Ratheesh