/dpt/ - Daily Programming Thread

>>59616097
I made this:
http://codereview.stackexchange.com/questions/159056/find-all-distinct-8x8-chessboards-where-5-queens-attack-occupy-every-square

I was inspired by this:
>>59601324 (Cross-thread)

Please help to improve

File: 10920129309123.png (107KB, 500x441px) Image search: [Google]

107KB, 500x441px

>>59616097
i'm learning lisp :3

>>59616172
you could concentrate better on the lap of a cute boy tho anon

Finishing touches on Google Summer of Code application. Nobody else (other than the maintainers) has submitted a PR to the project I'm submitting an application to in the past year, so I think I'm gonna make it.

>>59616179
p-pls you're making me blush >.<

ditched C for Elixir
muh concurrency

>>59616206
i'll turn your cheeks red indeed anon

>>59616208
You need portable threading? Use musl.

>>59616226
threads are more costly than elixir processes (which arent processes)

File: krautbbs_03a_02.webm (3MB, 832x1024px) Image search: [Google]

3MB, 832x1024px

Are there any good reads on how to port Python 2.7 scripts to Python 3? Is that even a good idea, or should I start webmrelated from scratch for 4chan?

File: lua.gif (5KB, 256x255px) Image search: [Google]

5KB, 256x255px

>>59616236

Fun fact:

Did you know that the creator of Lua is a Brazilian Jew?

>>59616236
What compatibility issues are you having specifically working in 2.7? Just keep developing in 2.7 if you like, set up a virtualenv (http://python-guide-pt-br.readthedocs.io/en/latest/dev/virtualenvs/) to make it easy for people to contribute and use.

File: 1307913409001.png (94KB, 500x500px) Image search: [Google]

94KB, 500x500px

>>59616221
oo stap

File: dpt.jpg (164KB, 800x1000px) Image search: [Google]

164KB, 800x1000px

>>59616097
As always.

I'm reading this book now
https://mitpress.mit.edu/decision-making-under-uncertainty

>>59616236
What the fuck am I looking at?
Anywhere there are modules that port 2.7 to 3 but it might result in your script being buggy. So why not just work in 2.7?

>>59616264
im afraid i cant hold myself back with you anon

>>59616248
Uh, ok, I guess.

>>59616261
No issues really, just want to have a fork ready for people who prefer Python 3.

>>59616273
A BBS simulator for the temporarily dead german imageboard Krautchan. It currently supports browsing all boards, viewing page 1 and the catalog as well as opening threads and downloading every attachment in a thread into a directory specified by the user. Was working on posting functionality but then KC got the Polizei treatment.
As for the porting: see above. Some people just prefer Python 3 so I want to please both Versionistas.

>>59616097
>>59616120
>>59616172
I don't really care much about this meme, but if there really is any crossdressing programmer. I would really like to give them a hug.

Anyhow. Guys I got a minor question, I haven't found any proper answer to it. The question is, what benefits are there in writing in a webservice in "java-ee" compared to "js/php".

>>59616324
Anyone who uses python 3 should also have virtualenv. The changes in syntax aren't hard to pick up on the fly if you want to change something but are annoying enough to change in the program itself that it's probably not worth updating.

Just google a guide on updating legacy code to 3, it might not be that hard for your use case. But if you have a library that depends on 2.7 you're probably fucked.

>>59616324
Do you have a git for that. I am interested? I amuse you are using basc_py4chan as your wrapper for 4chan API?

>>59616339
Thanks, will keep that in mind.

>>59616344
Yeah I have: https://gitgud.io/hackierendebernds/BBS-Simulator (you need to request group invitation first though).
Source can also be viewed here: https://pastebin.com/Yj0qmBTb (expires in a week; the 'heinrich' import is just a file that displays ASCII art at random)
As KC doesn't feature any API that I'm aware of, I used BeautifulSoup to scrape the website and print the content I needed. It's very crude as I just hacked this thing together rather quickly. For the 4chan version I'll probably rewrite everything from scratch anyway.

>>59616378
Have you seen this before? https://github.com/bibanon/BASC-py4chan. I use this to make a script to scrape 4chan to mass save images off /h/. It is pretty decently documented as well.

http://basc-py4chan.readthedocs.io/en/latest/index.html

>>59616407
Oh, this also allows for posting? Now that is something I'll definitely use. Thanks.

>>59616418
No it doesn't allow you to post it just makes scraping 4chan a lot more easier than using BS

>>59616325
>"js/php"
What do you mean by JS/PHP? Do you mean JS on backend or PHP on backend, or do you mean JS front-end/PHP backend?

A lot of services begin with PHP then later switch to compiled languages in order to make their service faster, more optimized. I don't know about writing web services in Java, but I assume the positive side is that you're not using PHP, ho ho ho!

If you mean JS on backend as well, then you don't really want that if your service needs any sort of heavier processing and/or I/O to files. JS is single threaded as well so that's not nice. I guess you could use workers to split it into processes, but I haven't really looked into that.

>>59616436
Oh that's a shame then. I'll take a look at it anyway during the weekend. But if posting is not possible, I'll probably stick to BS.

>>59616441
I meant in both way. Just wanted to know if anyone have concrete information about the question I had. Want to learn a bit about it, since I never know if I must work with it in the future(even the java-ee is old)

>>59616452
No the beauty of the wrapper is the simplification of lines in your code. Looking at the source here is an example.

Instead of

ThreadName = soup. Select ( 'div.thread_body div.postheader span.postsubject' ))

You use

ThreadName = thread.topic

Off topic but just curious how are you going to deal with captcha when posting?

>>59616325
i've worked with all 3. java webservices are great for performance (JVM) and making things very exact (strict typing) and structured (OOP). this makes it ideal for enteprise

I'm unemployed

My friend who has a job writing ENTERPRISE QUALITY C# code told me the other day "You know, I never really *got* recursion"

But I'm the one who's unemployed

Anybody who says markets allocate talent efficiently is full of shit

File: 1461353633402.jpg (139KB, 1280x720px) Image search: [Google]

139KB, 1280x720px

>>59616464
Fine then, the difference is that PHP is interpreted language and Java compiles to bytecode, so in theory, Java should be faster.

Just learn all you can.

>>59616533
>he thinks employers only care about how much you know
laughing_girls.webm

>>59616552
I've heard that the creator of php was a novice programmer, but are these things he's actually said?

>>59616533
That's because you're too autismal to get hired and you smell weird. People feel uncomfortable around you.

It's hard to get a job that way.

>>59616577
Check his twitter.

>>59616563
I don't
because they clearly don't

the idea that capitalism spurs individuals to better themselves is a big fat lie. We are only incentivized to grovel before those who hold the resources

>>59616577
I think half of it is made up. He may have said some those things. Or maybe none of them.

Still funny though, I really really like that image.

>>59616522
How easy/well is it to do test deployments with the java webservices?
>>59616552
Thanks I know that bit, just wanted to know the core benefits of using x instead of y and then a comparison of why use y than x. I am still googling about the matter and doing some readings but I am having my eye open here too, incase someone drops by and lends their knowledge

>>59616505
Yeah sure, the simplification is thanks to the library I have to import. To me it doesn't make any difference other than having to distribute less files. And the user will not notice any different either. I'll have a look, but seeing as I'd have to read through additional docs or at least read and understand the library I'll most likely stick to BS unless the library totally blows my mind.

Captcha is something I am currently struggling with. Which is why posting on 4chan will probably be a long stretch.
Downloading the images will probably not work. Thought about using the audio captcha.

>>59616586
His twitter seems pretty normal desu, nothing crazy in the past year of history at least.

>>59616594
tru

File: ojLcjvk.jpg (82KB, 800x600px) Image search: [Google]

82KB, 800x600px

>>59616221
>>59616206
>>59616172
>>59616179
>>59616264
>>59616275
Cute, but also degeneracy

>>59616615
Well I look forward to your project and wish you the best of luck. Hit me up with your git so I can keep up to date with it.

>>59616644
Let's have a consensus that as long as it's cute, degeneracy is allowed, yeah?

>>59616650
Will drop a link in a future dpt once I get to it. Cheers for helping so far.

File: STOP.jpg (15KB, 191x264px) Image search: [Google]

15KB, 191x264px

>>59616663
I'll concede that so long as it doesn't turn lewd

Is it normal after 3 months of learning to still not know what you're doing?

>>59616690
nah, /g/ stands for gay roleplay
i'll lewd anyone i want anon

>>59616723
if it's C or Rust, yea

Is anyone here good with cmake?
I have a .so I need to link to my project using cmake.

I have included it in a dir local to the project (src/LIB) relative to the CMakeLists.txt

SET(CMAKE_INSTALL_RPATH_USE_LINK_PATH TRUE)
INCLUDE_DIRECTORIES(src/LIB/include/)
LINK_DIRECTORIES(src/LIB/lib/x64/)
ADD_LIBRARY(${PROJECT_NAME} -lLIB

This worked well on one computer, but didn't work on the other computer I am testing on.
They both have the same library files.
Is there something else I can do?

File: 1374576513876.png (642KB, 726x1040px) Image search: [Google]

642KB, 726x1040px

>>59616663
No. Fuck you.
"Le trap meme XDDD" is not funny, cute, or endearing in any possibly way.
It's literally just a stupid forced meme.

>>59616846
the error is:

/usr/bin/ld: cannot find -lLIB

>>59616850
we aren't going to take your fetish seriously, sissy

>>59616760

You should put on a trip so that Anons who do not want to see your lewd roleplaying can filter it.

>>59616870
They can always start their own if they need a safe space from some gender dysphoric dudes roleplaying lol

>>59616857
>we
Who the hell is "we"?
Also, did you even comprehend my post in the slightest? What you said doesn't even make sense.

>>59616846

In CMake, add_library is used for building a new library. To tell your project to link against a library, you need to use target_link_libraries.

>>59616883

It's not about needing a safe space, it's about not wanting to be bothered by a bunch of off-topic cybersex when one is trying to discuss programming.

>>59616893
We is me and >>59616857

But that little chat between those two anons at the beginning of the thread was cute, so we allow that degeneracy. It was already conceded. You're late.

>>59616893
By we I meant the board, and you came across as some tranny who is butthurt someone's not taking them seriously by complaining about the meme.

I feel bad that I had to explain this to you, hope you get better soon my man

>>59616917
But the cybersex is to stimulate the dopamine release to aid concentration and brain performance anon, you want us to underperform?

>>59616919
>Calling traps a stupid fucking meme means that you're a trap
Are you fucking serious? Your logic baffles me.
>By we I meant the board
>I speak for /g/

>>59616907
I use these functions in Android native builds.

>>59616931
What, are you replying to the right person? You go right ahead, I'm well happy to let you go on with your cybersex.

>>59616940
>>Calling traps a stupid fucking meme means that you're a trap
>Are you fucking serious? Your logic baffles me.
So you aren't a tranny and still give so much shit about what people meme about? Get a hobby my man

>>59616949
fuck wanted to reply to ruby lmao, thanks for the permission daddy

>>59616973
You're welcome cutie ;)

>>59616976
;*

>>59616907
>In CMake, add_library is used for building a new library. To tell your project to link against a library, you need to use target_link_libraries.
my mistake, I wanted to cut out the irrelevant information, so it was easier to ask my question, I did a add_library, on the line above so I mixed them up when I wrote the question, I do use

target_link_libraries{${PROJECT_NAME} -lLIB}

It works on one machine, but ld cannot find the library.
as I have phrased it, the relative path is "src/LI B/lib/x64/libLIB.so"
There must be a way to get cmake to look for the file.

>>59616931

There are MANY ways to get the dopamine rush you need to perform better. Even if you do feel that the best way to do so is through cybersex, consider that you do not need to do so in the discussion thread, and that your continued actions may distract others trying to have any reasonable discussion, causing them to underperform.

I would like to repeat my suggestion that you adopt a trip, as not only would it allow others to filter you, it would allow you to identify each other so you can more easily fuck off to some private IRC channel somewhere, rather than grabbing a random Anon.

>>59616943

Right. And the same commands on desktop Linux, on Windows, and on Mac OS X, apply equally the same on Android.

add_library:
https://cmake.org/cmake/help/latest/command/add_library.html
>Adds a library target called <name> to be built from the source files listed in the command invocation. The <name> corresponds to the logical target name and must be globally unique within a project.

target_link_libraries:
https://cmake.org/cmake/help/latest/command/target_link_libraries.html
>Specify libraries or flags to use when linking a given target and/or its dependents

>>59616883
What this anon said x2 >>59616907

>>59617000
>There are MANY ways to get the dopamine rush you need to perform better. Even if you do feel that the best way to do so is through cybersex, consider that you do not need to do so in the discussion thread, and that your continued actions may distract others trying to have any reasonable discussion, causing them to underperform.
>I would like to repeat my suggestion that you adopt a trip, as not only would it allow others to filter you, it would allow you to identify each other so you can more easily fuck off to some private IRC channel somewhere, rather than grabbing a random Anon.

no
*teleports behind u and molests ur third chin erotically while u can do nothing about it*

i wanted to know if there was a way to "benchmark" a compiler so i came in here, but now i have a whyboner goddamnit

>>59617000
I just wrote it wrong here, it is correct in my CMakeLists.txt
I was more concerned with the LINK_DIRECTORIES path being correct.
I could set a environment variable on all systems, then load it with cmake so I write the full path, is this a better way?

Alright, I'm going to sleep after this post. Night, all!

>>59616992

Your syntax on target_link_libraries is also wrong. Don't think of it like what you'd pass to the compiler. You don't throw in any -l nonsense. You just put in the name of the library. CMake will then generate whatever flag is necessary for your linker to use that library.

>>59617038

http://lmgtfy.com/?q=compiler+benchmarks

File: 1438572076186.jpg (68KB, 700x700px) Image search: [Google]

68KB, 700x700px

>>59617059
>come to dpt
>get told to google it

>>59617073
>listening to tripfags
Just hide them.

File: Screenshot_2017-03-28_11-32-31.png (462KB, 2560x1409px) Image search: [Google]

462KB, 2560x1409px

>>59616097
>What are you working on, /g/?
Working on my NVMe driver for CUDA.

Attempting to do some multiple reads and writes and benchmark the performance of stuff. Right now I'm busy trying to weed out bugs though.

Some stuff I need to do is to set up MSI-X interrupts and in addition create and delete namespaces, in order to set block size to equal page size.

Btw, trying to come up with an idea to a workload to showcase when it makes sense, if anyone has an idea where it would make sense to control one or more disks from the GPU, let me know. From my perspective, the BAR size of the Quadro I have is really limiting, and the latency of the disk is too high to show any real performance improvement. I might choose to focus on the persistence aspect of the SSD.

    private String newID() {
        ArrayList<String> ids = new ArrayList<String>();
        
        for( int i = 0; i < nodes.size(); i++ ) {
            ids.add( nodes.get( i ).getID()  );
        }
        
        String id;
        int idNum = 1;
        id = "BH" + idNum;
        
        for( int i = 1; i < ids.size() + 2; i++ ) {
            if( ids.contains( id ) ) {
                idNum = i;
                id = "BH" + idNum;
            }
        }

        return id;
    }    

is this retarded enough that i should redo it?

Rewriting GNU coreutils in D. Dunno how is thr Rustfag doing. I finished ls, cat and wc. I found out writing manpages take a lot of time.
I also discovered freeBSD's cat and GNU cat are vastly different.
I got a question. How should I work with man-db? Amd how do I package manpages?

>>59617099
This is literally as memetacular as the "Rewriting x in Haskell" shit usually posted at the start of these threads.

>>59617115
Except im actually doing it, its a nice project to "git gud"

>>59617132
>Writing anything with some shitty deadlang
You may as well write it in Delphi or some shit.

>>59617147
Languages have no lives

>>59617147
D's not dead

>>59617158
Neither do you

File: 1407220426641.gif (2MB, 640x360px) Image search: [Google]

2MB, 640x360px

>>59617167
>D's not dead
That deserves a smug anime face.

>>59617186
It's not though

>>59617059
For some reason, my IDE was hiding the warnings, adding

cmake_policy(SET CMP0015 NEW)

helped, now it works for some reason....

>>59617186
Worst vivid.

Where can I learn C macroexpanding techniques, so I could write object-oriented C?

>>59617323
>object-oriented C
Don't bother. This is a terrible idea.

>>59617397
>https://soundcloud.com/recordsyourcomments/48-1

Holy fuck I'm dying.

>>59617323
read the docs and code on Gobject

>>59616097

(define (sums-even? field)
  (let ((bit-len (calc-bit-len (calc-field-width field)))
        (height (calc-field-height field)))
    (if (=
         (length
          (filter
           even?
           (accumulate-n
            + 0
            (map                
             convert-binary
             field 
             (generate height bit-len)))))
          bit-len)
        #t
        #f)))

Diving into Scheme.

I'm using Python and have to decode this string

0x55ba4283

Into

2015-07-30T15:28:03.000Z

But I can't manage to figure out how to do it

Documentation on the data is https://github.com/ethereum/wiki/wiki/JSON-RPC#hex-value-encoding

I tried with decode("utf-8") and ("hex") without success

>>59617323
https://pastebin.com/F2V852x7

>that one faggot who forces trap shit is still starting /dpt/'s
>node
>rust
This is the state of /dpt/ today.

>>59617435
0x55ba4283 == 1438270083
That's a Unix timestamp.

>>59617504
in 2008 it was
>traps
>RoR
>Visual Basic
lurk more

>>59617504
Rust isn't as bad for anything you would otherwise be forced to use C++ for. It also shares a lot of the problems of C++, so I just think of it as a memory safe C++. But of course it's not a fucking panacea, like the faggots would have you believe.

>>59617504
>This is the state of /dpt/ today.
Well, to be fair, there's also
>>59616378
>>59616107
>>59617079
>>59617436

So it's not all bad.

>>59617529
>memory safe C++
That's a giant meme, considering how C++ is merely a memory safe C ( + some metaprogramming shit)

>>59617099
Bump

>>59617551
Memory safe as long as you don't fuck up and don't want concurrency.

>>59617551
C++ is by no means memory safe. You can easily shoot yourself in the foot without realizing it.

>>59617563
>Memory safe as long as you don't fuck up and don't want concurrency.
So in other words, it's just like C++ with some new memory ownership concepts.

guys quick. whats the best way to do auto layout from a simple set of constraints, i know i could """just""" use the cassowary solver, but to gitgud id like to reinvent the wheel

>>59617504
>Things I don't like is the "current state" of /dpt/
I don't know may be you don't belong here

>>59617513
Thank you anon! Saved my day

>>59617577
Exactly. Like C++20 will probly have roughly Rustic concepts somewhere. Hopefully boost will be dead at that point.

>>59617504
You've forgotten:
>Don't write C

>>59617568
You can create unintentional memory leaks and trigger protection faults in Rust too, you know. See >>59617577

Rust is good because it has a new memory management concept to offer. That's what Rust is supposed to be.

It's *not* supposed to be a sepples replacement.

>>59617598
>hopefully boost will be dead at that point
I don't understand why boost is so disliked on /g/. 90% of the features added to C++11, 14 and 17 respectively originated in boost.

Boost is a good indicator of what is most likely coming in the next generation of C++.

>>59617602
Memory leaks are memory safe.

>>59617612
But protection faults (aka segmentation faults)? That's essentially the OS/hardware telling you that your application dun fucked up.

>>59617611
Boost added a lot of great things to C++ but some libs weren't particularly well made and the nonstandardness made it worse in the wild.

>>59617618
Source on safe Rust code causing protection faults?

>>59617625
Well, I agree for nonstandardness in general. It's always better to rely on standard features. However, boost today is pretty widely used. I agree that it wasn't always so in the past, though, and yeah, there are examples of badly made APIs that have made it until today because of legacy reasons.

>>59617637

https://github.com/rust-lang/rust/issues/28493

>>59617665
there was a PR and it was merged you lying nigger

>>59617665
That part is implemented in unsafe Rust.

>>59617618
and in the wonderful world of embedded development, you get even more fun debugging - the system bus starts to glitch out instead of segfaulting because there is no memory protection mechanism

>>59617700
>unsafe

use std::collections::BTreeSet;

fn main() {
    let _: BTreeSet<()> = BTreeSet::new();
}

If this is """""""""unsafe""""""""" Rust, then Rust truly is as useless as Haskell where doing any IO is considered impure.

What's the deal with prefixing your source filenames with project tag?
Like you are going to put headers in their own directory by project so filename clashing is not going happen.

>>59617715
>>59617700
Nah, it was just a bug. Probly something to do with jemalloc being a cunt.

>>59617715
Is that example supposed to back your claims?

>>59617715
https://github.com/rust-lang/rust/blob/6db1a0ea7c150c0d3ac361996a993cd9548f6daa/src/libcollections/btree/node.rs and Ctrl-F unsafe, retard.
>>59617728
See above.

>>59617732
That's the code from the issue that triggered a protection fault, you illiterate imbecile.

>>59617737
>standard collection is "unsafe"

That's the equivalent of C++ crashing in runtime when you use a std::vector

>>59617738
And it calls unsafe code, faggot.

>>59617750
So Rust is """safe""" as long as you don't use some (or any?) of the standard collection data structures?

wew what a language

>>59617748
Yes, and?

>>59617737
Right, a bug.
>This module represents all the internal representation and logic for a B-Tree's node with a safe interface

>>59617761
>>59617748
Rust generally guarantees you cannot make unsafe memory calls, not the standard library. If there's a bug, there's a bug. It's been fixed forever anyway.

>>59617764
I thought the point of Rust was that it was a safer replacement for C++.

But if you have to reimplement data structures and refrain from using the standard collections, you might as well be writing C code imo.

>>59617761
Rust, the language is safe. Any code written in Rust is only safe, if it explicitly does not disable these safety features.

>>59617775
>Any code written in Rust is only safe, if it explicitly does not disable these safety features
How does using the """safe"""" interface to BTreeSet "explicitly" disable those?

>>59617715
>>59617732
>>59617700
>>59617737
>>59617750
>>59617764
>>59617775

Wow, the "Rust fags will defend this" meme is actually true.

>>59617781
The unsafe code is encapsulated, people writing unsafe code can write bugs. People writing safe code can't write those kinds of bugs. It's fixed.

>>59617781
Because BTreeSet is not part of the language? Do you not understand the difference between a language and the standard library implementing it? If a bug in some ASM in the C++ standard library was discovered, would it be a bug in the C++ language?

>>59617795
Rust doesn't do a very good job at being """safe""" if it has a concept of safe and unsafe code in its language specification.

That's like Python specifying that standard library functions might not be safe to call because those are implemented in C. In reality, however, Python only specifies how its standard library should behave because an implementation that doesn't behave that way is not conforming to the standard.

>>59617811
As compared to what, every other language being purely unsafe? At some level (the domain of C/assembly) you need some non-safe code. This is not what people should normally be writing in for very obvious reasons. Rust can do the bare metal stuff but it's like writing inline assembly WITH the caveat that they ARE kept in separate unsafe blocks.

File: tmp_7067-1490599441762-49955412.jpg (43KB, 400x400px) Image search: [Google]

43KB, 400x400px

>>59617828
>At some level (the domain of C/assembly) you need some non-safe code
And isn't the point of Rust to have bare metal safe code?
If it's not, why even use it?

What's happening here?

>>59617099
Last bump

Holy shit rusters keep digging yourself deeper.

>>59617808
>Because BTreeSet is not part of the language?
std::collection

>Do you not understand the difference between a language and the standard library implementing it?
Standard library is part of a language. The C standard library has no bugs

>If a bug in some ASM in the C++ standard library was discovered, would it be a bug in the C++ language?
No, it would be a non-conforming implementation of C++. You, however, are arguing the opposite, that the standard library is NOT a part of the langauge specification.

>>59617840
No, the point is to be like C++ where you CAN do bare metal code but you also have high level abstractions at minimal cost. Bare metal isn't safe period, it's got no abstraction in place to protect you from yourself. Rust's job is to keep the abstractions safe from normal higher level code without a performance hit.

>>59617828
The argument for Rust is to have low-level, high-performance AND safe code. If you can only either have high performance OR safe code, then what's the fucking point?

>>59617860
>Rust's job is to keep the abstractions safe
But it doesn't, when fucking standard collection fails and the argument isn't that "this is a bug with the implementation", but actually Rustlers are saying "this is okay because it's unsafe code and unsafe code needn't be safe"

>>59617854
An _implementation_ of the standard library is not part of the language.

>>59617865
Rust has both pure-performance unsafe code and minimal-cost safetynet abstractions, plus the borrow checker which is a completely different level of protections in the design. That's what fucks most people up is the idea that variables aren't writing to memory. Somehow people can comprehend that interpreted languages aren't directly writing memory but not Rust's stuff.

>>59617884
Yeah, that was a bug. That bug was fixed. That's not an argument against Rust's design.

>>59617884
>unsafe code needn't be safe
No, I'm saying the guy, who wrote that, is a retard.

>>59617888
>An _implementation_ of the standard library is not part of the language.
No, but it should conform to the language specification. Obviously safe abstractions can fail (because they rely on inherently unsafe code in the implementation), thus Rust can never do what it promises to do; provide safeness guarantees.

>>59617904
What's it called when code fails to do the thing it's intending...? A bug, it was fixed.

>>59617901
Okay, good. Then we agree.

>>59617909
>What's it called when code fails to do the thing it's intending...? A bug, it was fixed.

Obnoxious Rustfags are going on and on in the daily programming threads about how C is unsafe and riddled with security flaws and buffer overflows because people write bugs in it.

I find it amusing that they would get so defensive over a bug in the reference implementation without seeing the irony.

https://www.nngroup.com/articles/response-times-3-important-limits/

>>59617904
You can formally prove the safety of any code labeled "unsafe".

>>59617933
Rust treats that as a bug
C treats that as a feature

>>59617933
That's exactly what C is though, unsafe code. Rust has unsafe code in the unsafe code ghetto where it will never impact your safe code. It might implode if the standard libarary is fucked up but no individual user code can do that without them explicitly telling it to be fucking unsafe. If pajeets could comprehend rust this is the language that they would be hired to write.

>>59616097
>What are you working on, /g/?
A raycasting game engine in Common Lisp.

>>59617947
>C treats that as a feature
Except it doesn't, any faulty implementation of C standard library would be called out as non-conforming.

>>59617944
The irony here is that this code was even labelled safe (the BTreeSet is a SAFE interface to unsafe implementation). So apparently, formal proofs are worth absolutely nothing.

>>59616248
>brazilians created Lua

They're 10/10 in my book

>>59617947
No, C treats it as undefined or unspecified behaviour.

>>59617949
>but no individual user code can do that without them explicitly telling it to be fucking unsafe.
Creating an instance of a BTreeSet is using a safe interface. It's NOT explicitly using unsafe code. Are you a fucking moron?

>get into IRC as angry C tards take /dpt/'s autism level to the maximum
>Ask about GUI development
>"hurr durr you need to understand the physcology first"
>"when it comes to psychology you basically have to study per usage, not cheap
and then you have the problem domain on top of that to map'
>"Human, Computer, Interaction: the field of study of how humans interact with computers (or machines/technology I suppose too)
it isn't a sub field, it is its own field of study, seperate from programming, psychology and social sciences"
Jesus christ I'm surrounded by autists this evening

>>59617964
>The irony here is that this code was even labelled safe
I'm not talking about Rust, though I wouldn't be surprised if the retards behind it fucked something up.
>formal proofs are worth absolutely nothing
Seems like you don't understand what a formal proof is. Also I'm pretty sure the rust compiler isn't formally verified.

>>59617973
Didn't even read the whole sentence, huh? That's literally what it takes to make something unsafe in Rust, a bug in the standard library's unsafe code. It was a bug, it was fixed.

>>59617973
>2017
>Undefined behavior
How about no?

>>59617964
It does.

>>59617987
Who are you quoting though?

>>59618000
Ur mum

>>59617996
>How about no?
How about yes.

Undefined behaviour means that the language is simple and easy to port. This is the reason why C is so immensely successful, several decades after its inception.

>>59618002
Just asked her and she didn't say that.

>>59618007
Undefined behaviors mean the behaviors are not defined and the language users are making a sorry excuse of """"portability""""

>>59617987
I don't follow, what did you ask about GUI development?

>>59617988
>Also I'm pretty sure the rust compiler isn't formally verified.
It isn't, but that doesn't stop Rust autists from attempting to formally verifying Rust standard library

https://kha.github.io/2016/07/22/formally-verifying-rusts-binary-search.html

Unittesting is a meme as it test your code, if your code fails and you cannot debug it might aswell hang yourself.
On the otherhand TryCatch helps you sort out/help users that makes stupid inputs.
Discuss

>>59618014
>Undefined behaviors mean the behaviors are not defined
Yes?

>and the language users are making a sorry excuse of """"portability""""
It's not the users, it's the creators.

Why is Java so loathed by C/C++ programmers?

>>59618022
Unit testing catches things that are not necessarily exceptions, unless you define exceptions for every possible not-quite-right value, in which case you might as well put the unit tests in to auto-build devops processes can take care of that for you.

>>59617988
>formally prove your code
>bit in RAM gets flipped by cosmic rays and your program crashes anyway

Yeah, formal proofs are so fucking useful in real world engineering issues.

>>59618032
I honestly kind of like the JVM, it's gotten pretty quick. Other than that though, I never got into any Java shit. Way too much boilerplate nonsense even compared to C++.

>>59618009
Keep asking

>>59618017
I didn't ask here, I asked in an IRC channel and a fucking autist ruined the conversation

I simply stated how I am more inclined to write small programs to apply my knowledge instead of reading more. He asked what are your end goal and I said I just want to make a desktop torrent client.
This was followed by 50 philosophical posts

>>59618032
Because Javafags actually have jobs

>>59618020
In what world is that supposed to be a bad thing?

>>59618022
Lack of Unit testing is one of the reasons why KDE is so buggy.

>>59618045
>I """""just""""" want to make a desktop torrent client
I don't think you understand how much work this entails to be done properly.

>>59618056
>In what world is that supposed to be a bad thing?
See >>59618043

Or in this case, the implementation of the standard library is buggy so even though your code is correct it still crashes.

>>59618070
Not nearly as much as creating your own X- window manager or making your own UI toolkit.

I am not trying to reinvent the wheel here, I just want to use the wheel for something useful

>>59618041
If you get the wrong value then there is something wrong with your code logic, then you spend 10minuts to fix that error instead of using 5mins to write the test to see the error and then 10mins to fix it. As the logic gets fixed ether way when why make the test at all?

>>59618073
To be fair ECC handles memory corrections. I personally think Rust's 1.0 release was way premature, by the time it's been around 1/10th the time C++ has, it'll probly be better. (Unless C++ implements borrow checking and stuff)

>>59618014
>sorry excuse of """"portability""""
That's a perfectly valid reason, though.
C not only wants to be be portable, it wants to be efficient.
Now imaging the situation where you want to support architectures with sign+magnitude, 1's compliment, and 2's compliment integers.
Considering the wildly different results that the max value + 1 gives, it would be completely infeasible for a language to specify something sane. Software emulation of 2's compliment is out, because we're trying to be efficient.
The only reasonable thing you can really do is say "fuck it, don't do that".

>>59618041
Only pajeets refuse to have good test coverage.

>>59618073
You do realize that post was probably a joke, right? Or are you really that retarded?
>the implementation of the standard library is buggy so even though your code is correct it still crashes.
Correct code can't crash by definition. If it crashes, then the retards making the compiler fucked up.

>>59618061
Care to explain? How would that help? IMO they just need fewer UX people and more propper programmers

>>59618084
>To be fair ECC handles memory corrections
Yeah, if you happen to run on a server. Most people have cheap desktop computers or laptops without ECC.

>>59618084
>Rust's 1.0 release was way premature
This. They wanted to grab some of the pie as Go has.

>>59618103
They have UX people??

>>59618081
>As the logic gets fixed ether way when why make the test at all?
Because a change could introduce the same issue later on, and a test written to catch this logic issue will catch it forever and not allow someone to go back and fuck up that code.

Get your mindset in the mode of >100 developers (where some of them might be shitty), rather than less than 10 developers that are all high-caliber.

>>59618092
>You do realize that post was probably a joke, right?
No, >>59618043 was sincere (I made it).

>Correct code can't crash by definition. If it crashes, then the retards making the compiler fucked up.
Which proves my point, code correctness is overrated and in no way a catch-all.

>>59618126
Except that it wasn't formally proven to be correct, it was just slapped together in a hurry (and fixed)

>>59618103
>Care to explain? How would that help?
Regression tests would have saved KDE4.

>>59618115
I am guessing, as a back-ender, if I had to make it, it would berly be basic WPF shit tier

>>59618103
KDE is fucking massive. It's very modular. However if you look at their sources, you will not find inconsistent version control system, some of the programs lack Travis CI, and some programs do not have unit testing.

For example Dolphin will share some of the modules of Ark. Dolphin has a very sophisticated Unit testing system, Ark has none.

Result: Developers around the globe have to test their shit on their own (if they actually do so) before publishing the programs.

Pretty sure turbo (((expert))) C/C++ kernel hackers don't need unit testing because they are unable to produce software bugs but in real world Unit testing is very important.

>>59618132
>Except that it wasn't formally proven to be correct
But binary search is, even though neither the compiler nor the library is.

>>59618147
>inconsistent
consistent*

>>59618149
I don't think that proof was completed yet either, was it? They're working on some proofs.

>>59618116
That is why the code mokeys make the shitty outline of the code, and the one that actually know what they are doing fixes it and refactors. At that point the mokeys should know to never touch that specific part again. Open for extention, closed for modificaion.

>>59618126
>I made it
In which case you have some kind of serious mental illness.
>Which proves my point
How does an impossible claim prove your point? You said correct code can somehow crash, which is by definition impossible.
>code correctness is overrated
What code correctness are you talking about? Rust isn't formally verified. The code its compiler produces isn't proven to be correct.

>>59618149
The """""""""""""proof""""""""""""" is worthless since the compiler itself isn't verified.

>>59618159
>I don't think that proof was completed yet either, was it?
It's not complete, because the underlying architecture can never be formally proved (because we exist in a nondiscrete physical world with quantum uncertainty that may cause a bit to get flipped). Hence my original statement, formal proofs are worth very little from an engineering point of view (aside from being an exercise in proof making).

>>59618166
Uh, yeah...let me just go stick a label to each employee and call them a monkey or a "know-what-you're-doing".

Then, I'll be able to guarantee that the 5% of programmers who are amazing programmers will never make a mistake, and this won't hamstring development time in the slightest!

>>59618178
Error correcting memory exists, taking that out takes out a system not the software.

>>59618147
So the moran of that story is that they never had working code to begin with? If the indiviual methods works there is no need for unit testing unless you plan on fucking up that part again. Testing if they work together is another thing completely and not even unittesting

>>59618174
>In which case you have some kind of serious mental illness.
Nice ad hominem attack.

>How does an impossible claim prove your point?
Reductio ad absurdum is a valid point.

>What code correctness are you talking about?
I'm making a general statement.

>Rust isn't formally verified.
I didn't say so, I'm making a general statement.

>The code its compiler produces isn't proven to be correct.
Any compiler output for an architecture other than a hypothetical Turing machine can never be proven to be correct. Hence my statement that formal proofs are useless when it comes to programming.

>>59618194
>Error correcting memory exists,
1) Not every computer uses ECC
2) ECC has a margin of error
3) Other errors exist, you might flip a bit when reading from the bus or your CPU floating point implementation may be buggy.

>>59618198
https://www.youtube.com/watch?v=S6mZGBEV7aY

File: tmp_20760-14905724194091824041501.gif (49KB, 200x210px) Image search: [Google]

49KB, 200x210px

>>59618086
It's nice to see intelligent people here.

>>59618188
"Open for extention, closed for modificaion."
Once a main part is made, if made well enough, the monkeys will use that to make code that does not work, but wont change the main part itself. Therefor there is no need to test the main part as that works, and no need to test the monkeys part as they will use longer time making the test that it would take to fix the problem.

>>59618223
Except UB is not relevant to that

>>59618240
Signed integer overflow is undefined behaviour in C.

>>59618227
I'm not sure if you're underage or just stupid.

You're assuming one or more of the following:
>one can identify "code monkeys" and still hire them anyway
>one is able to identify developers that are perfect, who will never make mistakes
>you can write the "main part" of something perfectly once and never touch it again ever
>you don't need to test any code that isn't this arbitrary cordoned off section
>you don't need to test this arbitrary cordoned off section either

>>59618214
So KDE is the new JS? '3'ed' party dpendicies all over the place. He does make a good point that you are using code from so many diffrent sources it can help in debugging if they fail. I might say that it would be more connen not for the dependendy to fail but more the inplmentation of it.

>>59618332
>'3'ed' party dpendicies all over the place.
What are you talking about?

File: config.png (127KB, 754x769px) Image search: [Google]

127KB, 754x769px

Hey guys, I'm running into errors with debugging on Visual Studio Code. It's throwing errors in 'launch.json'

Can anyone see anything wrong with my Config?

>>59618349
KDE has a lot of dependencies right? They are some from within and some are from other sources. As they cannot control all of these and some might break because of conflicting dependencies.

What's the best way to learn Prolog?

>>59618277
There si not enough 'great' programmers so yes they know when they are hireing someone not in the top 10%.
Ofc. no one is perfect, code can be perfect for acheving its perpose tho.
With a top down approch and OOP you can, the main part is just general enough that you can make your specific things under it, 1 tool for 1 job, dont edit it to make ti do 3.
If you use code from others and it has flaws then a) You should not use that code, find something else. or b) It is open source right? Fix it yourself.

What should I do to get all the object from a Map list in an ArrayList ?
The Map is initialised this way: Map<String, RunnableTask>
And in the arrayList I just want all the RunnableTask of the Map list.

>>59618406
Paradigms of AI Programming

>>59618819

for (Map.Entry<String, RunnableTask> e : your_map.entrySet())
{
  your_arraylist.add(e);
}

>>59618872
e.getValue, rather

File: 1429936002914.jpg (16KB, 291x291px) Image search: [Google]

16KB, 291x291px

ayyy how do I learn qt

>>59616533
What the fuck? Like he never understood why recursion worked and why it's useful? Or he just wasn't good at making up recursive functions?

Because I'm with him on #2 ;_;

>>59618944
You already asked this, and you already got an answer:
Become a loli.

unsafe is just a buzzword.
there is literally no difference between safe and unsafe code in the way that rust defines it

>>59618956
How is recursion so hard? You just define the base case (stopping point) then recursive case (loop otherwise).

>>59618944
Start with C++. As soon as you need something that's not in the standard library, look up for Qt classes.

>>59618985
>in the way that rust defines it
Then why even bother having the keywords?

>>59618944
Use Lisp.

>>59618872
>>59618881

ArrayList<RunnableTask> getTasksClasses() {
        Map<String, RunnableTask> beansOfType = appContext.getBeansOfType(RunnableTask.class);
        ArrayList<RunnableTask> nomClasses = new ArrayList();
        for(int i=0; i<beansOfType.size();i++){
            nomClasses.add(beansOfType.get(i));
        }
        return nomClasses;
    }

Well I have a weird issue.
beansOfType has all the objects I need but when I made an http request to get nomClasses, it gave me an array with th right number of object but they're all null.
where does this error might come from ?

If I want to write GUI app for desktop and mobile with ocaml what would be the best way to accomplish this?

>>59619084
>mobile
why

>>59617092
Wouldn't kill you. Some tips:
>first for loop could be foreach
>id could be created and initialised in one line
>you could
...Wait what the fuck is that last loop? It's not doing anything even...

Have you left something out for clarity? Is this bait?

>>59619105
Many people use mainly phones now days.

>>59616097
do you have to exclude the pivot of swapping when you quicksort?

In what language would you do the following application:

- Small desktop application GUI.
- Takes an input file, process it with an external application (would be better if I could implement the said ext app directly)
- From the output and from the given params, do some calculation and write them in a file

Done, quick and easy. I actually made it in C# before, when I was learning programming. But I want to rewrite it and why not choose another language for this.

What language would you go with? I'm willing to learn something new.

>>59619191
I'd do it all in Lisp.

>>59619210
Lisp is a social construct, its real code under the hood is assembly. C is more close to assembly.

>>59619233
Common Lisp compiles directly to assembly.

can someone post that chan browser in temple os

>>59619256
https://github.com/tramplersheikhs/uriel

>>59619004
Every recursive call litters a call stack, try to convert recursions to while loops, tail recursions are the simplest thing to convert. Recursion isn't a panacea, it is a social construct.

>>59619191
You could do this in any modern language that has GUI frameworks trivially. C# would have been my go-to for something like this, but you could do this in something like Python or Java easily enough.

Try making it web-enabled, or using something functional like F#.

>>59619065
bump.
Pretty sure the problem comes from the beansOfType.get(i) in my loop (I know you can't put the i there.
But whit what should I replace the i so it can put an object that is in the Map ?

>>59619276
Just convert all your recursion to tail-calls. Your language is deficient if it does not use TCO.

>>59619274
I'm interested in how they do the images.
Do they use img2txt or what?

>>59619084
Move to scala. Make GUI with javafx, backend with scala.

>>59618247
Finish your reasoning, you're almost at it. It's been repeated over and over why it's undefined behavior. Undefined behavior arises from implementation-defined behavior: signed integers are platform-specific, thus you can't predict what happens in case of overflow.
Of course, the behavior in case of overflow could have been implementation-defined too, meaning the behavior would be documented by implementators, but then you would have pretty as many possible behaviors as there would be implementations, and that's not the point of C. It's funny, because gcc has intrinsics to check for overflow in arithmetic operations.

>>59619299
>scala
eww

>>59619312
It removes A LOT of the java bloat. Come on, come to impure functional darkside. We have maps...

/g/uys, rec me some good C# books

>>59619325
https://www.youtube.com/watch?v=uiJycy6dFSQ

>>59619329
Practical Common Lisp

>>59619329
http://www.robmiles.com/c-yellow-book/

free PDF on page

10/10 if you're newish to programming in general, but you might skip a little into it if you've programmed before.

>>59619337
Thanks, i've been meaning to look into scala collections since I heard them discribed as "the longest suicide not in history".

>>59619337
Shit 45 mins? TLDR?

>>59619285
Why won't you use an iterator like the anon showed?

>>59619044
meme baiting.
its a meme language for meme people.
every piece of rust ive seen had unsafe every 3rd line. at that point the designers should ask themselves if maybe they fucked up somewhere

>>59619255
To machine code, actually.

>>59619368
Basic stuff in scala is not intuitive and there are bugs that will never be fixed.

>>59619210
>>59619340
Stop false-flagging, faggot.

Anyone know any good C libraries that will let me get an MP3 file in a sample buffer? I don't need to play the file, just get it's data

>>59619447
libsndfile

>>59619397
>adjusts glasses
ACKTUWALLY

im the guy who started daily prog threads on /g/ way back when

im disappointed to see that its now infested by faggots

now i remember why i left /g/ in the first place, eat shit and dont breed, fags

>>59619511
I don't like when people confuse assembly and machine code.

>>59619474
Doesn't support MP3

>>59619520
ask me how I know you have external sources of rage in your life, anon.

>>59619591
i dont ask anything of faggots, except that they eat excrement and never reproduce

>>59619004
>>59618956
>>59616533
I'm guessing he meant why/when you should use recursion, since the answer is almost never.

>>59619621
One of these days you are going to cut yourself on all that edge m8. If you were a real oldfag, you'd have grown out of this phase by now.

>>59616248
Isn't PUC Rio a Catholic university?

masturbating to trap shota mangas
feels gud man

File: china-fastest-superco_kuma759.jpg (36KB, 759x422px) Image search: [Google]

36KB, 759x422px

>>59616533
You've fallen for the /g/ meme.

What /g/ says you need to know to get a good job:

>Excellent programming skills
>Pure programming skills (can do anything in low level systems languages from scratch)
>Good Algo/Data Structure knowledge
>Know a functional Language
>Understand computational complexity
>Read SICP
>Essentially be a computer scientist

The reality. What you really need:

>People skills. You need to communicate well and work well in groups
>Knowledge of a practical language. This means one of three: Javascript, Java, C#
>Pre-calculus level math at best (trigonometry)
>Knowledge of common frameworks that you will end up gluing together for most of your life
>Probably gonna end up using some IDE w/ built in debugger
>Algos/data structures for INTERVIEWS ONLY. You will never need to implement data structures because your language has them built in, and you will never need to implement search algos because hash tables with buckets have replaced them.

The truth is that in your free time you will spend 10 hours a day straight programming something, but at your job you'll spend 2 hours in meetings to push 10 lines of code. #1 most useful skill is not being an autist.

>dig through math papers for a couple of days to solve a problem i was having
>find out a closed-form solution to my problem which according to math.stackexchange didn't exist
lol

>>59619798
I would be interested in the problem you were having. Could you tell more, please?

File: 2017-03-28-152550_93x95_scrot.png (1KB, 93x95px) Image search: [Google]

1KB, 93x95px

>computers in charge of handling basic arithmetic rules such as associativity

>>59619797
This post is absolutely the truth, generally speaking.

Yes, there are exceptions, but take it from someone who gets paid well and has a nice job programming: this is very accurate.

Worth noting that there's plenty of a market for Python, C++, and C, but the latter two are often going to require years of experience because you're fighting for positions with 30- and 40-somethings who have been using the languages for years and still want a job.

>>59619844
>you in charge of handling basic floating point arithmetic problems like inaccuracy

What new language should I learn? c++ is the only language i """know"""

>>59619891
graduate to c

>>59619797
You need both what you claim that /g/ says and "the reality" to be a good programmer

>>59619891
that's all you need to know, anyway.

File: french-girl.jpg (59KB, 540x960px) Image search: [Google]

59KB, 540x960px

>>59617079
>Working on my NVMe driver for CUDA.
For what purpose?

>>59619869
>holding the hand of my computer because he's too stupid

>>59619844
Who are you quoting, anon?

(let ((a 1)
      (b 1d100)
      (c -1d100))
  (values (+ (+ a b) c)
          (+ a (+ b c))))

0.0d0
1.0d0

>>59619934
>people who think the who are you quoting meme is funny

>>59619939
Who said this? I don't see "people who think the who are you quoting meme is funny" anywhere.

>>59619961
:))

File: mo0266lzj0my.jpg (49KB, 640x396px) Image search: [Google]

49KB, 640x396px

>>59619905
Being a good programmer is not a pre-requisite to getting a job

Who are you quoting?

>>59619967
It is to get a good job

>>59619974
Not really. I've seen plenty of people who can't code for shit with 90k/year salaries. Especially after you graduate from software engineering to leading teams and not doing a ton of actual coding yourself. Then you get the big bucks. In this case, having people skills and team leading skills is far more important than coding since you're likely to be replaced by a 22 year old eventually

>>59619994
>90k
>big bucks

>>59619994
90k is less than base salary for a good job

>>59619998
>>59620001
CoL comes into play here. It'd be pretty good starting pay for someone living in US midwest.

>>59620021
>midwest
There are like 3 development jobs in the midwest though.

>>59620021
You won't get a software engineer job at Google unless you're a pretty good programmer

>>59617079
How do I learn to be really good at CUDA and C/C++

>>59620031
Yeah, tell me about it.

>>59619817
i'll be a bit vague since i don't want to give it all way right now but i'll throw you a bone. it's to solve a specific definite integral and it works in a similar way as simpson's rule which is exact for 3rd-degree polynomials and also does approximate integration for other functions.

>>59619663
>If you were a real oldfag, you'd have grown out of this phase by now.

if you were a real faggot, you would have died of AIDS by now

>>59620052
I dunno, I learned C and CUDA in university. For C++, I picked up Efficient C++ and More Efficient C++ by Scott Meyers, but at that point I already had a pretty good grasp on C.

>>59619917
In order to access SSD directly from the GPU without involving the CPU. High parallelism combined with permanent storage for CUDA programs.

>>59620092
Do you do GPU programming for work?

>>59620198
Yes

>>59620206
>>59620198
Or, the correct answer to that is, sometimes.

>>59620206
Game development or machine learning or something else?

>>59620279
3D reconstruction

>>59620279
Pooping simulation to help Pajeets get accustomed to pooing in the loo

>>59620304
3D RECUSTRUCT MY ANUS!

>>59620323
>>>/prog/

is there any point in learning any of these fangled new javascript MVC/databinding frameworks if I'm just doing client side stuff?

I have a coding blog right now generated in jekyll. I like to embed code in the webpage that's editable and compiled live

Right now I'm using ace.js for the editor, and for simple one off fragment shaders it's workable but once I start trying to make a drag/drop shader graph the complexity increases and I get circular dependency issues

>>59620323
I'd rather not.

New thread:

>>59620334
>>59620334
>>59620334

http://paste.debian.net/924804/

Wrote a simple formula expander that applies distribute law of multiplication over addition. Currently rewriting it to make it cleaner and so that I can evaluate numbers.

>>59620341
It's dead

>>59620371
Don't remind me, I already suffer enough.

>>59619723
Yes it is.