terrible programmers: my god, it's full of chars

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > terrible programmers: my god, it's full of chars

«‹›1446 »

Venuz Patrol: Mar 27, 2011

fleshweasel posted:

c tp s: can someone help me

just use your hand

# ? Dec 3, 2016 08:35

Adbot: ADBOT LOVES YOU

# ? May 23, 2024 04:01

tef: May 30, 2004; -> some l-system crap ->

Bloody posted:

nah not super noice they generally wind up not hiring me lmao

me in an interview

"we want you to write a function that returns all possible combinations of a list of lists, any language"

so like, any? echo {a,b,c}{d,e,f} in bash stuff

"yes any language"

ok, here's it in prolog: comb([],[]). comb([H|T],[X|O]) :- member(X, H), comb(T,O).

i only found out later that one of them left the interview and asked around the office to find someone who knew prolog to check it

# ? Dec 3, 2016 09:01

tef: May 30, 2004; -> some l-system crap ->

"can u reverse a linked list"

ok here it is in python using tuples, i think the free-list + refcounting will ensure that it's effectively fixed memory usage

"what"

no seriously, ugh do you want it in another language

"yes"

fine here is a java one

# ? Dec 3, 2016 09:03

fart simpson: Jul 2, 2005; DEATH TO AMERICA

tef posted:

me in an interview

"we want you to write a function that returns all possible combinations of a list of lists, any language"

so like, any? echo {a,b,c}{d,e,f} in bash stuff

"yes any language"

ok, here's it in prolog: comb([],[]). comb([H|T],[X|O]) :- member(X, H), comb(T,O).

i only found out later that one of them left the interview and asked around the office to find someone who knew prolog to check it

lol

# ? Dec 3, 2016 09:09

anthonypants: May 6, 2007; by Nyc_Tattoo; Dinosaur Gum

tef posted:

me in an interview

"we want you to write a function that returns all possible combinations of a list of lists, any language"

so like, any? echo {a,b,c}{d,e,f} in bash stuff

"yes any language"

ok, here's it in prolog: comb([],[]). comb([H|T],[X|O]) :- member(X, H), comb(T,O).

i only found out later that one of them left the interview and asked around the office to find someone who knew prolog to check it

i'm going to guess from their reaction that you did not get the job, or hopefully that you did not accept that job

# ? Dec 3, 2016 09:30

eschaton: Mar 7, 2007; Don't you just hate when you wind up in a store with people who are in a socioeconomic class that is pretty obviously about two levels lower than your own?

ooh! ooh!

now do APL!

edit: Rosetta code to the rescue:

pre:

A[I]←1+I←(0⍷A)/⍳⍴A←('FIZZBUZZ' 'FIZZ� 'BUZZ' 0)[2⊥��(⊂3 5)|�1+⍳100]

eschaton fucked around with this message at 10:40 on Dec 3, 2016

# ? Dec 3, 2016 10:35

Xarn: Jun 26, 2015

c Java s: Netbeans just told me that I will have an infinite recursion and it was right. If it keeps being helpful, I might actually start liking it. :v:

# ? Dec 3, 2016 12:44

redleader: Aug 18, 2005; Engage according to operational parameters

recursion? uh, this is the terrible programmer thread

# ? Dec 3, 2016 12:56

Soricidus: Oct 21, 2010; freedom-hating statist shill

eschaton posted:

no see this should tell you that you're doing something entirely wrong by even thinking of using popen

this is kind of my point. you want to do a high level thing in c, and there's a convenient, obvious, high level function that does almost what you want, but not quite, and also it's got other serious flaws that mean you shouldn't use it anyway. so you're left flailing around in a sea of hard-to-use primitives, each of which introduces more exciting gotchas.

# ? Dec 3, 2016 13:14

tef: May 30, 2004; -> some l-system crap ->

anthonypants posted:

i'm going to guess from their reaction that you did not get the job, or hopefully that you did not accept that job

i needed money and worked there until i was made redundant after 11 months, during that time i had seven different managers and three written warnings (all dismissed)

http://thedailywtf.com/articles/The-Difference-Between-Better-and-Less-Bad may run very close to the original environment for screen scraping. i wrote a python replacement and somehow convinced a MSSQL/C#/Windows shop to deploy a linux/python web service

they still use my code, or at least in a ship of theseus way, almost 9 years on

# ? Dec 3, 2016 18:00

Arcsech: Aug 5, 2008

what does "key" mean in sql/innodb? like, not "primary key" or "foreign key", just "key". is it just an index?

specifically the ones here

this is loving impossible to google because of everything talking about primary or foreign or whatever keys

# ? Dec 3, 2016 18:07

Deep Dish Fuckfest: Sep 6, 2006; Advanced
Computer Touching; Toilet Rascal

from the mysql doc for the create table syntax

KEY | INDEX

KEY is normally a synonym for INDEX. The key attribute PRIMARY KEY can also be specified as just KEY when given in a column definition. This was implemented for compatibility with other database systems.

so yeah it seems it's just another way to say index. "normally" anyway, whatever that means

# ? Dec 3, 2016 18:14

Bloody: Mar 3, 2013

how do I find remote contract work I want to make side cash

# ? Dec 3, 2016 19:50

Luigi Thirty: Apr 30, 2006; Emergency confection port.

how is it this hard to save a drat 48x32 rectangle into a memory buffer and load it back again each frame with the Amiga blitter aaag

# ? Dec 3, 2016 22:24

Xarn: Jun 26, 2015

c Java s: Java enums are sweet. Not being able to make ArrayList<int> less so. Having to reason about this mess

Java code:

ArrayList<ArrayList<ArrayList<Integer>>> butts = new ArrayList<>();
ArrayList<ArrayList<ArrayList<Poop>>> poops = new ArrayList<>();
ArrayList<ArrayList<Integer>> turds = new ArrayList<>();

sucks rear end. Especially since the same index is used to index each of them at different places. :suicide:

# ? Dec 3, 2016 22:44

Soricidus: Oct 21, 2010; freedom-hating statist shill

Xarn posted:

c Java s: Java enums are sweet. Not being able to make ArrayList<int> less so. Having to reason about this mess
Java code:
ArrayList<ArrayList<ArrayList<Integer>>> butts = new ArrayList<>();
ArrayList<ArrayList<ArrayList<Poop>>> poops = new ArrayList<>();
ArrayList<ArrayList<Integer>> turds = new ArrayList<>();
sucks rear end. Especially since the same index is used to index each of them at different places.

ouch. I generally take deeply nested collections like this and write a goddamn class that represents the actual thing you're working with instead of the data structure used to store it. it's not 1996 any more, adding another class isn't going to hurt performance.

# ? Dec 3, 2016 22:51

fritz: Jul 26, 2003

redleader posted:

recursion? uh, this is the terrible programmer thread

'doing a recursive implementation when there's a perfectly good imperative solution' is primo terrible programmer territory

# ? Dec 3, 2016 23:22

fritz: Jul 26, 2003

eschaton posted:

I still don't understand how Python has taken off here rather than Maxima

maxima's for symbolic manipulation but these eigenvectors aren't going to find themselves

# ? Dec 3, 2016 23:25

ComradeCosmobot: Dec 4, 2004; USPOL July

Xarn posted:

c Java s: Java enums are sweet. Not being able to make ArrayList<int> less so. Having to reason about this mess
Java code:
ArrayList<ArrayList<ArrayList<Integer>>> butts = new ArrayList<>();
ArrayList<ArrayList<ArrayList<Poop>>> poops = new ArrayList<>();
ArrayList<ArrayList<Integer>> turds = new ArrayList<>();
sucks rear end. Especially since the same index is used to index each of them at different places.

isnt type erasure fun???

# ? Dec 4, 2016 00:58

The MUMPSorceress: Jan 6, 2012; ^SHTPSTS; Gary’s Answer

fritz posted:

'doing a recursive implementation when there's a perfectly good imperative solution' is primo terrible programmer territory

the recursive solution is sometimes easier to reason about and if your language has a non-lovely compiler they will get optimized into the same thing anyway.

# ? Dec 4, 2016 01:30

Jabor: Jul 16, 2010; #1 Loser at SpaceChem

LeftistMuslimObama posted:

the recursive solution is sometimes easier to reason about and if your language has a non-lovely compiler they will get optimized into the same thing anyway.

well, it will if it's tail-recursive. i don't know of many compilers that will transform a recursive algorithm into an iterative one with an explicit stack.

# ? Dec 4, 2016 02:22

kitten emergency: Jan 13, 2008; get meow this wack-ass crystal prison

i optimized my tail calls by putting your moms number on speed dial

# ? Dec 4, 2016 06:39

FamDav: Mar 29, 2008

uncurable mlady posted:

i optimized my tail calls by putting your moms number on speed dial

quoting a post from 2014

# ? Dec 4, 2016 06:50

Xarn: Jun 26, 2015

LeftistMuslimObama posted:

the recursive solution is sometimes easier to reason about and if your language has a non-lovely compiler they will get optimized into the same thing anyway.

If its able to optimize general recursion into iteration, thats some mighty fine compiler you've got there.

I needed to generate all bitmasks of length N with exactly k bits set.

# ? Dec 4, 2016 08:15

Mao Zedong Thot: Oct 16, 2008

uncurable mlady posted:

i optimized my tail calls by putting your moms number on speed dial

# ? Dec 4, 2016 08:26

Xarn: Jun 26, 2015

Soricidus posted:

ouch. I generally take deeply nested collections like this and write a goddamn class that represents the actual thing you're working with instead of the data structure used to store it. it's not 1996 any more, adding another class isn't going to hurt performance.

Thats what I am doing, but first I gotta unravel this mess.

--- edit ---
There are also 5 more similar collections in that class and that is after I split the original monster class into 3 logical units that aren't overly interconnected. :v:

Xarn fucked around with this message at 09:12 on Dec 4, 2016

# ? Dec 4, 2016 09:03

redleader: Aug 18, 2005; Engage according to operational parameters

Xarn posted:

Thats what I am doing, but first I gotta unravel this mess.

--- edit ---
There are also 5 more similar collections in that class and that is after I split the original monster class into 3 logical units that aren't overly interconnected.

how long until you get pulled up for spending time refactoring instead of hacking new features x, y and z in?

# ? Dec 4, 2016 10:17

redleader: Aug 18, 2005; Engage according to operational parameters

c++'s const correctness stuff seems pretty cool on the surface

# ? Dec 4, 2016 11:27

Xarn: Jun 26, 2015

redleader posted:

c++'s const correctness stuff seems pretty cool on the surface

Yup. Its infectious though, which can make introducing it in existing codebase painful.

Its also a massive pain to provide all the right overloads without using const cast, if the logic is nontrivial:

C++ code:

struct butt {
    void shart() & { ... }
    void shart() && { ... }
    void shart() const & { ... }
    void shart() const && { ... } // this one is basically nonsensical, but :shrug:
};

And as a bonus, const means that the object is either read-only, or just internally synchronized (ie it uses mutex to synchronize with mutating accesses).

# ? Dec 4, 2016 12:26

NihilCredo: Jun 6, 2011; iram omni possibili modo preme:
plus una illa te diffamabit, quam multæ virtutes commendabunt

FamDav posted:

quoting a post from 2014

mods never made it thread title... a shameful mods

# ? Dec 4, 2016 15:35

Xarn: Jun 26, 2015

c Java s: Erased generics can still go suck a dick.

# ? Dec 4, 2016 15:47

Illusive Fuck Man: Jul 5, 2004; RIP John McCain feel better xoxo ð ð; Taco Defender

whats the point of lvalue reference qualified member functions?

# ? Dec 4, 2016 15:58

Xarn: Jun 26, 2015

Illusive gently caress Man posted:

whats the point of lvalue reference qualified member functions?

So you can specify different behaviour for lvalues vs rvalues. :shrug:

In reality there are some use cases, but so far this is quite obscure, so I don't have any examples of people using this in their libraries handy.

# ? Dec 4, 2016 16:40

tef: May 30, 2004; -> some l-system crap ->

LeftistMuslimObama posted:

the recursive solution is sometimes easier to reason about and if your language has a non-lovely compiler they will get optimized into the same thing anyway.

lol

# ? Dec 4, 2016 16:52

Bloody: Mar 3, 2013

worrying about recursion compilation at the whiteboard is premature optimization

# ? Dec 4, 2016 17:15

Jerry Bindle: May 16, 2003

i have found that recursion in programming is great for programs about mathematics like induction with peano numbers and as an idiomatic way to process lists in scheme-like langs. one of the lambda papers notes that there is an isomorphism between peano numbers and s-exps, which explains why it feels natural to program that way in scheme. racket has better and more powerful ways to process lists that are named like they are imperitive things like for/list, for/set, the syntaxes aren't explicitly recursive but they might be macros over a recursive something. but thats my idiot spare time opinion, at work i use java and have only found one use-case for recursion: linearizing a list of javafx tree items entries. the tree represents a ui menu, so the assumption is that the walk won't take a significant portion of the heap.

# ? Dec 4, 2016 18:21

brap: Aug 23, 2004; Grimey Drawer

tps: what am I misunderstanding about std::thread such that running my workload with 1024 threads is faster than running it with 2 threads on a dual core machine? it's just computing a mandelbrot set by dividing up rows of pixels to compute amongst threads.

# ? Dec 4, 2016 19:46

Xarn: Jun 26, 2015

fleshweasel posted:

tps: what am I misunderstanding about std::thread such that running my workload with 1024 threads is faster than running it with 2 threads on a dual core machine? it's just computing a mandelbrot set by dividing up rows of pixels to compute amongst threads.

More info needed, there is nothing weird in std::thread and Mandelbrot set an IO bound task aint.

# ? Dec 4, 2016 20:06

brap: Aug 23, 2004; Grimey Drawer

I'm representing the bitmap with a std::vector<std::vector<int>> which could be bad. I pass a pointer to the vector in with the thread start arguments for each thread as well as the index into the outer vector that the thread should start inserting into the inner vector. I ran it with Instruments on macOS which made it look like half of my time is being spent in cabs though (absolute value of a complex number).

I feel like there's probably something really stupid going on causing threads to block each other or something. Running it with 2 threads is actually slower than my single threaded implementation without std::thread.

posting code because why the gently caress not

code:

#include <iostream>
#include <complex>
#include <vector>
#include <chrono>
#include <thread>
#include <SDL/SDL.h>
#include "sdl_helpers.h"
#include "mandelbrot_common.h"

struct thread_params
{
	std::vector<std::vector<int>> *set;
	int min_x;
	int max_x;
	int width;
	int height;
};

void mandelbrot_thread(thread_params *p)
{
	for (int i = p->min_x; i < p->max_x; i++) {
		std::vector<int> column(p->height);
		for (int j = 0; j < p->height; j++) {
			double x_norm = (double)i / p->width;
			double y_norm = (double)j / p->height;
			int iters = mandelbrot(std::complex<double>(x_norm, y_norm));
			column[j] = iters;
		}
		(*p->set)[i] = column;
	}
}

std::vector<std::vector<int>> mandelbrot_set(int width, int height)
{
	constexpr int num_threads = 1024;
	std::cout << "Number of threads: " << num_threads << "\n";

	auto set = new std::vector<std::vector<int>>(width);
	std::vector<std::thread*> threads;

	int slice = width / num_threads;
	for (int i = 0; i < num_threads; i++) {
		struct thread_params *p = new struct thread_params;
		p->set = set;
		p->min_x = slice * i;
		p->max_x = slice * (i + 1);
		p->width = width;
		p->height = height;

		std::thread *t = new std::thread(mandelbrot_thread, p);
		threads.push_back(t);
	}
	for (std::thread *t : threads) {
		t->join();
		delete t;
	}

	return *set;
}

int main(int argc, char **argv)
{
	std::chrono::system_clock clock;

	auto before = clock.now();
	auto set = mandelbrot_set(MAX_X, MAX_Y);
	auto after = clock.now();
	std::chrono::duration<double> diff = after - before;
	render_mandelbrot(set, diff.count());

	return 0;
}

# ? Dec 4, 2016 20:22

Adbot: ADBOT LOVES YOU

# ? May 23, 2024 04:01

Blinkz0rz: May 27, 2001; MY CONTEMPT FOR MY OWN EMPLOYEES IS ONLY MATCHED BY MY LOVE FOR TOM BRADY'S SWEATY MAGA BALLS

cases in which you need recursion:

traversal
mathy poo poo
????

# ? Dec 4, 2016 20:23

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > terrible programmers: my god, it's full of chars

«‹›1446 »