Teeworlds - general

Teeworlds

discord.gg/teeworlds / general

Teeworlds Discord Server.

Between 2020-06-19 00:00:00Z and 2020-06-20 00:00:00Z

the demo recording of the race will of course be toggleable?

Image attachment

yes

ah. so a client side option. nice

00:07

oh, and would it record only successful attempts, or all from beginning to end/kill?

good question

German joined the server. 2020-06-19 05:01:57Z

Hiya @German - nice to meet you

Deleted User joined the server. 2020-06-19 07:25:02Z

BulgarianTee joined the server. 2020-06-19 07:25:51Z

hi

hello c:

hello, question from a beginner coder what's the reasoning behind using else if here and not switch()? (edited)

Image attachment

none probably, it's not a big deal

Well it is possible that you might want to declare something in the block, which would not be allowed in a case

12:22

there are several things you are not allowed to jump over

@Learath2 you can just { } to declare things in a case block though

switch also has a bit weird semantic where fallthrough is the default. I guess a switch could also be used there

yeah, I don't often see a case where switch is significantly better performing or readable. maybe it's just me

I love switch fallthrus

they should be explicit, they are not, unfortunately

I spent a long time writing explicit fallthroughs because modern gcc freaks out over them

12:28

(rightfully imo)

yeah probably should have had a fallthru; instead of break

there is [[fallthrough]] or so iirc?

non-standard

standard in some C++xy

TIL

https://en.cppreference.com/w/cpp/language/attributes/fallthrough

12:29

C++17

and required for gcc not to shit walls of warnings

Well not in C mode

but tbh any code construct that makes programmers go "oh isn't there a bug there? ah... maybe not" isn't great

It's natural selection

12:30

survival of the keenest

at this point just use gotos, at least they name their labels

I love gotos aswell

12:32

I love all the "controversial" features

12:32

They all have great uses ruined by people that think they know better than you

12:34

I've seen people advocate for do { ...; break; ...; } while(false); instead of a goto because goto bad

12:34

that do-while has clearer RAII semantics to me than goto into/from blocks

At the cost of not being readable at all

12:37

Besides, you can read on the semantics if you have concerns, just like you learned the semantics of break

if the semantics aren't obvious or often used, I and other people will get them wrong

12:38

see e.g. for/else in python

The solution is to reintroduce these lost features to the community. Not pretend they dont exist

12:42

Honestly when I see a feature of C++ I'm not familiar with I open up a reference. Honestly I think this is also a skill missing from most new developers. As soon as they think they start to understand a language they think their training ends

12:43

The most important thing to learn is the ability to refer to authoritative resources on any subject you learn really

@heinrich5991 for else is a thing? til

@Dune yes, and it's semantics are non-intuitive at first

12:45

I predicted something completely different than what it actually is

12:46

its*

12:46

now I did it as well @Learath2, it's contagious

I got an educated guess at what it is, but I wouldnt rely on it at all

12:46

If I see a for else, I'm opening up the documentation, no question

12:47

I don't think I'll ever get rid of the habit of saying it's when I should say its

my guess was: it's executed when the for loop has no iterations

12:48

it really was: it's executed when the loop isn't broken out of

I thought it was the exact opposite of what it actually is

12:49

Executed when a loop is broken out of as that'd help with a common pattern

12:49

Alas, I wouldn't ever assume

when you read code for the first time, you can't look everything up

12:50

read a code base

12:50

you have to assume some stuff. the less unintuitive features/names there are, the better

12:50

e.g. I had to work on a huuge java project for university

@Learath2 why do you insist on authoritative sources? it's not like programming is such a political topic where anyone would want to feed you wrong information

I definitely couldn't look up what all the functions do

Why not? I don't read code in languages I don't know. And if I'm reading one in a language I'm not familiar with I always have documentation open

12:51

E.g. while browsing the ghidra codebase

12:52

@Dune there are a whole lot of bad resources out there, especially for C which I'm most familiar with

it looked hopeless, trying to understand everything

12:52

I still don't know how java annotations work internally, e.g.

Tutorialspoint saying arrays decay into pointers has us answering 10 questions a month

but they work in some sane way that I didn't have to look up

I don't read code in languages I don't know.

doesn't feel like a viable policy. e.g. I'm working on a C++ project, and it uses perl for testing, etc. sometimes you have to write languages you do'nt know

12:53

@Learath2 bad as in factually inaccurate or bad advice?

I've seen both

12:54

@Dune honestly, I wouldnt dare ship anything in a language I dont know

12:55

For personally toying around with a project sure

Not everything is critical

But I'd still have perl documentation up on my second screen

12:57

@Dune would you contribute to a go project if you dont know atleast a bit of go?

I would try to if I need to fix a bug that's affecting me

yeah ^

12:58

usually it's more that part of it is in go

12:58

e..g part of teeworlds is in python

12:58

part of it is in cmake slang too

12:58

gotta hack in both

I had to add sha support to libtw2 in rust, a language I dont know much about. I didnt even consider contributing it back to @heinrich5991's repo

adding new features is maybe a bit different

12:59

I'd argue a dirty bug fix is better that no bug fix, that does not hold with features in general

Besides, if you saw something like a => in js, wouldn't you look at documentation to see what it means?

13:00

I don't see how it'd be any different for for/else in python

don't we write for languages we don't know much of every day? I mean I write shell scripts, makefiles, cmakefiles, and I only know a bit of those

Even though I had a guess, I've never even seen anything similar in any other language. I really don't think I'd go with my assumption

That's true

13:01

What if C++ had a similar thing? Wouldn't you just assume the semantics?

Probably

13:03

I make it a point to defer to people who do know things as much as I can, I do know a bit of cmake but I ask the person who actually knows when I don't understand how a construct works

13:03

or just dig into the cmake docs, which are surprisingly decent actually

13:04

Actually, I'd like to weaken my position with "authoritative" sources, although I believe that is a very useful skill, I think the subset of it with referring to "known good" sources is also very important

13:05

Like cppreference for C/C++, it's actually quite useful

I just feel like for most applications, copy pasting anthony1234's stack overflow comment and checking that it works is safe enough

That I don't think I like, people really shouldn't refer to stackoverflow so much

I don't trust it, but sometimes any solution that works is a good solution

I've seen so many flat out wrong answers in stackoverflow especially when it comes to C and JS

I mean, if it's something that you can test easily it's fine

I guess yeah, when you want to hack together something, but when you are contributing to a project, I don't think you should be copy pasting stuff from stackoverflow and not even understanding it

if it's a hash function that could be wrong in a corner case, okay it's not fine

Like e.g. yesterday we were trying to help a guy trying to contribute to radare2, he was calling a function in a way that didn't make sense at all, and he had written the function which baffled me

13:08

Turns out he'd just copied over the implementation of the function from a stackoverflow answer and didn't really know how to call it

well reverse engineering tools are particularly sensitive because they make so many hypothesis and wrong results aren't immediately obvious, right

I mean even if it wasn't radare2, would you be inclined to trust someone that contributes a stackoverflow copypasta to teeworlds and can't even figure out how to call the function he stole?

right, no

Okay, I think my position on stackoverflow answers is, only okay if you actually understand what they do and it's not critical

1

what about if you understand what they do or you somewhat understand it and it's not critical?

you as in me or the guy who copied it over?

okay if …

13:14

but I guess you shouldn't copy it anyway due to licensing concerns

13:14

you can take the idea

Hm I think "somewhat understand" is where the most danger lies because you start to trust it too much

13:15

I've been bitten in the back a couple times thinking I understood an algorithm

Anyone has a link to that old teeworlds headless client?

@ChillerDragon

17:19

?

Exported 125 message(s)