this post sponsored by bonedad

Maybe he meant rivers of babylon

Moskau??

Bun bun bunsputin

Nightbun s

vip gitflow or tbd or what alternative do you recommend
we have no standards www might as well pick something
or maybe the best standard is not having one

Eh? We just use git, it does the things itself if you take the time to learn it.
w

Morning bunce

>>16486
and notably have someone who a large part of their job is wrangling git
because no automation makes // is as good a solution as someone just giving a shit

if you rebase your life is forefit

Sorry, I thought the purpose of git was rebasing and using git blame
I'll commit random crap, oh it doesn't work anymore, rebase, alright it works now, then carefully step forward til it's all working together
I guess my style isn't the best either but it's how I like to use git

horrifying
wwww I guess they drilled into my head the golden rule back in uni: "Don't commit code that doesn't work" wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww

👀
uhh y-yeah
I was never given a uni course on git
My coworkers don't understand git and still email me their code wwww

I mean, we didn't really have a course on git, there were just a few that had it as //learning it as a weeder.
>>16493
wwwwwwwwwwwwwwwwwwww
just, on it's own or like
as patches
which would be far more useful...

Usually just a zip file of the entire folder or sometimes, since it's 1 func per file, just the extra files and then the edited file.

>>16496
Matlab babyyyy

>1 func per file
The state of programming.

>>16495
Nothing but horror stories about that language wwww
everything is indexed wrong
amazing

I understand the benefit of matlab. The "oh the index starts at 1 and not 0" is such a stupid argument, who the hell cares about that
But things like matlab GUIs being unresponsive and having weird quirks that no one else does, or the licensing on various systems being a hassle... those are real issues I feel like

>>16499
I find both of those fairly mundane
Once matlab gets into callbacks and GUIs is where it falls apart
Using matlab just for quick math like matrix multiplication and then a nice plot of the data is where it shines big time
Global namespace is kinda annoying but not the end of the world

they start at 1 and go row-wise instead of column wise
awful

lots of weird quirks like that that really add up

Well, all of this is subjective www
It has benefits, it has downsides -- just like all languages :)
except rust

there are only downsides to rust
for loops are the only good thing man has ever invented, death to the slice cultists

wwwwwwwwwwww

Global namespace is fine, and where it isn't there are ample ways to deal with it, cracking headers or symbol level

Most times I do a slice loop, I realize later I also need the index and I really ought to have simply used a
for i in range(len(data))
instead of
for d in data
Just wish it was easier to type that range of len of

for(i=data[0];i!='\0';*i++)
undefeated
(do not do this)
>>16506
but what about ++i?
>>16506
...
1+3+4+4??
if it's 2 it's 2+4+5+5
although it depends on the compiler, I don't think it's actually in the spec that
preincrement returns at all

We get confused by i++ or i += 1 so we have to use i = i + 1

Okay vip vip

i = 1
x = i++ + ++i + ++i + i++
We could not fucking understand why it was the way it was
We had a thought and then we set
i = 2
and then it fell apart
Or we swapped the i++ vs ++i around and always failed
We spent like 2 hours on this (very productive, I know)
I dont ghave a computer so I cant verify but it was dumb

hmmmm well
Oh well

always remember
the compiler is your enemy
https://blog.cr.yp.to/20240803-clang.html

>>16486
trunk based development or gitflow i mean
how to handle pull requests
when there are 10 pull requests

there are only downsides to rust

>>16508
I mean, TBD is the closest to what you should be doing
there's not really any call to have long-lived version branches unless you CADT some whole subsystem or something wwwwwww
it's how git was originally envisioned! The whole reason there aren't a bunch of the old SVN 'features' www
although directory pulling is missed...........

i agree but then reviews take too long
but i also dont like the idea of having a master branch a development branch and smaller branches that are merged to develop
why overcomplicate

less merge = better
Never found reviews too much of a burden but I guess CI linting/builders helps with that
github makes you pay for that stuff though (it's not really hard to set up on your own git host for free though, but our company is allergic to not spending money)

>>16511
not a burden but more like
no one else really bothers
shit gets approved immediately

ok then my initial assumption was right
my assumptions and educated guesses are right 99% of the time
good thing i don't have any say in what we do

Warning: This function has 27 cyclomatic complexity which is above the allowed amount (10)
i fix ^ this shit instead
Hell

yeah, that's pretty universal, no one likes reading other people's code, and it doesn't show up well on metrics
that's the whole reason for those stupid code quality linters and the issue checkboxes every project has for PRs
>>16512
good god damn
>>16514
Make a startup together
inevitable failure in 3 months due 2 gaysex

you should become our lead dev instead
based on our minecraft gameplay we work well together

Cks Co.

I am a straight white male with 0 fetishes now
too bad!
do i count as white? maybe not even white

we have all the filters enabled

they just put the code in smaller functions

50 nested ifs? no problem, just put them all in a separate function so the quality checks think it's fine

Blade Runner I've seen things speech but it's me describing our codebase

wwww get in charge of the git and put a bunch of code quality filters that auto-reject everyone else's code
>>16514
pfffffffffffffffffffffffffffffft
straight as an inner-city rail line and as few fetishes as a catholic neighborhood

wwww maximum stack depth: 3
I get that they suck in Java, something something enums again, but good god why are people allergic to switch()

>>16515
i saw something like

if(){
if() {
if ( && && && && && ) {
if ( && && || && ) {
switch
case:
thing ? a : b
case:
thing2 ? a : c
}
}
}
}

>>16516
Yeah, that's a pretty standard Java pattern
The old minecraft source code pretty much enforces this kind of thing, fucking booleans. What was wrong with <<2 based enums?
Way more degenerate than aforementioned fetishes wwww
>>16519
But early returns are bad! Ignoring that that only applies to things which require constant time execution like password compare, and that if you do that you also need to ensure constant time execution...

at least it's not
try{
if (!bad){...}
else {throw new e}
} catch (e) {
return;
}

add 50 more ifs

>>16517
wwwwwwwwwwwwww

they are allergic to this design pattern:

if (bad)
return

/* thing when not bad */

instead they do

if (!bad) {
do shit
else
do nothing? do useless shit?

>>16517
>at least it's not
it's exactly that

pitching and catching being harmed by this try except shit
microsof teams worst program
at least it's not slack

i want to quit

wow me2

although I might have a chance to ascend to the plane of pet engineer that just goes to conferences all the time next month and doesn't have to actually use the tools anymore wwww
barring that just need to get my payout and run

wwwww hopefully get a job that gives me EU citizenship to do less annoying stuff

Kigu coding startup

if statements are complex
I always hafta draw state diagrams with testing nested if statements

>>16527
>>>/watch?v=m_Dp4nMr_AY

if(BooleanFactory<Boolean>.getFactoryInstance().createBoolean(new BooleanFactoryInfo(false)).testIsFalse((Boolean)caseObject.getBooleanValue(false))) || ...

>>16526
horriffic yet again

force everyone to make DRAKON state diagrams and mail them for approval before being allowed to write a line of code
www

wwwwwwwwww

I like matrices and matrix multiplication
These if statements and booleans are too complicated for meeeee
Just gimme the raw data

some day OpenCL will actually be something that can be used
wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww I kid
matrix SIMD and limited MIMD will always be locked down to NVIDIA cards with special drivers that have shit performance

I've been told GPUs have gotten much more useable for us in the past 20 years
Well, like, before GPUs were good at very large matrix multis but not many at the same time, like is done in graphics processing.
But now GPUs are powerful at small matrix multis but a lot repeatedly, which makes ML and BF very nice. :)

wwwwwww
a lot better at tessellation and fragment shading, which if you were a smart and clever programmer would make a lot of interesting possibilities and new ways to...
OR THEY COULD JUST FILL THE IO BUS WITH DMMS TO TRY AND BRUTE FORCE IT THE OLD WAY WOW
wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww
seriously, all the ML/AI GPUs are kinda shitty GPUs but with a bunch of fast flash memory on the IO port instead of making programmers grapple with hardware locality wwwww
why maintain state in the GPU when you can just PRETEND everything isn't IO bound!

I'm not that knowledgable about this honestly.
We have a hefty set up for beamforming and they all discuss the issue with IO, but I've never personally dealt with jt.
it*

well, 'problem with IO', the real issue is with data segmentation wwww
if you can cut up a problem into a bunch of sequential steps you can just pipeline and buffer swap a bunch and then get the final result

As my boss calls it "ping pong" the data :)

w

Second Cluddy

Buuuuuun
are you watching the gay sex with girls anime

Night boons

ok
gotta run, uh, be careful

WHY IS IT TRYING TO LOAD THE POSTS
ALL FROM ~SOMEWHERE~

A

ok, I think it works now
if you say something dumb and get autobanned DM me somewhere, but I'll look through it often as I can

night buns

Mornbuns

plop the fuck up
>>16575
no u

>>16567
faggot

My life is in shambles ;_;
nooooooooooo

wow
the price of faggotification

once you turn fag, you never turn back

That's a well known fact, inabun has never recovered.

hmm
not quite sure why redis.del() doesn't delete the thread...

e

Clud 2 clud

Night uns