Thread 106201862

Which is worse: pip or conda?
Or both?

Just use poetry.

Oh, and pipx for all your CLI tools.

imagine ur packaging system ruining the entire traction of your language lmao!

>>106201862 (OP)
pip is more convenient than building libraries from source in c++

>>106201862 (OP)
Imagine how excellent this language would be if it was actually batteries included instead of every little tiny thing being shoved into
>pip
>pipx
>conda
>uv
or any of another dozen of other retarded package managers!

>>106201862 (OP)
Fuck Conda bloated piece of shit.

>>106202203
external packages have advantages. Imagine having even more breakage with every python upgrade. Much better to just pin dependencies and save a ton of work.
Also, pytest is much better than what default unittest module

>>106201862 (OP)
Who cares? Python is a dalit lang.

python
more like scheithon (scheiße is shit in german)

>>106201862 (OP)
Both are shit. Use uv

>JUST MAKE ANOTHER ENVIRONMENT BRO! ARE YOU AFTRAID OF A LITTLE BLOAT? IT'S ONLY 5GB PER PYTORCH WHEEL
fucking aids

Why the fuck does anything python related break HARD even on a minor python version update? I think most of it is because it intentionally does explicit version checks and just goes legs up if it changes.

>>106201862 (OP)
Why can't they fucking fix it, just break the compatibility and make some npm/cargo-like package manager

>>106201862 (OP)
I have only ever used pip with requirements.txt I don't even know what the other things do

>>106203264
kek

>>106202203
>>106202176
>>106201862 (OP)
>pip install numpy
>(optional) pip install scipy
>never touch again
ez

what seems to be the issue, saars?

>>106204896
conda is like bizarro pip for itoddlers that pollutes your environment variables and downloads gigabytes of useless shit

it's used by lowcode biologer roasties and such to make horrible 3000 line spaghetti scripts using a gazillion one-liner packages wrapping some basic scipy functionality

why waste your energy hating?

>>106206636
It's nice to vent sometimes.

>>106201862 (OP)
Conda, everything about Windows' slow ass IO will be felt with extreme intensity when touching this thing.

>>106206625
Also AI shit.

>>106206636
Need my anime girls slop.

>>106207124
I work with AI a lot and I not only do not use conda, I forbid anyone in my lab from even touching it.

>>106202203
debian ruined python because they made python packages installable only through the distro package manager but didn't include 80% of them.
yes i know you can use virtual env, but why?

Why can’t I download a sus exe and just double click it? Or some equivalent of a bat file?

Why must I jump through hoops to run anything python?

Why can’t I download a sus exe and just double click it? Or some equivalent of a bat file?
Why must I jump through hoops to run anything python??

>>106201862 (OP)
This language is only useful for retards and people that don't know how to code, even coding in Javascript has more dignity than this crap. My job is translating python notebooks made by boomers into something more maintainable, and I can assure you the userbase of this language is just basedboys, niggers and boomers

>>106201862 (OP)
Package managers for languages are all equally shit. If I want a library I'll go to its respective source and download it

>>106202111
it's fascinating how in every ecosystem the "just use" advice is always followed by the sloppiest most bloated approach imaginable

it always comes from a standpoint of comfort with mediocrity and complete disregard for everything wrong with the approach

god forbid someone builds something to last a decade or more, instead of "just use" pointless abstraction/reimplementation/reinvention that's popular in $CURRENT_YEAR and will be completely forgotten for the next $CURRENT_YEAR thing which will of course be yet another reinvention of the same wheel

between like 1980 and 2020 everything has been about software software software, all trust was into software and by 2020 it had failed - and it failed precisely because of irresponsible short sighted thinking like that

/rant

imagine not being a uvchad in 2025
shit just works

As I was installing some linux packages I saw python scripts showing up often among the downloads, not even microsoft has stooped this low lmao

>>106209480
https://developer.microsoft.com/en-us/python

>>106201862 (OP)
conda can kiss my ass for their stupid fucking licensing bullshit. Work locked down everyone installing anaconda because of it. You just put together a bunch of other people's work into a pretty package why the fuck is that in a restrictive license?

>>106201862 (OP)
>t.nocoder

>>106207124
>Also AI shit.
Yeah, some AI stuff recommends Conda but you're not forced to use it, you can just make a venv and use pip.

I still don't understand what conda etc do that pip doesn't. I always assume conda is just monkeys copying what other monkeys do.

>>106211428
If you don't follow their installation instructions exactly, you're in for a bad time.

>>106211521
conda also ships non-python packages. It's mightier that pip but """lighter""" than docker or nix

>>106211565
No, I've got stuff that had Conda in the install instructions running using a venv and pip. You just have to make sure you get the correct version of python and the requirements.txt takes care of the rest.

>>106211609
>You just have to make sure you get the correct version of python
Do you even know how retarded this statement is?

>>106212070
pyjeets don't know what other langs are like. they just don't know anon

>>106209075
java:
>just copy paste the maven repository and the library link
>want to have it offline? no problem, get the jar file and put it in the libs folder
>just works
why can't they make it like this anymore?

>>106211609
too bad you have the incorrect os version

>>106212070
what? Some dependencies only work on 3.12 for example.

>>106212194
Details? It may happen that a package uses a newer language feature or relies on a recent bug fix, so it may work only on python 3.12 or newer, for example.
But actual Python dependencies are like "you need 3.12.3 and ANYTHING ELSE, even 3.12.4, will fucking break the shit out of it".
There's even this "problem" that venvs explicitly break if you update the systems Python interpreter by a minor version bump.
This is bullshit. Fucking bullshit.

>>106212461
There's a newer version of the dependency that works on 3.13 but the one this program uses only works 3.12.

>>106202111
... if you want slow deprecated garbage

>>106212489
>3 but the one this program uses only works 3.12.
Why? Normally shit is backward compatible.

Python dependency hell is something I'd not even wish on my greatest enemy

Is there any solution for using two different versions of a package in a python script?

>>106212769
Why would you use two different versions in a single script? At that point just update the code to work with the latest version.

UV is cancer.

i absolutely fucking hate python

>>106213434
>uv radiation exposure can cause cancer
You probably don't mean that. Tell your story.

>>106207679
>python notebooks
that's like one edge case, an educational tool that makes use of python's ease of use.

if you've got a linux system that ships with a system python and don't mind setting up a venv or two, python is the best tool bar none for quickly bashing together data processing pipelines for scientific and engineering purposes. it's also useful as a more capable alternative to shell scripting. and since it is a full-featured programming language you can start with your bashed together proof of concept demonstrator and get a working piece of software out of it, up to a limit.

and finally, it's infinitely easier to write a high performance C++ application after you've ironed out all the issues in your python version than starting from scratch, before you even know what exactly it is you're trying to build.

it's not a good language for standalone deployments but it's often good enough. and it's great for prototypes and research.

>>106211521
it has a gui for roasties and itoddlers. it comes with infinite bloat for the jeets.

>>106211565
skill issue. with neural networks in torch, especially if you only plan on doing inference, you can literally rewrite the code from scratch however you like and read the weights in yourself. you don't need their federated learning bloatware if all you've got is a single GPU.

>>106212769
not having terminal skill issue. there's a 90% chance you don't need either package and a 100% chance you don't need both. how hard is it to adjust to API changes?

>>106201862 (OP)
Conda
I do everything in venv. Who the fuck needs conda

>>106212769
git clone twice and rename them

>>106209075
But I am better than everyone else so I build my own $Current_Thing because fuck everyone else.
Now pls hire

>>106207499
>yes i know you can use virtual env, but why?
no one doing serious python development uses a global python install to do it. you use uv or pyenv to create per-project venvs so you don't have a mess of dependency conflicts all the time. if you literally only use python to fuck around with the occasional script or exclusively work on one thing I guess you could get away with the global install, but I would never bother. I use direnv + pyenv so the correct python install is automatically put into my path when I cd into a project directory.

>>106209421
Link to milf?

>>106213994
>dependency conflicts
Don't exist if everything is just backwards compatible, which is proper software engineering.

>>106214024
>someone makes a practical post about real world software development
>nocoder unemployed response dismissing it
few sets of 2 posts could sum up this board better

>>106214064
I'm neither nocoder nor unemployed. I'm used to tools, libraries, and languages that are not shit and where monkeys don't just change API or introduce unfixable bugs in every minor release.
What do you fucking faggots do when you have to update a dependency? Cry?

>>106213326
I just want to. I can do it in Rust.

>>106213994
>so you don't have a mess of dependency conflicts all the time
What if like Python jeets would finally like learn to code and don't depend their shit on a tree of a million different libraries?

>>106214848
Careful what you wish for. You'll only get that if everyone takes up reinventing the wheel. (Badly, of course.) We know this from the big-dependency-graph problems that turn up in other languages, such as the npm ecosystem for JS.

>>106211521
Conda also manages platform wheels (as is, including the required GCC and libraries to compile them). Pip does this poorly. Also for a long time conda was the only place some packages were hosted.

>>106201862 (OP)
One is a baseline ubiquitous tool, the other is a trial version of proprietary garbage that can't even make clean envs.

>>106207679
TFW 80% of my team works exclusively in notebooks and commit them with results to git. They complained to superiors when I added a pre-commit hook to clean it up (the repo grew 50Mb per day otherwise).
I hate python and python "devs" so much it's unreal.

>>106216406
It's marginally better than mouthbreathers wanting to version control Excel files. Yeah that's a thing. O365 made it less popular, thank god.

>>106212489
that's the updoooters fault, they keep refactoring their stuff over and over and over again each minor version, and then this happens. autism ruined computing

>>106201862 (OP)
I use pip within conda envs and I'm the only one in the company not constantly fumbling with the source .venv/activate shit that they never understand how it actually works, you'll see them just typing it into random folders and when it errors they type it again

>>106212489
funniest shit is when one library supports say python 3.14 but not 3.12 and depends on another library that works in 3.12 and 3.13 but not 3.14

>>106201862 (OP)
>>106202203
conda is amazing for creating testing/dev environment specific to a project (that is a project that needs constant package upkeep and changes, like data analytics in progress)
venv is used for when project is done, so you don't need a huge overhead of conda (like for example that data-analytics project you have done in conda, you pip freezed the list of libraries and now have a requirements.txt that you install on venv, usually for in production)
pip is used inside conda and venv
pipx for standalone cli tools without shitting in global library install
uv is great to use for asynchronous web apps and has nothing to do with package managment

I generally use a mix of these whenever coding, but I can't shake the feeling that you retards don't know when to use what and would use an excavator for what a shovel can do and then whine about the overhead

Like, did you niggers ever come across that some libraries just work on python 2.7, or that older libraries are better and want to segregate shit?
no, because you faggots never coded shit in your life

>>106218414
>you see, having multiple binaries of python with different or same minor versions in multiple nested envs is actually sane!
Your tech is garbage.

SQUEAK

>>106201862 (OP)
Imagine the kind of creature that thinks they are partaking in the art of writing programs when they are "programming" in Python. Fucking kek. Even Java jeets are superior to Pyjeets.

>>106218458
>nested
who said anything about nested you retard, cant you read?

and c/c++ have cmake, conan and update-alternatives
why is it hard to believe that you want different versions and libraries for different projects?

>>106218510
>cmake
Worst build system ever, but unfortunately widely used.
>conan
Literally who uses that?
>update-alternatives
Saying PATH and friends would have made you look less retarded.

>>106201862 (OP)
conda is nice, you can manage nvidia drivers through it.

i didnt really understand the conda/venv/uv aspect of python until i started working in it regularly

now i have 500 different conda envs with 2000 different cuda configuration, 80% of my nvme drives are cuda installs in different environments and i couldnt be happier :)

>>106207197
what alternatives do have?

>>106201862 (OP)
It could be worse.

>>106218664
What am I looking at?

>>106213994
Damn I didn't know about pyenv. I always activate the venv manually

>>106213994
And people think Python is a simple, easy to use language. This is worse than jumping through CMAKE fagatory.

>>106214009
https://xhamster.com/videos/japanese-milf-hairy-pussy-eager-for-pov-creampie-xh02aSD
https://xhamster.com/videos/sensual-japanese-women-fumika-12761936

enjoy king

>>106218414
>>uv is great to use for asynchronous web apps and has nothing to do with package managment

dumb bot you're confusing libuv with uv the package manager for python

>>106220325
So he was actually thinking about node? That's hilarious.

>>106216437
>my colleague commits excel files GENERATED BY THE CODE IN THE REPO
>when reminded we have releases he says "I always forget it"

>>106221100
>generating excel files
I'd kill myself.

>>106201862 (OP)
>Deploy in CloudPlatform1ᵀᴹ©.io.ai
>"BUILD FAILED: niggerPy version==9/11.420.69 does not exist in the PyPI registry"
>Deploy in CloudPlatform2ᵀᴹ©.io.ai
>"Deployed successfully!"
>Suddenly niggerPy version==9/11.420.69 is installed

How the fuck is this possible

i am brand new to coding. id rather learn python. ill never catch up with c++ or rust and it seems like theres no point in learning it these days. theres a sense that everythings already been done and figured out. i just dabble in open source AI and python seems sufficient

Python just works. Autists think anyone cares about memory and types and performance, meanwhile python just makes everything a fucking object and becomes by far the most powerful high level language.

What else am I meant to use for data analytics & visualizations? R?

>>106222214
You use nothing
Data science is just a scam to impress managers, just use your eyes nigga

>>106222249
I say this as someone who is aware it's a jeetfield. That's a dumb fucking thing you just said.

>>106201862 (OP)
pip my conda anaconda in you mouth ahahaha

>>106222015
Whats the name of the boy(left) ?

>>106209421
you have access to so many gorgeous Japanese women and you pick an average hag like that?

>>106222112
C is 20 keywords and you will feel comfortable in any language after learning it. Python is 9001 keywords and you learn how to jeetcode in only one language.

>>106222188
>Python just works
until it doesn't
>most powerful high level language
this is b8

>>106222214
just use libtorch directly

>>106202111
Just use ev

>>106202203
>if it was actually batteries included
I'm old enough to remember when python's claim to fame was how much useful stuff it had in it's std lib.

>>106212769
I had to do this once. After trying to refactor stuff with grep and AI I gave up and did the following:
>main venv with version a of package
>venv inside main venv with version b of package
>main.py starts two processes: one with version a and one with version b
>two processes exchange data through pipes

It sucked and I felt dirty afterwards, but in the end I got the functionality that I wanted.

>>106213326
For example, here >>106223448
I needed the discord.py package and the discord.py-self package. The second one is a fork of the first one with userbotting features, and the new one has features that the fork for some reason is missing. I needed both at once.

>>106223204
I remember a pythontard saying "the stdlib is where code goes die"

>>106220723
>>106220325
yea, didn't even know there was uv package manager, thanks fellas for making me learn something

>>106224177
I don't really understand what this means desu.

>>106225895
nvm, it's about how slow the stdlib is to change. I see I see

>>106222702
Kek'd

>>106222865
I like average hags. Imagine how good she must smell like. Like a cute granny.

>pip shitpiss
>"Ok"
>pip cumshit
>"NO VERSION COMPATIBLE IS COMPATIBLE SHITPISS MODULE, DELETING ENTIRE HARD DRIVE IN 3, 2, 1..."

everyone uses uv now old man

>>106201862 (OP)
A gentle reminder that if you use jupyter notebooks for anything but exploratory throwaway code you deserve to get your balls crushed. Slowly.

>>106227870
This thread is literally the first time I heard about it. Is it better than pip or conda? Does it depend on any of them?

>>106221100
dude, learn what a .gitignore file is

>>106218819
It's easy to learn for people who have never written code in their entire life.

i mostly write java, but i did some python today and it wasn't bad. I;m thinking about thos Beans

>>106201862 (OP)
both are trash shit should be written in pure assembly or holyC like a white men proper software.

>>106229224
more like holyChit, lol!

uv is better

>>106228150
You think that stops him? Oooh no. I tried. It is not an accident, he deliberately adds them to the repo, claiming he might need their old versions later. Picrel is me seeing that shit appear in the log.

>>106231867
Most people think it's normal and good to add garbage to git repos.
>binaries
>those OSX meta files
>project files for your favorite IDE (contents change randomly because the formats were invented pre-git and nobody tried to noise reduce them)
>flavor of the month CI pipeline files
All that shit doesn't belong into a git used for source code.

>>106232706
Makes my blood boil holy shit

>>106222015
Can you tell me whats the name of the boy(left) please!!!!

>>106231867
>>106234090
escalate to tech lead or project manager
explain that its wasting company time, resources and money

>>106201862 (OP)
>>106202111
>>106202121
>>106202203
>>106206719

Pip + Venv are the standard library tools. Everything else is bloat. Conda was made by Anaconda to sell subscriptions. UV was made by Rust trannies to be "relevant". Pipx and poetry are trying to recreate pip+venv for no reason whatsoever.

JUST USE PIP+VENV
IT WAS ALWAYS THIS WAY AND ALWAYS BE THIS WAY

>but pip and venv is le hard
NO IT'S NOT, TRANNIES TELL YOU THAT TO SELL THEIR TRANNY SLOP AND YOU FELL FOR IT

>>106234113
bruv you are down bad

>>106223448
That's beautiful for some definition of beautiful

Python is a great language for up-and-coming stars of the coding world who don't know or NEED to know boomer things like grep.

>>106235796
They don't know WHAT?

>>106234964
based
but even venv is bloat

>>106235331
Please i need to know.

>>106223007
>C is 20 keywords and you will feel comfortable in any language after learning it.

Only if "any language" = "any procedural, weakly typed language with direct memory access through pointers"

meanwhile ruby has the easiest dep management system, whoops.

>>106236719
>Ruby_logo.svg.png
This is exactly what I expect webdevs to be like

How do i become Python wizard?

>>106237194
Practice writing pythonic code.

>>106222015
god please i need the name

>>106234886
Escalated to tech lead, he said I need to suck it up since I did not work at google for 500k.

>>106235796
If you do ANY kind of refactoring in p*thon and don't grep (or equivalent) you deserve to be cucked by black bvlls

>>106202176
vcpkg install fmt :)

>>106237194
open a python console, write the following line, and read carefully the output
import this

>>106239143
Does this make mustard gas?

>>106228001
It has more features than pip, dunno about conda. It doesn't depend on either of them but it has an pip option that pretty much translates pip options into uv native ones.
Main reason I use uv is global caching throughout my various python projects, automatic handling of python environments, not too dissimilar to pipenv but, apparently, faster and more reliable

>>106236035
3.12 kind made venvs mandatory

So what I like to do is make a single venv and have each project use it
*evil grin*

>>106201862 (OP)
what kind of obsessed nocoder are you to spam those anti-python threads for so long?

>>106239426
If they want everyone to use venvs they could at least make them a little easier to use. I don't know if I'm missing something but I would expect it to be one command to start a new venv in a directory, and then automagically whenever running python in that directory it uses that one. Apparently that's too difficult and it takes 3 different tools and config files to do this. I thought uv might fix this but it doesn't.
So I just type
>--break-system-packages
each time idc

>>106239473
yes, there are steps to it, but a bash alias could work

does break system packages still work though? I thought I tried it and it was like "mmmm, sweaty, nooo"

>>106202111
Don't use poetry. Don't even use venv. Never run tests. Literally requirements.txt packages with no specified versions and slap that shit in a gigabloated docker image. I don't pay the AWS bill, I don't care. Everything is AI generated garbage anyway. Who gives a fuck anymore.

>>106239426
What did 3.12 change?

>>106239473
>--break-system-packages
I really don't understand what the fuck they're thinking. Doesn't that make pip completely fucking useless? Fucking retards.

>>106239558
>with no specified versions
As it should be. Pinning versions is just admitting that every author of every dependency is a retard who doesn't know how to do proper software dev.

>>106239522
>does break system packages still work though? I thought I tried it and it was like "mmmm, sweaty, nooo"
Idk, I haven't updated my manjaro in a while because I didn't want to break some python project I was working on that I hadn't used venv for. Lel.

>>106239577
I wonder why pinning versions is a common practice for node niggers

>>106239843
I don't know, but my personal impression is that node folks are afraid of supply chain attacks.
What really pisses me off that gtihub has default-on bots that make pull requests only for bumping dependency versions. These devs are so desanitized tio bad practices that they love having their git log full of bot commits that only bump versions.

>>106236719
>ruby slippers
You'll never convince us that Princess Qajar was pretty.

>>106222188
>Python just works.
Until you start dealing with dunders (with the possible exception of __init__). Then the jaws of hell itself yawn wide, as you find that Python's a pretty mask on top of a pile of shit on top of a pile of shit on top of a pile of shit.

Python is for my real Brahmin niggers. All I see is a bunch of seething Dalit scum in this thread. Python won, India won. The rest of you lost.

>>106201862 (OP)
Are you a time traveler from 10 years ago? Just use Poetry or uv, problem solved.

>>106209421
is she male?

>>106201862 (OP)
I hate how the crippled pip and now you are forced to use a venv to install libraries, I wish we had some way to install python libraries in a user local folder instead.

>>106222015
name of the big lips boy?

>>106240110
obvious skill issue

I HATE PYTORCH I HATE PYTHON I HATE PIP I HATE VENVS
I HATE THIS SHITASS SCRIPTING LANG WHERE EVERY VERSION IS INEXPLICABLY INCOMPATIBLE WITH EVERY OTHER VERSION
HOW CAN YOU FUCK UP SO COMPLETELY AND UTTERLY, GODDAMN JAVA HAS MORE CROSS COMPATIBILITY THAN YOUR AVERAGE PYTHON SHITWARE

>>106240501
This.

uv won. everything else sucks.

>>106203264
Haha, funny.

>>106227979
>deserve to get your balls crushed. Slowly.
With a rusty pickaxe, if i may suggest. Also i'm quitting anything that still uses Anaconda in 2025 unless they're paying a big fat bonus beforehand for a complete rewrite. People bitched their souls out for Perl and PHP and ended using this garbage. For what purpose? Le pretty syntax?

>>106231867
If you for some reason can't quit this trash project, file a case to your boss(es) explaining them why it may dragging the team's budget. Cause i really doubt that Github Actions/Collab bill is cheap if they have to fill +300mbs of impossible to cache slop per hour BEFORE the actual dependencies.

>>106241557
Is anaconda really this bad? Personally I hate even the language itself. It looks simple, but it's actually incredibly complex. The syntax is actually ugly, the indentation annoying, and it even makes it hard to write functional style code.
Python is still my go to scripting language, but now I'm only using it because I know my way around. I wonder what I should switch to. Maybe Go?

>>106240431
>>106237726
kek ur still at it. It's a woman (Taya Sour), you're welcome

>>106202111
>just install yet another piece of software just to be able to have a bit less atrocious development experience
>>106239558
This. C/C++ never needed any of this, and for decades people somehow managed to write the code together. Imagine your project did not compile because you updated glibc or gcc LMAO.

python is great and all but it's stupid to use it on a web server.

>>106243696
>. It's a woman (Taya Sour)
I don't care anymore....

>>106244125
Why?

>>106242087
At least is notoriously slow and can hoards GBs upon GBs of shit per project. Pure Python doesn't need it, is tremendously overkill; It's a hack to work with platform wheels (those that need to be compiled in the target machine like the torch stack for ML; it pulls custom GCC versions and libraries) but a container is better suited for the task.

>>106245372
they don't know. The rule of thumb is, if they don't specify in their first disillusioned post, they have no clue why they're against something (very common with anti semitism, they can't explain themselves clearly)

pip install skimage

>>106201862 (OP)
I am a heathen who pip installs into my conda env even when conda installs exist.

>>106236719
Ruby deserved better

>>106201862 (OP)
The entire python ecosystem is so bad and all the fanboys defend it with the usual shit where people come with “solutions” to problems to defend shit that just show they are desperately trying to defend.
Last time their “solution” to the problem was
>Hurr durr, just don't update
Wow, would be cool if python doesn't Eol security updates in like 4 years for each interpreter version.

Interpreter updates are simply a disaster no matter how you twist it in that environment and yes you need to do it because they have no real LTS.

>>106250064
It works on all my machines, and I've never spent more than 5 minutes figuring out venv environments.

The things I waste time on are the stupid poetry, uv, etc. solutions to problems that aren't problems.

I don't understand what your problem with python is. Maybe im sheltered, and C++ and rust offer something better? I know js isn't any better with yarn, npm, and all their dependency hell.

>>106202111
Dont

Use uv

>>106250882
>environments
needing more than one is a problem. there is no backwards compatibility in mind and the bloat is encourages. bloat is hypocrisy to the pythonic method yet it is a pillar of the language nobody bothered to get rid of. using python to make some small script or do a quick prototype you can show a team isn't the issue either, it's using python as the final solution a senior has to deal with later down the line when things bloat to hell. fuck python

Why can't they just make minor interpreter releases backward compatible?

>>106251033
it kind of was like that in v2 but they jeeted everything for v3

Why can't python on windows just use the correct version of python for the program instead of having the user manually change their path variable?
Truly one of the dumbest designs

>>106251085
because every python program doesn't use the same version of python so you have to have multiple python installs. os devs have no control over that so it's just the python maintainers and python developers that continue the cancerous bloat

>>106250975
True, 3.9 type hints can't work in 3.8 and below. But don't all other evolving languages have this issue? Like cpp 20 probably doesn't work with cpp 05

>>106251152
nobody uses cpp 05. cpp 11 is the lowest I've seen but it has no issues getting pulled up to v20. the cpp committee is retarded but at the very least they deprecate old stuff a considerate number of versions down the line. cpp just has an issue adding features to convince people it's worth upgrading. personally, I use 17 since it's still in a sweet spot for compatibility

>>106251152
>>106251152
just to add one more thing, every version of c++ is compatible with C and always will be which is it's strength

>>106251152
In a sane language, you may get problems if something uses newer language features, but your compiler or interpreter is old.
In Python, you absolute will get problems if your interpret version doesn't match exactly your bullshit venv or dependency rat tail or whatever the fuck, even though it could in theory.

>>106251236
Everything you said is retarded and you haven't addressed anyone's question even a bit.
Stop making up convenient bullshit, read the question, and try again.

>>106251290
But you're wrong, you idiot. Give an example of these stupid issues your retarded ass has encountered, and let us laugh at them.

>>106251308
Using a venv. Upgrading system interpreter. The end.

>>106250882
>It works on all my machines, and I've never spent more than 5 minutes figuring out venv environments.
You still need to update inside of the venv, why do people act like venvs solve the update problem.
You're still faced with the same shit that updating may break some things because fuck backwards compatibiity and most of all that you manually now need to also update all the libraries because they're single version only but now that everything has its own venv, you need to do this everywhere.

In sane interpreters, the same library can be installed and used for any version but they can't do that here because even their own core library and interpreter sometimes breaks with minor version updates so what's the point of version numbers to begin with then?

With Rust, you can update the compiler to a new version and just compile your old code without issue typically because Rust takes semver seriously and endeavours very hard to not break things and if something does break, then that's a bug that will be fixed.

>>106251360
>With Rust, you can update the compiler to a new version and just compile your old code without issue typically
That's any language ever that isn't a complete joke.

>>106251033
Because nothing is specified and there are no clear interfaces, this is the issue with Guido's “we're all adults here” thing.

It's really entirely unclear what part of a function interface is actually the stable interface and what part is a leaked implementation detail in Python so people rely on things that they later find out are “just interpretation details we changed now” but they leak because in python every little implementation detail leaks down to formal paramatre names. If you define a function like def foo ( x ): and you change that x to a y it might very well break something because someone tried to call it as foo(x=3) instead of foo(3).

Then comes the duck typing in the sense that some function that was meant to only be called with lists just works with strings because internally all it uses is the ability to iterate over members and someone uses that and passes it a string and then they change it in the new implementation that it errors out because now they do actually use some list-specific function that doesn't apply to strings, which would be cool if they were actually meticulous about documenting this but they aren't either, the documentation is very vague about these things in many places.

>>106251291
the difference is everything doesn't fucking break every version update in any other language. yes they deprecate shit but at a sane pace. cope faggot and kill yourself for defending this shitty "language"

>>106251395
Yes, and Python isn't. Your script can easily break when your interpreter jumps from 3.8 to 3.12 or whatever and the issue is that seeing this coming is so hard because it's a dynamically typed language so it often only breaks on one obscure code path that's barely ever used.

>>106251337
Recreate the venv if you change the FUCKING INTERPRETER.
What exactly do you expect to happen? Also, your incompetent ass may not be aware, but you actually CAN continue using the same venv provided that no C/C++/Rust extensions are installed in that venv.
In summary, if your venv breaks after a Python upgrade then it's due to extensions written in anything but Python.

>>106251429
>Your script can easily break when your interpreter jumps from 3.8 to 3.12
no it can't

>>106251546
>Recreate the venv if you change the FUCKING INTERPRETER.
In C/C++ I don't even need to rebuild my object files if the compiler updated. Your language is a joke.

>>106251429
>when your interpreter jumps from 3.8 to 3.12
How does that "happen" to you? Are you to stupid to download the version you need, or to package your script onto an application?
Actually, no need to answer, I know you are since you're the "scripting" kind.

>>106251576
>confuses compiler and runtime
A retard, as expected.

>>106251601
I don't need to rebuild anything if the runtime changes in C.

>>106251559
Yes it is, basic example:
>Changed in version 3.4: The filename attribute is now the original file name passed to the function, instead of the name encoded to or decoded from the filesystem encoding and error handler. Also, the filename2 constructor argument and attribute was added.
Your code could suddenly bug out if you relied on analysing the filename from this exception for instance. Or:
>Changed in version 3.3: Raise OverflowError instead of ValueError if the timestamp is out of the range of values supported by the platform C localtime() function. Raise OSError instead of ValueError on localtime() failure.
This language encourages exception for flow control.

This is another thing by the way, exceptions constantly change between versions but it's pretty unworkable in many cases to do things with this language without analysing exceptions, except the relevant data to analyse is often poorly documented and an implementation detail that may or may not change next time around.

>>106251586
Anon, again, you just need to update because these things are only supported for three years. This
>Muh just never update
Shit is how 4chan got hacked.
You can't just stick at 3.6 forever in 2025 or something.

>>106201862 (OP)
>python
language. systems are developed to address problems. stupid people make more problems requiring more solutions increasing complexity which creates more problems which cause stupid people to make solutions. learn to code indeed

>>106251586
work uses 3.8 to 3.11 in prod for different applications. It's not my choice

>>106251664
>This language encourages exception for flow control.
That's fucking stupid.

we should start a nonprofit for eradicating python altogether

>>106251744
It is, especially when it's again not clear whether these exceptions are a stable interface or implementation detail as they simply propagate whatever exception is called inside.

I'm forced to rely on catching exceptions multiple times and which to catch is honestly just trial and error reverse engineering as it really isn't documented in a lot of cases and even if the type of error is documented, various things you rely upon like some kind of keyword inside or specific content of the msg or reason payload of the exception is undocumented and can change in a new version or just in some random case be different.

>>106251642
If the runtime changes significantly, you most certainly do need to rebuild, and libc major version updates tend to be highly painful. But keep on pretending otherwise, larper.
The word "significantly" is doing heavy lifting. The maintainers of libc tend to try to avoid doing incompatible changes; there are other languages that try for similar levels of long-term compatibility in their runtime, but Python isn't one of them.

Rebuilding a venv typically isn't difficult. The tools can spit out a list of what to install based on the old one.

I like Python and I use it to connect my modules, which I write in C++.

>>106251964
>and libc major version updates tend to be highly painful.
It's fixed part of the ABI, dumbass. The last time libc updates caused problems (outside of bugs) was when distros upgraded from an older libc to glibc.
Besides, versioned library binary files and versioned symbols keep binary compatibility, so that NO update can break shit.
You can also look at Windows, where .exe files compiled 30 years ago still work.
You're a retard who knows nothing.

Why can't python jeets just learn about compilers already. their entire world is interpreter for some reason it's so fucking frustrating to speak to these brainlets

>>106252043
can you share some? or tell us what they do?

>>106201862 (OP)
>5 days ago
Even the hate thread is slow...

>>106254343
what else is there to say about the best shittiest lang?

>>106252586
I recently made one that applies majority filter or connected component size filter to trinary images, which is something I happened to need for another project.
I don't publish them.

>>106254343
I'd rather say the hate is very persistent.

>>106201862 (OP)
I don't use python if I can avoid it

>>106201862 (OP)
>have lib1.py and lib2.py in the same dir
>lib2 has "import lib1"
>import lib2 from another dir
>error lib1 not found
>OK fine do "from . import lib1"
>__main__ block in lib2 is unusable now because... uhhh... we said so
FUCK YOU FUCK YOU FUCK YOU FUCK YOU FUCK YOU FUCK YOU
>uhhh just throw more software at it (install the project) ((don't forget a venv BRO)) and use absolute imports
>haha oops you wanted to move the dir with lib2? now update all your lib2 imports ;) you didn't forget the editable install RIGHT? ;P
AAAAAAAAA FUCKING DIE IN HELL YOU PIECE OF SHIT NIGGERLICIOUS ABOMINATION DIE DIE DIE DIE DIE DIE DIIIIIIIIIIIIE

>>106255265
Before doing anything you should always set up a virtual environment, it's not that hard.

>>106255513
>dude just use a venv
>dude just use pip
>dude just roll a docker instance
>dude just get a separate computer
>dude just buy this specific hardware
>dude just do *a growing list of things* because we were too dumb to design our language properly SORRY (not really)

>>106255513
Doesn't solve many of the issues and create new ones, like that they aren't portable between machines officially so sharing scripts between machines becomes a mess, also the first time I set one up was hilarious because my shell prompt runs a python script to generate itself so it did not work inside of the venv of course, kek.

>>106255572
Yeah fanboys to deny problems always come with these
>Just do...
solutions that are often more involved than the problem itself in order to still maintain like there is no problem. One time someone here actually came with a problem that involved about 350 euro to deal with some issue with bluetooth to keep maintaining it was superior to a cable in every single way.

“We're all adults here” simply failed as a philosophy. It was a bad idea and in the end very specific contracts that make it clear what is implementation and what is stable pay back in the long run.

>>106255707
>Doesn't solve many of the issues and create new ones, like that they aren't portable between machines officially so sharing scripts between machines becomes a mess,

It doesn't create any new issues, and it's the only way to port scripts between machines. When you want to export your project, run pip freeze > requirements.txt and then you just port your project files plus the requirements.txt file to the new machine. On the new machine your first time you run pip install -f requirements.txt and you're done.

>also the first time I set one up was hilarious because my shell prompt runs a python script to generate itself so it did not work inside of the venv of course, kek.
create a virtual environment for your shell prompt generating script, call it with the specific python interpreter for that virtual environment, it's in ./venv/bin/python

>>106255572
>>dude just use a venv
>>dude just use pip
Literally just stop there. If you believed companies like Anaconda or Docker or whoever is trying to sell you something, then that's your own retardation for not being able to realize you're being sold something. If you're complaining because Python isn't for retards, well you are right about that at least.

>>106239454
I agree that Python is living rent-free in her head and there are plenty enough needlessly hateful threads as it is already. Though I feel somewhat better knowing other people who have to work with Python also have actual griefs against this shit heap

>>106256378
If the python 2->3 fiasco did not teach you that python is for retards, by retards, I don't know what can

>>106256911
Meant for >>106256306

>>106255265
>import lib from dir
I used to work with such a specimen, he called himself "a researcher", even had a PhD.
Just like you, he couldn't grasp the notion of package, module, or import path. Just couldn't, PhD and all.
Everything was a file to him, and every file was in a directory, period.
He couldn't deal with classes either, everything **had** to be limited to one encapsulation level or his brain would explode (file-in-dir, function-in-file).
He always had issues, and those issues could never be his own making – the package was updated, the python was upgraded, the server was unstable, the user name has changed, the password had to be committed to git for the script to work, the 5GB data file was included by git add, etc.
Literally never his mistake, always blamed something else and was convinced of it due to his victim complex.

I see the pattern.

>>106257103
Normal languages (C) do not require you to do five backflips so that it can find your code, dumbass. Python was designed to be a scripting lang for retards with 500 LOC max that fit in a script and maybe a module in the same location. It is literally was not designed with complex hierarchies in mind.
>a PhD reduced the setup to the most basic level: files and directories
BASED
A genius admires simplicity, a retard admires complexity

>venv bad
devops skill issue

>>106255572
If you don't have at least one spare bare metal machine and an array of virtual machines at your disposal you shouldn't even be bothering to write code in any language

>>106257103
>dude DUDE this is NOT a file, alright? it's a "MODULE"
>and that? NOT a directory! a "NAMESPACE" (unless there's a __init__.py inside then it's a MODULE (I guess? (depends on the python version (probably))))

>>106257302
I see the pattern.

>>106257202
>you need a 4 blades rack with kubernetes setup to provision dev environments just to neetcode a fizzbuzz in python
BASADO

>>106255265
I honestly keep forgetting how python import resolution works. I know in nodejs or esm, relative paths just work thankfully. Rust has rust-analyzer to just figure it out for you in terms of self:: or crate:: or super:: bullshit so I don't care.

RIP pyniggas. Meme language desu.

>>106257382
>hyperbole
All you need is fucking virtualbox and a random old garbage laptop with ethernet and a vga port

>>106255572
>just learn this particular functional programming language to manage dependencies
Seriously though, flakes kinda solve a whole lot of packaging issues, but you'll never convince your team to use them.

>>106256911
>the python 2->3 fiasco
It wasn't a fiasco, it was supremely useful.

>>106257433
Something like kubernetes kind or even just bare podman is better than VMs. Less bullshit, less wasted memory and prepares you for real "production."

>>106257188
It is. uv takes care of all that bullshit anyway so there is no reason to bother.

>>106257433
>>106257548
It is a hyperbole but it's not that far from truth if you think about it. If you're learning python chances are you'll be the one running the infra, so you'll have to learn containers and orchestration anyway.

>>106257548
VMs are good for testing different platforms for random shit

>>106257641
What do you mean by that? Python is interpreted so architecture should be irrelevant and python in non-linux context is a meme anyways.

>>106257668
When I said any language I meant any language. I personally use python scripts on wangblows regularly thougheverbeit. With virtual machines you can pick and choose choose platforms to test different os api calls, save machine states(!!), fuck with the file system as much as you want, etc.

Obfuscated java code for example has different class, method, and field names across platforms (and versions), so having a virtual machine to more quickly test that reflection code is platform agnostic is useful in this case also

>>106257668
Back in the 90s and early 00s python's first killer app was being a sysadmin lang because you can write OS independent alternative to bash scripts.

>>106257480
>supremely useful
in redpilling the world that python devs don't give a shit about stability and have 0 foresight, and your code might need a complete rewrite a few years from now

>>106257939
It won't need a complete rewrite, they promised not to do it again. Which is fine, because they got things right with Python3.

>>106257559
>please PLEASE install our cuck-licensed rust toy project

>>106257957
>they promised
yeah and an alcoholic promises not to touch booze again
a fuckup like that must have destroyed all trust in the developers, and it is stunning that it have not

>>106257957
>they got things right with Python3.
Now they simply break everything with every minor or micro release.

>>106257963
>toy project
no coder detected.

>>106257847
That was Perl. Python was neither popular, nor used in any scripting capacity before Python 2.6.
It was always a programming language first and scriptards hated it for this, they all used Perl in anger.

>>106256303
>Wow just repeat all these steps for every single little venv to update periodically to a new interpreter which by the way still doesn't work if you have any locally made libraries for which the source code doesn't conveniently reside any more on the machine for which there should be no need after installation.

Anon, no, this is exactly the issue people talk about. The solutions to a problem that shouldn't exist are extremely involved and make multiple assumptions about a user's system when really in a sanely designed langauge all that is necessary is

>python3 symlink now points to python3.13 instead of python3.12
>everything just werks because python3.13 is backwards compatible with python3.12 in every way barring some bugs that will be fixed when discovered.

That's what minor version numbering means. Anything that works with 3.12 should work with 3.13 as well without this involved mess. Most languages do not recommend setting up venvs, the process of which itself makes a multitude of assumptions about how people do their things because they don't need them because new interpreter versions are just backwards compatible. I currently have /both/ 3.12 and 3.13 installed, not because I choose to but my distribution deems it necessary with their packages. That should never happen and indeed Gentoo is required to multislot python interpreters to make this possible because without an explicit multislot declaration the package manager rejects the possibility of multiple different minor versions of something coexisting because there should not be a need, until there be, such as with Python. python3.13 should be strictly superior to python3.12 in an ideal world.

>>106257103
just wanna say, good story. Not enough of them on here, meanwhile where's like little bukowskis pumping out content every night

>>106243696
>It's a woman
my disappointment is immeasurable and my day is ruined

>>106251337
Is recreating the venv really that painful? It's not like you upgrade the interpreter every day or even every month. It's not ideal but I'd say it's a very small price to avoid CMake or the npm ecosystem clusterfuck.

>>106251429
Has that actually bitten you in practice?

>>106260358
This is the changelog for the debian package that provides bin/python3:

https://metadata.ftp-master.debian.org/changelogs//main/p/python3-defaults/python3-defaults_3.13.5-1_changelog

It's not every month, but often enough. Probably every time upstream python makes a release.

>>106260368
Yes it has, every time it was some part of an exception that I had to inspect that suddenly changed so my test which was a complete hack, but a complete hack you have to rely upon in python failed and because it's exceptions for flow control, you don't even error out drastically but it just takes a code path it shouldn't so it chugs on for a while before you notice it.

>>106258738
>That's what minor version numbering means.
Where, in some unrelated versioning scheme?
>Anything that works with 3.12 should work with 3.13 as well
That's not how Python versioning works. You have some misplaced expectations, perhaps this is the source of your frustration with Python.

>>106258192
Me and our team uses venv + pip little troon bro

If the world were pythonic, you'd die if you tried to use a newer street with your outdated cat-env. Just make a new car-env, bro.

>>106260846
>That's not how Python versioning works
https://packaging.python.org/en/latest/discussions/versioning/
it mostly is. minor updates after the first decimal, patches after the second. things can also break between patches too btw

>python jeet doesn't understand versioning
how do these people even function?

>>106261664
That link discusses valid versioning for *packages*, and it does not prescribe any strategy.
This is the Python release versioning: https://docs.python.org/3/faq/general.html#how-does-the-python-version-numbering-scheme-work

I see the pattern.

>>106203264
I loled

>>106258549
Perl had some good ideas that Python later stole, like CPAN and Perl-style regular expressions.
The problem with Perl is that it's best wizards showed off their stuff by writing cryptic one-liners that were very powerful. Python's best wizards wrote amazingly clear code that anyone could understand even if they barely knew Python. The latter it turns out is much more useful (as the old adage goes, programmers spend 10% of their time writing code and 90% of their time reading it).

>>106258738
>>Wow just repeat all these steps for every single little venv to update periodically to a new interpreter
There is no reason to update a venv to a new interpreter. One project - one venv - one interpreter.

>>106262692
>just have hundreds of copies of full python installs bro
You are mentally ill.

>>106262782
Why do you have hundreds of projects? Feel free to have one misc project that you shit all your stupid ideas in to. You can have a misc project for each python version too if you want. That's gonna be like 3 venvs max

>>106262692
That tard thinks that every micro version increase is incompatible, which is simply false.
He also thinks that the system Python (the one used by the OS) is somehow imposed on programmers who have no choice but to use that version and nothing else, so when the distro version is upgraded then everything needs to be recreated and fixed. As in – he doesn't comprehend that he can download any Python version and use it for any project in perpetuity.

To me this smells of perltardation, I know of no other community worshiping /usr/bin/ as much as they do.

>>106260846
Which is what makes Python absolutely not to be trusted for any serious production work because interpreters go eol after 5 years.

There's a reason that anything that takes itself seriously takes backwards compatibility and long term support seriously.

>>106262692
Anon, Python interpreters only get 3 years of bugfix support and 5 years of critical security fix support.
This is how 4chan got hacked.

>>106263966
>>106263978
>only 5 years
You're grasping at straws, it's pathetic.

>>106263772
>he thinks the package manager providing an interpreter is an inconvenience
>he wants to manually install interpreters and keep dozens of them around for every project
What an utterly retarded attitude. A pyttitude.

>>106216406
Why do you care? it's almost entirely certain that your job is 99.9999% useless bullshit.
>muh data analytics
who cares nigga, your job is a jewish daycare, it doesn't matter at all.
Quit, go surfing instead.

>>106202111
like pottery :*

How do you deal with dependencies that need a compilation step (on the top of my head, Pillow, weasyprint, uwsgi, and other more specific libs) if you don't use the interpreter and packages from your distribution? Provided you don't use a specific Docker image for each Python interpreter version.

I love you guys and I love this thread. I fucking hate python

>>106264214
I have python code on my machine I wrote before python 3 even came out that as originally in written in 2 that I had to port to 3...
Am I like talking to a 19 year old for whom 5 years feels so long it might as well be an eternity?
Python itself recommends updating to the latest interpreter as soon as it comes out, a different venv for every little script only makes this more and more involved.

>>106267415
Dependencies that require compilation during installation are virtually non-existent, wheels have solved this problem many years ago.
Where wheels aren't available (say, for an unreleased version of Python), pip compiles the native code during installation. This requires a compiler and whatever headers the library depends on, if any.
No different than any C/C++/Rust build.

>>106267617
>Python itself recommends updating to the latest interpreter as soon as it comes out
You don't have to use the Python build provided by the PSF. There are many Python vendors, among them Canonical and Red hat who provide 10+ years of support.
Python 2.7 is fully supported on Ubuntu 18.04, which you can install for free today.
> a different venv for every little script only makes this more and more involved.
1. Why would you write incompatible scripts? Surely you can target one Python release for most of your scripts?
2. Why do you keep ignoring the part about packaging your "script" into a standalone application? It takes one minute and removes the need to have Python on the machine altogether.
https://pyinstaller.org/en/stable/

>>106267778
>You don't have to use the Python build provided by the PSF. There are many Python vendors, among them Canonical and Red hat who provide 10+ years of support.
>Python 2.7 is fully supported on Ubuntu 18.04, which you can install for free today.

>Just... SWITCH OPERATING SYSTEM BRO
This has to take the crown for the most retarded “solutions” proposed to justify that there supposedly is no problem. If the fundamental language design requires this kind of stuff which isn't required with other interpreters then it's just poorly designed from the ground up, which it is. “We're all adults here.” was a mistake.

>1. Why would you write incompatible scripts? Surely you can target one Python release for most of your scripts?
New features come out and you want to use them in new code because they make life easier?

Would be great if shit were fully backwards compatible wouldn't it?

>2. Why do you keep ignoring the part about packaging your "script" into a standalone application? It takes one minute and removes the need to have Python on the machine altogether.
https://pyinstaller.org/en/stable/
Because again, then you don't update the interpreter which is only supported for 5 years and again if you just want to change some small thing about this script you have to repackage it again on every machine it's deployed at rather than a simple case of scp like it works with every other language.

Again, your solutions are half as complex as the problem.

>>106267807
Dunno, man. I kind of sympathise with you since it's hard to see retards suffering.
Let's hope that you can find peace one day.

>>106267896
Yeah, no argument.
Fanboys in software who will defend something they like from all flaws to the death are just so oddly common and bizarre.

It's just a basic truth that most interpreters don't have this issue because “We're all adults here.” and exceptions for flow control was never a design idea there because most people know it sucks and among other things can lead to that.

The reality is that my distribution right now has me with 3 different versions of the system interpreter of Python3 and one of Python2; it doesn't even multislot perl, ruby or Guile because it doesn't need to.

This is on Python that packages apparently need to depend on a specific version of the interpreter rather than on “3.6 and anything higher” like it happens in any other interpreter. This is just a reality. Other interpreters
>Do not require venvs to work with
>Do not require distributions to maintain their own version and backport fixes to it

>>106267617
>I have python code on my machine I wrote before python 3 even came out that as originally in written in 2 that I had to port to 3...
You didn't *have* to port to 3. You are free to download a Python2 interpreter and have your code run just as it did before.

>>106267807
>>Just... SWITCH OPERATING SYSTEM BRO
>This has to take the crown for the most retarded “solutions” proposed to justify that there supposedly is no problem. If the fundamental language design requires this kind of stuff which isn't required with other interpreters then it's just poorly designed from the ground up, which it is. “We're all adults here.” was a mistake.
Are you Zed Shaw? Your irrational hate of Python3 despite it doing nothing to you is bizarre

>>106268468
>my distribution right now has me with 3 different versions of the system interpreter of Python3 and one of Python2
We have already established that you're a retard, so I call a big fat lie on the quoted text.

I've used python for decades for small scripts, even wrote some bigger projects in it. I think on the surface, it's a nice scripting language, that just works, with batteries included. Especially I can just print() any data structure, instead of jumping through hoops like in JS or Lua.
So what do I switch to?

>>106268998
>So what do I switch to?
you don't, as you said it's a nice scripting language

>>106269010
There are many things about it that frustrate me. Significant whitespace is a bad idea. It's slow. Exceptions are inelegant, but increasingly used for flow control. (Why the fuck does list.index throw an exception if not found, but string.find returns -1?) It shits .pyc files everywhere. Installing libs with pip is increasingly a pain. It's slow as shit. The language keeps getting more complex and arcane.

>>106214192
Which is why Cargo is an even bigger clusterfuck than pypi.
Protip: languages should not have a package manager.

>>106267415
Idk, I use Poetry and for the last few years, it has been just werking for me. It seems that in most cases the packages simply include compiled binaries and the rest of the time, pip finds a C compiler on my system and compiles stuff automatically.

>>106269614
It's the cool thing to do now. Zoomers and "pro coders" will complain if your language has no package manager.

>>106236719
Why isn't Ruby used more? It's known only for Rails framework but the language itself is much better than Python in my opinion. Pity that Ruby didn't replace Python and is not mainstream thing. Also is Ruby already dead? Does anyone use Ruby for some system scripts or anything else than webdev (which I can't stand).

>>106271909
https://harmful.cat-v.org/software/ruby/rails/is-a-ghetto

>>106271926
https://www.youtube.com/watch?v=i35QJVw0RBM

>>106271926
What is this page about? I've never seen it before, why should I care about what these people think?

>>106271909
Ruby is just Python but worse in every way except package management.

>>106271991
>Zed A. Shaw is a software developer best known for creating the Learn Code the Hard Way series of programming tutorials, as well as for creating the Mongrel web server for Ruby web applications.[1]

>>106269200
>There are many things about it that frustrate me.
Alright let's see these stupid complaints...
>Significant whitespace is a bad idea.
Actually it's objectively good, one of the few objectively good features that a programming language can have. All other languages are worse for not having this feature.
>It's slow.
Skill issue. Use dictionaries and iterators more.
>Exceptions are inelegant,
Tf does this mean
>but increasingly used for flow control.
It's called duck-typing, and it allows you to program the intended behavior first without thinking about exceptions, and then if one of your assumptions is violated you can immediately kick it to an exception.
>(Why the fuck does list.index throw an exception if not found, but string.find returns -1?)
Because list.index is looking for elements while string.find is looking for substrings. Substrings are not "elements" of a string, if you want to turn your string into a list then you need to tokenize it, which is application specific. There's nothing application specific about checking an element existing in a list or not.
>It shits .pyc files everywhere.
add *.pyc to your .gitignore and stop crying
>Installing libs with pip is increasingly a pain.
It's not, see >>106256303
>It's slow as shit.
Skill issue. Use dictionaries and iterators more.
>The language keeps getting more complex and arcane.
No it doesn't.

>>106274140
>Alright let's see these stupid complaints...
Alright let's see these stupid responses...
>Actually it's objectively good
No. Significant whitespace makes it harder to quickly restructure code for testing, like adding an additional if (you would have to reindent the whole shit). It's also easy to completely change the logic by accidentally reindent too many or too little lines. Bracket matching in editors is superior for visual cues.
>Skill issue. Use dictionaries and iterators more.
That shit will make it even slower.
>Tf does this mean
Sorry that you're too stupid to understand. I even gave an example. Exceptions make it harder to use a concise functional programming style, because suddenly you need awkward blocks to catch exceptions.
>It's called duck-typing, and it allows you to program the intended behavior first without thinking about exceptions, and then if one of your assumptions is violated you can immediately kick it to an exception.
Exceptions have nothing to do with duck typing, you fucking idiot. Enough examples have been given in this thread how exceptions make python APIs even more fragile than they would be in a language with duck typing but without exceptions.
>Because list.index is looking for elements while string.find is looking for substrings. Substrings are not "elements" of a string,
God fucking dammit, it's like I'm talking to an AI with this trivial, meaningless drivel that's getting shit out for the sake of having an answer.
>add *.pyc to your .gitignore and stop crying
Why can't it just stop shitting completely useless files everywhere?
>It's not, see >>106256303
See >>106258738
>Skill issue. Use dictionaries and iterators more.
Pythons legendary slowness lead Google to try to optimize it. Despite putting the best people on it, they couldn't achieve anything significant, because Python as languiage and CPython as interpreter are over-complicated trainwrecks.
>No it doesn't.
They keep adding new language features all the time.

Anyone arguing about programming language syntax, or comparing languages by syntax, is retarded and junior level at most. There are so many important things.

>>>106274308
>harder to quickly restructure code for testing, like adding an additional if (you would have to reindent the whole shit).
Imagine programming in an editor that doesn't let you select a range of lines and reindent with one or two key presses.

>>106274463
>Anyone arguing about programming language syntax, or comparing languages by syntax, is retarded and junior level at most. There are so many important things.
Midwit statement. Syntax is not everything, but it matters.
>Imagine programming in an editor that doesn't let you select a range of lines and reindent with one or two key presses.
Of course. It still sucks.

>>106274308
>Exceptions are INELEGANT\
>Exceptions are AWWWKKKWAAARDDDDD
these literally don't mean anything. do an awkward penguin dance and move on with life

>>106275080
>Exceptions make it harder to use a concise functional programming style, because suddenly you need awkward blocks to catch exceptions.
>Enough examples have been given in this thread how exceptions make python APIs even more fragile
But I get that pyjeets can only make dumb posts without any real content like you do.