What's your backend web tech stack Jow Forums?

What's your backend web tech stack Jow Forums?

you aren't just a PHP pleb, right Jow Forums?

Attached: download (4).jpg (279x181, 8K)

Java, SQL, real boring shit

Node aka MEME stack

>Node aka MEME stack
>mongodb
>graphql
>if it can be done in JS, its there

Is flask actually worth learning? I tried out the megatutorial for it and got about halfway through before I just gave up because I don't know python very well.

Instead I grabbed a textbook on PHP7, Mariadb and JS which I've been enjoying a lot more. I would like to focus on shit that gets me a job though so advice would be appreciated.

Attached: DwvlCK7W0AMjDIc.jpg (750x751, 67K)

Spring because I'm employed

python, flask and Go
flask is for someone who knows exactly what he wants, if you are newbie it is better to learn django as it will guide you to do things right

don't use either, python is too slow to be webscale

based

Flask, SQLAlchemy, PostgreSQL and in-house extensions. More time writing business logic and less time writing glue code.

Got it in production and ETL queries in Postgres are far more expensive than the application itself.

Not flask. If I want a real ass micro python webshite, I'd use bottle. But given how easy it is to install whatever you want, why would you limit yourself to python? It's so mediocre of a language and has lazy designers.

I wrote flask code in 2 out of 3 jobs so far as a data engineer

>flask is for someone who knows exactly what he wants, if you are newbie it is better to learn django as it will guide you to do things right

Uh, no. Django is too complicated for newbies (trust me, I've tried teaching it). Flask is more accessible.

Porn Hub P

Attached: IMG_20190518_083521.jpg (1002x942, 80K)

PHP 7, MariaDB. Pretty cozy once you learn PHP's quirks.

asp.net,
node memes stack

Attached: 1524306010115.png (424x412, 276K)

I'm using MySQL, Node JS, Express JS, and React

MESN
SEMN

Spring is really pleasant to use, once you are done with security configuration imo

What Python/Flask does better than PHP though?

it's a snake instead of an elephant

> Work
.net / java / node / everything aws & docker

> Side Hustle
php (laravel) / mysql / manual deloyment

Want to get into Flask, but I can't find any good tutorial

Ubuntu Server LTS
Docker
Node.js + Express
PostgreSQL

What's wrong with reading the docs?
It's what I did.
Flask's docs are pretty good too.

BCHS

>being this dump

You can scalate horizontally with any decent language

Vue/ Ruby / Postgres / Redis / Docker

Flask doesn't even scale with my hobbyist uses.
It's shit.

Which wsgi are you using?

uwsgi

isn't pornhub really old at this point? they probably started writing when php was the only (((sane))) choice. nowadays you'd have to pick whatever has the best support i reckon.

i serve dynamic web content entirely using c

Attached: 1550685162267.gif (270x270, 930K)

Reading docs is like reading manpages
Way more info than needed to get things to work

PHP-FPM 7, MariaDB, Redis, and Nginx.

Attached: computer_glasses.gif (320x240, 883K)

Linux
Nginx
Postgres/Redis
Python/Django

LNPRPD doesn't roll of the tongue like LAMP

>you aren't just a php pleb
>post just as pleb technology
>implies webdev isn't pleb

But good god is the initial setup soul crushing. Then it's mostly smooth sailing but if you get stuck, you're really gone be stuck.

Java (Quarkus/Vert.x on GraalVM) and PostreSQL

>All this dynamically typed, interpreted language baby shit

Attached: snap-framework.png (412x282, 49K)

Read flask tutorial docs

kotlin + ktor + jdbi

if youtube and instagram can make it scale you are doing it wrong

Attached: ktor.png (380x133, 4K)

Recently joined a small team that uses django. It's a decade years old code base, very dense and maze-like.
It's dirty, require constant maintenance, extremely slow and doesn't spark joy.

I would have to say it's opposite. For newbies I would do flask then Django.

php. it does the job very well. unless you need to know something more, there's little reason not to use it.

Spring provides a highly predictable structure for generic REST API projects. Once you get the hang of it, even when you jump into an existing project (given that the code was written by professionals, not pajeets), the structure is most likely intuitive and easy to get a handle on.

Go + Mariadb at work.

Forget flask, lool at Quart. Same interface, but asynchronous and with built-in support for websockets.

Python... the only language slower than PHP, which also uses the same retardedly inefficient "let's build everything up and tear everything down again EVER SINGLE FUCKING REQUEST".

I like PHP and haven't experienced performance related problems that couldn't be fixed with better code.

Python is great for more complicated tasks, but only because its libraries make them pretty straightforward. The syntax is atrocious, performance is laughable.

Python for web? What's the fucking point.

I just released the first version of my webapp which uses flask. I liked using it very much, i would reccomend it.

java/angular 6/mysql

hehe.. SEMEN

Elixir, Phoenix, PostgreSQL. Super easy to learn and Elixir is an awesome language.

>let's build everything up and tear everything down again EVER SINGLE FUCKING REQUEST

lmao, what are you using? python cgi scripts? that's not how it works retard

language performance is utterly irrelevant for crud web apps. all the time is spent talking to the database

Flask with custom CPython extensions
Postgres (long term data persistency) and Mongo (session storage and cache)
uwsgi + Nginx

Attached: chadpad.jpg (758x644, 64K)

Based enterprise chad

>if multi-billion dollar companies can make it scale you are doing it wrong

Laravel (legacy), in house aiohttp server, R, postgres, docker, react, and a semi-truck full of aws boilerplate. Apparently some of the php was ported to django first, but that code was a disaster so the in house server library was rolled.

Can't beat .Net core and MariaDB

What gives ENTERPRISE such a boner for AWS? It's expensive as fuck, makes even IBM seem cheap.

Status

AWS has a bunch of features that let large companies move a ton of in house shit beyond webservers to public cloud. The only other cloud that comes close on features is Azure, but being built on a a Microsoft base, Azure's performance is crap.

Mongo
Express
Meteor
Ember

It's another step on the dev ops meme. Instead of hiring admins for each aspect of the stack (db, sre, network, etc) there's a dev ops guy or 2 that write all the boiler plate to make everything "just werk". Except in reality you pay a crap ton of money because the dev ops guys fucked up and didn't make things flexible and now everyone has to debug something that should be simple, but was done in a back-assward complicated way.

Linux
Apache
PHP + Laravel
MySQL

I have only tried Kotlin with Javalin because I hate MVC boilerplate

Can recommend Javalin as the Flask for the JVM. 8/10 would use it again

I've used Elixir, even Erlang

I still have no idea why the fuck they chose to restrict 1 struct to each module. It feels like a crippling design. Same with Erlang records.

In fact I hate the idea of records and structs in Elixir so much that I switched over to other languages. If you're gonna do types, at least make them statically checked and not some dialyzer bullshit.

Furthermore, the syntax is ugly as fuck for both structs and records.

Flask or Go?

>uwsgi

UwUsgi

>company hires senior cloud engineer
>cloud engineer proceeds to write 15000 of chef and puppet to deploy a monolithic app to AWS that gets less that 1000 enterprise users a month
>web app is less than 15 pages and basically data tables
>mfw more devops glue code than actual code
>mfw we just migrate it all back to our colo and make it deployable with a single shell script

Devops engineers, not even once.

>devops engineers are useless
>single shell script for deployment is tops lol
t. has only seen garage-grade software

I do full stack but I've done devops too, and regardless whether the cloud engineer in question was incompetent or not, it's clear you're a fucking peon and you have no idea what you're talking about.

>t. triggered

No need to have an epeen contest bud. It's about picking the right tool for the job and the point of my post is pointy hair managers from wharton and these business schools read a few articles in infoworld and medium etc, then balance their staff with a bunch of devops people, and before you know it the ops stack is larger than the core product in such a disproportionate way that your paying more people more money than would have been cheaper with classic IT and good ITLM processes in place.

I see a lot of your types expecting to build the next Google or Twitter when in reality its yet another corparate erp tool that will get hcl'd/mahindra'd/tcs'd anyways.

Honestly it sounds like you're stuck in like early 2000's vis-a-vis deployment automation. I agree, there's no point in bloating your ops stack to the point where it's larger than the core product. However to dismiss modern tech completely in this regard is to force yourself into a corner where in the very near future you'll have no place to go except shitty, dust mite sized organizations that you're in, or smaller.

The age of IT guy's custom k00l sh3ll skr1pts is over. There are standards now, and those automations provide features BY DEFAULT that your small mind could have never even conceived of. Does your shell script run tests on every commit? Does it automatically do integrations tests and dry runs for db migrations before deploying the new package? Does it do zero downtime deployments? Can it do rollbacks? No, because it's a fucking shell script that you and your dumbass colleagues wrote.

Again, if you ever leave the shithole you're in, you're going to realize everything's getting automated because there's money to be made, and because even if the automations don't always work perfectly, at least now there's a framework that'll allow the IT guys and devs to become more dispensable.

The future is now, 30-year-old boomer.

>Buttmad dec ops "artisan" mad his job can be boiled down to a few shell scripts
I'll absolutely agree that automation technology advancements have helped out a lot, but the frameworks you refer to don't even reduce time on tasks. Instead spending time managing deployments, time is spent managing the system (which will have just as many problems as any other deployment system) that manages the deployments. Then when some other framework comes along that makes something else easier you have to rewrite everything because none of your framework shit is compatible. Meanwhile the chad scripters just edit a few lines and move on to making some other cool thing to deploy.

I already told you, I'm not a devops guy, I do fullstack.
>he keeps implying non garage-tier software could and/or should be deployed with a shell script, completely ignoring the requirements I explicitly listed
Fucking plebs.

Node(typescript+react+express) for small web services, Rust for real time stuff like games. Also PostgreSQL.
The comfiest stack.

Spring annotations are a bitch and the run time issues you can encounter from anonymous classes and reflections can be really nerve wrecking. But reading the documentation can prevent this from happening.

Spring Boot makes this incredibly easy. Not sure what you mean by "initial setup is soul crushing" unless you're trying to set up your own tomcat server and everything.

This is exactly correct. I can understand any Spring project within 5 minutes if the pajeets didn't fuck it up. My boss has made me write some FUCKED code because he doesn't understand Spring (e.g. he had me rename a CustomerDAO to "DataConnect" because he wanted to explain it to business people).

Node, mariadb, p comfy

>he had me rename a CustomerDAO to "DataConnect" because he wanted to explain it to business people
Why couldn't he just rename it in the slides and leave the codebase alone? Fucking idiots.

>Does your shell script run tests on every commit?

post receive hook calls the script, appropriate test get run on appropriate dev/prod instance based on branch/tag

>Does it automatically do integrations tests and dry runs for db migrations before deploying the new package?

chef/ansible/puppet/docker/kubernetes/aws/azure don't do those things either. that's just glue code my friend. a couple if-thens calling some migration test scaffolding can achieve the same result with return code checking and a few conditional blocks.

>Does it do zero downtime deployments?

a small bit of nginx and proper multi-site storage/database replication mitigate this

>Can it do rollbacks?

if it needs to. just re-deploy an earlier commit and roll-back the changes that can be rolled back without data loss

Nothing you describe can't be achieved with early 2000's unix tier tooling. It's just nobody read the man pages and now there are 10,000 wrappers, CI pipelining tools, etc. Hell, most circle ci, jenkinsfiles, etc that i've seen are just wrappers for some shell/perl/ somewhere.

There is absolutely no need for miles of tooling for simple corporate web-apps, and these koolaid drinkers convinced otherwise are literal drains on companies.

The most depressing thing is that those complex ass tools are still called with a shell script.

>Does it automatically do integrations tests and dry runs for db migrations before deploying the new package?
>chef/ansible/puppet/docker/kubernetes/aws/azure don't do those things either
Integration tests can absolutely automatically run in any CI pipeline software.
>no dry runs for db migrations
>cannot into flyway
>Nothing you describe can't be achieved with early 2000's unix tier tooling.
I know. The thing is, you just gave me the list of things you need to build MANUALLY - AND THAT IS MY WHOLE POINT. Even with Jenkins you get all that shit already configured, for free. You made my point for me. Instead of your artisanal unix tooling, you could use shit that works out of the box and isn't indecipherable garbage to the next person who has the misfortune of replacing you (whenever you move on).

I love the kind of pants-on-head retarded elitism in believing that wrappers with auto-configurations are bad. Yeah, more repetitive configuration work is bad, that's why IT came to be in the first place. No, wait...

That's a very good question. I'm going to just silently change it to CustomerDAO anyways because, knowing him, he'll forget or stop caring about it. He's made us choose some weird architectural decisions because of what he wants to justify on a slide or to business people. Like we used eclipse link for filtering (instead of Hibernate) but we don't even use filtering because we filter on the serialization level.

Djangooooooooo

>you just gave me the list of things you need to build MANUALLY - AND THAT IS MY WHOLE POINT

see pxe booted images. for smb this is a solved problem. you get nothing with jenkins except for a bloated plugin ecosystem that at the end of the day results in gigantic jenkinsfiles that are just clever wrappers for shell scripts with the only net positive being a 90's era UI for viewing the results, or 2000's era if you use blue ocean. anybody who has used jenkins for more than 1 hour and still fangirls it must have brain damage.

i'm not saying doing devops is bad. i am saying people who put this mountain of devops tooling above the core business are egocentric retards who believe their god tier tooling is solving problems that haven't already been solved, over. and over. and over. just with a different name, moniker, acronym, each and every time.

jesus fucking christ, imagine using python for the backend

ruby on rails

nodejs and mongodb is my go to, convince me why it shouldn't be

> you get nothing with jenkins except for a bloated plugin ecosystem that at the end of the day results in gigantic jenkinsfiles that are just clever wrappers for shell scripts with the only net positive being a 90's era UI for viewing the results,
If your jenkinsfiles are gigantic you're doing it wrong. You reveal your own ignorance in these ridiculous, painfully out of touch analogies you push.

Your hand-crafted shit doesn't scale. I worked in an environment where there were 50+ packages/artifacts being compiled, built and tested at the same time. The fucking thing ran on two separate clusters. +10M lines of actively developed code with 80% test coverage. It's blatantly clear to me you've never seen setups like that, and you have no fucking idea of the scale of work that kind of territory comes with.

You don't use wrappers in enterprise tier software development processes for the fun of it, you do it out of necessity.

GNU/Debian
Nginx reverse proxy for Kestrel
Postgres + Dapper
ASP Core 2.1
it's really comfy and has so many jobs

idk what ur doing i usually read doc + source code. are u completly new to webdev or something?
watch a seminar on flask if u wanna get the gist of the famewrok fast

>t. jenkins marketing department

"thing has lots of things therefore it must be complex to build and deploy"

you guys are an anathema to simplicity and bring cancer to everything you touch.

>Plebthon
enjoy shit performance

Node.js or Go

Yet somehow out of the two of us I'm the only one who has been given the responsibility to develop and maintain such extremely expensive, business-critical systems. I guess I just fell for marketing, and you're working in a shitty basement because you're so fucking l33t.

So much cope.

flask and django and node and postgreSQL and mongoDB

Java with Eclipse Vert.x and MariaDB

>t. still triggered and has to fall back on epeen adhominems

you fell for the memes senpai, just admit it

>you fell for memes
Yeah, memes that make me +$100k a year outside SF/NY/etc. Enough said.