I don’t understand what you are asking.

i think exercism.io gives you a test suite

Anyone else going to AWS Re-invent this year? I can bring a +1 to the big party at the end for $300. Must be a hell of a party lol.

Hacker rank gives you a base case or 2, but wheh you submit it runs a bunch. But if one fails, it doesn't show you the result set that caused a failure (unless you pay hackos, whatever the **** that is).

Code wars doesn't show them either but you can hack out print statements. I'd just like to be shown the full test suite

Hackos are just activity points that you get when you use hackerrank. I don’t think you can even buy them. You get them simply from solving problems, logging in and participating in contests.

Also if your algorithm is correct, it could possibly be because of overflow problems. So check for that.

Random rant. Please stop using all caps variables in shell scripts to, I guess represent globals? Or constants? I can see an argument for constants but I never see them used that way. All vars in bash are global unless you specify otherwise. Most are not even written in a way where it matters (mine can, I write mine in a function oriented style with a main).

Caps variables are meant to represent environment/system vars like $PATH or $HOME. Stop, please.

This reminds me of a problem I once did on a site like that (ran your code against public test cases, then also against private ones). I thought it might be numeric overflow (it was something like mathematical operations on large numbers), so I implemented something like, converting integers to string and then doing multiplication on the two strings the way you learned in school.

I then later realized that if I switched to Python it just worked, lol.

Also reminds me of doing a similar thing (except, no public test cases) that I kept failing. I noticed it would report certain crash states in the test output (like, it would print something different for segfault vs. stack overflow vs. exit code 0) so I started using intentional crashes as a version of printfs, hiding them in conditionals and re-running to see exactly what was causing my program to misbehave. Eventually realized I had an incredibly dumb, obvious bug that I had just been working on the code too long to notice.

The latter is kind of what I would be worried about. The idea its a practice thing but they won't show you the set you are practicing on is kind of annoying.

Overall I really like Hacker rank since they have a focus on efficiency too, but Jesus just seeing hackos tilts me. What is this, EA sports?

Pretty much actually. I am ok with that. I don’t think it is a site for interactive type learning/development.

Any opinions on Auth0 vs roll your own?

Our sign in is already working but biz side wants linkedin login, which I think is stupid because we're going to compete with them and it's going to let them know all the users that are using our system.

So for the about the same amount of work we can use Auth0 and then get linkedin and another social sign in, or just add linkedin sign in to what we already have.

It's not a banking app or anything so I'm not overly worried about someone breaking in.

auth0 for sure if you want FB, Google, Amazon login.

If not and you feel comfortable incorporating linked in login manually then I'm not sure auth0 adds a lot.

What do you mean roll your own? Never used Auth0 which after 2 minutes of googling I don't get why you'd ever use/pay for. Obviously just use OAuth 2 with whatever backend you have. It took me like a day or two to figure it out from scratch (i.e. npm packages specifically passport) for node/express.

With Auth0 you can have them manage all your users and authentication, and integrate with social media authentication pretty easily.

For some that makes a lot more sense than saving users locally, and writing all their user management, authentication and authorization code from scratch. Forgot password, MFA, registration, verification - all that crap ends up being a huge chunk of your work just to create something totally standard and unsexy that doesn't differentiate your app from any others.

Build our own system to authenticate and save user settings.

I'm in the same boat as you Grue, seems like a waste of money. One of my developers mentioned it as an alternative to writing the code to add linkedin as a login for customers, I just wanted to ask here to see if someone had some reason it was way better than what we already have, and it doesn't seem like there is any great reason to switch.

If you've already written registration, sign in, forgot password etc. and have your own user management system then yeah it doesn't add a lot.

The thing I think you want auth0 for is if you have several different apps of your own, and you'd like shared signon to all of them. It makes managing that kind of thing a lot easier. I used to work for Underarmour Connected Fitness, which owned mapmyrun/mapmyride/etc, myfitnesspal, endomondo and a few other properties. They probably could have used auth0, but ended up writing their own SSO that got used across all the websites.

But if you have a single website with a single userbase, imo it's just a lot easier to manage logins yourself. And adding login-via-whatever these days is almost always oauth1 or oauth2, which has good implementations in every major language.

I also take offense to some of the implementation details of auth0. I suspect for example that they maintain 2 different databases of user data, which are synced somehow - like say a SQL database and elastic search. They have a bunch of different APIs which in the short term did not agree with each other - like you might be able to create a new user and query their profile, but logins wouldn't work for some random interval from 15-60s. Deleting a user might work, partially work, or not work. We'd get into a state with say 100 deleted users where you couldn't log in with them, but they'd show up in a search, but if you tried to click on them you'd get an error. We'd call the auth0 guys and they'd "fix it" somehow.

We had some automated tests that needed to create new auth0 users, do something, then tear them down again - this generally would only work with very generous timeouts and even then like above we'd get into states where the tests stopped working because some database **** on their end got borked.

I tore out all the auth0 stuff, said **** it, and never looked back. Good riddance.

Oauth2 is pretty simple. There should be plenty of libraries to pick from or diy.

Random q of the day: when someone asks you what you do for a living, what is your answer?

Usually just say programmer. or software developer.

POOTERS!

Whatever term you use, you have to do the keyboard typing pantomime at the same time. It's the law.

I say programmer because software engineer sounds douchey but programmer sounds like I'm 23 wearing shorts to work which isn't entirely far off but still

i say software developer because it's right in the middle of those two.

Really? No one is claiming code ninja status??

I just say I’m an engineering manager