Every Python developers favourite issue

The number of times I was GAing for an intro C++ class where a student would submit something, I'd almost reflexively send a "does not compile, 0" out for it, only to stop myself, throw it at the compiler, and then have to spend a solid 1-2 hours figuring out what the fuck the code actually did so I could explain exactly why the compiler didn't say "fuck off" and why the output (which was drastically wrong) was what it was.

That sounds like.. Fun

At least with C, unless you have some sort of overflow or are doing pointer magic, it's hyper-literal.

Agreed, that's what's appealing about it

I will say I mostly enjoy working with C#. .NET's builtins are trashy though

And tbh, the elegance of it

SQL parameter with a database type of TIME, you feed it a DateTime, it fucking crashes because the system can only ship a TimeSpan as an SQL TIME value. Never mind that when you pull that TIME back from SQL you'll be handed a DateTime again.

Then comes in the beauty of user made libraries

Shit like that is all over the place

And you import a library you've found online

And it just.. *works*

>importing a lib

Fuck that I wrote my own

lol

>it just works.

Just importing libs is the 'Python' way

"Eh, someone's already wrote it"

Okay Todd Howard.

It's super lazy

https://www.youtube.com/watch?v=EWMPVn1kgIQ

And I had to go digging for a lot of the shit.
Deepest arcane depths I've gone was figuring out what the fuck the W meant at the end of a function imported from AdvAPI32.dll

https://www.youtube.com/watch?v=zRXl8Ow2bUs

If I had just used the code I found I'd have had a bad memory leak and an even worse security risk due to keeping multiple user credentials in memory long past their logout.

Reason I hate using other people's code tbh

Unless it's from a known competent

If I'm doing anything involving security, then I'm probably going to look for other peoples, open source code

or someone I'm already implicitly trusting

Writing security related code, generally isn't a good thing

Writing your own crypto isn't a good thing. Your own security-related stuff isn't a problem so long as you actually do the legwork

When chances are, there's an opensource library which has been vetted and checked for 2+ years out there

It depends on the scenario I suppose

But then I've gone pretty deep on that end of things as well and have an idea of what to look for when it comes to even timing attacks.

Yeah, experience comes through actually doing the thing

You'll never learn if you don't do it I suppose

But even with that argument- still never roll your own crypto if you need actual crypto

Agreed

That is the one thing I don't care how experienced you are

Since it's so easy to mess up..

Open Source ONLY on crypto

You can *contribute* to said open source