Well, I find myself going through library code most times. I've had a situation where the documentation of a library wasn't sufficient for my solution, after trying for hours, I figured it out after going through their code.

Done this for CPUs/hardware too, reading the Verilog source or schematic was easier than interpreting gaps in the documentation... Having non black box layers is a good thing.

Hmm this is something I've never thought of doing. It does make sense, no matter how docs are, they can't cover the essence of a library the way the code can

🤔

that's part of why Delphi was brilliant. Could Ctrl+click all the way from your own code through the library source files down to OS API calls

Yes. "Go to definition" is one of my favourite things.

I may have to steal this :), one of many reason to stick with open source, where I can read code as opposed to struggling with doc. I will send it to my management when they want me to use any closed source layer.

I barely use libraries at all and instead write my own utilities. It all compiles in seconds, is compatible with every system & in many cases even runs faster. It put me off when I once saw some simulation code needing 2 hours to compile because it used a library for every crap.

Tooooootally. But it takes time to get started

What were the other big growth moments you experienced?