As a Rust programmer, I approve this message. Tumbling through a turbine repeatedly would be less stressful than working on a large python/js codebase.
I’m a Linux System Engineer and was the only one in my team that knew Go. I decided to update our mess of old shell scripts for post-provisioning and my boss suggested that I do it in Python so it can easily by edited/fixed by anyone on the team. I spent like two days attempting to do it in Python and then gave up because it would mean transferring a bunch of source code around, installing dependencies and just general annoyances.
In the end the Go project ended up being about 1300 lines of code across a few source files, but it could act as both the client and server (necessary for our hosts in our DMZ to hit our AWX server) with a single binary and no additional dependencies. It was also only like 10 MB.
If you store data in a struct, you want that struct to have ownership of that data. So, avoid storing references in structs.
If you need to pass data into a function, you usually want to pass it as a reference.
This makes it so you have your data stored in some place with the ownership and from there you just pass data down into functions as references. It forces you to structure your program like a tree (which is often a very good idea to begin with).
As a Rust programmer, I approve this message. Tumbling through a turbine repeatedly would be less stressful than working on a large python/js codebase.
Plus you have plenty of time to tumble once or twice while your large codebase compiles.
I’m a Linux System Engineer and was the only one in my team that knew Go. I decided to update our mess of old shell scripts for post-provisioning and my boss suggested that I do it in Python so it can easily by edited/fixed by anyone on the team. I spent like two days attempting to do it in Python and then gave up because it would mean transferring a bunch of source code around, installing dependencies and just general annoyances.
In the end the Go project ended up being about 1300 lines of code across a few source files, but it could act as both the client and server (necessary for our hosts in our DMZ to hit our AWX server) with a single binary and no additional dependencies. It was also only like 10 MB.
Sounds better than coming up with the most mindfucking ways to please the borrowchecker in a large rust codebase
(Skill issue probably, i know)
Rule of thumb, which I feel gets you 80% there:
If you store data in a struct, you want that struct to have ownership of that data. So, avoid storing references in structs.
If you need to pass data into a function, you usually want to pass it as a reference.
This makes it so you have your data stored in some place with the ownership and from there you just pass data down into functions as references. It forces you to structure your program like a tree (which is often a very good idea to begin with).
Yeah, it took me a bit to wrap my head around it. It’s worth it to avoid subtle, weird, and hard to diagnose bugs later on.