DEV Community: bajubullet

Binary Tree in Rust

bajubullet — Wed, 07 May 2025 13:11:53 +0000

I am not new to rust. I have dabbled in axum, tonic, bevy and ratatui for some personal and work projects as well. Even then when someone asked me to write a Binary Search Tree(BST) in rust, I got scared. I didn't express this at the time but it kept bugging me inside.

Anyway later on whenever I used to get time I used to ask AI tools to help me teach rust and write a BST specifically. I used Gemini and ChatGPT and both gave solutions as well. I could understand their code but still couldn't understand how they got there. They used Rc, RefCell etc. which I understand theoretically but never understood when and how to use it properly.

After I didn't feel satisfied using AI answers I thought let me just read code. Rust actually provides a BST in standard library, you can find it here. Its very well written and probably the most efficient implementation. I went through it, understood it, but again the same problem its great but there is no explanation of why things are the way they are. Why for example its using 3 structs BinarySearchTree, Node and Tree. I am sure people smarter than me can decipher the reasons but for my average brained self it looks like magic.

Anyway the fear inside me started becoming dread. So I finally gave up and decided to overcome it by actually implementing it. Opened my editor, played some music and sat looking at the screen, the cursor blinking, my mind thinking and my hands printing chars on the screen.

I didn't use all the structs described in the core rust or AI implementations. I wanted to implement the way I used to implement BST when I learned it, basically one Node class with an insert method calling itself recursively. I decided maybe when I am implementing it this way I'll get stuck and then I will understand why these intermediate structs are necessary.

I started with the Node struct:

#[derive(Debug)]
struct Node {
    val: i32,
    left: Option<Node>,
    right: Option<Node>,
}

then I decide I wanted to make it generic, so updated:

#[derive(Debug)]
struct Node<T: Ord> {
    val: T,
    left: Option<Node<T>>,
    right: Option<Node<T>>,
}

Ord because I want to the generic type to be comparable. Then rust-analyser started complaining about recursive types so I added Box, which basically ensures the value is spawned on the heap and not the stack.

#[derive(Debug)]
struct Node<T: Ord> {
    val: T,
    left: Option<Box<Node<T>>>,
    right: Option<Box<Node<T>>>,
}

The AI implementations always used Rc<RefCell<>> but I decided not to use it until I actually need it. Also reading through the standard library implementation told me that it is possible to do this with Box only. Anyway now its the meat of the program: the insert function. I was absolutely sure that this will destroy me and I'll just close the laptop lid and drop it.I kept writing till I got this:

#[derive(Debug)]
struct Node<T: Ord> {
    val: T,
    left: Option<Box<Node<T>>>,
    right: Option<Box<Node<T>>>,
}

impl<T: Ord> Node<T> {
    fn new(val: T) -> Self {
        Self {
            val,
            left: None,
            right: None,
        }
    }

    fn insert(&mut self, val: T) {
        match self.val.cmp(&val) {
            std::cmp::Ordering::Greater => {
                if let Some(ref mut node) = self.left {
                    node.insert(val);
                } else {
                    self.left = Some(Box::new(Node::new(val)));
                }
            }
            std::cmp::Ordering::Equal => todo!(),
            std::cmp::Ordering::Less => {
                if let Some(ref mut node) = self.right {
                    node.insert(val);
                } else {
                    self.right = Some(Box::new(Node::new(val)));
                }
            }
        }
    }
}

At least I managed to silence the rust-analyser till the point it was not showing any errors in my editor. Still I was not sure it will work, so I wrote the main and used the example in Rust standard lib implementation:

fn main() {
    let mut head = Node::new(5);
    head.insert(7);
    head.insert(3);
    head.insert(4);
    head.insert(8);
    head.insert(6);
    head.insert(1);

    println!("final tree: {:#?}", head);
}

and that I finally ran cargo run -q. I was surprised to see it worked perfectly. I was so astonished that ended up writing this post, just to share my feeling of elation and the confidence I feel. I know its not much but the kind of happiness you get from conquering your fears is unexplainable. I hope this journey also inspires someone to conquer their fears and do something they have been putting off, you never know you might end up solving it in first go or maybe after few iterations.

iTerm superpower: type in all tabs/panes

bajubullet — Thu, 23 Sep 2021 12:07:28 +0000

One thing that I always liked about tumx is that you can open multiple panes and send the same input to all the panes. Today I figured out you can do something similar with iTerm as well.

Just open multiple panes or tabs and press cmd+option+i (for all panes) or cmd+shift+i(for all windows) and type away. Your inputs will be broadcasted to all panes/windows.

This is especially useful in cases where you have to do a lot of setup/logging in before in all terminals before you can tail logs or run commands.

For more detail open iterm and go to Shell > Broadcast input in the menubar.

invalid server address: ad user sync error

bajubullet — Wed, 22 Sep 2021 11:53:16 +0000

I wasted a lot of time today debugging an ad sync error. The message I was getting from my ad sync library was invalid server address. This was making me thing that something was wrong in the server address(ip, host, protocol etc). But after some time i figured out that was not the case because other ad functions like group search etc were working fine.

This was very weird to me, how come some actions are working and others are throwing invalid server address error. This was crashing my brain 🤯.

After some debugging and carefully inspecting every attribute I was providing to the ad library I finally found the issue which was I added an extra m in the base dn I provided e.g. xyz.comm instead of xyz.com. This was a bad case of wrong error message and shows how a small assumption or bad error message can make a simple error seem like a mind breaking error.

I like the new ios 15 safari

bajubullet — Tue, 21 Sep 2021 11:38:51 +0000

The new ios 15 safari UI seems to be very controversial, the majority of videos on youtube suggest changing it back to older view. They seem to like the way things are and want to stick with them.
I on the other hand love the new UI. I don't care so much about the way it looks and yes there is a learning curve to it. But the additional swipe functionality is definitely worth it.

Swipe left/right to switch between tabs.
Swipe up to go to your open tabs.

Using this swipe gestures feels great. Plus to get the added convenience of a reachable address bar which is great. I have a iphone 11 and average sized fingers and I had to mostly resort to the reachability swipe down shortcut to edit the address bar if i am using a single hand which to be frank is most of the time.
I still think there is a lot that Apple can do to improve safari like adding better PWA support, supporting new APIs so websites look nicer but as far as the new address bar location is concerned, I am very happy with it and feel that its an improvement.

Why use websockets over TCP sockets

bajubullet — Tue, 14 Sep 2021 05:23:52 +0000

I always wondered why we use websockets internally to communicate between services when we are not even using web browsers to connect with them. After some digging around I found the explanation. Keeping it here incase I forget.

from StackOverflow: https://stackoverflow.com/questions/2681267/what-is-the-fundamental-difference-between-websockets-and-pure-tcp?lq=1

It's easier to communicate via TCP sockets when you're working within an intranet boundary, since you likely have control over the machines on that network and can open ports suitable for making the TCP connections.

Over the internet, you're communicating with someone else's server on the other end. They are extremely unlikely to have any old socket open for connections. Usually they will have only a few standard ones such as port 80 for HTTP or 443 for HTTPS. So, to communicate with the server you are obliged to connect using one of those ports.

Given that these are standard ports for web servers that generally speak HTTP, you're therefore obliged to conform to the HTTP protocol, otherwise the server won't talk to you. The purpose of web sockets is to allow you to initiate a connection via HTTP, but then negotiate to use the web sockets protocol (assuming the server is capable of doing so) to allow a more "TCP socket"-like communication stream.