Sarat Chandra

Using Mullvad VPN with Tailscale

Mon, 23 Mar 2026 00:00:00 +0000

I wanted to use Mullvad VPN</a> and Tailscale</a> together on my Android phone. The problem is Android only allows one active VPN at a time. Tailscale does have a built-in option to purchase Mullvad through them, but that's not available in India.

So I built a small tool to work around this: mullvad-tailscale</a>.

It's a Docker container that takes a Mullvad WireGuard tunnel and exposes it as a Tailscale exit node. Once it's running, it shows up as a node in your tailnet with exit node enabled. Any device on your tailnet can then route traffic through it and out via Mullvad.

It also comes with a small web UI called mullvad-switcher that lets you change the Mullvad server location on the fly.

Check out the repo</a> for setup details.

Using WaitGroups in Go without Leaky Abstractions

Thu, 25 May 2023 00:00:00 +0000

In many Go codebases, I come across patterns like this:

func</span> main</span>() {</span></span>
    var</span> wg</span> sync</span>.</span>WaitGroup</span></span>
    wg.</span>Add</span>(</span>1</span>)</span></span>
    go</span> func1</span>(</span>&</span>wg)</span></span>
    wg.</span>Wait</span>()</span></span>
}</span></span></code></pre>
In this pattern, the goal is to let a function call the Done()</code> method once it is finished, synchronizing multiple goroutines with a sync.WaitGroup</code>. However, there's a disadvantage: now the function needs to take a WaitGroup as one of its arguments. Consequently, we force the caller to pass in a waitgroup, which might not always be necessary. This issue is a leaky abstraction provoking unwanted dependencies.</p>
A better way to achieve the same result without affecting the function signature is:</p>
func</span> main</span>() {</span></span>
    var</span> wg</span> sync</span>.</span>WaitGroup</span></span>
    wg.</span>Add</span>(</span>1</span>)</span></span>
    go func</span>() {</span></span>
        defer</span> wg.</span>Done</span>()</span></span>
        func1</span>()</span></span>
</span>
        // Any cleanup code can be run here.</span></span>
    }()</span></span>
    wg.</span>Wait</span>()</span></span>
}</span></span></code></pre>
By using an anonymous function, we encapsulate the WaitGroup handling without affecting function signatures. This way, we don't force the caller to pass in a waitgroup unnecessarily, and func1 doesn't need to be aware of our goroutine orchestration. It achieves the same goal but in a cleaner and more elegant way.</p>
This small post serves as a documentation, so I can easily reference it whenever I come across the earlier mentioned pattern and suggest this improved approach.</p>


Fuzzy Search with fzf
Fri, 28 Apr 2023 00:00:00 +0000
fzf</a> is a powerful tool for fuzzy searching, and can greatly enhance your workflow. Here are some ways that I use fzf</code>:</p>
1. Changing directories with ease</h3>
To quickly change directories, I use a bash alias that combines the speed of fd</code> (an alternative to find</code>) and the convenience of fzf</code>.</p>
cdh</span> () {</span></span>
        cd</span> $(</span>fd</span> --type</span> d .</span> "</span>/home/sarat</span>"</span> |</span> fzf</span>)</span></span>
}</span></span></code></pre>
This command lists all directories under the /home/sarat directory (replace this with your desired directory) and pipes them to fzf, allowing you to select the desired directory with fuzzy search.</p>
For example, if I need to search for the personal directory under $HOME/projects/</code>, I just run cdh, search for personal, and hit enter. This command will change the current directory to $HOME/projects/personal/</code>.</p>