Emacs, The Leetcode Archive Webserver

There's been a gap in publishing posts after my post on setting up a PHP environment in emacs. It's because I've been working on https://leetcode-archive.onrender.com/. I've been making room for the past 2 months on this project consistently with some breaks between. Honestly I never thought I'd get to the point where you'd be reading about this. It's weird any angle you look at it.

Here's what's weird about it:

• Requests are served via emacs.
  
  • The server backend reads your request in emacs lisp.
    
  • HTML is generated through emacs lisp. I don't even have html templates that aren't lisp.
    
  • Yes, records are backed by a non-lisp database, but interfacing with the database is done in emacs lisp.
    
• Did I mention it's emacs? Inside a docker container??? Oooohh…!
  
  • And I did that because I have no money for an AWS EC2?
    
  • And that I could not find any where to host an emacs app? Because…in the web there are approx. 0 emacs apps, so why would a cloud platform think that someone would ship an emacs app and have that as part of their product offering?
    
  • And that only a few cloud providers allow shipping a docker container for free, but with resource restrictions?¹
    
  • And I cannot afford a hosted db? Or that a free hosted db would have no usable data retention?
    
  • And that my container starts cold, and thus I have data loss issues every time the container starts?
    
  • And in this purgatory of suffering, I had to make sure the data auto backfills itself, and on every UTC midnight, I grab the new daily problem? Before data gets wiped clean again?
    
  • And that I did this with emacs timers and graphql requests from emacs?
    
  • And that you can get your name/alias in the project sponsors section if you'd like to fund the project?
    
• It uses transparent, non-opaque cursors for paginating data. I spend as little time in the DB grabbing those records for you as I can². This was overkill for scaling, but I feel like I have to tell every star in the sky and every ancient civilization that once inhabited them: "Hey, I implemented cursor pagination for (almost) arbitrary, user-defined sort keys and sort orders! It's way faster than using LIMIT + OFFSET."
  
• It's self-exposing--it shows you the source code it's running: https://leetcode-archive.onrender.com/src/
  
• It's the only resource online where you can query for dates when a specific problem was a daily problem. (Ok, fine, that is a bit niche usecase but you can query for a bunch of stuff).
  
• Did I mention? Cursor based pagination?³ Yes I did. And I auto generate each SQL query using recursion to get the right sort order for the right sort keys, whether it's forward OR backward pagination?
  
• And, maybe not so great, it's running HTTP 1.1. HTTP 2 was a massive upgrade over 1. Fund my work, and I'll do whatever I can to upgrade the elnode package to use HTTP 2 or higher.
  
• And it's all public domain code. Go. Steal it. Make your own thing. Or make the project better.
  
• And that it's a site that requires absolutely no javascript?
  

Upon the sky, it is
written,
that on this day,
4025-01-26,
cursor based pagination was implemented in the application server for
https://leetcode-archive.onrender.com/.

The stars will speak the names of my sponsors for eternity.

- jwow0

And if you like the work I do,

• from reading my blog,
  
• benefiting from the very niche software projects I write,
  
• or enjoy my videos
  

sponsor more my work at https://www.patreon.com/patreon/jwow0 and have your name appear in the timeless credits in a project if your choice⁴. See you soon!

---