Our great sponsors
-
WorkOS
The modern identity platform for B2B SaaS. The APIs are flexible and easy-to-use, supporting authentication, user identity, and complex enterprise features like SSO and SCIM provisioning.
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
-
tracks
A bare-bones Ruby HTTP server that talks Rack and uses a thread per connection model of concurrency. (by matsadler)
The other major thing you'll run into is performance. WEBrick's HTTP parsing code may look hairy but it is fast. Mongrel was the first Ruby HTTP server to implement it's HTTP parser in C using Ragel, which Thin, Unicorn, and Puma all copied; although there's a bug in the original Mongrel HTTP parser where it does not combine the values duplicate HTTP headers (yes, HTTP Header names can actually be repeated).
The other major thing you'll run into is performance. WEBrick's HTTP parsing code may look hairy but it is fast. Mongrel was the first Ruby HTTP server to implement it's HTTP parser in C using Ragel, which Thin, Unicorn, and Puma all copied; although there's a bug in the original Mongrel HTTP parser where it does not combine the values duplicate HTTP headers (yes, HTTP Header names can actually be repeated).
WEBrick. Written entirely in Ruby, this is where you should start your examination, but not where it should end. WEBrick is easy to understand, because it is all Ruby, but it is not suitable for hosting websites. It is not robust or performant enough.
For example, I enjoy sim racing, and some of my games provide a network API for things like telemetry data. So I wrote a simple telemetry logger that I use to gather data, which I then mess around with using R Studio. Ruby worked exceptionally well for this because of a cool little library called BinData.
You don’t need to implement a server from scratch. Here’s an exercise I wrote about wiring the HTML generating bits of Ruby code together you might find interesting https://github.com/schneems/ruby_view_server
Yes, I actually have. Implementing a basic HTTP parser isn't that difficult, once you implement the correct parsing rules from RFC 9110. If you want to fully support the HTTP protocol, you also have to implement things such as Chunked Transfer Encoding or Gzip Deflate.
The other major thing you'll run into is performance. WEBrick's HTTP parsing code may look hairy but it is fast. Mongrel was the first Ruby HTTP server to implement it's HTTP parser in C using Ragel, which Thin, Unicorn, and Puma all copied; although there's a bug in the original Mongrel HTTP parser where it does not combine the values duplicate HTTP headers (yes, HTTP Header names can actually be repeated).
My server is on GitHub here: https://github.com/matsadler/tracks
And the HTTP parser I wrote for it: https://github.com/matsadler/http_tools
The spec for Ruby's Rack is another good reference for how a Ruby webserver is expected to work.
Maybe check out EventMachine. You can roll your own using sockets if you don't want to use a library.