Skip to content

Architecture

Jump to bottom
Joshua Thijssen edited this page Aug 8, 2023 · 1 revision

This page describes the high-level architecture of the browser.

As of this point in time, we haven't got any idea what I want, but I want to be able to do at least the following:

  • have a UI that can run under windows and/or linux and/or mac. Preferably not electron.

  • have the UI separated from the engine so we can easily create different UI's and still use the same engine underneath.

  • have the UI do at least the following:

    • have a address bar where we can type in urls
    • have the UI load the url and display the page
    • have a "tabbed" system, where each tab is a separate page, with separate history, cookies and such
    • support cookies, localstorage (is this engine?)
    • support javascript (v8) (is this engine?)
    • communicate with the engine by using an API (DLL/SO, not an HTTP api)

  • have an engine that can do the following:

    • receive commands from the UI (ie: load page)
    • generate the canvas to show in the UI
    • load a URL through HTTP 1.0 / HTTP 1.1 or HTTP 2.0
    • load a URL through HTTPS
    • parse the received HTML headers / body into HTML tokens
    • parse the HTML tokens into a DOM (might need javascript v8 for injecting new elements during parsing)
    • parse CSS into rules

Few questions we need to ask ourselves:

  • Where does the painting / canvas generation reside? Is it the UI that receives the DOM or is the engine that generates the picture for the UI?
  • If it's the engine, how to deal with textmode generation (text only browsing)
  • cookies / localstorage etc.. where is this handled?
  • what is the bare minimum we can get away with?
  • Which language should we use? (GoLang, C++, Rust?)

Interesting pages:

Clone this wiki locally