D Donkeh docs
GitHub

Implementation for Nerds

The user documentation explains how to build with Donkeh. This area is the other half: how each package works internally, why the boundaries are where they are, which data structures carry the hot path, and what was rejected along the way.

Written from the implementation Each package page now covers the problem it solves, its core data structures, the hot path, the decisions that were difficult, and the limitations that are still open. Start wherever your curiosity points.

Reading order

The list follows the data flow rather than the solution order. Start at the parser if you want the whole pipeline; jump straight to a package if you are investigating one layer.

What each page covers

SectionQuestion it answers
Problem and boundaryWhat does this package own, and what is deliberately left to its neighbours?
Core modelWhich types, tables, buffers and invariants carry the package?
Hot pathWhat runs every frame, every layout or every token, and why is it shaped that way?
Trade-offsWhat was simplified, approximated or rejected, and what would it cost to change it?
Code tourWhere to start reading, which files matter, and which ones are generated or legacy.
Bring the hard questions This area is meant to be opinionated. Performance numbers, failed designs, ugly edge cases and “we tried the obvious thing first” stories are all fair game.