Preprocess webgl shader code before browser/drivergpu compilation:

• lint

• minimize

• obfuscate

• optimize

• format

• Tokens subAST API:

  • token: type, depth, scope
  • identifiers instances/declaration
  • variables usage range (reuse/prune)

• error and warnings

  • invalid
  • vague "perfromance token to cycles"
  • uniform flow warning
  • unused uniform/varying/variable/function

• prune

  • comment
  • unused
  • preprocessor branched out code
  • whitespace

• format

  • guarantee consistent style

• perf oriented

  • performance/allocation oriented
  • tokenisation and strip at same passe
  • tokens Actions in least pass possible
  • export to string and minify/stylify in one pass

TODO: - worker sample - faster - token based editor (Edit sub token array update from diff. (shader edit a 'view' of token tree)) - try to work on array/typedarray uint8 instead of js strings (immutable js string makes for a lot of allocations...)

Fork/Code/Inspiration From:

• https://github.com/glslify/glsl-tokenizer
• https://www.npmjs.com/browse/depended/glsl-tokenizer

MIT, see LICENSE.md for further information.