Mirror of forum.nim-lang.org

10367 :: Chame - an HTML5 parser library in Nim

• nrk (orginal) [2023-08-01T23:18:40+02:00] view original

  A preview version of Chame, the HTML5 parser used by Chawan, has been released. (Previously discussed here.)

  Repository: https://git.sr.ht/~bptato/chame

  Unlike the stdlib or fusion HTML parsers, Chame aims for 1:1 compatibility with how modern web browsers parse HTML. More precisely, it implements WHATWG's HTML parsing algorithm as described here with an API very similar to Servo's html5ever parser.

  The parser itself is, for the most part, complete (and has been for quite a long time now). Nevertheless, I am calling it a preview version for several reasons:

  1. As noted in the readme file, the current API is not final. Some obvious optimizations are still missing (e.g. replacing the remove() calls with a single theoretical moveChildren call), and I am also considering replacing the C-style object-of-callbacks API with generic mixins (some benchmarking will have to be done to determine if this is worth the effort/ugliness).
  2. The library is still missing tests. This situation will be remedied soon(TM), or at least before the 1.0 release.
  3. decoderstream is currently included as a mandatory dependency; this is problematic both for projects that do not need arbitrary charset decoding support, and also for projects that would like to depend on decoderstream outside the HTML parser. Therefore, it will have to be split into its own package before 1.0.

  To get a rough idea of how this library is supposed to be used, check out the minidom included with Chame, or the less simplistic chadombuilder of Chawan. The parser also has documentation of the individual callbacks, which you can access here.

  Comments, suggestions (esp. concerning the API), etc. are welcome.

---