An XML DOM-style parser, that only parses a subset of XML, but is designed to be fast. In particular:
- Entities, e.g.
&
, are not expanded. - Not all the validity conditions are checked.
- No support for
<!DOCTYPE
related features.
The name "hexml" is a combination of "Hex" (a curse) and "XML". The "X" should not be capitalised because the parser is more curse and less XML.
Hexml may be suitable if you want to quickly parse XML, from known sources, and a full XML parser has been shown to be a bottleneck. As an alternative to hexml, which supports things like entities but is still pretty fast, see Pugixml (with a Haskell binding - but be aware the Haskell binding of Pugixml can segfault, in contrast to Hexml which does not).
Hexml is tested with AFL.
If you want lenses for Hexml, see hexml-lens.
The optimisation work around Hexml spawned Xeno, a Haskell-only alternative to Hexml. There is a talk covering the performance tricks of Hexml and Xeno.