xmlbf provides FromXML and ToXML instances intended to be used as the
familiar FromJSON and ToJSON from the aeson package.
Build by running nix-build in the top-level directory. Alternatively, running
nix-shell in the top-level directory puts you in a shell where all of the
dependencies of this project are available. There you can use cabal and/or
./Setup.hs as usual.
Changes
0.7
COMPILER ASSISTED BREAKING CHANGE. Parser is now a type synonym
for ParserT Identity.
COMPILER ASSISTED BREAKING CHANGE. runParser renamed to parse.
COMPILER ASSISTED BREAKING CHANGE. Text, text and text' are
now named TextLazy, textLazy and textLazy. The previous names
are now used with strict Text.
encode now renders attributes in alphabetical order.
0.6.1
Documentation improvements.
Compatible with GHC 8.8.
0.6
COMPILER ASSISTED BREAKING CHANGE. pFail was removed in favour of
fail from Monad of MonadFail.
Added instances for Parser: Semigroup, Monoid, Selective,
MonadFix, MonadZip.
Documentation improvements.
Added dependency on selective.
Removed unnecessary test dependencies.
0.5
COMPILER ASSISTED BREAKING CHANGE. element now returns [Node].
This makes the library safer. It’s not possible to construct a
malformed Element anymore using the names exported by the library.
It is also ergonomic, since most functions with which you would want
to use a manually constructed Element expect a [Node] anyway.
COMPILER ASSISTED BREAKING CHANGE. element' now returns
Either String Node.
COMPILER ASSISTED BREAKING CHANGE. text now returns [Node].
COMPILER ASSISTED BREAKING CHANGE. Use lazy Text inside Text
nodes. This improves Text concatenation performance, performed
internally by Xmlbf, and makes more intelligent use of memory when
dealing with long texts.
COMPILER ASSISTED BREAKING CHANGE. Removed pRead. You are encouraged
to use pFail or mzero if you want to write a failing parser.
BREAKING CHANGE. pText now skips empty text nodes.
Added node.
Added pFail.
Added text'.
Added pChildren.
Added pAnyElement.
Added pName.
Added NFData instance for Node.
encode doesn’t render self-closing tags anymore. Instead, each element has
its corresponding closing tag.
0.4.1
Generalized type of pRead.
0.4
BREAKING CHANGE. pElement now skips leading whitespace before an element.
Fixed nested element parsing (#6)
0.3
BREAKING CHANGE. Renamed df and dfM to dfpos and dfposM respectively.
Added dfpre and dfpreM.
Improved Show instance for Node.
Added element'.
0.2
Text constructor hidden in favor of a text function plus a Text pattern
synonym, just like with element and Element.
