- Show instance for ParseOutput
- Correctly return remaining data for direct String parser (doesn't return correct data for direct Number/Bool/Null parser)
- lifted upper bounds on aeson
- user unsafeCoerce to existential wrapper
- fix bad parsing with .| on object
- fixed testing suite
- objectOf parser for faster one-pass JSON object parsing
- set minimum base to 4.11
- aeson-2.1
- added objectKeyValues (Dylan Simon)
- optimization for reading ASCII strings
- added text 2.0 compatibility
- added conduit interface behind a flag (Dylan Simon)
- added manyReverse (Dylan Simon)
- added valueWith (Dylan Simon)
- Aeson 2.0 compatibility
- Added support for raw bytestring
Fix compiling with new ghc.
Fix 32-bit number parsing.
Speed optimization of many
and aeson object.
Added Semigroup instance, compatibility with base-4.11
Renamed _js_decode_string
function to avoid conflict with aeson
.
Added support for GHC 8.2.
Fix windows build.
Slightly more strictness in arrayOf.
Fixed memory leak in arrayOf.
Added aeson-compatibile encode/decode functions.
Breaking changes (this could really break your code):
- Changed
<|>
to<>
(Monoid
is better for 'appending' thanAlternative
) - Changed
>^>
to<|>
- (Alternative
now really means alternative) - Changed
toList
tomany
(Use existingAlternative
function instead of a custom one) - Added
some
function (Alternative, default implementation won't work) - C-lexer now supports parsing numbers up to 18 digits (E-notation is not optimized yet)
- Completely rewritten text unescapes based on text decodeUtf8; fixes some surprising crashes, speed improvements.
- Changed string parsing; parsing of escaped strings is now very fast
- Removed bytestring parser
- Fixed bug in safestring
- Fixed test so it doesn't depend on versions of other packages
- Added sax-like parsers
- Fixed wrong size of C structure in FFI that was causing a segfault.