v8.0.0 (#1797)

This version represents a significant refactoring of nom to reduce the amount of code generated by parsers, and reduce the API surface. As such, it comes with some breaking changes, mostly around the move from closure based combinators to trait based ones. In practice, it means that instead of writing combinator(arg)(input), we now write combinator(arg).parse(input).

This release also marks the introduction of the nom-language crate, which will hold tools more focused on language parsing than the rest of nom, like the VerboseError type and the newly added precedence parsing combinators.

Thanks

• @cky
• @5c077m4n
• @epage
• @Fumon
• @jtracey
• @OliveIsAWord
• @Xiretza
• @flier
• @cenodis
• @Shadow53
• @@jmmaa
• @terror
• @zanedp
• @atouchet
• @CMDJojo
• @ackxolotl
• @xmakro
• @tfpk
• @WhyNotHugo
• @brollb
• @smheidrich
• @glittershark
• @GuillaumeGomez
• @LeoDog896
• @fmiras
• @ttsugriy
• @McDostone
• @superboum
• @rruppy
• @thssuck
• @Chasing1020
• @thatmarkenglishguy
• @ambiso
• @boxdot
• @krtab
• @code10129
• @manunio
• @stuarth
• @mindeng
• @JonathanPlasse
• @nabilwadih
• @phoenixr-codes
• @gdennie
• @art049
• @kstrohbeck

Added

• Parser::map_res
• Parser::map_opt
• many and fold combinators using ranges
• many can collect into other types than Vec
• Error and VerboseError can be converted to owned versions

Removed

• nom::bits::* is no longer re-exported at the crate root. This export caused frequent confusion, since e.g. nom::complete::tag referred to nom::bits::complete::tag instead of the much more commonly used nom::bytes::complete::tag. To migrate, change any imports of nom::{complete::*, streaming::*, bits, bytes} to nom::bits::[...].
• parse combinator
• InputIter, InputTakeAtPosition, InputLength, InputTake and Slice are now merged in the Input trait

Changed

• Parser::map and Parser::flat_map now take a FnMut as argument

This versions represents a significant refactoring of nom to reduce the amount of code generated by prsers, and reduce the API surface. As such, it comes with some breaking changes, mostly around the move from closure based combinators to trait based ones. In practice, it means that instead of writing combinator(arg)(input), we now write combinator(arg).parse(input).

This release also marks th introduction of the nom-language crate, which will hold tools more focused on language parsing than the rest of nom, like the VerboseError type and the newly added precedence parsing combinators.

Thanks

• @cky
• @5c077m4n
• @epage
• @Fumon
• @jtracey
• @OliveIsAWord
• @Xiretza
• @flier
• @cenodis
• @Shadow53
• @@jmmaa
• @terror
• @zanedp
• @atouchet
• @CMDJojo
• @ackxolotl
• @xmakro
• @tfpk
• @WhyNotHugo
• @brollb
• @smheidrich
• @glittershark
• @GuillaumeGomez
• @LeoDog896
• @fmiras
• @ttsugriy
• @McDostone
• @superboum
• @rruppy
• @thssuck
• @Chasing1020
• @thatmarkenglishguy
• @ambiso
• @boxdot
• @krtab
• @code10129
• @manunio
• @stuarth
• @mindeng
• @JonathanPlasse
• @nabilwadih
• @phoenixr-codes
• @gdennie
• @art049
• @kstrohbeck

Added

• Parser::map_res
• Parser::map_opt
• many and fold combinators using ranges
• many can collect into other types than Vec
• Error and VerboseError can be converted to owned versions

Removed

• nom::bits::* is no longer re-exported at the crate root. This export caused frequent confusion, since e.g. nom::complete::tag referred to nom::bits::complete::tag instead of the much more commonly used nom::bytes::complete::tag. To migrate, change any imports of nom::{complete::*, streaming::*, bits, bytes} to nom::bits::[...].
• parse combinator
• InputIter, InputTakeAtPosition, InputLength, InputTake and Slice are now merged in the Input trait

Changed

• Parser::map and Parser::flat_map now take a FnMut as argument

v8.0.0-alpha2 (#1758)

v7.1.2 (#1605)

Thanks

• @joubs
• @Fyko
• @LoganDark
• @darnuria
• @jkugelman
• @barower
• @puzzlewolf
• @epage
• @cky
• @wolthom
• @w1ll-i-code

Changed

• documentation fixes
• tests fixes
• limit the initial capacity of the result vector of many_m_n to 64kiB
• bits parser now accept Parser implementors instead of only functions

Added

• implement Tuple parsing for the unit type as a special case
• implement ErrorConvert on the unit type to make it usable as error type for bits parsers
• bool parser for bits input

Thanks

• @ThomasdenH
• @@SphinxKnight
• @irevoire
• @doehyunbaek
• @pxeger
• @punkeel
• @max-sixty
• @Xiretza
• @5c077m4n
• @erihsu
• @TheNeikos
• @LoganDark
• @nickelc
• @chotchki
• @ctrlcctrlv

Changed

• documentation fixes
• more examples

Thanks

• @nickelc
• @Stargateur
• @NilsIrl
• @clonejo
• @Strytyp
• @schubart
• @jihchi
• @nipunn1313
• @Gungy2
• @Drumato
• @Alexhuszagh
• @Aehmlo
• @homersimpsons
• @dne
• @epage
• @saiintbrisson
• @pymongo

Changed

• documentation fixes
• Ci fixes
• the move to minimal-lexical for float parsing introduced bugs that cannot be resolved right now, so this version moves back to using the standard lib' parser. This is a performance regression*. If you have specific requirements around float parsing, you are strongly encouraged to use recognize_float and another library to convert to a f32 or f64

Added

• alt now works with 1 elment tuples

This release fixes dependency compilation issues and strengthen the minimum supported Rust version (MSRV) policy. This is also the first release without the macros that were used since nom's beginning.

Thanks

• @djc
• @homersimpsons
• @lo48576
• @myrrlyn
• @RalXYZ
• @nickelc
• @cenodis

Added

• take_until1 combinator
• more to_owned implementations
• fail: a parser that always fail, useful as default condition in other combinators
• text to number parsers: in the character::streaming and character::complete modules, there are parsers named i8, u16, u32, u64, u128 and u8 ,u16, u32, u64, u128 that recognize decimal digits and directly convert to a number in the target size (checking for max int size)

Removed

• now that function combinators are the main way to write parsers, the old macro combinators are confusing newcomers. THey have been removed
• the BitSlice input type from bitvec has been moved into the nom-bitvec crate. nom does not depend on bitvec now
• regex parsers have been moved into the nom-regex crate. nom does not depend on regex now
• ErrorKind::PArseTo was not needed anymore

Changed

• relax trait bounds
• some performance fixes
• split_at_position* functions should now be guaranteed panic free
• the lexical-core crate used for float parsing has now been replaced with minimal-lexical: the new crate is faster to compile, faster to parse, and has no dependencies

Fixed

• infinite loop in escaped combinator
• many_m_n now fails if min > max

Thanks

This release was done thanks to the hard work of (by order of appearance in the commit list):

• @homersimpsons

Fixed

• fix documentation building

Thanks

This release was done thanks to the hard work of (by order of appearance in the commit list):

• @DavidKorczynski
• @homersimpsons
• @kornelski
• @lf-
• @lewisbelcher
• @ronan-d
• @weirane
• @heymind
• @marcianx
• @Nukesor

Added

• nom is now regularly fuzzed through the OSSFuzz project

Changed

• lots of documentation fixes
• relax trait bounds
• workarounds for dependency issues with bitvec and memchr

Changed

• Fix cargo feature usage in previous release

Thanks

This release was done thanks to the hard work of (by order of appearance in the commit list):

• @nickelc

Changed

• Fix dependenciy incompatibilities: Restrict the bitvec->funty dependency to <=1.1

Thanks

This release was done thanks to the hard work of (by order of appearance in the commit list):

• @sachaarbonel
• @vallentin
• @Lucretiel
• @meiomorphism
• @jufajardini
• @neithernut
• @drwilco

Changed

• readme and documentation fixes
• rewrite of fold_many_m_n
• relax trait bounds on some parsers
• implement std::error::Error on VerboseError

Thanks

This release was done thanks to the hard work of (by order of appearance in the commit list):

• @Leonqn
• @nickelc
• @toshokan
• @juchiast
• @shssoichiro
• @jlkiri
• @chifflier
• @fkloiber
• @Kaoet
• @Matthew Plant

Added

• ErrorConvert implementation for VerboseError

Changed

• CI fixes
• fold_many* now accept FnMut for the accumulation function
• relaxed input bounds on length_count

Fixed

• documentation fixes
• the #[deprecated] attribute was removed from traits because it does not compile anymore on nightly
• bits and bytes combinators from the bits modules are now converted to use FnMut

Thanks

This release was done thanks to the hard work of (by order of appearance in the commit list):

• @chifflier
• @shepmaster
• @amerelo
• @razican
• @Palladinium
• @0ndorio
• Sebastian Zivota
• @keruspe
• @devonhollowood
• @parasyte
• @nnt0
• @AntoineCezar
• @GuillaumeGomez
• @eijebong
• @stadelmanma
• @sphynx
• @snawaz
• @fosskers
• @JamesHarrison
• @calebsander
• @jthornber
• @ahmedcharles
• @rljacobson
• @benkay86
• @georgeclaghorn
• @TianyiShi2001
• @shnewto
• @alfriadox
• @resistor
• @myrrlyn
• @chipsenkbeil
• @ruza-net
• @fanf2
• @jameysharp
• @FallenWarrior2k
• @jmg-duarte
• @ericseppanen
• @hbina
• Andreas Molzer
• @nickelc
• @bgourlie

Thanks

• @Alexhuszagh for float fixes
• @AlexanderEkdahl, @JoshOrndorff, @akitsu-sanae for docs fixes
• @ignatenkobrain: dependency update
• @derekdreery: map implementation for errors
• @Lucretiel for docs fixes and compilation fixes
• adytzu2007: warning fixes
• @lo48576: error management fixes

Fixed

• C symbols compilation errors due to old lexical-core version

Added

• Err now has a map function

Changed

• Make error::context() available without alloc feature

Thanks

• @Hywan, @nickmooney, @jplatte, @ngortheone, @ejmg, @SirWindfield, @demurgos, @spazm, @nyarly, @guedou, @adamnemecek, for docs fixes
• @Alxandr for error management bugfixes
• @Lucretiel for example fixes and optimizations
• @adytzu2007 for optimizations
• @audunhalland for utf8 fixes

Fixed

• panic in convert_error
• compile_error macro usage

Added

• std::error::Error, std::fmt::Display, Eq, ToOwned implementations for errors
• inline attribute for ToUsize

Changed

• convert_error optimization
• alt optimization

Thanks

• @waywardmonkeys, @phaazon, @dalance for docs fixes
• @kali for many0_m_n fixes
• @ia0 for macros fixes

Fixed

• many0_m_n now supports the n=1 case
• relaxed trait requirements in cut
• peek! macro reimplementation
• type inference in value!

This version comes with a complete rewrite of nom internals to use functions as a base for parsers, instead of macros. Macros have been updated to use functions under the hood, so that most existing parsers will work directly or require minimal changes.

The CompleteByteSlice and CompleteStr input types were removed. To get different behaviour related to streaming or complete input, there are different versions of some parsers in different submodules, like nom::character::streaming::alpha0 and nom::character::complete::alpha0.

The verbose-errors feature is gone, now the error type is decided through a generic bound. To get equivalent behaviour to verbose-errors, check out nom::error::VerboseError

Thanks

• @lowenheim helped in refactoring and error management
• @Keruspe helped in refactoring and fixing tests
• @pingiun, @Songbird0, @jeremystucki, @BeatButton, @NamsooCho, @Waelwindows, @rbtcollins, @MarkMcCaskey for a lot of help in rewriting the documentation and adding code examples
• @GuillaumeGomez for documentation rewriting and checking
• @iosmanthus for bug fixes
• @lo48576 for error management fixes
• @vaffeine for macros visibility fixes
• @webholik and @Havvy for escaped and escaped_transform fixes
• @proman21 for help on porting bits parsers

Added

• the VerboseError type accumulates position info and error codes, and can generate a trace with span information
• the lexical-core crate is now used by default (through the lexical compilation feature) to parse floats from text
• documentation and code examples for all functions and macros

Changed

• nom now uses functions instead of macros to generate parsers
• macros now use the functions under the hood
• the minimal Rust version is now 1.31
• the verify combinator's condition function now takes its argument by reference
• cond will now return the error of the parser instead of None
• alpha*, digit*, hex_digit*, alphanumeric* now recognize only ASCII characters

Removed

• deprecated string parsers (with the _s suffix), the normal version can be used instead
• verbose-errors is not needed anymore, now the error type can be decided when writing the parsers, and parsers provided by nom are generic over the error type
• AtEof, CompleteByteSlice and CompleteStr are gone, instead some parsers are specialized to work on streaming or complete input, and provided in different modules
• character parsers that were aliases to their *1 version: eol, alpha, digit, hex_digit, oct_digit, alphanumeric, space, multispace
• count_fixed macro
• whitespace::sp can be replaced by character::complete::multispace0
• method combinators are now in the nom-methods crate
• take_until_either, take_until_either1, take_until_either_and_consume and take_until_either_and_consume1: they can be replaced with is_not (possibly combined with something else)
• take_until_and_consume, take_until_and_consume1: they can be replaced with take_until combined with take
• sized_buffer and length_bytes!: they can be replaced with the length_data function
• non_empty, begin and rest_s function
• cond_reduce!, cond_with_error!, closure!, apply, map_res_err!, expr_opt!, expr_res!
• alt_complete, separated_list_complete, separated_nonempty_list_complete

Fixed

• add missing build.rs file to the package
• fix code comparison links in changelog

Fixed

• regression in do_parse macro import for edition 2018

Fixed

• macro expansion error in do_parse due to compile_error macro usage

Thanks

• @JoshMcguigan for unit test fixes
• @oza for documentation fixes
• @wackywendell for better error conversion
• @Zebradil for documentation fixes
• @tsraom for new combinators
• @hcpl for minimum Rust version tests
• @KellerFuchs for removing some unsafe uses in float parsing

Changed

• macro import in edition 2018 code should work without importing internal macros now
• the regex parsers do not require the calling code to have imported the regex crate anymore
• error conversions are more ergonomic
• method combinators are now deprecated. They might be moved to a separate crate
• nom now specifies Rust 1.24.1 as minimum version. This was already the case before, now it is made explicit

Added

• many0_count and many1_count to count applications of a parser instead of accumulating its results in a Vec

Fixed

• overflow in the byte wrapper for bit level parsers
• f64 parsing does not use transmute anymore

Fixed

• compilation issue in verbose-errors mode for add_return_error

Thanks

• @xfix for fixing warnings, simplifying examples and performance fixes
• @dvberkel for documentation fixes
• @chifflier for fixing warnings
• @myrrlyn for dead code elimination
• @petrochenkov for removing redundant test macros
• @tbelaire for documentation fixes
• @khernyo for fixing warnings
• @linkmauve for documentation fixes
• @ProgVal for documentation fixes, warning fixes and error management
• @Nemo157 for compilation fixes
• @RReverser for documentation fixes
• @xpayn for fixing warnings
• Blas Rodriguez Irizar for documentation fixes
• @badboy for documentation fixes
• @kyrias for compilation fixes
• @kurnevsky for the rest_len parser
• @hjr3 for new documentation examples
• @fengalin for error management
• @ithinuel for the pcap example project
• @phaazon for documentation fixes
• @juchiast for documentation fixes
• @jrakow for the u128 and i128 parsers
• @smarnach for documentation fixes
• @derekdreery for pub(crate) support
• @YaLTeR for map_res_err!

Added

• rest_len parser, returns the length of the remaining input
• parse_to has its own error code now
• u128 and i128 parsers in big and little endian modes
• support for pub(crate) syntax
• map_res_err! combinator that appends the error of its argument function in verbose errors mode

Fixed

• lots of unused imports warnings were removed
• the bytes combinator was not compiling in some cases
• the big and little endian combinators now work without external imports
• CI is now faster and uses less cache
• in add_return_error, the provided error code is now evaluated only once

Changed

• fold_many1 will now transmit a Failure instead of transforming it to an Error
• float and double now work on all of nom's input types (&[u8], &str, CompleteByteSlice, CompleteStr and any type that implements the required traits). float_s and double_s got the same modification, but are now deprecated
• CompleteByteSlice and CompleteStr get a small optimization by inlining some functions

Thanks

• @jsgf for the new AtEof trait
• @tmccombs for fixes on escaped* combinators
• @s3bk for fixes around non Copy input types and documentation help
• @kamarkiewicz for fixes to no_std and CI
• @bheisler for documentation and examples
• @target-san for simplifying the InputIter trait for &[u8]
• @willmurphyscode for documentation and examples
• @Chaitanya1416 for typo fixes
• @fflorent for input_len() usage fixes
• @dbrgn for typo fixes
• @iBelieve for no_std fixes
• @kpp for warning fixes and clippy fixes
• @keruspe for fixes on FindToken
• @dtrebbien for fixes on take_until_and_consume1
• @Henning-K for typo fixes
• @vthriller for documentation fixes
• @federicomenaquintero and @veprbl for their help fixing the float parsers
• @vmchale for new named_args versions
• @hywan for documentation fixes
• @fbenkstein for typo fixes
• @CAD97 for catching missing trait implementations
• @goldenlentils for &str optimizations
• @passy for typo fixes
• @ayrat555 for typo fixes
• @GuillaumeGomez for documentation fixes
• @jrakow for documentation fixes and fixes for switch!
• @phlosioneer for documentation fixes
• @creativcoder for typo fixes
• @derekdreery for typo fixes
• @lucasem for implementing Deref on CompleteStr and CompleteByteSlice
• @lowenheim for parse_to! fixes
• @myrrlyn for trait fixes around CompleteStr and CompleteByteSlice
• @NotBad4U for fixing code coverage analysis
• @murarth for code formatting
• @glandium for fixing build in no_std
• @csharad for regex compatibility with CompleteStr
• @FauxFaux for implementing AsRef<str> on CompleteStr
• @jaje for implementing std::Error on nom:Err
• @fengalin for warning fixes
• @@khernyo for doc formatting

Special thanks to @corkami for the logo :)

Breaking changes

• the IResult type now becomes a Result from the standard library
• Incomplete now returns the additional data size needed, not the total data size needed
• verbose-errors is now a superset of basic errors
• all the errors now include the related input slice
• the arguments from error_position and other such macros were swapped to be more consistent with the rest of nom
• automatic error conversion: to fix error type inference issues, a custom error type must now implement std::convert::From<u32>
• the not! combinator returns unit ()
• FindToken's calling convention was swapped
• the take_* combinators are now more coherent and stricter, see commit 484f6724ea3ccb for more information
• many0 and other related parsers will now return Incomplete if the reach the end of input without an error of the child parser. They will also return Incomplete on an empty input
• the sep! combinator for whitespace only consumes whitespace in the prefix, while the ws! combinator takes care of consuming the remaining whitespace

Added

• the AtEof trait for input type: indicate if we can get more input data later (related to streaming parsers and Incomplete handling)
• the escaped* parsers now support the &strinput type
• the Failure error variant represents an unrecoverable error, for which alt and other combinators will not try other branches. This error means we got in the right part of the code (like, a prefix was checked correctly), but there was an error in the following parts
• the CompleteByteSlice and CompleteStr input types consider there will be no more refill of the input. They fixed the Incomplete related issues when we have all of the data
• the exact!() combinator will fail if we did not consume the whole input
• the take_while_m_n! combinator will match a specified number of characters
• ErrorKind::TakeUntilAndConsume1
• the recognize_float parser will match a float number's characters, but will not transform to a f32 or f64
• alpha and other basic parsers are now much stricter about partial inputs. We also introduce the *0 and *1 versions of those parsers
• named_args can now specify the input type as well
• HexDisplay is now implemented for &str
• alloc feature
• the InputTakeAtposition trait allows specialized implementations of parsers like take_while!

Removed

• the producers and consumers were removed
• the error_code and error_node macros are not used anymore

Fixed

• anychar! now works correctly with multibyte characters
• take_until_and_consume1! no longer results in "no method named `find_substring`" and "no method named `slice`" compilation errors
• take_until_and_consume1! returns the correct Incomplete(Needed) amount
• no_std compiles properly, and nom can work with alloc too
• parse_to! now consumes its input

Changed

• alt and other combinators will now clone the input if necessary. If the input is already Copy there is no performance impact
• the rest parser now works on various input types
• InputIter::Item for &[u8] is now a u8 directly, not a reference
• we now use the compile_error macro to return a compile time error if there was a syntax issue
• the permutation combinator now supports optional child parsers
• the float numbers parsers have been refactored to use one common implementation that is nearly 2 times faster than the previous one
• the float number parsers now accept more variants

Thanks

• @ordian for alt_complete fixes
• @friedm for documentation fixes
• @kali for improving error management

Fixed

• there were cases where alt_complete could return Incomplete

Added

• an into_error_kind method can be used to transform any error to a common value. This helps when the library is included multiple times as dependency with different feature sets

Thanks

• @jedireza for documentation fixes
• @gmorenz for the bytes combinator
• @meh for character combinator fixes for UTF-8
• @jethrogb for avoiding move issues in separated_list

Changed

• new layout for the main page of documentation
• anychar can now work on any input type
• length_bytes is now an alias for length_data

Fixed

• one_of, none_of and char will now index correctly UTF-8 characters
• the compiler_error macro is now correctly exported

Added

• the bytes combinator transforms a bit stream back to a byte slice for child parsers

Thanks

• @sdroege: implementing be_i24 and le_i24
• @Hywan: integrating faster substring search using memchr
• @nizox: fixing type issues in bit stream parsing
• @grissiom: documentation fixes
• @doomrobo: implementing separated_list_complete and separated_nonempty_list_complete
• @CWood1: fixing memchr integration in no_std
• @lu_zero: integrating the compiler_error crate
• @dtolnay: helping debug a type inference issue in map

Changed

• memchr is used for substring search if possible
• if building on nightly, some common syntax errors will display a specific error message. If building no stable, display the documentation to activate those messages
• count no longer preallocates its vector

Fixed

• better type inference in alt_complete
• alt should now work with whitespace parsing
• map should not make type inference errors anymore

Added

• be_i24 and le_i24, parsing big endian and little endian signed 24 bit integers
• separated_list_complete and separated_nonempty_list_complete will treat incomplete from sub parsers as error

Thanks

• Chris Pick for some Incomplete related refactors
• @dbrgn for documentation fixes
• @valarauca for adding be_u24
• @ithinuel for usability fixes
• @evuez for README readability fixes and improvements to IResult
• @s3bk for allowing non-Copy types as input
• @keruspe for documentation fixes
• @0xd34d10cc for trait fixes on InputIter
• @sdleffler for lifetime shenanigans on named_args
• @chengsun for type inference fixes in alt
• @iBelieve for adding str to no_std
• @Hywan for simplifying code in input traits
• @azerupi for extensive documentation of alt and alt_complete

Breaking Changes

• escaped, separated_list and separated_nonempty_list can now return Incomplete when necessary
• InputIter does not require AsChar on its Item type anymore
• the core feature that was putting nom in no_std mode has been removed. There is now a std feature, activated by default. If it is not activated, nom is in no_std
• in verbose-errors mode, the error list is now stored in a Vec instead of a box based linked list
• chain! has finally been removed

Changed

• Endianness now implements Debug, PartialEq, Eq, Clone and Copy
• custom input types can now be cloned if they're not Copy
• the infamous 'Cannot infer type for E' error should happen less often now
• str is now available in no_std mode

Fixed

• FileProducer will be marked as Eof on full buffer
• named_args! now has lifetimes that cannot conflict with the lifetimes from other arguments

Added

• be_u24: big endian 24 bit unsigned integer parsing
• IResult now has a unwrap_or method

Thanks

• @Victor-Savu for formatting fixes in the README
• @chifflier for detecting and fixing integer overflows
• @utkarshkukreti for some performance improvements in benchmarks

Changed

• when calculating how much data is needed in IResult::Incomplete, the addition could overflow (it is stored as a usize). This would apparently not result in any security vulnerability on release code

Thanks

• @seppo0010 for fixing named_args
• @keruspe for implementing or() on IResult, adding the option of default cases in switch!, adding support for cargo-travis
• @timlyo for documentation fixes
• @JayKickliter for extending hex_u32
• @1011X for fixing regex integration
• @Kerollmops for actually marking chain! as deprecated
• @joliss for documentation fixes
• @utkarshkukreti for tests refactoring and performance improvement
• @tmccombs for documentation fixes

Added

• IResult gets an or() method
• take_until1, take_until_and_consume1, take_till1! and take_till1_s! require at least 1 character

Changed

• hex_u32 accepts uppercase digits as well
• the character based combinators leverage the input traits
• the whitespace parsers now work on &str and other types
• take_while1 returns Incomplete on empty input
• switch! can now take a default case

Fixed

• named_args! now imports IResult directly
• the upgrade to regex 0.2 broke the regex combinators, they work now

Thanks

• @nickbabcock for documentation fixes
• @derekdreery for documentation fixes
• @DirkyJerky for documentation fixes
• @saschagrunert for documentation fixes
• @lucab for documentation fixes
• @hyone for documentation fixes
• @tstorch for factoring Slice
• @shepmaster for adding crate categories
• @antoyo for adding named_args!

Added

• verify! uses a first parser, then applies a function to check that its result satisfies some conditions
• named_args! creates a parser function that can accept other arguments along with the input
• parse_to! will use the parse method from FromStr to parse a value. It will automatically translate the input to a string if necessary
• float, float_s, double, double_s can recognize floating point numbers in text

Changed

• escaped! will now return Incomplete if needed
• permutation! supports up to 20 child parsers

Bugfix release

Warning: there is a small breaking change, add_error! is renamed to add_return_error!. This was planned for the 2.0 release but was forgotten. This is a small change in a feature that not many people use, for a release that is not yet widely in use, so there will be no 3.0 release for that change.

Thanks

• @nickbabcock for catching and fixing the add_error! mixup
• @lucab for documentation fixes
• @jtdowney for noticing that tag_no_case! was not working at all for byte slices

Fixed

• add_error! has been renamed to add_return_error!
• the not! combinator now accepts functions
• tag_no_case! is now working as accepted (before, it accepted everything)