Skip to content

Latest commit

 

History

History
1324 lines (1005 loc) · 53.7 KB

CHANGELOG.md

File metadata and controls

1324 lines (1005 loc) · 53.7 KB

Change Log

[Unreleased][unreleased]

Thanks

Changed

6.0.1 - 2020-11-24

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

6.0.0 - 2020-10-31

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

Notable changes

This release is a more polished version of nom 5, that came with a focus on function parsers, by relaxing the requirements: combinators will return a impl FnMut instead of impl Fn, allowing closures that change their context, and parsers can be any type now, as long as they implement the new Parser trait. That parser trait also comes with a few helper methods.

Error management was often a pain point, so a lot of work went into making it easier. Now it integrates with std:error::Error, the IResult::finish() method allows you to convert to a more usable type, the into combinator can convert the error type if there's a From implementation, and there are more specific error traits like ContextError for the context combinator, and FromExternalError for map_res. While the VerboseError type and its convert_error function saw some changes, not many features ill be added to it, instead you are encouraged to build the error type that corresponds to your needs if you are building a language parser.

This version also integrates with the excellent bitvec crate for better bit level parsing. This part of nom was not great and a bit of a hack, so this will give better options for those parsers.

At last, documentation! There are now more code examples, functions and macros that require specific cargo features are now clearly indicated, and there's a new recipes module containing example patterns.

Breaking changes

  • the minimal Rust version is now 1.44 (1.37 if building without the alloc or std features)
  • streaming parsers return the number of additional bytes they need, not the total. This was supposed to be the case everywhere, but some parsers were forgotten
  • removed the regexp_macros cargo feature
  • the context combinator is not linked to ParseError anymore, instead it come with its own ContextError trait
  • Needed::Size now contains a NonZeroUsize, so we can reduce the structure's size by 8 bytes. When upgrading, Needed::Size(number) can be replaced with Needed::new(number)
  • there is now a more general Parser trait, so parsers can be something else than a function. This trait also comes with combinator methods like map, flat_map, or. Since it is implemented on Fn* traits, it should not affect existing code too much
  • combinators that returned a impl Fn now return a impl FnMut to allow parser closures that capture some mutable value from the context
  • separated_list is now separated_list0
  • removed the deprecated methods module
  • removed the deprecated whitespace module
  • the default error type is now a struct (nom::error::Error) instead of a tuple
  • the FromExternalError allows wrapping the error returned by the function in the map_res combinator
  • renamed the dbg! macro to avoid conficts with std::dbg!
  • separated_list now allows empty elements

Added

  • function version of regex parsers
  • fill: attempts to fill the output slice passed as argument
  • success: returns a value without consuming the input
  • satisfy: checks a predicate over the next character
  • eof function combinator
  • consumed: returnes the produced value and the consumed input
  • length_count function combinator
  • into: converts a parser's output and error values if From implementations are available
  • IResult::finish(): converts a parser's result to Result<(I, O), E> by removing the distinction between Error and Failure and panicking on Incomplete
  • non macro versions of u16, i32, etc, with configurable endianness
  • is_newline function
  • std::error::Error implementation for nom's error types
  • recipes section of the documentation, outlining common patterns in nom
  • custom errors example
  • bitstream parsing with the BitSlice type from the bitvec crate
  • native endianness parsers
  • github actions for CI

Changed

  • allows lexical-core 0.7
  • number parsers are now generic over the input type
  • stabilized the alloc feature
  • convert_error accepts a type that derefs to &str
  • the JSON example now follows the spec better

Fixed

  • use fold_many0c in the fold_many0 macro

5.1.1 - 2020-02-24

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

5.1.0 - 2020-01-07

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

5.0.1 - 2019-08-22

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!

5.0.0 - 2019-06-24

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

4.2.3 - 2019-03-23

Fixed

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

4.2.2 - 2019-03-04

Fixed

  • regression in do_parse macro import for edition 2018

4.2.1 - 2019-02-27

Fixed

  • macro expansion error in do_parse due to compile_error macro usage

4.2.0 - 2019-01-29

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

4.1.1 - 2018-10-14

Fixed

  • compilation issue in verbose-errors mode for add_return_error

4.1.0 - 2018-10-06

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

4.0.0 - 2018-05-14

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

3.2.1 - 2017-10-27

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

3.2.0 - 2017-07-24

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

3.1.0 - 2017-06-16

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

3.0.0 - 2017-05-12

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

2.2.1 - 2017-04-03

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

2.2.0 - 2017-03-20

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

2.1.0 - 2017-01-27

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

2.0.1 - 2016-12-10

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)

2.0 - 2016-11-25

The 2.0 release is one of the biggest yet. It was a good opportunity to clean up some badly named combinators and fix invalid behaviours.

Since this version introduces a few breaking changes, an upgrade documentation is available, detailing the steps to fix the most common migration issues. After testing on a set of 30 crates, most of them will build directly, a large part will just need to activate the "verbose-errors" compilation feature. The remaining fixes are documented.

This version also adds a lot of interesting features, like the permutation combinator or whitespace separated formats support.

Thanks

  • @lu-zero for license help
  • @adamgreig for type inference fixes
  • @keruspe for documentation and example fixes, for the IResult => Result conversion work, making AsChar's method more consistent, and adding many_till!
  • @jdeeny for implementing Offset on &str
  • @vickenty for documentation fixes and his refactoring of length_value! and length_bytes!
  • @overdrivenpotato for refactoring some combinators
  • @taralx for documentation fixes
  • @keeperofdakeys for fixing eol behaviour, writing documentation and adding named_attr!
  • @jturner314 for writing documentation
  • @bozaro for fixing compilation errors
  • @uniphil for adding a crates.io badge
  • @badboy for documentation fixes
  • @jugglerchris for fixing take_s!
  • @AndyShiue for implementing Error and Display on ErrorKind and detecting incorrect UTF-8 string indexing

Added

  • the "simple" error management system does not accumulates errors when backtracking. This is a big perf gain, and is activated by default in nom 2.0
  • nom can now work on any type that implement the traits defined in src/traits.rs: InputLength, InputIter, InputTake, Compare, FindToken, FindSubstring, Slice
  • the documentation from Github's wiki has been moved to the doc/ directory. They are markdown files that you can build with cargo-external-doc
  • whitespace separated format support: with the ws! combinator, you can automatically introduce whitespace parsers between all parsers and combinators
  • the permutation! combinator applies its child parsers in any order, as long as they all succeed once, and return a tuple of the results
  • do_parse! is a simpler alternative to chain!, which is now deprecated
  • you can now transform an IResult in a std::result::Result
  • length_data! parses a length, and returns a subslice of that length
  • tag_no_case! provides case independent comparison. It works nicely, without any allocation, for ASCII strings, but for UTF-8 strings, it defaults to an unsatisfying (and incorrect) comparison by lowercasing both strings
  • named_attr! creates functions like named! but can add attributes like documentation
  • many_till! applies repeatedly its first child parser until the second succeeds

Changed

  • the "verbose" error management that was available in previous versions is now activated by the "verbose-errors" compilation feature
  • code reorganization: most of the parsers were moved in separate files to make the source easier to navigate
  • most of the combinators are now independent from the input type
  • the eof function was replaced with the eof! macro
  • error! and add_error! were replaced with return_error! and add_return_error! to fix the name conflict with the log crate
  • the offset() method is now in the Offset trait
  • length_value! has been renamed to length_count!. The new length_value! selects a slice and applies the second parser once on that slice
  • AsChar::is_0_to_9 is now AsChar::is_dec_digit
  • the combinators with configurable endianness now take an enum instead of a boolean as parameter

Fixed

  • the count!, count_fixed! and length_*! combinator calculate incomplete data needs correctly
  • eol, line_ending and not_line_ending now have a consistent behaviour that works correctly with incomplete data
  • take_s! didn't correctly handle the case when the slice is exactly the right length

1.2.4 - 2016-07-20

Thanks

  • @Phlosioneer for documentation fixes
  • @sourrust for fixing offsets in take_bits!
  • @ChrisMacNaughton for the XFS crate
  • @pwoolcoc for rest_s
  • @fitzgen for more IResult methods
  • @gtors for the negative lookahead feature
  • @frk1 and @jeandudey for little endian float parsing
  • @jethrogb for fixing input usage in many1
  • @acatton for beating me at nom golf :D

Added

  • the rest_s method on IResult returns the remaining &str input
  • unwrap_err and unwrap_inc methods on IResult
  • not! will peek at the input and return Done if the underlying parser returned Error or Incomplete, without consuming the input
  • le_f32 and le_f64 parse little endian floating point numbers (IEEE 754)

Fixed

  • documentation fixes
  • take_bits! is now more precise
  • many1 inccorectly used the len function instead of input_len
  • the INI parser is simpler
  • recognize! had an early return that is removed now

1.2.3 - 2016-05-10

Thanks

  • @lu-zero for the contribution guidelines
  • @GuillaumeGomez for fixes on length_bytes and some documentation
  • @Hywan for documentation and test fixes
  • @Xirdus for correct trait import issues
  • @mspiegel for the new AST example
  • @cholcombe973 for adding the cond_with_error! combinator
  • @tstorch for refactoring many0!
  • @panicbit for the folding combinators
  • @evestera for separated_list! fixes
  • @DanielKeep for correcting some enum imports

Added

  • Regular expression combinators starting with re_bytes_ work on byte slices
  • example parsing arithmetic expressions to an AST
  • cond_with_error! works like cond! but will return None if the condition is false, and Some(value) if the underlying parser succeeded
  • fold_many0!, fold_many1! and fold_many_m_n! will take a parser, an initial value and a combining function, and fold over the successful applications of the parser

Fixed

  • length_bytes! converts the result of its child parser to usize
  • take_till! now imports InputLength instead of assuming it's in scope
  • separated_list! and separated_nonempty_list! will not consume the separator if there's no following successfully parsed value
  • no more warnings on build

Changed

  • simpler implementation of many0!

1.2.2 - 2016-03-09

Thanks

  • @conradev for fixing take_until_s!
  • @GuillaumeGomez for some documentation fixes
  • @frewsxcv for some documentation fixes
  • @tstorch for some test refactorings

Added

  • nom::Err now implements std::error::Error

Fixed

  • hex_u32 does not parses more than 8 chars now
  • take_while! and take_while1! will not perturb the behaviour of recognize! anymore

1.2.1 - 2016-02-23

Thanks

  • @sourrust for adding methods to IResult
  • @tstorch for the test refactoring, and for adding methods to IResult and Needed
  • @joelself for fixing the method system

Added

  • mapping methods over IResult and Needed

Changed

  • apply_rf is renamed to apply_m. This will not warrant a major version, since it is part missing from the methods feture added in the 1.2.0 release
  • the regexp_macros feature that used regex! to precompile regular expressions has been replaced by the normal regex engine combined with lazy_static

Fixed

  • when a parser or combinator was returning an empty buffer as remaining part, it was generating one from a static empty string. This was messing with buffer offset calculation. Now, that empty slice is taken like this: &input[input.len()..].
  • The regexp_macros and no_std feature build again and are now tested with Travis CI

1.2.0 - 2016-02-08

Thanks

  • @zentner-kyle for type inference fixes
  • @joelself for his work on &str parsing and method parsers
  • @GuillaumeGomez for implementing methods on IResult
  • @dirk for the alt_complete! combinator
  • @tstorch for a lot of refactoring work and unit tests additions
  • @jansegre for the hex digit parsers
  • @belgum for some documentation fixes
  • @lwandrebeck for some documentation fixes and code fixes in hex_digit

Added

  • take_until_and_consume_s! for consumption of string data until a tag
  • more function patterns in named!. The error type can now be specified
  • alt_complete! works like the alt! combinator, but tries the next branch if the current one returned Incomplete, instead of returning directly
  • more unit tests for a lot of combinators
  • hexadecimal digit parsers
  • the tuple! combinator takes a list of parsers as argument, and applies them serially on the input. If all of them are successful, it willr eturn a tuple accumulating all the values. This combinator will (hopefully) replace most uses of chain!
  • parsers can now be implemented as a method for a struct thanks to the method!, call_m! and apply_rf! combinators

Fixed

  • there were type inference issues in a few combinators. They will now be easier to compile
  • peek! compilation with bare functions
  • &str parsers were splitting data at the byte level, not at the char level, which can result in inconsistencies in parsing UTF-8 characters. They now use character indexes
  • some method implementations were missing on IResult<I,O,E> (with specified error type instead of implicit)

1.1.0 - 2016-01-01

This release adds a lot of features related to &str parsing. The previous versions were focused on &[u8] and bit streams parsing, but there's a need for more text parsing with nom. The parsing functions like alpha, digit and others will now accept either a &[u8] or a &str, so there is no breaking change on that part.

There are also a few performance improvements and documentation fixes.

Thanks

  • @Binero for pushing the work on &str parsing
  • @meh for fixing Option and Vec imports
  • @hoodie for a documentation fix
  • @joelself for some documentation fixes
  • @vberger for his traits magic making nom functions more generic

Added

  • string related parsers: tag_s!, take_s!, is_a_s!, is_not_s!, take_while_s!, take_while1_s!, take_till_s!
  • value! is a combinator that always returns the same value. If a child parser is passed as second argument, that value is returned when the child parser succeeds

Changed

  • tag! will now compare even on partial input. If it expects "abcd" but receives "ef", it will now return an Error instead of Incomplete
  • many0! and others will preallocate a larger vector to avoid some copies and reallocations
  • alpha, digit, alphanumeric, space and multispace now accept as input a &[u8] or a &str. Additionally, they return an error if they receive an empty input
  • take_while!, take_while1!, take_while_s!, take_while1_s! wilreturn an error on empty input

Fixed

  • if the child parser of many0! or many1! returns Incomplete, it will return Incomplete too, possibly updating the needed size
  • Option, Some, None and Vec are now used with full path imports

1.0.1 - 2015-11-22

This releases makes the 1.0 version compatible with Rust 1.2 and 1.3

Thanks

  • @steveklabnik for fixing lifetime issues in Producers and Consumers

1.0.0 - 2015-11-16

Stable release for nom. A lot of new features, a few breaking changes

Thanks

  • @ahenry for macro fixes
  • @bluss for fixing documentation
  • @sourrust for cleaning code and debugging the new streaming utilities
  • @meh for inline optimizations
  • @ccmtaylor for fixing function imports
  • @soro for improvements to the streaming utilities
  • @breard-r for catching my typos
  • @nelsonjchen for catching my typos too
  • @divarvel for hex string parsers
  • @mrordinaire for the length_bytes! combinator

Breaking changes

  • IResult::Error can now use custom error types, and is generic over the input type
  • Producers and consumers have been replaced. The new implementation uses less memory and integrates more with parsers
  • nom::ErrorCode is now nom::ErrorKind
  • filter! has been renamed to take_while!
  • chain! will count how much data is consumed and use that number to calculate how much data is needed if a parser returned Incomplete
  • alt! returns Incomplete if a child parser returned Incomplete, instead of skipping to the next parser
  • IResult does not require a lifetime tag anymore, yay!

Added

  • complete! will return an error if the child parser returned Incomplete
  • add_error! will wrap an error, but allow backtracking
  • hex_u32 parser

Fixed

  • the behaviour around Incomplete is better for most parsers now

0.5.0 - 2015-10-16

This release fixes a few issues and stabilizes the code.

Thanks

  • @nox for documentation fixes
  • @daboross for linting fixes
  • @ahenry for fixing tap! and extending dbg! and dbg_dmp!
  • @bluss for tracking down and fixing issues with unsafe code
  • @meh for inlining parser functions
  • @ccmtaylor for fixing import of str::from_utf8

Fixed

  • tap!, dbg! and dbg_dmp! now accept function parameters

Changed

  • the type used in count_fixed! must be Copy
  • chain! calculates how much data is needed if one of the parsers returns `Incomplete
  • optional parsers in chain! can return Incomplete

0.4.0 - 2015-09-08

Considering the number of changes since the last release, this version can contain breaking changes, so the version number becomes 0.4.0. A lot of new features and performance improvements!

Thanks

  • @frewsxcv for documentation fixes
  • @ngrewe for his work on producers and consumers
  • @meh for fixes on chain! and for the rest parser
  • @daboross for refactoring many0! and many1!
  • @aleksander for the switch! combinator idea
  • @TechnoMancer for his help with bit level parsing
  • @sxeraverx for pointing out a bug in is_a!

Fixed

  • count_fixed! must take an explicit type as argument to generate the fixed-size array
  • optional parsing behaviour in chain!
  • count! can take 0 elements
  • is_a! and is_not! can now consume the whole input

Added

  • it is now possible to seek to the end of a MemProducer
  • opt! returns Done(input, None) if the child parser returned Incomplete`
  • rest will return the remaining input
  • consumers can now seek to and from the end of input
  • switch! applies a first parser then matches on its result to choose the next parser
  • bit-level parsers
  • character-level parsers
  • regular expression parsers
  • implementation of take_till!, take_while! and take_while1!

Changed

  • alt! can return Incomplete
  • the error analysis functions will now take references to functions instead of moving them
  • performance improvements on producers
  • performance improvement for filter!
  • performance improvement for count!: a Vec of the right size is directly allocated

0.3.11 - 2015-08-04

Thanks

  • @bluss for remarking that the crate included random junk lying non commited in my local repository

Fixed

  • cleanup of my local repository will ship less files in the crates, resulting in a smaller download

0.3.10 - 2015-08-03

Added

  • bits! for bit level parsing. It indicates that all child parsers will take a (&[u8], usize)as input, with the second parameter indicating the bit offset in the first byte. This allows viewing a byte slice as a bit stream. Most combinators can be used directly under bits!
  • take_bits! takes an integer type and a number of bits, consumes that number of bits and updates the offset, possibly by crossing byte boundaries
  • bit level parsers are all written in src/bits.rs

Changed

  • Parsers that specifically handle bytes have been moved to src/bytes.rs. This applies to tag!, is_not!, is_a!, filter!, take!, take_str!, take_until_and_consume!, take_until!, take_until_either_and_consume!, take_until_either!`

0.3.9 - 2015-07-20

Thanks

  • @badboy for fixing filter!
  • @idmit for some documentation fixes

Added

  • opt_res! applies a parser and transform its result in a Result. This parser never fails
  • cond_reduce! takes an expression as parameter, applies the parser if the expression is true, and returns an error if the expression is false
  • tap! pass the result of a parser to a block to manipulate it, but do not affect the parser's result
  • AccReader is a Read+BufRead that supports data accumulation and partial consumption. The consume method must be called afterwardsto indicate how much was consumed
  • Arithmetic expression evaluation and parsing example
  • u16!, u32!, u64!, i16!, i32!, i64! take an expression as parameter, if the expression is true, apply the big endian integer parser, if false, the little endian version
  • type information for combinators. This will make the documentation a bit easier to navigate

Fixed

  • map_opt! and map_res! had issues with argument order due to bad macros
  • delimited! did not compile for certain combinations of arguments
  • filter! did not return a byte slice but a fixed array

0.3.8 - 2015-07-03

Added

  • code coverage is now calculated automatically on Travis CI
  • Stepper: wrap a Producer, and call the method step with a parser. This method will buffer data if there is not enough, apply the parser if there is, and keep the rest of the input in memory for the next call
  • ReadProducer: takes something implementing Read, and makes a Producer out of it

Fixed

  • the combinators separated_pair! and delimited! did not work because an implementation macro was not exported
  • if a MemProducer reached its end, it should always return Eof
  • map! had issues with argument matching

0.3.7 - 2015-06-24

Added

  • expr_res! and expr_opt! evaluate an expression returning a Result or Opt and convert it to IResult
  • AsBytes is implemented for fixed size arrays. This allows tag!([41u8, 42u8])

Fixed

  • count_fixed! argument parsing works again

0.3.6 - 2015-06-15

Added

  • documentation for a few functions
  • the consumer trait now requires the failed(&self, error_code) method in case of parsing error
  • named! now handles thge alternative named!(pub fun_name<OutputType>, ...)

Fixed

  • filter! now returns the whole input if the filter function never returned false
  • take! casts its argument as usize, so it can accepts any integer type now

0.3.5 - 2015-06-10

Thanks

  • @cmr for some documentation fixes

Added

  • count_fixed! returns a fixed array

Fixed

  • count! is back to the previous behaviour, returning a Vec for sizes known at runtime

Changed

  • functions and traits exported from nom::util are now directly in nom::

0.3.4 - 2015-06-09

Thanks

  • @andrew-d for fixes on cond!
  • @keruspe for features in chain!

Added

  • chain! can now have mutable fields

Fixed

  • cond! had an infinite macro recursion

Changed

  • chain! generates less code now. No apprent compilation time improvement

0.3.3 - 2015-06-09

Thanks

  • @andrew-d for the little endian signed integer parsers
  • @keruspe for fixes on count!

Added

  • le_i8, le_i16, le_i32, le_i64: little endian signed integer parsers

Changed

  • the alt! parser compiles much faster, even with more than 8 branches
  • count! can now return a fixed size array instead of a growable vector

0.3.2 - 2015-05-31

Thanks

  • @keruspe for the take_str parser and the function application combinator

Added

  • take_str!: takes the specified number of bytes and return a UTF-8 string
  • apply!: do partial application on the parameters of a function

Changed

  • Needed::Size now contains a usize instead of a u32

0.3.1 - 2015-05-21

Thanks

  • @divarvel for the big endian signed integer parsers

Added

  • be_i8, be_i16, be_i32, be_i64: big endian signed integer parsers
  • the core feature can be passed to cargo to build with no_std
  • colored hexdump can be generated from error chains

0.3.0 - 2015-05-07

Thanks

  • @filipegoncalves for some documentation and the new eof parser
  • @CrimsonVoid for putting fully qualified types in the macros
  • @lu_zero for some documentation fixes

Added

  • new error types that can contain an error code, an input slice, and a list of following errors
  • error! will cut backtracking and return directly from the parser, with a specified error code
  • eof parser, successful if there is no more input
  • specific error codes for the parsers provided by nom

Changed

  • fully qualified types in macros. A lot of imports are not needed anymore

Removed

  • FlatMap, FlatpMapOpt and Functor traits (replaced by map!, map_opt! and map_res!)

0.2.2 - 2015-04-12

Thanks

  • @filipegoncalves and @thehydroimpulse for debugging an infinite loop in many0 and many1
  • @thehydroimpulse for suggesting public named parsers
  • @skade for removing the dependency on the collections gate

Added

  • named! can now declare public functions like this: named!(pub tst, tag!("abcd"));
  • pair!(X,Y) returns a tuple (x, y)
  • separated_pair!(X, sep, Y) returns a tuple (x, y)
  • preceded!(opening, X) returns x
  • terminated!(X, closing) returns x
  • delimited(opening, X, closing) returns x
  • separated_list(sep, X) returns a Vec<X>
  • separated_nonempty_list(sep, X) returns a Vec<X> of at list one element

Changed

  • many0! and many1! forbid parsers that do not consume input
  • is_a!, is_not!, alpha, digit, space, multispace will now return an error if they do not consume at least one byte

0.2.1 - 2015-04-04

Thanks

  • @mtsr for catching the remaining debug println!
  • @jag426 who killed a lot of warnings
  • @skade for removing the dependency on the core feature gate

Added

  • little endian unsigned int parsers le_u8, le_u16, le_u32, le_u64
  • count! to apply a parser a specified number of times
  • cond! applies a parser if the condition is met
  • more parser development tools in util::*

Fixed

  • in one case, opt! would not compile

Removed

  • most of the feature gates are now removed. The only one still needed is collections

0.2.0 - 2015-03-24

works with rustc 1.0.0-dev (81e2396c7 2015-03-19) (built 2015-03-19)

Thanks

  • Ryman for the AsBytes implementation
  • jag426 and jaredly for documentation fixes
  • eternaleye on #rust IRC for his help on the new macro syntax

Changed

  • the AsBytes trait improves readability, no more b"...", but "..." instead
  • Incomplete will now hold either Needed;;Unknown, or Needed::Size(u32). Matching on Incomplete without caring for the value is done with Incomplete(_), but if more granularity is mandatory, Needed can be matched too
  • alt! can pass the result of the parser to a closure
  • the take_* macros changed behaviour, the default case is now not to consume the separator. The macros have been renamed as follows: take_until! -> take_until_and_consume!, take_until_and_leave! -> take_until!, take_until_either_and_leave! -> take_until_either!, take_until_either! -> take_until_either_and_consume!

Added

  • peek! macro: matches the future input but does not consume it
  • length_value! macro: the first argument is a parser returning a n that can cast to usize, then applies the second parser n times. The macro has a variant with a third argument indicating the expected input size for the second parser
  • benchmarks are available at https://github.com/Geal/nom_benchmarks
  • more documentation
  • Unnamed parser syntax: warning, this is a breaking change. With this new syntax, the macro combinators do not generate functions anymore, they create blocks. That way, they can be nested, for better readability. The named! macro is provided to create functions from parsers. Please be aware that nesting parsers comes with a small cost of compilation time, negligible in most cases, but can quickly get to the minutes scale if not careful. If this happens, separate your parsers in multiple subfunctions.
  • named!, closure! and call! macros used to support the unnamed syntax
  • map!, map_opt! and map_res! to combine a parser with a normal function, transforming the input directly, or returning an Option or Result

Fixed

  • is_a! is now working properly

Removed

  • the o! macro does less than chain!, so it has been removed
  • the fold0! and fold1! macros were too complex and awkward to use, the many* combinators will be useful for most uses for now

0.1.6 - 2015-02-24

Changed

  • consumers must have an end method that will be called after parsing

Added

  • big endian unsigned int and float parsers: be_u8, be_u16, be_u32, be_u64, be_f32, be_f64
  • producers can seek
  • function and macros documentation
  • README documentation

Fixed

  • lifetime declarations
  • tag! can return Incomplete

0.1.5 - 2015-02-17

Changed

  • traits were renamed: FlatMapper -> FlatMap, Mapper -> FlatMapOpt, Mapper2 -> Functor

Fixed

  • woeks with rustc f1bb6c2f4

0.1.4 - 2015-02-17

Changed

  • the chaining macro can take optional arguments with '?'

0.1.3 - 2015-02-16

Changed

  • the chaining macro now takes the closure at the end of the argument list

0.1.2 - 2015-02-16

Added

  • flat_map implementation for <&[u8], &[u8]>
  • chaining macro
  • partial MP4 parser example

0.1.1 - 2015-02-06

Fixed

  • closure syntax change

Compare code