[initial import John Meacham **20060819204331] [lots of changes John Meacham **20060820045154] [make real packrat parser fly John Meacham **20060820081431] [rename some data types John Meacham **20060820083229] [rename project to 'frisby', move thingss around appropriatly John Meacham **20060820093347] [clean up code a bunch. move out testing stuff, rename things, add parsec compatability funcitons John Meacham **20060820120556] [add test cases John Meacham **20060820121429] [clean up comments John Meacham **20060820122258] [add grep test case John Meacham **20060821032412] [fix up regex parser, make grep example show all matches John Meacham **20060821042556] [add character classes to regex parsing. John Meacham **20060821050924] [add more normalization optimizations, make star matching more lazy John Meacham **20060821060620] [don't pass around the phantom quantified type inside implementation John Meacham **20060821062951] [add docs directory and rules for making haddock documentation. clean up export list and code some John Meacham **20060821093115] [code cleanups John Meacham **20060821095125] [add some more escape sequences to regular expression parser John Meacham **20060821100551] [add lots of documentation John Meacham **20060821114430] [add some more docs John Meacham **20060821115338] [update documentation some more John Meacham **20060822005324] [don't overflow Aaron Denney **20060822075742] [spelling and grammar fixes Aaron Denney **20060822075003] [add <++> operator, add regex normalization, code cleanups John Meacham **20060822085242] [add Grep and Lazy test cases John Meacham **20060822085349] [Fix haddock for mkRule Samuel Bronson **20060821173930] [Fix docstring for parseFailure Samuel Bronson **20060822192442] [make regular expressions allocated in the 'PM' monad. add some documentation for them. clean up some code. John Meacham **20060823011152] [Handle Char cases seperately in newRule Sven Moritz Hallberg **20070301182718 This is to satisfy GHC 6.6 the typechecker. In 'newRule', it complains when the internal function 'f' returns something of type 'PE Char' because it then assumes the whole of 'f' to return a 'PM s (P s Char)'. Removing these two cases ('Any' and 'Char') from the definition of 'f' and handling them seperately as pattern matches on the top level of 'newRule' makes GHC notice that we're dealing with a GADT and it accepts the code. ] [Silence warning about ambiguous reference to 'Any' Sven Moritz Hallberg **20070301183245] [Add -fglasgow-exts OPTIONS pragma Sven Moritz Hallberg **20070301183315] [Cabalize Sven Moritz Hallberg **20070301183438 Add generic Setup.lhs and frisby.cabal. Note that license is still set to AllRightsReserved and license-file is not set. Also of note is that -fglasgow-exts is required for the module to compile which is not given in this cabal file, so it should be in an OPTIONS pragma in the source. ] [update comment for why -fglasgow-exts Isaac Dupree **20070616193101] [remove unused import Random, System Isaac Dupree **20070616193147] [towards standardized unsafeCoerce Isaac Dupree **20070616193516] [add comment explaining unused 's' type-parameter Isaac Dupree **20070616193617] [use newtype deriving instead of explicit Functor P Isaac Dupree **20070616193851] [add Applicative interface with minimal interface-breakage Isaac Dupree **20070616193928] [remove redundant g (Slash x (Unit res)) case Isaac Dupree **20070617093334 It just had refactorable code that became very close to the g (Slash x y) case after refactoring, and I don't see a good reason to look for that Unit first - unless there's a reason to be "explicitly" strict in the "y" PE here? ] [add some explanatory comments in runPeg Isaac Dupree **20070617101927] [emptyD-->emptyDAt considering how it's used Isaac Dupree **20070617102737 It seems sounder to put the call of f together with the replacement of n since those just happen to work consistently together. (f could be called directly with argument n, but is the sharing important?) ] [more runPeg commenting Isaac Dupree **20070617105742] [comments related to doesConsumeInput, and name->may* Isaac Dupree **20070617110840] [rearrange P/PE Monoid instance Isaac Dupree **20070617111318] [Narrative explaining why Isaac Dupree **20070617120843 what we're doing is so difficult in Haskell - mdo, GADT, unsafeCoerce - the resulting P structure represents a polytypic cyclic graph, and graphs have always been difficult in Haskell (they're not inductive); we have manipulations including repeated memoization... ] [more comments, renaming, some alist->array Isaac Dupree **20070617130326 Neato! The patched version appears slightly faster than the original now! (with ghc-6.6.1 -O2 for each) (as I slightly suspected... it only needs to fmap an Array to generate each array now, rather than interpret the indices with 'array' on each input character) (also the array index is more simply Token now) ] [rearrange some code & shorten my obscenely long names Isaac Dupree **20070617133429] [improve comment describing extensions used Isaac Dupree **20070617133857] [move fromP definition to the newtype definition Isaac Dupree **20070618114600] [comment: instance Functor Results "should be" derived Isaac Dupree **20070618114748] [change licence to 3BSD John Meacham **20070921234551] [Add Unicode character parsers Felix Martini **20070923085951] [Export between Robin Green **20070405125111] [Typo in license field Robin Green **20070925132244] [add GetPos case in normalize routine. thanks to Roshan James for pointing this out John Meacham **20081120023323] [Some spelling fixes. Antonio Regidor García **20081011080929] [Make Frisby build on recent Cabal taralx@gmail.com**20100210214906 Ignore-this: 6f9f4d8c81b4ff86acb89fb18a4dda1c ]