50a5cc9...master · rowtype-yoga/purescript-parsing-dataview

git.iko.soy/rowtype-yoga/purescript-parsing-dataview

on master help sign in

6 files changed+39−12

since splitsnapshots

unifiedsplit

CHANGELOG.md+3−1 README.md+7−6 bower.json+1−1 spago.dhall+1−1 src/Parsing/DataView.Basic.purs+5−1 test/Main.purs+22−2

CHANGELOG.md+3−1

@@ -1,9 +1,11 @@

11# Changelog

3−## Unreleased

3+## v3.2.0

5+- Build for PureScript v0.15.4

56- New combinator: `anyTill`

67- New parser: `anyCodePointUTF8`

8+- Move to rowtype-yoga org

810## v3.1.0

911

README.md+7−6

@@ -1,7 +1,8 @@

11# purescript-parsing-dataview

3−[![CI](https://github.com/jamesdbrock/purescript-parsing-dataview/workflows/CI/badge.svg?branch=master)](https://github.com/jamesdbrock/purescript-parsing-dataview/actions)

3+[![CI](https://github.com/rowtype-yoga/purescript-parsing-dataview/workflows/CI/badge.svg?branch=master)](https://github.com/rowtype-yoga/purescript-parsing-dataview/actions)

44[![Pursuit](http://pursuit.purescript.org/packages/purescript-parsing-dataview/badge)](http://pursuit.purescript.org/packages/purescript-parsing-dataview/)

5+[![Maintainer: jamesdbrock](https://img.shields.io/badge/maintainer-jamesdbrock-teal.svg)](https://github.com/jamesdbrock)

67Primitive parsers for

78`DataView`s on JavaScript `ArrayBuffer`s with the package

@@ -13,7 +14,7 @@ is similar to the built-in stream support of [__Megaparsec__](https://hackage.ha

1314| Stream type | parsing | Megaparsec |

1415|----|-----|----|

1516| UTF-16 strings | String | Text < v2.0|

16−| UTF-8 strings | | Text ≥ v2.0 |

17+| UTF-8 strings | __DataView__ | Text ≥ v2.0 |

1718| Listy strings | Token | String |

1819| Binary blobs | __DataView__ | ByteString |

1920

@@ -28,8 +29,8 @@ Parse values out of a `dataview :: Data.ArrayBuffer.Types.DataView`. All

2829Parse two big-endian IEEE 754 double-precision `Number`s.

2930

3031```purescript

31−import Text.Parsing.Parser (runParserT)

32−import Text.Parsing.Parser.DataView (anyFloat64be)

32+import Parsing (runParserT)

33+import Parsing.DataView (anyFloat64be)

3334

3435do

3536 result <- runParserT dataview do

@@ -43,8 +44,8 @@ do

4344Parse an array of `n` 32-bit big-endian signed `Int`s.

4445

4546```purescript

46−import Text.Parsing.Parser (runParserT)

47−import Text.Parsing.Parser.DataView (anyUint32be)

47+import Parsing (runParserT)

48+import Parsing.DataView (anyUint32be)

4849import Data.Unfoldable (replicateA)

4950

5051do

bower.json+1−1

@@ -5,7 +5,7 @@

55 ],

66 "repository": {

77 "type": "git",

8− "url": "https://github.com/jamesdbrock/purescript-parsing-dataview"

8+ "url": "https://github.com/rowtype-yoga/purescript-parsing-dataview"

99 },

1010 "ignore": [

1111 "**/.*",

spago.dhall+1−1

@@ -19,5 +19,5 @@

1919, packages = ./packages.dhall

2020, sources = [ "src/**/*.purs" ]

2121, license = "BSD-2-Clause"

22−, repository = "https://github.com/jamesdbrock/purescript-parsing-dataview"

22+, repository = "https://github.com/rowtype-yoga/purescript-parsing-dataview"

2323}

src/Parsing/DataView.Basic.purs+5−1

@@ -29,19 +29,23 @@ anyCodePointUTF8 = tryRethrow do

2929 c0 <- toInt <$> anyUint8

3030 if c0 < 128 then do

3131 liftMaybe (\_ -> "Invalid CodePoint " <> show c0) (toEnum c0)

32+ else if c0 < 192 then fail "Invalid UTF-8 encoding"

3233 else if c0 < 224 then do

3334 c1 <- toInt <$> anyUint8

35+ if c1 < 128 || c1 > 191 then fail "Invalid UTF-8 encoding" else pure unit

3436 let codepoint = (64 * (c0 `mod` 32)) + (c1 `mod` 64)

3537 liftMaybe (\_ -> "Invalid CodePoint " <> show codepoint) (toEnum codepoint)

3638 else if c0 < 240 then do

3739 c1 <- toInt <$> anyUint8

3840 c2 <- toInt <$> anyUint8

41+ if c1 < 128 || c1 > 191 || c2 < 128 || c2 > 191 then fail "Invalid UTF-8 encoding" else pure unit

3942 let codepoint = (4096 * (c0 `mod` 16)) + (64 * (c1 `mod` 64)) + (c2 `mod` 64)

4043 liftMaybe (\_ -> "Invalid CodePoint " <> show codepoint) (toEnum codepoint)

4144 else if c0 < 248 then do

4245 c1 <- toInt <$> anyUint8

4346 c2 <- toInt <$> anyUint8

4447 c3 <- toInt <$> anyUint8

48+ if c1 < 128 || c1 > 191 || c2 < 128 || c2 > 191 || c3 < 128 || c3 > 191 then fail "Invalid UTF-8 encoding" else pure unit

4549 let codepoint = (262144 * (c0 `mod` 8)) + (4096 * (c1 `mod` 64)) + (64 * (c2 `mod` 64)) + (c3 `mod` 64)

4650 liftMaybe (\_ -> "Invalid CodePoint " <> show codepoint) (toEnum codepoint)

47− else fail "Invalid first code point"

51+ else fail "Invalid UTF-8 encoding"

test/Main.purs+22−2

@@ -7,6 +7,8 @@ import Control.Monad.Trans.Class (lift)

77import Data.ArrayBuffer.ArrayBuffer as AB

88import Data.ArrayBuffer.Cast (fromUint8Array, toUint8Array)

99import Data.ArrayBuffer.DataView as DataView

10+import Data.ArrayBuffer.Typed (fromArray)

11+import Data.ArrayBuffer.Types (Uint8Array)

1012import Data.Either (Either(..))

1113import Data.Enum (fromEnum)

1214import Data.List (fromFoldable)

@@ -19,7 +21,7 @@ import Effect (Effect)

1921import Effect.Class (liftEffect)

2022import Effect.Console (logShow)

2123import Effect.Exception (catchException, message)

22−import Parsing (ParserT, Position(..), fail, liftExceptT, parseErrorPosition, runParserT)

24+import Parsing (ParseError(..), ParserT, Position(..), fail, liftExceptT, parseErrorPosition, runParserT)

2325import Parsing.Combinators (many, manyTill)

2426import Parsing.DataView (anyInt8, anyTill, eof, satisfyInt8)

2527import Parsing.DataView as DV

@@ -205,7 +207,25 @@ main = do

205207 eof

206208 pure [ c0, c1, c2, c3, c4 ]

207209

208− assertEqual' "anyCodePointUTF8"

210+ assertEqual' "anyCodePointUTF8 basic"

209211 { expected: Right [ 97, 955, 0x6708, 0x1F453, 97 ]

210212 , actual: result

211213 }

214+

215+ do

216+ test :: Uint8Array <- fromArray (map fromInt [ 0x00BF ])

217+ testv <- fromUint8Array test

218+ result <- runParserT testv anyCodePointUTF8

219+ assertEqual' "anyCodePointUTF8 invalid 1"

220+ { expected: Left (ParseError "Invalid UTF-8 encoding" (Position { index: 0, column: 1, line: 1 }))

221+ , actual: result

222+ }

223+

224+ do

225+ test :: Uint8Array <- fromArray (map fromInt [ 0x00C0, 0 ])

226+ testv <- fromUint8Array test

227+ result <- runParserT testv anyCodePointUTF8

228+ assertEqual' "anyCodePointUTF8 invalid 2"

229+ { expected: Left (ParseError "Invalid UTF-8 encoding" (Position { index: 0, column: 1, line: 1 }))

230+ , actual: result

231+ }