ec88d47 · typeable/pandoc

git.iko.soy/typeable/pandoc

Correctly implement capitalisation.

Using `map toUpper` to capitalise text is wrong, as e.g. “Straße” should be converted to “STRASSE”, which is 1 character longer. This commit adds a `capitalize` function and replaces 2 identical implementations in different modules (`toCaps` and `capitalize`) with it.

commitec88d47f23d6761cf2120f76e45ca23cdc478e6c authorArtyom Kazak <yom@artyom.me> dateSun 3 Aug 2014, 16:48 (12 years ago) parent842c705

5 files changed+42−15

compare with master

unifiedsplit

src/Text/Pandoc/Shared.hs+13 src/Text/Pandoc/Writers/FB2.hs+3−8 src/Text/Pandoc/Writers/Markdown.hs+3−7 tests/Tests/Writers/Plain.hs+21 tests/test-pandoc.hs+2

src/Text/Pandoc/Shared.hs+13

@@ -59,6 +59,7 @@ module Text.Pandoc.Shared (

5959 normalizeBlocks,

6060 removeFormatting,

6161 stringify,

62+ capitalize,

6263 compactify,

6364 compactify',

6465 compactify'DL,

@@ -122,6 +123,7 @@ import qualified Data.ByteString.Char8 as B8

122123import Text.Pandoc.Compat.Monoid

123124import Data.ByteString.Base64 (decodeLenient)

124125import Data.Sequence (ViewR(..), ViewL(..), viewl, viewr)

126+import qualified Data.Text as T (toUpper, pack, unpack)

125127

126128#ifdef EMBED_DATA_FILES

127129import Text.Pandoc.Data (dataFiles)

@@ -527,6 +529,17 @@ stringify = query go . walk deNote

527529 deNote (Note _) = Str ""

528530 deNote x = x

529531

532+-- | Bring all regular text in a pandoc structure to uppercase.

533+--

534+-- This function correctly handles cases where a lowercase character doesn't

535+-- match to a single uppercase character – e.g. “Straße” would be converted

536+-- to “STRASSE”, not “STRAßE”.

537+capitalize :: Walkable Inline a => a -> a

538+capitalize = walk go

539+ where go :: Inline -> Inline

540+ go (Str s) = Str (T.unpack $ T.toUpper $ T.pack s)

541+ go x = x

542+

530543-- | Change final list item from @Para@ to @Plain@ if the list contains

531544-- no other @Para@ blocks.

532545compactify :: [[Block]] -- ^ List of list items (each a list of blocks)

src/Text/Pandoc/Writers/FB2.hs+3−8

@@ -28,7 +28,7 @@ module Text.Pandoc.Writers.FB2 (writeFB2) where

2828import Control.Monad.State (StateT, evalStateT, get, modify)

2929import Control.Monad.State (liftM, liftM2, liftIO)

3030import Data.ByteString.Base64 (encode)

31−import Data.Char (toUpper, toLower, isSpace, isAscii, isControl)

31+import Data.Char (toLower, isSpace, isAscii, isControl)

3232import Data.List (intersperse, intercalate, isPrefixOf)

3333import Data.Either (lefts, rights)

3434import Network.Browser (browse, request, setAllowRedirects, setOutHandler)

@@ -44,8 +44,7 @@ import qualified Text.XML.Light.Cursor as XC

4444

4545import Text.Pandoc.Definition

4646import Text.Pandoc.Options (WriterOptions(..), HTMLMathMethod(..), def)

47−import Text.Pandoc.Shared (orderedListMarkers, isHeaderBlock)

48−import Text.Pandoc.Walk

47+import Text.Pandoc.Shared (orderedListMarkers, isHeaderBlock, capitalize)

4948

5049-- | Data to be written at the end of the document:

5150-- (foot)notes, URLs, references, images.

@@ -421,10 +420,6 @@ indent = indentBlock

421420 indentLines ins = let lns = split isLineBreak ins :: [[Inline]]

422421 in intercalate [LineBreak] $ map ((Str spacer):) lns

423422

424−capitalize :: Inline -> Inline

425−capitalize (Str xs) = Str $ map toUpper xs

426−capitalize x = x

427−

428423-- | Convert a Pandoc's Inline element to FictionBook XML representation.

429424toXml :: Inline -> FBM [Content]

430425toXml (Str s) = return [txt s]

@@ -434,7 +429,7 @@ toXml (Strong ss) = list `liftM` wrap "strong" ss

434429toXml (Strikeout ss) = list `liftM` wrap "strikethrough" ss

435430toXml (Superscript ss) = list `liftM` wrap "sup" ss

436431toXml (Subscript ss) = list `liftM` wrap "sub" ss

437−toXml (SmallCaps ss) = cMapM toXml $ walk capitalize ss

432+toXml (SmallCaps ss) = cMapM toXml $ capitalize ss

438433toXml (Quoted SingleQuote ss) = do -- FIXME: should be language-specific

439434 inner <- cMapM toXml ss

440435 return $ [txt "‘"] ++ inner ++ [txt "’"]

src/Text/Pandoc/Writers/Markdown.hs+3−7

@@ -39,7 +39,7 @@ import Text.Pandoc.Writers.Shared

3939import Text.Pandoc.Options

4040import Text.Pandoc.Parsing hiding (blankline, blanklines, char, space)

4141import Data.List ( group, isPrefixOf, find, intersperse, transpose, sortBy )

42−import Data.Char ( isSpace, isPunctuation, toUpper )

42+import Data.Char ( isSpace, isPunctuation )

4343import Data.Ord ( comparing )

4444import Text.Pandoc.Pretty

4545import Control.Monad.State

@@ -672,10 +672,6 @@ escapeSpaces (Str s) = Str $ substitute " " "\\ " s

672672escapeSpaces Space = Str "\\ "

673673escapeSpaces x = x

674674

675−toCaps :: Inline -> Inline

676−toCaps (Str s) = Str (map toUpper s)

677−toCaps x = x

678−

679675-- | Convert Pandoc inline element to markdown.

680676inlineToMarkdown :: WriterOptions -> Inline -> State WriterState Doc

681677inlineToMarkdown opts (Span attrs ils) = do

@@ -693,7 +689,7 @@ inlineToMarkdown opts (Emph lst) = do

693689inlineToMarkdown opts (Strong lst) = do

694690 plain <- gets stPlain

695691 if plain

696− then inlineListToMarkdown opts $ walk toCaps lst

692+ then inlineListToMarkdown opts $ capitalize lst

697693 else do

698694 contents <- inlineListToMarkdown opts lst

699695 return $ "**" <> contents <> "**"

@@ -716,7 +712,7 @@ inlineToMarkdown opts (Subscript lst) = do

716712inlineToMarkdown opts (SmallCaps lst) = do

717713 plain <- gets stPlain

718714 if plain

719− then inlineListToMarkdown opts $ walk toCaps lst

715+ then inlineListToMarkdown opts $ capitalize lst

720716 else do

721717 contents <- inlineListToMarkdown opts lst

722718 return $ tagWithAttrs "span"

tests/Tests/Writers/Plain.hsnew file+21

@@ -0,0 +1,21 @@

1+{-# LANGUAGE OverloadedStrings #-}

2+module Tests.Writers.Plain (tests) where

4+import Test.Framework

5+import Text.Pandoc.Builder

6+import Text.Pandoc

7+import Tests.Helpers

8+import Tests.Arbitrary()

10+

11+infix 4 =:

12+(=:) :: (ToString a, ToPandoc a)

13+ => String -> (a, String) -> Test

14+(=:) = test (writePlain def . toPandoc)

15+

16+

17+tests :: [Test]

18+tests = [ "strongly emphasized text to uppercase"

19+ =: strong "Straße"

20+ =?> "STRASSE"

21+ ]

tests/test-pandoc.hs+2

@@ -17,6 +17,7 @@ import qualified Tests.Writers.HTML

1717import qualified Tests.Writers.Docbook

1818import qualified Tests.Writers.Native

1919import qualified Tests.Writers.Markdown

20+import qualified Tests.Writers.Plain

2021import qualified Tests.Writers.AsciiDoc

2122import qualified Tests.Shared

2223import qualified Tests.Walk

@@ -33,6 +34,7 @@ tests = [ testGroup "Old" Tests.Old.tests

3334 , testGroup "HTML" Tests.Writers.HTML.tests

3435 , testGroup "Docbook" Tests.Writers.Docbook.tests

3536 , testGroup "Markdown" Tests.Writers.Markdown.tests

37+ , testGroup "Plain" Tests.Writers.Plain.tests

3638 , testGroup "AsciiDoc" Tests.Writers.AsciiDoc.tests

3739 ]

3840 , testGroup "Readers"