d6c58eb · typeable/pandoc

git.iko.soy/typeable/pandoc

on master help sign in

Docx reader: don't strip out empty paragraphs.

d6c58ebJohn MacFarlane9 years ago

src/Text/Pandoc/Readers/Docx.hs+1−3

compare with master

unifiedsplit

@@ -534,9 +534,7 @@ bodyPartToBlocks (Paragraph pPr parparts)

534 then do modify $ \s -> s { docxDropCap = ils' }

535 return mempty

536 else do modify $ \s -> s { docxDropCap = mempty }

537− return $ case isNull ils' of

538− True -> mempty

539− _ -> parStyleToTransform pPr $ para ils'

537+ return $ parStyleToTransform pPr $ para ils'

540bodyPartToBlocks (ListItem pPr numId lvl (Just levelInfo) parparts) = do

538bodyPartToBlocks (ListItem pPr numId lvl (Just levelInfo) parparts) = do

541 let

539 let

542 kvs = case levelInfo of

540 kvs = case levelInfo of

test/Tests/Readers/Docx.hs+20−7

@@ -10,6 +10,7 @@ import Test.Tasty

10import Test.Tasty.HUnit

11import Tests.Helpers

12import Text.Pandoc

13+import Text.Pandoc.Shared (stripEmptyParagraphs)

13import qualified Text.Pandoc.Class as P

14import qualified Text.Pandoc.Class as P

14import Text.Pandoc.MediaBag (MediaBag, lookupMedia, mediaDirectory)

15import Text.Pandoc.MediaBag (MediaBag, lookupMedia, mediaDirectory)

15import Text.Pandoc.UTF8 as UTF8

16import Text.Pandoc.UTF8 as UTF8

@@ -37,20 +38,23 @@ instance ToString NoNormPandoc where

37instance ToPandoc NoNormPandoc where

38instance ToPandoc NoNormPandoc where

38 toPandoc = unNoNorm

39 toPandoc = unNoNorm

41+compareOutput :: Bool

40−compareOutput :: ReaderOptions

42+ -> ReaderOptions

41− -> FilePath

43+ -> FilePath

42− -> FilePath

44+ -> FilePath

43− -> IO (NoNormPandoc, NoNormPandoc)

45+ -> IO (NoNormPandoc, NoNormPandoc)

44−compareOutput opts docxFile nativeFile = do

46+compareOutput strip opts docxFile nativeFile = do

45 df <- B.readFile docxFile

47 df <- B.readFile docxFile

46 nf <- UTF8.toText <$> BS.readFile nativeFile

48 nf <- UTF8.toText <$> BS.readFile nativeFile

47 p <- runIOorExplode $ readDocx opts df

49 p <- runIOorExplode $ readDocx opts df

48 df' <- runIOorExplode $ readNative def nf

50 df' <- runIOorExplode $ readNative def nf

49− return $ (noNorm p, noNorm df')

51+ return $ (noNorm (if strip

52+ then stripEmptyParagraphs p

53+ else p), noNorm df')

51testCompareWithOptsIO :: ReaderOptions -> String -> FilePath -> FilePath -> IO TestTree

55testCompareWithOptsIO :: ReaderOptions -> String -> FilePath -> FilePath -> IO TestTree

52testCompareWithOptsIO opts name docxFile nativeFile = do

56testCompareWithOptsIO opts name docxFile nativeFile = do

53− (dp, np) <- compareOutput opts docxFile nativeFile

57+ (dp, np) <- compareOutput True opts docxFile nativeFile

54 return $ test id name (dp, np)

58 return $ test id name (dp, np)

56testCompareWithOpts :: ReaderOptions -> String -> FilePath -> FilePath -> TestTree

60testCompareWithOpts :: ReaderOptions -> String -> FilePath -> FilePath -> TestTree

@@ -71,6 +75,11 @@ testForWarningsWithOpts :: ReaderOptions -> String -> FilePath -> [String] -> Te

71testForWarningsWithOpts opts name docxFile expected =

75testForWarningsWithOpts opts name docxFile expected =

72 unsafePerformIO $ testForWarningsWithOptsIO opts name docxFile expected

76 unsafePerformIO $ testForWarningsWithOptsIO opts name docxFile expected

78+testCompareNoStrip :: String -> FilePath -> FilePath -> TestTree

79+testCompareNoStrip name docxFile nativeFile = unsafePerformIO $ do

80+ (dp, np) <- compareOutput False defopts docxFile nativeFile

81+ return $ test id name (dp, np)

82+

74-- testForWarnings :: String -> FilePath -> [String] -> TestTree

83-- testForWarnings :: String -> FilePath -> [String] -> TestTree

75-- testForWarnings = testForWarningsWithOpts defopts

84-- testForWarnings = testForWarningsWithOpts defopts

@@ -257,6 +266,10 @@ tests = [ testGroup "inlines"

257 "dropcap paragraphs"

266 "dropcap paragraphs"

258 "docx/drop_cap.docx"

267 "docx/drop_cap.docx"

259 "docx/drop_cap.native"

268 "docx/drop_cap.native"

269+ , testCompareNoStrip

270+ "empty paragraphs without stripping"

271+ "docx/drop_cap.docx"

272+ "docx/drop_cap_nostrip.native"

260 ]

273 ]

261 , testGroup "track changes"

274 , testGroup "track changes"

262 [ testCompare

275 [ testCompare

test/docx/0_level_headers.native+8−8

@@ -1,15 +1,15 @@

1[Table [] [AlignDefault] [0.0]

2 [[]]

3− [[[]]

3+ [[[Plain []]]

4 ,[[Plain [Str "User\8217s",Space,Str "Guide"]]]

5− ,[[]]

5+ ,[[Plain []]]

6− ,[[]]

6+ ,[[Plain []]]

7− ,[[]]

7+ ,[[Plain []]]

8 ,[[Plain [Str "11",Space,Str "August",Space,Str "2017"]]]

9− ,[[]]

9+ ,[[Plain []]]

10− ,[[]]

10+ ,[[Plain []]]

11− ,[[]]

11+ ,[[Plain []]]

12− ,[[]]]

12+ ,[[Plain []]]]

13,Para [Str "CONTENTS"]

14,Para [Strong [Str "Section",Space,Str "Page"]]

15,Para [Str "FIGURES",Space,Str "iv"]

test/docx/comments.native+1−1

@@ -1,4 +1,4 @@

1[Para [Str "I",Space,Str "want",Space,Span ("",["comment-start"],[("id","0"),("author","Jesse Rosenthal"),("date","2016-05-09T16:13:00Z")]) [Str "I",Space,Str "left",Space,Str "a",Space,Str "comment."],Str "some",Space,Str "text",Space,Str "to",Space,Str "have",Space,Str "a",Space,Str "comment",Space,Span ("",["comment-end"],[("id","0")]) [],Str "on",Space,Str "it."]

2,Para [Str "This",Space,Str "is",Space,Span ("",["comment-start"],[("id","1"),("author","Jesse Rosenthal"),("date","2016-05-09T16:13:00Z")]) [Str "A",Space,Str "comment",Space,Str "across",Space,Str "paragraphs."],Str "a",Space,Str "new",Space,Str "paragraph."]

3,Para [Str "And",Space,Str "so",Span ("",["comment-end"],[("id","1")]) [],Space,Str "is",Space,Str "this."]

4−,Para [Str "One",Space,Span ("",["comment-start"],[("id","2"),("author","Jesse Rosenthal"),("date","2016-05-09T16:14:00Z")]) [Str "This",Space,Str "one",Space,Str "has",Space,Str "multiple",Space,Str "paragraphs.",Space,Str "\182",Space,Str "See?"],Str "more",Span ("",["comment-end"],[("id","2")]) [],Str ".",Space,Str "And",Space,Str "this",Space,Str "is",Space,Str "one",Space,Str "with",Space,Str "a",Space,Span ("",["comment-start"],[("id","3"),("author","Jesse Rosenthal"),("date","2016-06-22T14:35:00Z")]) [Str "Do",Space,Str "something."],Span ("",["comment-start"],[("id","4"),("author","Jesse Rosenthal"),("date","2016-06-22T14:36:00Z")]) [Str "Do",Space,Str "something",Space,Str "else."],Str "comment",Space,Str "in",Space,Str "a",Space,Str "comment",Span ("",["comment-end"],[("id","3")]) [Span ("",["comment-end"],[("id","4")]) []],Str "."]]

4+,Para [Str "One",Space,Span ("",["comment-start"],[("id","2"),("author","Jesse Rosenthal"),("date","2016-05-09T16:14:00Z")]) [Str "This",Space,Str "one",Space,Str "has",Space,Str "multiple",Space,Str "paragraphs.",Space,Str "\182",Space,Str "\182",Space,Str "See?"],Str "more",Span ("",["comment-end"],[("id","2")]) [],Str ".",Space,Str "And",Space,Str "this",Space,Str "is",Space,Str "one",Space,Str "with",Space,Str "a",Space,Span ("",["comment-start"],[("id","3"),("author","Jesse Rosenthal"),("date","2016-06-22T14:35:00Z")]) [Str "Do",Space,Str "something."],Span ("",["comment-start"],[("id","4"),("author","Jesse Rosenthal"),("date","2016-06-22T14:36:00Z")]) [Str "Do",Space,Str "something",Space,Str "else."],Str "comment",Space,Str "in",Space,Str "a",Space,Str "comment",Span ("",["comment-end"],[("id","3")]) [Span ("",["comment-end"],[("id","4")]) []],Str "."]]

test/docx/drop_cap_nostrip.nativenew file+9

@@ -0,0 +1,9 @@

1+[Para [Str "Drop",Space,Str "cap."]

2+,Para []

3+,Para [Str "Next",Space,Str "paragraph."]

4+,Para []

5+,Para [Str "Drop",Space,Str "cap",Space,Str "in",Space,Str "margin."]

6+,Para []

7+,Para []

8+,Para []

9+,Para [Str "Drop",Space,Str "cap",Space,Str "(not",Space,Str "really)."]]