e279175 · typeable/pandoc

git.iko.soy/typeable/pandoc

Options: Changed `writerSourceDir` to `writerSourceURL` (now a Maybe).

Previously we used to store the directory of the first input file, even if it was local, and used this as a base directory for finding images in ODT, EPUB, Docx, and PDF. This has been confusing to many users. It seems better to look for images relative to the current working directory, even if the first file argument is in another directory. writerSourceURL is set to 'Just url' when the first command-line argument is an absolute URL. (So, relative links will be resolved in relation to the first page.) Otherwise, 'Nothing'. The ODT, EPUB, Docx, and PDF writers have been modified accordingly. Note that this change may break some existing workflows. If you have been assuming that relative links will be interpreted relative to the directory of the first file argument, you'll need to make that the current directory before running pandoc. Closes #942.

commite279175ea517e2df65fe5d716bc02e383b04fc36 authorJohn MacFarlane <fiddlosopher@gmail.com> dateSun 11 Aug 2013, 15:58 (13 years ago) parent6f736df

7 files changed+34−41

compare with master

unifiedsplit

pandoc.hs+7−5 src/Text/Pandoc/Options.hs+2−2 src/Text/Pandoc/PDF.hs+3−3 src/Text/Pandoc/Shared.hs+10−10 src/Text/Pandoc/Writers/Docx.hs+1−2 src/Text/Pandoc/Writers/EPUB.hs+7−14 src/Text/Pandoc/Writers/ODT.hs+4−5

pandoc.hs+7−5

@@ -1034,13 +1034,15 @@ main = do

10341034 return $ Just csl { CSL.styleAbbrevs = abbrevs }

10351035 else return Nothing

10361036

1037− let sourceDir = case sources of

1038− [] -> "."

1037+ let sourceURL = case sources of

1038+ [] -> Nothing

10391039 (x:_) -> case parseURI x of

10401040 Just u

10411041 | uriScheme u `elem` ["http:","https:"] ->

1042− show u{ uriPath = "", uriQuery = "", uriFragment = "" }

1043− _ -> takeDirectory x

1042+ Just $ show u{ uriPath = "",

1043+ uriQuery = "",

1044+ uriFragment = "" }

1045+ _ -> Nothing

10441046

10451047 let readerOpts = def{ readerSmart = smart || (texLigatures &&

10461048 (laTeXOutput || "context" `isPrefixOf` writerName'))

@@ -1074,7 +1076,7 @@ main = do

10741076 writerColumns = columns,

10751077 writerEmailObfuscation = obfuscationMethod,

10761078 writerIdentifierPrefix = idPrefix,

1077− writerSourceDirectory = sourceDir,

1079+ writerSourceURL = sourceURL,

10781080 writerUserDataDir = datadir,

10791081 writerHtml5 = html5,

10801082 writerHtmlQTags = htmlQTags,

src/Text/Pandoc/Options.hs+2−2

@@ -286,7 +286,7 @@ data WriterOptions = WriterOptions

286286 , writerEmailObfuscation :: ObfuscationMethod -- ^ How to obfuscate emails

287287 , writerIdentifierPrefix :: String -- ^ Prefix for section & note ids in HTML

288288 -- and for footnote marks in markdown

289− , writerSourceDirectory :: FilePath -- ^ Directory path of 1st source file

289+ , writerSourceURL :: Maybe String -- ^ Absolute URL + directory of 1st source file

290290 , writerUserDataDir :: Maybe FilePath -- ^ Path of user data directory

291291 , writerCiteMethod :: CiteMethod -- ^ How to print cites

292292 , writerBiblioFiles :: [FilePath] -- ^ Biblio files to use for citations

@@ -329,7 +329,7 @@ instance Default WriterOptions where

329329 , writerColumns = 72

330330 , writerEmailObfuscation = JavascriptObfuscation

331331 , writerIdentifierPrefix = ""

332− , writerSourceDirectory = "."

332+ , writerSourceURL = Nothing

333333 , writerUserDataDir = Nothing

334334 , writerCiteMethod = Citeproc

335335 , writerBiblioFiles = []

src/Text/Pandoc/PDF.hs+3−3

@@ -65,17 +65,17 @@ makePDF :: String -- ^ pdf creator (pdflatex, lualatex, xelatex)

6565 -> Pandoc -- ^ document

6666 -> IO (Either ByteString ByteString)

6767makePDF program writer opts doc = withTempDir "tex2pdf." $ \tmpdir -> do

68− doc' <- handleImages (writerSourceDirectory opts) tmpdir doc

68+ doc' <- handleImages (writerSourceURL opts) tmpdir doc

6969 let source = writer opts doc'

7070 tex2pdf' tmpdir program source

7171

72−handleImages :: String -- ^ source directory/base URL

72+handleImages :: Maybe String -- ^ source base URL

7373 -> FilePath -- ^ temp dir to store images

7474 -> Pandoc -- ^ document

7575 -> IO Pandoc

7676handleImages baseURL tmpdir = walkM (handleImage' baseURL tmpdir)

7777

78−handleImage' :: String

78+handleImage' :: Maybe String

7979 -> FilePath

8080 -> Inline

8181 -> IO Inline

src/Text/Pandoc/Shared.hs+10−10

@@ -612,18 +612,18 @@ readDataFileUTF8 userDir fname =

612612

613613-- | Fetch an image or other item from the local filesystem or the net.

614614-- Returns raw content and maybe mime type.

615−fetchItem :: String -> String

615+fetchItem :: Maybe String -> String

616616 -> IO (Either E.SomeException (BS.ByteString, Maybe String))

617−fetchItem sourceDir s =

618− case s of

619− _ | isAbsoluteURI s -> openURL s

620− | isAbsoluteURI sourceDir -> openURL $ sourceDir ++ "/" ++ s

621− | otherwise -> E.try $ do

617+fetchItem sourceURL s

618+ | isAbsoluteURI s = openURL s

619+ | otherwise = case sourceURL of

620+ Just u -> openURL (u ++ "/" ++ s)

621+ Nothing -> E.try readLocalFile

622+ where readLocalFile = do

622623 let mime = case takeExtension s of

623− ".gz" -> getMimeType $ dropExtension s

624− x -> getMimeType x

625− let f = sourceDir </> s

626− cont <- BS.readFile f

624+ ".gz" -> getMimeType $ dropExtension s

625+ x -> getMimeType x

626+ cont <- BS.readFile s

627627 return (cont, mime)

628628

629629-- | Read from a URL and return raw data and maybe mime type.

src/Text/Pandoc/Writers/Docx.hs+1−2

@@ -728,8 +728,7 @@ inlineToOpenXML opts (Image alt (src, tit)) = do

728728 case M.lookup src imgs of

729729 Just (_,_,_,elt,_) -> return [elt]

730730 Nothing -> do

731− let sourceDir = writerSourceDirectory opts

732− res <- liftIO $ fetchItem sourceDir src

731+ res <- liftIO $ fetchItem (writerSourceURL opts) src

733732 case res of

734733 Left (_ :: E.SomeException) -> do

735734 liftIO $ warn $ "Could not find image `" ++ src ++ "', skipping..."

src/Text/Pandoc/Writers/EPUB.hs+7−14

@@ -55,7 +55,7 @@ import Text.Pandoc.UUID

5555import Text.Pandoc.Writers.HTML

5656import Text.Pandoc.Writers.Markdown ( writePlain )

5757import Data.Char ( toLower )

58−import Network.URI ( isAbsoluteURI, unEscapeString )

58+import Network.URI ( unEscapeString )

5959import Text.Pandoc.MIME (getMimeType)

6060#if MIN_VERSION_base(4,6,0)

6161#else

@@ -93,7 +93,6 @@ writeEPUB opts doc@(Pandoc meta _) = do

9393 then MathML Nothing

9494 else writerHTMLMathMethod opts

9595 , writerWrapText = False }

96− let sourceDir = writerSourceDirectory opts'

9796 let mbCoverImage = lookup "epub-cover-image" vars

9897

9998 -- cover page

@@ -117,10 +116,10 @@ writeEPUB opts doc@(Pandoc meta _) = do

117116 -- handle pictures

118117 picsRef <- newIORef []

119118 Pandoc _ blocks <- walkM

120− (transformInline opts' sourceDir picsRef) doc

119+ (transformInline opts' picsRef) doc

121120 pics <- readIORef picsRef

122121 let readPicEntry entries (oldsrc, newsrc) = do

123− res <- fetchItem sourceDir oldsrc

122+ res <- fetchItem (writerSourceURL opts') oldsrc

124123 case res of

125124 Left _ -> do

126125 warn $ "Could not find image `" ++ oldsrc ++ "', skipping..."

@@ -414,19 +413,13 @@ showDateTimeISO8601 :: UTCTime -> String

414413showDateTimeISO8601 = formatTime defaultTimeLocale "%FT%TZ"

415414

416415transformInline :: WriterOptions

417− -> FilePath

418416 -> IORef [(FilePath, FilePath)] -- ^ (oldpath, newpath) images

419417 -> Inline

420418 -> IO Inline

421−transformInline opts sourceDir picsRef (Image lab (src,tit))

422− | isAbsoluteURI src = do

423− raw <- makeSelfContained Nothing

424− $ writeHtmlInline opts (Image lab (src,tit))

425− return $ RawInline (Format "html") raw

426− | otherwise = do

419+transformInline opts picsRef (Image lab (src,tit)) = do

427420 let src' = unEscapeString src

428421 pics <- readIORef picsRef

429− let oldsrc = sourceDir </> src'

422+ let oldsrc = maybe src' (</> src) $ writerSourceURL opts

430423 let ext = takeExtension src'

431424 newsrc <- case lookup oldsrc pics of

432425 Just n -> return n

@@ -435,11 +428,11 @@ transformInline opts sourceDir picsRef (Image lab (src,tit))

435428 modifyIORef picsRef ( (oldsrc, new): )

436429 return new

437430 return $ Image lab (newsrc, tit)

438−transformInline opts _ _ (x@(Math _ _))

431+transformInline opts _ (x@(Math _ _))

439432 | WebTeX _ <- writerHTMLMathMethod opts = do

440433 raw <- makeSelfContained Nothing $ writeHtmlInline opts x

441434 return $ RawInline (Format "html") raw

442−transformInline _ _ _ x = return x

435+transformInline _ _ x = return x

443436

444437writeHtmlInline :: WriterOptions -> Inline -> String

445438writeHtmlInline opts z = trimr $

src/Text/Pandoc/Writers/ODT.hs+4−5

@@ -62,8 +62,7 @@ writeODT opts doc@(Pandoc meta _) = do

6262 readDataFile datadir "reference.odt"

6363 -- handle pictures

6464 picEntriesRef <- newIORef ([] :: [Entry])

65− let sourceDir = writerSourceDirectory opts

66− doc' <- walkM (transformPic sourceDir picEntriesRef) doc

65+ doc' <- walkM (transformPic opts picEntriesRef) doc

6766 let newContents = writeOpenDocument opts{writerWrapText = False} doc'

6867 epochtime <- floor `fmap` getPOSIXTime

6968 let contentEntry = toEntry "content.xml" epochtime $ fromStringLazy newContents

@@ -111,9 +110,9 @@ writeODT opts doc@(Pandoc meta _) = do

111110 let archive'' = addEntryToArchive metaEntry archive'

112111 return $ fromArchive archive''

113112

114−transformPic :: FilePath -> IORef [Entry] -> Inline -> IO Inline

115−transformPic sourceDir entriesRef (Image lab (src,_)) = do

116− res <- fetchItem sourceDir src

113+transformPic :: WriterOptions -> IORef [Entry] -> Inline -> IO Inline

114+transformPic opts entriesRef (Image lab (src,_)) = do

115+ res <- fetchItem (writerSourceURL opts) src

117116 case res of

118117 Left (_ :: E.SomeException) -> do

119118 warn $ "Could not find image `" ++ src ++ "', skipping..."