98c2a52 · typeable/pandoc

git.iko.soy/typeable/pandoc

Clean up BibTeX parsing.

Previously there was a messy code path that gave strange results in some cases, not passing through raw tex but trying to extract a string content. This was an artefact of trying to handle some special bibtex-specific commands in the BibTeX reader. Now we just handle these in the LaTeX reader and simplify parsing in the BibTeX reader. This does mean that more raw tex will be passed through (and currently this is not sensitive to the `raw_tex` extension; this should be fixed). Closes #7049.

commit98c2a52b4ee6c833c0a2f2652386cec024e377eb authorJohn MacFarlane <jgm@berkeley.edu> dateTue 26 Jan 2021, 22:45 (6 years ago) parent12bc662

4 files changed+25−37

compare with master

unifiedsplit

src/Text/Pandoc/Citeproc/BibTeX.hs+1−32 src/Text/Pandoc/Readers/LaTeX.hs+18 test/command/biblatex-cotton.md+3−3 test/command/biblatex-murray.md+3−2

src/Text/Pandoc/Citeproc/BibTeX.hs+1−32

@@ -750,41 +750,10 @@ blocksToInlines bs =

750750 _ -> B.fromList $ Walk.query (:[]) bs

751751

752752adjustSpans :: Lang -> Inline -> Inline

753−adjustSpans lang (RawInline (Format "latex") s)

754− | s == "\\hyphen" || s == "\\hyphen " = Str "-"

755− | otherwise = parseRawLaTeX lang s

753+adjustSpans lang (Span ("",[],[("bibstring",s)]) _) = Str $ resolveKey' lang s

756754adjustSpans _ SoftBreak = Space

757755adjustSpans _ x = x

758756

759−parseRawLaTeX :: Lang -> Text -> Inline

760−parseRawLaTeX lang t@(T.stripPrefix "\\" -> Just xs) =

761− case parseLaTeX lang contents of

762− Right [Para ys] -> f command ys

763− Right [Plain ys] -> f command ys

764− Right [] -> f command []

765− _ -> RawInline (Format "latex") t

766− where (command', contents') = T.break (\c -> c =='{' || c =='\\') xs

767− command = T.strip command'

768− contents = T.drop 1 $ T.dropEnd 1 contents'

769− f "mkbibquote" ils = Span nullAttr [Quoted DoubleQuote ils]

770− f "mkbibemph" ils = Span nullAttr [Emph ils]

771− f "mkbibitalic" ils = Span nullAttr [Emph ils]

772− f "mkbibbold" ils = Span nullAttr [Strong ils]

773− f "mkbibparens" ils = Span nullAttr $

774− [Str "("] ++ ils ++ [Str ")"]

775− f "mkbibbrackets" ils = Span nullAttr $

776− [Str "["] ++ ils ++ [Str "]"]

777− -- ... both should be nestable & should work in year fields

778− f "autocap" ils = Span nullAttr ils

779− -- TODO: should work in year fields

780− f "textnormal" ils = Span ("",["nodecor"],[]) ils

781− f "bibstring" [Str s] = Str $ resolveKey' lang s

782− f "adddot" [] = Str "."

783− f "adddotspace" [] = Span nullAttr [Str ".", Space]

784− f "addabbrvspace" [] = Space

785− f _ ils = Span nullAttr ils

786−parseRawLaTeX _ t = RawInline (Format "latex") t

787−

788757latex' :: Text -> Bib [Block]

789758latex' t = do

790759 lang <- gets localeLang

src/Text/Pandoc/Readers/LaTeX.hs+18

@@ -940,6 +940,24 @@ inlineCommands = M.union inlineLanguageCommands $ M.fromList

940940 , ("uline", underline <$> tok)

941941 -- plain tex stuff that should just be passed through as raw tex

942942 , ("ifdim", ifdim)

943+ -- bibtex

944+ , ("mkbibquote", spanWith nullAttr . doubleQuoted <$> tok)

945+ , ("mkbibemph", spanWith nullAttr . emph <$> tok)

946+ , ("mkbibitalic", spanWith nullAttr . emph <$> tok)

947+ , ("mkbibbold", spanWith nullAttr . strong <$> tok)

948+ , ("mkbibparens",

949+ spanWith nullAttr . (\x -> str "(" <> x <> str ")") <$> tok)

950+ , ("mkbibbrackets",

951+ spanWith nullAttr . (\x -> str "[" <> x <> str "]") <$> tok)

952+ , ("autocap", spanWith nullAttr <$> tok)

953+ , ("textnormal", spanWith ("",["nodecor"],[]) <$> tok)

954+ , ("bibstring",

955+ (\x -> spanWith ("",[],[("bibstring",x)]) (str x)) . untokenize

956+ <$> braced)

957+ , ("adddot", pure (str "."))

958+ , ("adddotspace", pure (spanWith nullAttr (str "." <> space)))

959+ , ("addabbrvspace", pure space)

960+ , ("hyphen", pure (str "-"))

943961 ]

944962

945963accent :: PandocMonad m => Char -> Maybe Char -> LP m Inlines

test/command/biblatex-cotton.md+3−3

@@ -43,9 +43,9 @@ Cotton, F. A., Wilkinson, G., Murillio, C. A., & Bochmann, M. (1999).

4343---

4444nocite: "[@*]"

4545references:

46−- annote: A book entry with author authors and an edition field. By

47− default, long author and editor lists are automatically truncated.

48− This is configurable

46+- annote: A book entry with `\arabic{author}`{=latex} authors and an

47+ edition field. By default, long author and editor lists are

48+ automatically truncated. This is configurable

4949 author:

5050 - family: Cotton

5151 given: Frank Albert

test/command/biblatex-murray.md+3−2

@@ -55,8 +55,9 @@ properties as a function of core size. *Langmuir*, *14*(1), 17–30.

5555---

5656nocite: "[@*]"

5757references:

58−- annote: An article entry with author authors. By default, long author

59− and editor lists are automatically truncated. This is configurable

58+- annote: An article entry with `\arabic{author}`{=latex} authors. By

59+ default, long author and editor lists are automatically truncated.

60+ This is configurable

6061 author:

6162 - family: Hostetler

6263 given: Michael J.