From 7b11cdee49ec491a33dcade6ea68c3084a425992 Mon Sep 17 00:00:00 2001 From: John MacFarlane Date: Wed, 2 Dec 2020 10:46:23 -0800 Subject: Citeproc: ensure that BCP47 lang codes can be used. We ignore the variants and just use the base lang code and country code when passing off to citeproc. --- src/Text/Pandoc/Citeproc.hs | 19 +++++++++++++++++-- 1 file changed, 17 insertions(+), 2 deletions(-) (limited to 'src/Text/Pandoc/Citeproc.hs') diff --git a/src/Text/Pandoc/Citeproc.hs b/src/Text/Pandoc/Citeproc.hs index a48f97c3b..0de2882ae 100644 --- a/src/Text/Pandoc/Citeproc.hs +++ b/src/Text/Pandoc/Citeproc.hs @@ -16,6 +16,7 @@ import Text.Pandoc.Citeproc.BibTeX (readBibtexString, Variant(..)) import Text.Pandoc.Citeproc.MetaValue (metaValueToReference, metaValueToText) import Text.Pandoc.Readers.Markdown (yamlToRefs) import Text.Pandoc.Class (setResourcePath, getResourcePath, getUserDataDir) +import qualified Text.Pandoc.BCP47 as BCP47 import Data.ByteString (ByteString) import qualified Data.ByteString.Lazy as L import Text.Pandoc.Definition as Pandoc @@ -91,8 +92,9 @@ processCitations (Pandoc meta bs) = do case styleRes of Left err -> throwError $ PandocAppError $ prettyCiteprocError err Right style -> return style{ styleAbbreviations = mbAbbrevs } - let mblang = parseLang <$> - ((lookupMeta "lang" meta <|> lookupMeta "locale" meta) >>= metaValueToText) + mblang <- maybe (return Nothing) bcp47LangToIETF + ((lookupMeta "lang" meta <|> lookupMeta "locale" meta) >>= + metaValueToText) let locale = Citeproc.mergeLocales mblang style let getCiteId (Cite cs _) = Set.fromList $ map B.citationId cs getCiteId _ = mempty @@ -579,3 +581,16 @@ removeFinalPeriod ils = isRightQuote "\8217" = True isRightQuote "\187" = True isRightQuote _ = False + +bcp47LangToIETF :: PandocMonad m => Text -> m (Maybe Lang) +bcp47LangToIETF bcplang = + case BCP47.parseBCP47 bcplang of + Left _ -> do + report $ InvalidLang bcplang + return Nothing + Right lang -> + return $ Just + $ Lang (BCP47.langLanguage lang) + (if T.null (BCP47.langRegion lang) + then Nothing + else Just (BCP47.langRegion lang)) -- cgit v1.2.3