Docx parser: implement PAGEREF fields

These fields, often used in tables of contents, can be a hyperlink.

2 files changed, 26 insertions, 0 deletions

diff --git a/src/Text/Pandoc/Readers/Docx.hs b/src/Text/Pandoc/Readers/Docx.hs
index 462e3c679..5c8f20c18 100644
--- a/src/Text/Pandoc/Readers/Docx.hs
+++ b/src/Text/Pandoc/Readers/Docx.hs
@@ -448,6 +448,7 @@ parPartToInlines' (PlainOMath exps) =
 parPartToInlines' (Field info children) =
   case info of
     HyperlinkField url -> parPartToInlines' $ ExternalHyperLink url children
+    PagerefField fieldAnchor True -> parPartToInlines' $ InternalHyperLink fieldAnchor children
     _ -> smushInlines <$> mapM parPartToInlines' children
 parPartToInlines' NullParPart = return mempty
 
diff --git a/src/Text/Pandoc/Readers/Docx/Fields.hs b/src/Text/Pandoc/Readers/Docx/Fields.hs
index 442bc3466..5f090b6be 100644
--- a/src/Text/Pandoc/Readers/Docx/Fields.hs
+++ b/src/Text/Pandoc/Readers/Docx/Fields.hs
@@ -21,8 +21,11 @@ import Text.Parsec
 import Text.Parsec.Text (Parser)
 
 type URL = T.Text
+type Anchor = T.Text
 
 data FieldInfo = HyperlinkField URL
+                -- The boolean indicates whether the field is a hyperlink.
+               | PagerefField Anchor Bool
                | UnknownField
                deriving (Show)
 
@@ -33,6 +36,8 @@ fieldInfo :: Parser FieldInfo
 fieldInfo =
   try (HyperlinkField <$> hyperlink)
   <|>
+  try ((uncurry PagerefField) <$> pageref) 
+  <|>
   return UnknownField
 
 escapedQuote :: Parser T.Text
@@ -72,3 +77,23 @@ hyperlink = do
               ("\\l", s) : _ -> farg <> "#" <> s
               _              -> farg
   return url
+
+-- See §17.16.5.45
+pagerefSwitch :: Parser (T.Text, T.Text)
+pagerefSwitch = do
+  sw <- string "\\h"
+  spaces
+  farg <- fieldArgument
+  return (T.pack sw, farg)
+
+pageref :: Parser (Anchor, Bool)
+pageref = do
+  many space
+  string "PAGEREF"
+  spaces
+  farg <- fieldArgument
+  switches <- spaces *> many pagerefSwitch
+  let isLink = case switches of
+              ("\\h", _) : _ -> True 
+              _              -> False
+  return (farg, isLink)