From 2befeaa29f9589f81a3d6cd394d88f29089e7338 Mon Sep 17 00:00:00 2001
From: John MacFarlane <jgm@berkeley.edu>
Date: Sat, 2 Oct 2021 22:13:03 -0700
Subject: Remove splitSentences from T.P.Shared [API change].

We used to attempt automatic sentence splitting in man and ms
output, since sentence-ending periods need to be followed by
two spaces or a newline in these formats.

But it's difficult to do this reliably at the level of
`[Inline]`.
---
 src/Text/Pandoc/Shared.hs      | 28 ----------------------------
 src/Text/Pandoc/Writers/Man.hs |  5 ++---
 src/Text/Pandoc/Writers/Ms.hs  |  5 ++---
 3 files changed, 4 insertions(+), 34 deletions(-)

(limited to 'src/Text/Pandoc')

diff --git a/src/Text/Pandoc/Shared.hs b/src/Text/Pandoc/Shared.hs
index 7bb830d0e..6e1f29fb1 100644
--- a/src/Text/Pandoc/Shared.hs
+++ b/src/Text/Pandoc/Shared.hs
@@ -68,7 +68,6 @@ module Text.Pandoc.Shared (
                      makeMeta,
                      eastAsianLineBreakFilter,
                      htmlSpanLikeElements,
-                     splitSentences,
                      filterIpynbOutput,
                      -- * TagSoup HTML handling
                      renderTags',
@@ -709,33 +708,6 @@ eastAsianLineBreakFilter = bottomUp go
 htmlSpanLikeElements :: Set.Set T.Text
 htmlSpanLikeElements = Set.fromList ["kbd", "mark", "dfn"]
 
--- | Returns the first sentence in a list of inlines, and the rest.
-breakSentence :: [Inline] -> ([Inline], [Inline])
-breakSentence [] = ([],[])
-breakSentence xs =
-  let isSentenceEndInline (Str ys)
-        | Just (_, c) <- T.unsnoc ys = c == '.' || c == '?'
-      isSentenceEndInline LineBreak  = True
-      isSentenceEndInline _          = False
-      (as, bs) = break isSentenceEndInline xs
-  in  case bs of
-        []             -> (as, [])
-        [c]            -> (as ++ [c], [])
-        (c:Space:cs)   -> (as ++ [c], cs)
-        (c:SoftBreak:cs) -> (as ++ [c], cs)
-        (Str ".":Str s@(T.uncons -> Just (')',_)):cs)
-          -> (as ++ [Str ".", Str s], cs)
-        (x@(Str (T.stripPrefix ".)" -> Just _)):cs) -> (as ++ [x], cs)
-        (LineBreak:x@(Str (T.uncons -> Just ('.',_))):cs) -> (as ++[LineBreak], x:cs)
-        (c:cs)         -> (as ++ [c] ++ ds, es)
-          where (ds, es) = breakSentence cs
-
--- | Split a list of inlines into sentences.
-splitSentences :: [Inline] -> [[Inline]]
-splitSentences xs =
-  let (sent, rest) = breakSentence xs
-  in  if null rest then [sent] else sent : splitSentences rest
-
 -- | Process ipynb output cells.  If mode is Nothing,
 -- remove all output.  If mode is Just format, select
 -- best output for the format.  If format is not ipynb,
diff --git a/src/Text/Pandoc/Writers/Man.hs b/src/Text/Pandoc/Writers/Man.hs
index 45516ea06..89789d3c5 100644
--- a/src/Text/Pandoc/Writers/Man.hs
+++ b/src/Text/Pandoc/Writers/Man.hs
@@ -108,10 +108,9 @@ blockToMan :: PandocMonad m
            -> StateT WriterState m (Doc Text)
 blockToMan opts (Div _ bs) = blockListToMan opts bs
 blockToMan opts (Plain inlines) =
-  liftM vcat $ mapM (inlineListToMan opts) $ splitSentences inlines
+  inlineListToMan opts inlines
 blockToMan opts (Para inlines) = do
-  contents <- liftM vcat $ mapM (inlineListToMan opts) $
-    splitSentences inlines
+  contents <- inlineListToMan opts inlines
   return $ text ".PP" $$ contents
 blockToMan opts (LineBlock lns) =
   blockToMan opts $ linesToPara lns
diff --git a/src/Text/Pandoc/Writers/Ms.hs b/src/Text/Pandoc/Writers/Ms.hs
index 055324448..719407ac1 100644
--- a/src/Text/Pandoc/Writers/Ms.hs
+++ b/src/Text/Pandoc/Writers/Ms.hs
@@ -142,7 +142,7 @@ blockToMs opts (Div (ident,cls,kvs) bs) = do
        setFirstPara
        return $ anchor $$ res
 blockToMs opts (Plain inlines) =
-  liftM vcat $ mapM (inlineListToMs' opts) $ splitSentences inlines
+  inlineListToMs' opts inlines
 blockToMs opts (Para [Image attr alt (src,_tit)])
   | let ext = takeExtension (T.unpack src) in (ext == ".ps" || ext == ".eps") = do
   let (mbW,mbH) = (inPoints opts <$> dimension Width attr,
@@ -165,8 +165,7 @@ blockToMs opts (Para [Image attr alt (src,_tit)])
 blockToMs opts (Para inlines) = do
   firstPara <- gets stFirstPara
   resetFirstPara
-  contents <- liftM vcat $ mapM (inlineListToMs' opts) $
-    splitSentences inlines
+  contents <- inlineListToMs' opts inlines
   return $ literal (if firstPara then ".LP" else ".PP") $$ contents
 blockToMs _ b@(RawBlock f str)
   | f == Format "ms" = return $ literal str
-- 
cgit v1.2.3