From be2d7921cbf33a3aa839cd54a0b3ec0a7dfc4a9b Mon Sep 17 00:00:00 2001
From: danse <f.occhipinti@gmail.com>
Date: Mon, 23 Apr 2018 08:54:06 +0200
Subject: RST reader: remove support for nested inlines.

RST does not allow nested emphasis, links, or other inline
constructs.

Closes #4581, double parsing of links with URLs as
link text.  This supersedes the earlier fix for #4581
in 6419819b46c0d69c7024ba8aa4a6381cb311341c.

Fixes #4561, a bug parsing with URLs inside emphasis.

Closes #4792.
---
 src/Text/Pandoc/Readers/RST.hs | 32 +++++++++++++++-----------------
 test/Tests/Readers/RST.hs      | 11 +++++++++++
 test/command/4581.md           |  6 ------
 3 files changed, 26 insertions(+), 23 deletions(-)
 delete mode 100644 test/command/4581.md

diff --git a/src/Text/Pandoc/Readers/RST.hs b/src/Text/Pandoc/Readers/RST.hs
index 2a36ca1f1..f9752a83c 100644
--- a/src/Text/Pandoc/Readers/RST.hs
+++ b/src/Text/Pandoc/Readers/RST.hs
@@ -45,7 +45,6 @@ import Data.Maybe (fromMaybe, isJust)
 import Data.Sequence (ViewR (..), viewr)
 import Data.Text (Text)
 import qualified Data.Text as T
-import Text.Pandoc.Walk (walk)
 import Text.Pandoc.Builder (Blocks, Inlines, fromList, setMeta, trimInlines)
 import qualified Text.Pandoc.Builder as B
 import Text.Pandoc.Class (PandocMonad, fetchItem, readFileFromDirs)
@@ -1314,19 +1313,24 @@ table = gridTable False <|> simpleTable False <|>
 
 inline :: PandocMonad m => RSTParser m Inlines
 inline = choice [ note          -- can start with whitespace, so try before ws
-                , whitespace
                 , link
-                , str
                 , endline
                 , strong
                 , emph
                 , code
                 , subst
                 , interpretedRole
-                , smart
-                , hyphens
-                , escapedChar
-                , symbol ] <?> "inline"
+                , inlineContent ] <?> "inline"
+
+-- strings, spaces and other characters that can appear either by
+-- themselves or within inline markup
+inlineContent :: PandocMonad m => RSTParser m Inlines
+inlineContent = choice [ whitespace
+                       , str
+                       , smart
+                       , hyphens
+                       , escapedChar
+                       , symbol ] <?> "inline content"
 
 parseInlineFromString :: PandocMonad m => String -> RSTParser m Inlines
 parseInlineFromString = parseFromString' (trimInlines . mconcat <$> many inline)
@@ -1369,11 +1373,11 @@ atStart p = do
 
 emph :: PandocMonad m => RSTParser m Inlines
 emph = B.emph . trimInlines . mconcat <$>
-         enclosed (atStart $ char '*') (char '*') inline
+         enclosed (atStart $ char '*') (char '*') inlineContent
 
 strong :: PandocMonad m => RSTParser m Inlines
 strong = B.strong . trimInlines . mconcat <$>
-          enclosed (atStart $ string "**") (try $ string "**") inline
+          enclosed (atStart $ string "**") (try $ string "**") inlineContent
 
 -- Note, this doesn't precisely implement the complex rule in
 -- http://docutils.sourceforge.net/docs/ref/rst/restructuredtext.html#inline-markup-recognition-rules
@@ -1480,8 +1484,8 @@ explicitLink :: PandocMonad m => RSTParser m Inlines
 explicitLink = try $ do
   char '`'
   notFollowedBy (char '`') -- `` marks start of inline code
-  label' <- removeLinks . trimInlines . mconcat <$>
-             manyTill (notFollowedBy (char '`') >> inline) (char '<')
+  label' <- trimInlines . mconcat <$>
+             manyTill (notFollowedBy (char '`') >> inlineContent) (char '<')
   src <- trim <$> manyTill (noneOf ">\n") (char '>')
   skipSpaces
   string "`_"
@@ -1495,12 +1499,6 @@ explicitLink = try $ do
                           _      -> return ((src, ""), nullAttr)
   return $ B.linkWith attr (escapeURI src') tit label''
 
-removeLinks :: B.Inlines -> B.Inlines
-removeLinks = B.fromList . walk (concatMap go) . B.toList
-  where go :: Inline -> [Inline]
-        go (Link _ lab _) = lab
-        go x              = [x]
-
 citationName :: PandocMonad m => RSTParser m String
 citationName = do
   raw <- citationMarker
diff --git a/test/Tests/Readers/RST.hs b/test/Tests/Readers/RST.hs
index 906ed4ff9..540c5d45a 100644
--- a/test/Tests/Readers/RST.hs
+++ b/test/Tests/Readers/RST.hs
@@ -188,4 +188,15 @@ tests = [ "line block with blank line" =:
             ] =?>
               para ("foo" <> note (para "bar"))
           ]
+        , testGroup "inlines"
+          [ "links can contain an URI without being parsed twice (#4581)" =:
+            "`http://loc <http://loc>`__" =?>
+            para (link "http://loc" "" "http://loc")
+          , "inline markup cannot be nested" =:
+            "**a*b*c**" =?>
+            para (strong "a*b*c")
+          , "bare URI parsing disabled inside emphasis (#4561)" =:
+            "*http://location*" =?>
+            para (emph (text "http://location"))
+          ]
         ]
diff --git a/test/command/4581.md b/test/command/4581.md
deleted file mode 100644
index b38ebaaf0..000000000
--- a/test/command/4581.md
+++ /dev/null
@@ -1,6 +0,0 @@
-```
-% pandoc -f rst -t native
-`http://loc <http://loc>`__
-^D
-[Para [Link ("",[],[]) [Str "http://loc"] ("http://loc","")]]
-```
-- 
cgit v1.2.3