diff --git a/CHANGELOG.md b/CHANGELOG.md
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,3 +1,8 @@
+## MMark 0.0.4.2
+
+* Made parsing of emphasis-like markup more flexible and forgiving, see
+  `README.md` for more information.
+
 ## MMark 0.0.4.1
 
 * This version uses `megaparsec-6.4.0` and `parser-combinators-0.4.0` and
diff --git a/LICENSE.md b/LICENSE.md
--- a/LICENSE.md
+++ b/LICENSE.md
@@ -1,4 +1,4 @@
-Copyright © 2017 Mark Karpov
+Copyright © 2017–2018 Mark Karpov
 
 All rights reserved.
 
diff --git a/README.md b/README.md
--- a/README.md
+++ b/README.md
@@ -126,38 +126,58 @@
 
 I decided to make parsing of emphasis, strong emphasis, and similar
 constructs like strikethrough, subscript, and superscript more symmetric and
-less ad-hoc. This is a work in progress and I'm not fully satisfied with the
-current approach as it does not allow to express some combinations of
-characters and markup, but in 99% of practical cases it is identical to
-Common Mark, and normal markdown intuitions will work OK for the users.
+less ad-hoc. In 99% of practical cases it is identical to Common Mark, and
+normal markdown intuitions will work OK for the users.
 
-Let's start by dividing all characters into three groups:
+Let's start by dividing all characters into four groups:
 
+* **Space characters**, including space, tab, newline, carriage return, and
+  other characters like non-breaking space.
+
 * **Markup characters**, including the following: `*`, `~`, `_`, `` ` ``,
   `^`, `[`, `]`. These are used for markup and whenever they appear in a
   document, they must form valid markup constructions. To be used as
   ordinary punctuation characters they must be backslash escaped.
 
-* **Space characters**, including space, tab, newline and carriage return.
+* **Punctuation characters**, which include all punctuation characters that
+  are not **markup characters**.
 
 * **Other characters**, which include all characters not falling into the
-  two groups described above.
+  three groups described above.
 
-**Markup characters** can be “converted” to **other characters** via
-backslash escaping. We'll see how this is useful in a few moments.
+Next, let's assign *levels* to all groups but **markup characters**:
 
-We'll call **markdown characters** placed between **space characters** and
-**other characters** *left-flanking delimiter run*. These markup characters
-sort of hang on the left hand side of a word.
+* **Space characters**—level 0
+* **Punctuation characters**—level 1
+* **Other characters**—level 2
 
-Similarly we'll call **markdown characters** placed between **other
-characters** and **space characters** *right-flanking delimiter run*. These
-hang on the right hand side of a word.
+When **markup characters** or **punctuation characters** are escaped with
+backslash they become **other characters**.
 
-Emphasis markup (and other similar things like strikethrough, which we won't
-mention explicitly anymore for brevity) can start only as left-flanking
-delimiter run and end only as right-flanking delimiter run.
+We'll call **markdown characters** placed between a character of level `L`
+and a character of level `R` *left-flanking delimiter run* if and only if:
 
+```
+level(L) < level(R)
+```
+
+These **markup characters** sort of hang on the left hand side of a word.
+
+Similarly we'll call **markdown characters** placed between a character of
+level `L` and a character of level `R` *right-flanking delimiter run* if and
+only if:
+
+```
+level(L) > level (R)
+```
+
+These **markup characters** hang on the right hand side of a word.
+
+*Emphasis markup* (and other similar things like strikethrough, which we
+won't mention explicitly anymore for brevity) can start only as
+*left-flanking delimiter run* and end only as *right-flanking delimiter
+run*.
+
 This produces a parse error:
 
 ```
@@ -171,40 +191,24 @@
 __foo__bar
 ```
 
-This means that inter-word emphasis is not supported by this approach. (This
-is a pity, maybe I should adjust something to allow it.)
+This means that inter-word emphasis is not supported by this approach.
 
-There is one more tricky thing. In some cases we want to end emphasis and
-have full stop or other punctuation right after it:
+The next example is OK because `s` is an **other character** and `.` is a
+**punctuation character**, so `level('s') > level('.')`.
 
 ```
 Here it *goes*.
 ```
 
-You can see that the closing `*` is not in right-flanking position here, and
-so it's a parse error. To avoid this, some punctuation characters that
-normally appear outside of markup were made “transparent” and thus they are
-regarded as white space, so the example above parses correctly and works as
-expected. To put a transparent character inside emphasis, backslash escaping
-is necessary:
-
-```
-We *\(can\)* have it.
-```
-
-Here `(` and `)` are transparent punctuation characters, just like `.`, so
-they must be turned into **other characters** to go inside the emphasis.
-This is a corner case and should not be common in practice.
-
-So far the main limitation of this approach is the pains with inter-word
-markup, as in this example:
+In some rare cases backslash escaping can help get the right result:
 
 ```
-**We started to work on the *issue*.**
+Here goes *(something\)*.
 ```
 
-Should we escape `.` here? On one hand we should, to close `**`. But if we
-do, the closing `*` won't be in right-flanking position anymore. God dammit.
+We escaped the closing parenthesis `)` so it becomes an **other character**
+with level 2 and so its level is greater than the level of plain punctuation
+character `.`.
 
 ### Other differences
 
@@ -265,7 +269,7 @@
 Library             | Parsing library     | Execution time | Allocated   | Max residency
 --------------------|---------------------|---------------:|------------:|-------------:
 `cmark-0.5.6`       | Custom C code       |       311.8 μs |     228,440 |         9,608
-`mmark-0.0.4.1`     | Megaparsec          |       7.169 ms |  32,203,304 |        37,856
+`mmark-0.0.4.2`     | Megaparsec          |       7.757 ms |  32,207,640 |        37,856
 `cheapskate-0.1.1`  | Custom Haskell code |       10.22 ms |  44,686,272 |       799,200
 `markdown-0.1.16` † | Attoparsec          |       13.51 ms |  69,261,816 |       699,656
 `pandoc-2.0.5`      | Parsec              |       36.01 ms | 141,868,840 |     1,471,080
@@ -288,6 +292,6 @@
 
 ## License
 
-Copyright © 2017 Mark Karpov
+Copyright © 2017–2018 Mark Karpov
 
 Distributed under BSD 3 clause license.
diff --git a/Text/MMark.hs b/Text/MMark.hs
--- a/Text/MMark.hs
+++ b/Text/MMark.hs
@@ -1,6 +1,6 @@
 -- |
 -- Module      :  Text.MMark
--- Copyright   :  © 2017 Mark Karpov
+-- Copyright   :  © 2017–2018 Mark Karpov
 -- License     :  BSD 3 clause
 --
 -- Maintainer  :  Mark Karpov <markkarpov92@gmail.com>
diff --git a/Text/MMark/Extension.hs b/Text/MMark/Extension.hs
--- a/Text/MMark/Extension.hs
+++ b/Text/MMark/Extension.hs
@@ -1,6 +1,6 @@
 -- |
 -- Module      :  Text.MMark.Extension
--- Copyright   :  © 2017 Mark Karpov
+-- Copyright   :  © 2017–2018 Mark Karpov
 -- License     :  BSD 3 clause
 --
 -- Maintainer  :  Mark Karpov <markkarpov92@gmail.com>
diff --git a/Text/MMark/Parser.hs b/Text/MMark/Parser.hs
--- a/Text/MMark/Parser.hs
+++ b/Text/MMark/Parser.hs
@@ -1,6 +1,6 @@
 -- |
 -- Module      :  Text.MMark.Parser
--- Copyright   :  © 2017 Mark Karpov
+-- Copyright   :  © 2017–2018 Mark Karpov
 -- License     :  BSD 3 clause
 --
 -- Maintainer  :  Mark Karpov <markkarpov92@gmail.com>
@@ -591,7 +591,7 @@
                takeWhile1P Nothing (== '`') <|>
                takeWhile1P Nothing (/= '`'))
       finalizer
-  r <$ lastOther
+  r <$ lastChar OtherChar
 
 -- | Parse a link.
 
@@ -602,7 +602,7 @@
   txt <- disallowLinks (disallowEmpty pInlines)
   void (char ']')
   (dest, mtitle) <- pLocation pos txt
-  Link txt dest mtitle <$ lastOther
+  Link txt dest mtitle <$ lastChar OtherChar
 
 -- | Parse an image.
 
@@ -610,7 +610,7 @@
 pImage = do
   (pos, alt)    <- emptyAlt <|> nonEmptyAlt
   (src, mtitle) <- pLocation pos alt
-  Image alt src mtitle <$ lastOther
+  Image alt src mtitle <$ lastChar OtherChar
   where
     emptyAlt = do
       pos <- getPosition
@@ -638,7 +638,7 @@
           Just email ->
             ( nes (Plain email)
             , URI.makeAbsolute mailtoScheme uri' )
-  Link txt uri Nothing <$ lastOther
+  Link txt uri Nothing <$ lastChar OtherChar
 
 -- | Parse inline content inside an enclosing construction such as emphasis,
 -- strikeout, superscript, and\/or subscript markup.
@@ -668,7 +668,7 @@
   eol
   notFollowedBy eof
   sc'
-  lastSpace
+  lastChar SpaceChar
   return LineBreak
 
 -- | Parse plain text.
@@ -677,37 +677,36 @@
 pPlain = fmap (Plain . bakeText) . foldSome $ do
   ch <- lookAhead (anyChar <?> "inline content")
   let newline' =
-        (('\n':) . dropWhile isSpace) <$ eol <* sc' <* lastSpace
+        (('\n':) . dropWhile isSpace) <$ eol <* sc' <* lastChar SpaceChar
   case ch of
     '\\' -> (:) <$>
-      ((escapedChar <* lastOther) <|>
-        try (char '\\' <* notFollowedBy eol <* lastOther))
+      ((escapedChar <* lastChar OtherChar) <|>
+        try (char '\\' <* notFollowedBy eol <* lastChar OtherChar))
     '\n' ->
       newline'
     '\r' ->
       newline'
     '!' -> do
       notFollowedBy (string "![")
-      (:) <$> char '!'
+      (:) <$> char '!' <* lastChar PunctChar
     '<' -> do
       notFollowedBy pAutolink
-      (:) <$> char '<'
+      (:) <$> char '<' <* lastChar PunctChar
     '&' -> choice
       [ (:) <$> numRef
       , (++) . reverse <$> entityRef
-      , (:) <$> char '&' ]
+      , (:) <$> char '&' ] <* lastChar PunctChar
     _ ->
-      (:) <$> pOther ch
-  where
-    pOther ch
-      | isSpace ch = char ch <* lastSpace
-      | isTrans ch = char ch <* lastSpace
-      | isOther ch = char ch <* lastOther
-      | otherwise  = failure
-          (Just . Tokens . nes $ ch)
-          (E.singleton . Label . NE.fromList $ "inline content")
-    isTrans x = isTransparentPunctuation x && x /= '!'
-    isOther x = not (isMarkupChar x) && x /= '\\' && x /= '!' && x /= '<'
+      (:) <$>
+        if Char.isSpace ch
+          then char ch <* lastChar SpaceChar
+          else if isSpecialChar ch
+                 then failure
+                   (Just . Tokens . nes $ ch)
+                   (E.singleton . Label . NE.fromList $ "inline content")
+                 else if Char.isPunctuation ch
+                        then char ch <* lastChar PunctChar
+                        else char ch <* lastChar OtherChar
 
 ----------------------------------------------------------------------------
 -- Auxiliary inline-level parsers
@@ -823,9 +822,9 @@
       failNow = do
         setPosition pos
         (customFailure . NonFlankingDelimiterRun . toNesTokens) dels
-  isLastOther >>= flip when failNow
-  rch <- lookAhead (optional anyChar)
-  when (maybe True isTransparent rch) failNow
+  lch <- getLastChar
+  rch <- getNextChar OtherChar
+  when (lch >= rch) failNow
   return st
 
 -- | Parse a closing markup sequence corresponding to given 'InlineFrame'.
@@ -842,13 +841,25 @@
   let failNow = do
         setPosition pos
         (customFailure . NonFlankingDelimiterRun . toNesTokens) dels
-      goodAfter x =
-        isTransparent x || isMarkupChar x
-  isLastSpace >>= flip when failNow
-  rch <- lookAhead (optional anyChar)
-  unless (maybe True goodAfter rch) failNow
+  lch <- getLastChar
+  rch <- getNextChar SpaceChar
+  when (lch <= rch) failNow
   return frame
 
+-- | Get 'CharType' of the next char in the input stream.
+
+getNextChar
+  :: CharType          -- ^ What we should consider frame constituent characters
+  -> IParser CharType
+getNextChar frameType = lookAhead (option SpaceChar (charType <$> anyChar))
+  where
+    charType ch
+      | isFrameConstituent ch = frameType
+      | Char.isSpace       ch = SpaceChar
+      | ch == '\\'            = OtherChar
+      | Char.isPunctuation ch = PunctChar
+      | otherwise             = OtherChar
+
 ----------------------------------------------------------------------------
 -- Parsing helpers
 
@@ -959,25 +970,19 @@
 eol' = option False (True <$ eol)
 
 ----------------------------------------------------------------------------
--- Other helpers
-
-slevel :: Pos -> Pos -> Pos
-slevel a l = if l >= ilevel a then a else l
-
-ilevel :: Pos -> Pos
-ilevel = (<> mkPos 4)
+-- Char classification
 
 isSpace :: Char -> Bool
 isSpace x = x == ' ' || x == '\t'
 
 isSpaceN :: Char -> Bool
-isSpaceN x = isSpace x || x == '\n' || x == '\r'
+isSpaceN x = isSpace x || isNewline x
 
-notNewline :: Char -> Bool
-notNewline x = x /= '\n' && x /= '\r'
+isNewline :: Char -> Bool
+isNewline x = x == '\n' || x == '\r'
 
-isBlank :: Text -> Bool
-isBlank = T.all isSpace
+notNewline :: Char -> Bool
+notNewline = not . isNewline
 
 isFrameConstituent :: Char -> Bool
 isFrameConstituent = \case
@@ -996,6 +1001,9 @@
       '`' -> True
       _   -> False
 
+isSpecialChar :: Char -> Bool
+isSpecialChar x = isMarkupChar x || x == '\\' || x == '!' || x == '<'
+
 isAsciiPunctuation :: Char -> Bool
 isAsciiPunctuation x =
   (x >= '!' && x <= '/') ||
@@ -1003,26 +1011,17 @@
   (x >= '[' && x <= '`') ||
   (x >= '{' && x <= '~')
 
-isTransparentPunctuation :: Char -> Bool
-isTransparentPunctuation = \case
-  '!' -> True
-  '"' -> True
-  '(' -> True
-  ')' -> True
-  ',' -> True
-  '-' -> True
-  '.' -> True
-  ':' -> True
-  ';' -> True
-  '?' -> True
-  '{' -> True
-  '}' -> True
-  '–' -> True
-  '—' -> True
-  _   -> False
+----------------------------------------------------------------------------
+-- Other helpers
 
-isTransparent :: Char -> Bool
-isTransparent x = Char.isSpace x || isTransparentPunctuation x
+slevel :: Pos -> Pos -> Pos
+slevel a l = if l >= ilevel a then a else l
+
+ilevel :: Pos -> Pos
+ilevel = (<> mkPos 4)
+
+isBlank :: Text -> Bool
+isBlank = T.all isSpace
 
 assembleCodeBlock :: Pos -> [Text] -> Text
 assembleCodeBlock indent ls = T.unlines (stripIndent indent <$> ls)
diff --git a/Text/MMark/Parser/Internal.hs b/Text/MMark/Parser/Internal.hs
--- a/Text/MMark/Parser/Internal.hs
+++ b/Text/MMark/Parser/Internal.hs
@@ -1,6 +1,6 @@
 -- |
 -- Module      :  Text.MMark.Parser.Internal
--- Copyright   :  © 2017 Mark Karpov
+-- Copyright   :  © 2017–2018 Mark Karpov
 -- License     :  BSD 3 clause
 --
 -- Maintainer  :  Mark Karpov <markkarpov92@gmail.com>
@@ -29,12 +29,11 @@
   , isLinksAllowed
   , disallowImages
   , isImagesAllowed
-  , isLastSpace
-  , isLastOther
-  , lastSpace
-  , lastOther
+  , getLastChar
+  , lastChar
   , lookupReference
   , Isp (..)
+  , CharType (..)
     -- * Reference and footnote definitions
   , Defs
     -- * Other
@@ -49,7 +48,7 @@
 import Data.Ratio ((%))
 import Data.Text (Text)
 import Data.Text.Metrics (damerauLevenshteinNorm)
-import Lens.Micro (Lens', (^.), (.~), set, over, to)
+import Lens.Micro (Lens', (^.), (.~), set, over)
 import Lens.Micro.Extras (view)
 import Text.MMark.Parser.Internal.Type
 import Text.Megaparsec hiding (State)
@@ -186,27 +185,16 @@
 isImagesAllowed :: IParser Bool
 isImagesAllowed = gets (view istAllowImages)
 
--- | Ask whether the last seen char type was space.
-
-isLastSpace :: IParser Bool
-isLastSpace = gets $ view (istLastChar . to (== SpaceChar))
-
--- | Ask whether the last seen char type was “other” (not space).
-
-isLastOther :: IParser Bool
-isLastOther = gets $ view (istLastChar . to (== OtherChar))
-
--- | Register that the last seen char type is space.
+-- | Get type of the last parsed character.
 
-lastSpace :: IParser ()
-lastSpace = modify' $ set istLastChar SpaceChar
-{-# INLINE lastSpace #-}
+getLastChar :: IParser CharType
+getLastChar = gets (view istLastChar)
 
--- | Register that the last seen char type is “other” (not space).
+-- | Register type of the last parsed character.
 
-lastOther :: IParser ()
-lastOther = modify' $ set istLastChar OtherChar
-{-# INLINE lastOther #-}
+lastChar :: CharType -> IParser ()
+lastChar = modify' . set istLastChar
+{-# INLINE lastChar #-}
 
 -- | Lookup a link\/image reference definition.
 
diff --git a/Text/MMark/Parser/Internal/Type.hs b/Text/MMark/Parser/Internal/Type.hs
--- a/Text/MMark/Parser/Internal/Type.hs
+++ b/Text/MMark/Parser/Internal/Type.hs
@@ -1,6 +1,6 @@
 -- |
 -- Module      :  Text.MMark.Parser.Internal.Type
--- Copyright   :  © 2017 Mark Karpov
+-- Copyright   :  © 2017–2018 Mark Karpov
 -- License     :  BSD 3 clause
 --
 -- Maintainer  :  Mark Karpov <markkarpov92@gmail.com>
@@ -124,8 +124,9 @@
 
 data CharType
   = SpaceChar          -- ^ White space or a transparent character
+  | PunctChar          -- ^ Punctuation character
   | OtherChar          -- ^ Other character
-  deriving (Eq)
+  deriving (Eq, Ord, Show)
 
 ----------------------------------------------------------------------------
 -- Reference and footnote definitions
diff --git a/Text/MMark/Render.hs b/Text/MMark/Render.hs
--- a/Text/MMark/Render.hs
+++ b/Text/MMark/Render.hs
@@ -1,6 +1,6 @@
 -- |
 -- Module      :  Text.MMark.Render
--- Copyright   :  © 2017 Mark Karpov
+-- Copyright   :  © 2017–2018 Mark Karpov
 -- License     :  BSD 3 clause
 --
 -- Maintainer  :  Mark Karpov <markkarpov92@gmail.com>
diff --git a/Text/MMark/Type.hs b/Text/MMark/Type.hs
--- a/Text/MMark/Type.hs
+++ b/Text/MMark/Type.hs
@@ -1,6 +1,6 @@
 -- |
 -- Module      :  Text.MMark.Type
--- Copyright   :  © 2017 Mark Karpov
+-- Copyright   :  © 2017–2018 Mark Karpov
 -- License     :  BSD 3 clause
 --
 -- Maintainer  :  Mark Karpov <markkarpov92@gmail.com>
diff --git a/Text/MMark/Util.hs b/Text/MMark/Util.hs
--- a/Text/MMark/Util.hs
+++ b/Text/MMark/Util.hs
@@ -1,6 +1,6 @@
 -- |
 -- Module      :  Text.MMark.Util
--- Copyright   :  © 2017 Mark Karpov
+-- Copyright   :  © 2017–2018 Mark Karpov
 -- License     :  BSD 3 clause
 --
 -- Maintainer  :  Mark Karpov <markkarpov92@gmail.com>
diff --git a/mmark.cabal b/mmark.cabal
--- a/mmark.cabal
+++ b/mmark.cabal
@@ -1,5 +1,5 @@
 name:                 mmark
-version:              0.0.4.1
+version:              0.0.4.2
 cabal-version:        >= 1.18
 tested-with:          GHC==7.10.3, GHC==8.0.2, GHC==8.2.2
 license:              BSD3
diff --git a/tests/Text/MMarkSpec.hs b/tests/Text/MMarkSpec.hs
--- a/tests/Text/MMarkSpec.hs
+++ b/tests/Text/MMarkSpec.hs
@@ -100,7 +100,7 @@
         let s = "_ _ _ _ a\n\na------\n\n---a---\n"
         in s ~-> errFancy posI (nonFlanking "_")
       it "CM26" $
-        " *\\-*" ==-> "<p><em>-</em></p>\n"
+        " *-*" ==-> "<p><em>-</em></p>\n"
       it "CM27" $
         "- foo\n***\n- bar" ==->
           "<ul>\n<li>\nfoo\n</li>\n</ul>\n<hr>\n<ul>\n<li>\nbar\n</li>\n</ul>\n"
@@ -983,10 +983,10 @@
         in s ~-> err (posN 8 s) (ueib <> etok '*' <> eic)
       it "CM348" $
         let s = "*(*foo)\n"
-        in s ~-> errFancy posI (nonFlanking "*")
+        in s ~-> err (posN 7 s) (ueib <> etok '*' <> eic)
       it "CM349" $
-        let s = "*(*foo*)*\n"
-        in s ~-> errFancy posI (nonFlanking "*")
+        "*(*foo*)*" ==->
+          "<p><em>(<em>foo</em>)</em></p>\n"
       it "CM350" $
         let s = "*foo*bar\n"
         in s ~-> errFancy (posN 4 s) (nonFlanking "*")
@@ -994,11 +994,11 @@
         let s = "_foo bar _\n"
         in s ~-> errFancy (posN 9 s) (nonFlanking "_")
       it "CM352" $
-        let s = "_(_foo)\n"
-        in s ~-> errFancy posI (nonFlanking "_")
+        let s = "_(_foo)"
+        in s ~-> err (posN 7 s) (ueib <> etok '_' <> eic)
       it "CM353" $
-        let s = "_(_foo_)_\n"
-        in s ~-> errFancy posI (nonFlanking "_")
+        "_(_foo_)_" ==->
+          "<p><em>(<em>foo</em>)</em></p>\n"
       it "CM354" $
         let s = "_foo_bar\n"
         in s ~-> errFancy (posN 4 s) (nonFlanking "_")
@@ -1009,7 +1009,7 @@
         let s = "_foo_bar_baz_\n"
         in s ~-> errFancy (posN 4 s) (nonFlanking "_")
       it "CM357" $
-        "_\\(bar\\)_.\n" ==-> "<p><em>(bar)</em>.</p>\n"
+        "_(bar\\)_.\n" ==-> "<p><em>(bar)</em>.</p>\n"
       it "CM358" $
         "**foo bar**\n" ==-> "<p><strong>foo bar</strong></p>\n"
       it "CM359" $
@@ -1045,19 +1045,19 @@
         "__foo, __bar__, baz__" ==->
           "<p><strong>foo, <strong>bar</strong>, baz</strong></p>\n"
       it "CM370" $
-        "foo-__\\(bar\\)__" ==-> "<p>foo-<strong>(bar)</strong></p>\n"
+        "foo-__\\(bar)__" ==-> "<p>foo-<strong>(bar)</strong></p>\n"
       it "CM371" $
         let s = "**foo bar **\n"
         in s ~-> errFancy (posN 10 s) (nonFlanking "**")
       it "CM372" $
         let s = "**(**foo)\n"
-        in s ~-> errFancy posI (nonFlanking "**")
+        in s ~-> err (posN 9 s) (ueib <> etoks "**" <> eic)
       it "CM373" $
-        let s = "*(**foo**)*\n"
-        in s ~-> errFancy posI (nonFlanking "*")
-      xit "CM374" $ -- FIXME doesn't pass with current approach
+        "*(**foo**)*" ==->
+          "<p><em>(<strong>foo</strong>)</em></p>\n"
+      it "CM374" $
         "**Gomphocarpus (*Gomphocarpus physocarpus*, syn.\n*Asclepias physocarpa*)**" ==->
-        "<p><strong>Gomphocarpus (<em>Gomphocarpus physocarpus</em>, syn.\n<em>Asclepias physocarpa</em>)</strong></p>\n"
+          "<p><strong>Gomphocarpus (<em>Gomphocarpus physocarpus</em>, syn.\n<em>Asclepias physocarpa</em>)</strong></p>\n"
       it "CM375" $
         "**foo \"*bar*\" foo**" ==->
           "<p><strong>foo &quot;<em>bar</em>&quot; foo</strong></p>\n"
@@ -1069,10 +1069,10 @@
         in s ~-> errFancy (posN 10 s) (nonFlanking "__")
       it "CM378" $
         let s = "__(__foo)\n"
-        in s ~-> errFancy posI (nonFlanking "__")
+        in s ~-> err (posN 9 s) (ueib <> etoks "__" <> eic)
       it "CM379" $
-        let s = "_(__foo__)_\n"
-        in s ~-> errFancy posI (nonFlanking "_")
+        "_(__foo__)_" ==->
+          "<p><em>(<strong>foo</strong>)</em></p>\n"
       it "CM380" $
         let s = "__foo__bar\n"
         in s ~-> errFancy (posN 5 s) (nonFlanking "__")
@@ -1083,7 +1083,7 @@
         "__foo\\_\\_bar\\_\\_baz__" ==->
           "<p><strong>foo__bar__baz</strong></p>\n"
       it "CM383" $
-        "__\\(bar\\)__." ==->
+        "__(bar\\)__." ==->
           "<p><strong>(bar)</strong>.</p>\n"
       it "CM384" $
         "*foo [bar](/url)*" ==->
@@ -1862,7 +1862,7 @@
         let s = "#My header\n\nSomething goes __here __.\n"
         s ~~->
           [ err (posN 1 s) (utok 'M' <> etok '#' <> ews)
-          , errFancy (posN 34 s) (nonFlanking "__") ]
+          , err (posN 37 s) (ueib <> etoks "__" <> eic) ]
       describe "every block in a list gets its parse error propagated" $ do
         context "with unordered list" $
           it "works" $ do