Make auto_identifiers work as advertised.

Replace emojis with aliases.
jgm · Jun 17, 2024 · b7f0ca3 · b7f0ca3
1 parent 097ba1f
commit b7f0ca3
Show file tree

Hide file tree

Showing 3 changed files with 16 additions and 2 deletions.
diff --git a/commonmark-extensions/commonmark-extensions.cabal b/commonmark-extensions/commonmark-extensions.cabal
@@ -62,7 +62,7 @@ library
     , network-uri
     , commonmark >= 0.2.4.1 && < 0.3
     -- for extensions:
-    , emojis >= 0.1 && < 0.2
+    , emojis >= 0.1.4 && < 0.2
   exposed-modules:
       Commonmark.Extensions
       Commonmark.Extensions.Smart

diff --git a/commonmark-extensions/src/Commonmark/Extensions/AutoIdentifiers.hs b/commonmark-extensions/src/Commonmark/Extensions/AutoIdentifiers.hs
@@ -16,6 +16,7 @@ import Data.Char (isSpace, isAlphaNum, isAscii, isMark,
 import Data.Dynamic
 import qualified Data.Map as M
 import qualified Data.Text as T
+import qualified Text.Emoji as Emoji
 import Text.Parsec
 
 autoIdentifiersSpec :: (Monad m, IsBlock il bl, IsInline il, ToPlainText il)
@@ -49,7 +50,8 @@ addId ascii bd
       Nothing  -> do
         contents <- runInlineParser
                     (removeIndent . mconcat . reverse . blockLines $ bd)
-        let ident = makeIdentifier ascii (toPlainText contents)
+        let ident = makeIdentifier ascii
+                     (Emoji.replaceEmojis emojiToAlias $ toPlainText contents)
         counterMap <- counters <$> getState
         let key = "identifier:" <> ident
         cnt <- case M.lookup key counterMap of
@@ -64,6 +66,10 @@ addId ascii bd
       Just _ -> return $! bd
   | otherwise = return $! bd
 
+emojiToAlias :: T.Text -> [T.Text] -> T.Text
+emojiToAlias t [] = t
+emojiToAlias _ (a:_) = a
+
 makeIdentifier :: Bool -> T.Text -> T.Text
 makeIdentifier ascii = toIdent . T.toLower
   where

diff --git a/commonmark-extensions/test/auto_identifiers.md b/commonmark-extensions/test/auto_identifiers.md
@@ -82,3 +82,11 @@ Hi
 <h1 id="hi-2">Hi</h1>
 <h1 id="hi-3">Hi</h1>
 ````````````````````````````````
+
+Emojis should be replaced with aliases:
+
+```
+# Introduction ❤️💯
+.
+<h1 id="introduction-heart-100">Introduction ❤️💯</h1>
+```