Script tags can contain less-than symbol

snoyberg · Oct 18, 2018 · 3e1bd12 · 3e1bd12
1 parent 92f68f4
commit 3e1bd12
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 3 deletions.
diff --git a/html-conduit/src/Text/HTML/TagStream.hs b/html-conduit/src/Text/HTML/TagStream.hs
@@ -167,7 +167,7 @@ tillScriptEnd open =
       chunk <- takeTill (== '<')
       let acc' = acc <> B.fromText chunk
           finish = pure [open, Text $ L.toStrict $ B.toLazyText acc', TagClose "script"]
-          hasContent = (string "/script>" *> finish) <|> loop acc'
+          hasContent = (string "/script>" *> finish) <|> loop (acc' <> "<")
       (char '<' *> hasContent) <|> finish
 
 tokens :: Parser [Token]

diff --git a/html-conduit/test/main.hs b/html-conduit/test/main.hs
@@ -120,9 +120,9 @@ main = hspec $ do
 
     describe "script tags" $ do
       it "ignores funny characters" $
-        let html = "<script>hello > world</script>"
+        let html = "<script>hello <> world</script>"
             doc = X.Document (X.Prologue [] Nothing []) root []
-            root = X.Element "script" Map.empty [X.NodeContent "hello > world"]
+            root = X.Element "script" Map.empty [X.NodeContent "hello <> world"]
          in H.parseLBS html @?= doc
 
       {-