Merge branch 'master' of https://github.com/fsharp/FSharp.Data

fsprojects · Sep 24, 2019 · 936194f · 936194f
2 parents 993cf47 + 75bd4df
commit 936194f
Show file tree

Hide file tree

Showing 3 changed files with 38 additions and 7 deletions.
diff --git a/src/FSharp.Data/FSharp.Data.fsproj b/src/FSharp.Data/FSharp.Data.fsproj
@@ -9,6 +9,9 @@
     <DisableImplicitFSharpCoreReference>true</DisableImplicitFSharpCoreReference>
     <GenerateAssemblyInfo>false</GenerateAssemblyInfo>
   </PropertyGroup>
+  <PropertyGroup Condition="'$(Configuration)|$(TargetFramework)|$(Platform)'=='Debug|netstandard2.0|AnyCPU'">
+    <Tailcalls>true</Tailcalls>
+  </PropertyGroup>
   <ItemGroup>
     <Compile Include="..\Net\Http.fs" />
     <Compile Include="..\CommonRuntime\IO.fs" />

diff --git a/src/Html/HtmlParser.fs b/src/Html/HtmlParser.fs
@@ -700,7 +700,7 @@ module internal HtmlParser =
             | '&' ->
                 assert (state.ContentLength = 0)
                 state.InsertionMode := InsertionMode.CharRefMode
-                attributeValueCharRef ['/'; '>'] attributeValueUnquoted state
+                attributeValueUnquotedCharRef  ['/'; '>'] state
             | _ -> state.ConsAttrValue(); attributeValueUnquoted state
         and attributeValueUnquotedSlash state =
             match state.Peek() with
@@ -713,23 +713,38 @@ module internal HtmlParser =
             | '&' ->
                 assert (state.ContentLength = 0)
                 state.InsertionMode := InsertionMode.CharRefMode
-                attributeValueCharRef [quote] (attributeValueQuoted quote) state
+                attributeValueQuotedCharRef quote state
             | _ -> state.ConsAttrValue(); attributeValueQuoted quote state
-        and attributeValueCharRef stop continuation (state:HtmlState) =
+        and attributeValueQuotedCharRef quote state =
             match state.Peek() with
             | ';' ->
                 state.Cons()
                 state.EmitToAttributeValue()
-                continuation state
+                attributeValueQuoted quote state
             | TextParser.EndOfFile _ ->
                 state.EmitToAttributeValue()
-                continuation state
+                attributeValueQuoted quote state
+            | c when c = quote ->
+                state.EmitToAttributeValue()
+                attributeValueQuoted quote state
+            | _ ->
+                state.Cons()
+                attributeValueQuotedCharRef quote state
+        and attributeValueUnquotedCharRef stop state =
+            match state.Peek() with
+            | ';' ->
+                state.Cons()
+                state.EmitToAttributeValue()
+                attributeValueUnquoted state
+            | TextParser.EndOfFile _ ->
+                state.EmitToAttributeValue()
+                attributeValueUnquoted state
             | c when List.exists ((=) c) stop ->
                 state.EmitToAttributeValue()
-                continuation state
+                attributeValueUnquoted state
             | _ ->
                 state.Cons()
-                attributeValueCharRef stop continuation state
+                attributeValueUnquotedCharRef stop state
         and afterAttributeValueQuoted state =
             match state.Peek() with
             | TextParser.Whitespace _ -> state.Pop(); state.NewAttribute(); afterAttributeValueQuoted state

diff --git a/tests/FSharp.Data.Tests/HtmlParser.fs b/tests/FSharp.Data.Tests/HtmlParser.fs
@@ -129,6 +129,19 @@ let ``Can handle attributes with no value``() =
         ]
     node.Attributes() |> should equal expected
 
+[<Test>]
+let ``Can handle long html encoded attributes without StackOverflow``() = 
+    let html =
+        HtmlNode.Parse (
+            let sb = new System.Text.StringBuilder()
+            sb.Append "<html><body><p attrib=\"" |> ignore
+            for i in 0 .. 50000 do sb.Append "&lt;br/&gt;" |> ignore
+            sb.Append "\">Test</p></html></body>" |> ignore
+            sb.ToString()
+        )
+    let result = html.Head.InnerText()
+    result |> should equal "Test"
+
 [<TestCase("var r = \"</script>\"")>]
 [<TestCase("var r = '</script>'")>]
 [<TestCase("var r = /</g")>]