From aa2243bad97b755171dd3cadf7a09b7dd0cd7967 Mon Sep 17 00:00:00 2001
From: "github-actions[bot]"
 <41898282+github-actions[bot]@users.noreply.github.com>
Date: Sun, 2 Aug 2026 15:44:07 +0000
Subject: [PATCH] perf: HtmlParser reuses CurrentTag and Content StringBuilders
 across tokens

Instead of allocating a new { Contents = StringBuilder() } on every
EmitTag, EmitSelfClosingTag, EmitToAttributeValue, and Emit call,
reuse the existing CharList by calling .Clear(). This eliminates two
StringBuilder allocations per HTML element (one for CurrentTag, one
for Content) - for a typical HTML document with hundreds of elements
this removes hundreds of short-lived heap objects.

The CharList.Clear() method already existed for this purpose.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
---
 src/FSharp.Data.Html.Core/HtmlParser.fs | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/src/FSharp.Data.Html.Core/HtmlParser.fs b/src/FSharp.Data.Html.Core/HtmlParser.fs
index 2940216ea..af6ce78f3 100644
--- a/src/FSharp.Data.Html.Core/HtmlParser.fs
+++ b/src/FSharp.Data.Html.Core/HtmlParser.fs
@@ -171,7 +171,7 @@ module internal HtmlParser =
         member x.EmitSelfClosingTag() =
             let name = x.CurrentTag.ToString().Trim()
             let result = Tag(true, name, x.GetAttributes())
-            x.CurrentTag <- { Contents = StringBuilder() }
+            x.CurrentTag.Clear()
             x.InsertionMode <- DefaultMode
             x.Attributes <- []
             x.Tokens <- result :: x.Tokens
@@ -217,7 +217,7 @@ module internal HtmlParser =
                 else
                     DefaultMode
 
-            x.CurrentTag <- { Contents = StringBuilder() }
+            x.CurrentTag.Clear()
             x.Attributes <- []
             x.Tokens <- result :: x.Tokens
 
@@ -228,7 +228,7 @@ module internal HtmlParser =
             for c in content do
                 x.ConsAttrValue c
 
-            x.Content <- { Contents = StringBuilder() }
+            x.Content.Clear()
             x.InsertionMode <- DefaultMode
 
         member x.Emit() : unit =
@@ -252,7 +252,7 @@ module internal HtmlParser =
                 | DocTypeMode -> DocType content
                 | CDATAMode -> CData(content.Replace("<![CDATA[", "").Replace("]]>", ""))
 
-            x.Content <- { Contents = StringBuilder() }
+            x.Content.Clear()
             x.InsertionMode <- DefaultMode
 
             match result with