From 454b6e7bdd2b1a83625b723d3a5ba1a28e08f66a Mon Sep 17 00:00:00 2001
From: John Koster <john@stillat.com>
Date: Fri, 4 Sep 2026 10:50:40 -0500
Subject: [PATCH] Fix Antlers line and column numbers after multibyte text or a
 leading newline

---
 .../Language/Parser/DocumentParser.php        | 23 +++++--------
 tests/Antlers/Parser/BasicNodeTest.php        | 33 +++++++++++++++++++
 2 files changed, 42 insertions(+), 14 deletions(-)

diff --git a/src/View/Antlers/Language/Parser/DocumentParser.php b/src/View/Antlers/Language/Parser/DocumentParser.php
index ee5f47b2adc..f6d0bd154ac 100644
--- a/src/View/Antlers/Language/Parser/DocumentParser.php
+++ b/src/View/Antlers/Language/Parser/DocumentParser.php
@@ -328,24 +328,19 @@ private function processInputText($input)
         $this->inputLen = mb_strlen($this->content);
 
         // The document content was normalized, so we can search for "\n".
-        preg_match_all('/\n/', $this->content, $documentNewLines, PREG_OFFSET_CAPTURE);
-        $newLineCountLen = count($documentNewLines[0]);
-
         $currentLine = $this->seedStartLine;
-        $lastOffset = null;
-        for ($i = 0; $i < $newLineCountLen; $i++) {
-            $thisNewLine = $documentNewLines[0][$i];
-            $thisIndex = $thisNewLine[1];
-            $indexChar = $thisIndex;
-
-            if ($lastOffset != null) {
-                $indexChar = $thisIndex - $lastOffset;
-            } else {
-                $indexChar = $indexChar + 1;
+        $lastOffset = -1;
+        $thisIndex = -1;
+
+        foreach (explode("\n", $this->content) as $line) {
+            $thisIndex += mb_strlen($line) + 1;
+
+            if ($thisIndex >= $this->inputLen) {
+                break;
             }
 
             $this->documentOffsets[$thisIndex] = [
-                self::K_CHAR => $indexChar,
+                self::K_CHAR => $thisIndex - $lastOffset,
                 self::K_LINE => $currentLine,
             ];
 
diff --git a/tests/Antlers/Parser/BasicNodeTest.php b/tests/Antlers/Parser/BasicNodeTest.php
index 8e89e89fa83..74c3bcf53d7 100644
--- a/tests/Antlers/Parser/BasicNodeTest.php
+++ b/tests/Antlers/Parser/BasicNodeTest.php
@@ -31,6 +31,39 @@ public function test_it_returns_nodes()
         $this->assertInstanceOf(LiteralNode::class, $nodes[3]);
     }
 
+    public function test_multibyte_literals_keep_character_based_region_boundaries_and_lines()
+    {
+        $nodes = $this->parseNodes("caf\u{00E9}\n\u{65E5}\u{672C} {{ first }}\n\u{1F389} {{ second }} tail");
+
+        $this->assertSame("caf\u{00E9}\n\u{65E5}\u{672C} ", $nodes[0]->content);
+        $this->assertSame(' first ', $nodes[1]->content);
+        $this->assertSame("\n\u{1F389} ", $nodes[2]->content);
+        $this->assertSame(' second ', $nodes[3]->content);
+        $this->assertSame(' tail', $nodes[4]->content);
+
+        $this->assertSame(8, $nodes[1]->startPosition->offset);
+        $this->assertSame(2, $nodes[1]->startPosition->line);
+        $this->assertSame(4, $nodes[1]->startPosition->char);
+        $this->assertSame(22, $nodes[3]->startPosition->offset);
+        $this->assertSame(3, $nodes[3]->startPosition->line);
+        $this->assertSame(3, $nodes[3]->startPosition->char);
+    }
+
+    public function test_columns_are_correct_after_a_document_initial_newline()
+    {
+        foreach (["\n", "\r", "\r\n"] as $newline) {
+            $nodes = $this->parseNodes($newline."\u{65E5}\u{672C}{{ first }}".$newline.'ab{{ second }}');
+
+            $this->assertSame(3, $nodes[1]->startPosition->offset);
+            $this->assertSame(2, $nodes[1]->startPosition->line);
+            $this->assertSame(3, $nodes[1]->startPosition->char);
+
+            $this->assertSame(17, $nodes[3]->startPosition->offset);
+            $this->assertSame(3, $nodes[3]->startPosition->line);
+            $this->assertSame(3, $nodes[3]->startPosition->char);
+        }
+    }
+
     public function test_it_doesnt_trim_off_content_start()
     {
         $nodes = $this->parseNodes('{{ meta_title ?? title ?? "No Title Set" }}');