From e7ba03ddd109dda3f0d1cdac4a8053a4753bfbf0 Mon Sep 17 00:00:00 2001
From: Andreas Kling <kling@serenityos.org>
Date: Thu, 24 Nov 2022 14:57:20 +0100
Subject: [PATCH] AK: Add
 Utf8View::iterator_at_byte_offset_without_validation()

Unlike iterator_at_byte_offset(), this function assumes the provided
byte offset is a valid offset into the UTF-8 character stream.

This avoids walking the stream from the start.
---
 AK/Utf8View.cpp | 5 +++++
 AK/Utf8View.h   | 2 ++
 2 files changed, 7 insertions(+)

diff --git a/AK/Utf8View.cpp b/AK/Utf8View.cpp
index 7e7788a0ff..5ebfccedfd 100644
--- a/AK/Utf8View.cpp
+++ b/AK/Utf8View.cpp
@@ -24,6 +24,11 @@ Utf8CodePointIterator Utf8View::iterator_at_byte_offset(size_t byte_offset) cons
     return end();
 }
 
+Utf8CodePointIterator Utf8View::iterator_at_byte_offset_without_validation(size_t byte_offset) const
+{
+    return Utf8CodePointIterator { reinterpret_cast<u8 const*>(m_string.characters_without_null_termination()) + byte_offset, m_string.length() - byte_offset };
+}
+
 size_t Utf8View::byte_offset_of(Utf8CodePointIterator const& it) const
 {
     VERIFY(it.m_ptr >= begin_ptr());
diff --git a/AK/Utf8View.h b/AK/Utf8View.h
index 19dd62ff1d..6c84ce42ab 100644
--- a/AK/Utf8View.h
+++ b/AK/Utf8View.h
@@ -80,6 +80,8 @@ public:
     Utf8CodePointIterator end() const { return { end_ptr(), 0 }; }
     Utf8CodePointIterator iterator_at_byte_offset(size_t) const;
 
+    Utf8CodePointIterator iterator_at_byte_offset_without_validation(size_t) const;
+
     unsigned char const* bytes() const { return begin_ptr(); }
     size_t byte_length() const { return m_string.length(); }
     size_t byte_offset_of(Utf8CodePointIterator const&) const;