From 461a96ea163b144ea2898d088efe65fce311d5be Mon Sep 17 00:00:00 2001
From: Mike Dalessio <mike.dalessio@gmail.com>
Date: Fri, 15 Mar 2024 11:03:26 -0400
Subject: [PATCH] fix: Reader#read sets @encoding if it is unset

This allows Reader#encoding to remain unchanged while the libxml2
implementation of encoding reporting has changed in v2.12.6.
---
 CHANGELOG.md              | 5 +++++
 ext/nokogiri/xml_reader.c | 1 +
 2 files changed, 6 insertions(+)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index b70cf13543..da6e227e65 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -11,6 +11,11 @@ Nokogiri follows [Semantic Versioning](https://semver.org/), please see the [REA
 * [CRuby] Vendored libxml2 is updated to [v2.12.6](https://gitlab.gnome.org/GNOME/libxml2/-/releases/v2.12.6) from v2.12.5. (@flavorjones)
 
 
+### Changed
+
+* [CRuby] `XML::Reader` sets the `@encoding` instance variable during reading if it is not passed into the initializer. Previously, it would remain `nil`. The behavior of `Reader#encoding` has not changed. This works around changes to how libxml2 reports the encoding used in v2.12.6.
+
+
 ## v1.16.2 / 2024-02-04
 
 ### Security
diff --git a/ext/nokogiri/xml_reader.c b/ext/nokogiri/xml_reader.c
index 0520f8f7d5..c987e2bd3d 100644
--- a/ext/nokogiri/xml_reader.c
+++ b/ext/nokogiri/xml_reader.c
@@ -537,6 +537,7 @@ read_more(VALUE self)
     if (RTEST(constructor_encoding)) {
       c_document->encoding = xmlStrdup(BAD_CAST StringValueCStr(constructor_encoding));
     } else {
+      rb_iv_set(self, "@encoding", NOKOGIRI_STR_NEW2("UTF-8"));
       c_document->encoding = xmlStrdup(BAD_CAST "UTF-8");
     }
   }