From c8332d90c01f6558e1c13fd5ebb96e85818c71d6 Mon Sep 17 00:00:00 2001
From: Dan Griffin <github.mk@xiragon.com>
Date: Wed, 24 May 2023 22:48:35 +0500
Subject: [PATCH 1/2] Add tests for comments and processing instructions in XML
 prolog

See <https://www.w3.org/TR/xml11/#NT-prolog>

failures (2):
    xml_prolog::comments
    xml_prolog::pi
---
 tests/serde-de.rs | 74 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 74 insertions(+)

diff --git a/tests/serde-de.rs b/tests/serde-de.rs
index 7b12fd02..554abac4 100644
--- a/tests/serde-de.rs
+++ b/tests/serde-de.rs
@@ -6496,3 +6496,77 @@ mod resolve {
         );
     }
 }
+
+/// Tests for https://github.com/tafia/quick-xml/pull/603.
+///
+/// According to <https://www.w3.org/TR/xml11/#NT-prolog> comments,
+/// processing instructions and spaces are possible after XML declaration or DTD.
+/// Their existence should not break deserializing
+///
+/// ```text
+/// [22] prolog ::= XMLDecl Misc* (doctypedecl Misc*)?
+/// [27] Misc   ::= Comment | PI | S
+/// ```
+mod xml_prolog {
+    use super::*;
+    use pretty_assertions::assert_eq;
+    use std::collections::HashMap;
+
+    #[test]
+    fn spaces() {
+        assert_eq!(
+            from_str::<HashMap<(), ()>>(
+                r#"
+        <?xml version="1.1"?>
+
+        <!DOCTYPE dict>
+
+        <doc>
+        </doc>
+        "#
+            )
+            .unwrap(),
+            HashMap::new()
+        );
+    }
+
+    #[test]
+    fn comments() {
+        assert_eq!(
+            from_str::<HashMap<(), ()>>(
+                r#"
+        <?xml version="1.1"?>
+        <!-- comment between xml declaration and doctype -->
+        <!-- another comment -->
+        <!DOCTYPE dict>
+        <!-- comment between doctype and root element -->
+        <!-- another comment -->
+        <doc>
+        </doc>
+        "#,
+            )
+            .unwrap(),
+            HashMap::new()
+        );
+    }
+
+    #[test]
+    fn pi() {
+        assert_eq!(
+            from_str::<HashMap<(), ()>>(
+                r#"
+        <?xml version="1.1"?>
+        <?pi?>
+        <?another pi?>
+        <!DOCTYPE dict>
+        <?pi?>
+        <?another pi?>
+        <doc>
+        </doc>
+        "#,
+            )
+            .unwrap(),
+            HashMap::new()
+        );
+    }
+}

From d49f2d54287542fcafac2943d39fb9be86581259 Mon Sep 17 00:00:00 2001
From: Dan Griffin <github.mk@xiragon.com>
Date: Sat, 10 Jun 2023 00:14:34 +0500
Subject: [PATCH 2/2] Trim `Text` events after DOCTYPE so spaces does not
 produce an event.

Otherwise consequent `Text` events (which is possible if their delimited by
Comment or PI events, which is skipped) will be merged but not trimmed.
That will lead to returning a `Text` event when try to call `deserialize_struct`
or `deserialize_map` which will trigger `DeError::ExpectedStart` error.

The incorrect trim behavior was introduced in #581, when DocType event began to be processed
---
 Changelog.md  | 6 ++++++
 src/de/mod.rs | 2 +-
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/Changelog.md b/Changelog.md
index e0a01192..f1bf2801 100644
--- a/Changelog.md
+++ b/Changelog.md
@@ -18,9 +18,15 @@
 
 ### Bug Fixes
 
+- [#603]: Fix a regression from [#581] that an XML comment or a processing
+  instruction between a <!DOCTYPE> and the root element in the file brokes
+  deserialization of structs by returning `DeError::ExpectedStart`
+
 ### Misc Changes
 
+[#581]: https://github.com/tafia/quick-xml/pull/581
 [#601]: https://github.com/tafia/quick-xml/pull/601
+[#603]: https://github.com/tafia/quick-xml/pull/603
 [#606]: https://github.com/tafia/quick-xml/pull/606
 
 
diff --git a/src/de/mod.rs b/src/de/mod.rs
index cc0d59d7..20017d00 100644
--- a/src/de/mod.rs
+++ b/src/de/mod.rs
@@ -2880,7 +2880,7 @@ impl StartTrimmer {
     #[inline(always)]
     fn trim<'a>(&mut self, event: Event<'a>) -> Option<PayloadEvent<'a>> {
         let (event, trim_next_event) = match event {
-            Event::DocType(e) => (PayloadEvent::DocType(e), false),
+            Event::DocType(e) => (PayloadEvent::DocType(e), true),
             Event::Start(e) => (PayloadEvent::Start(e), true),
             Event::End(e) => (PayloadEvent::End(e), true),
             Event::Eof => (PayloadEvent::Eof, true),