From 683ee75703f9bde9ecf34ae3ec6ab2c3b68b52b2 Mon Sep 17 00:00:00 2001
From: Jeroen van Rijn <Kelimion@users.noreply.github.com>
Date: Fri, 28 Jul 2023 15:53:39 +0200
Subject: Fix #2684

---
 core/encoding/xml/debug_print.odin         | 16 +++++++-----
 core/encoding/xml/example/xml_example.odin |  6 ++---
 core/encoding/xml/helpers.odin             | 29 ++++++++++++---------
 core/encoding/xml/xml_reader.odin          | 42 +++++++++++-------------------
 4 files changed, 44 insertions(+), 49 deletions(-)

(limited to 'core/encoding/xml')

diff --git a/core/encoding/xml/debug_print.odin b/core/encoding/xml/debug_print.odin
index e9a1cb160..0736e8893 100644
--- a/core/encoding/xml/debug_print.odin
+++ b/core/encoding/xml/debug_print.odin
@@ -65,19 +65,21 @@ print_element :: proc(writer: io.Writer, doc: ^Document, element_id: Element_ID,
 
 	if element.kind == .Element {
 		wprintf(writer, "<%v>\n", element.ident)
-		if len(element.value) > 0 {
-			tab(writer, indent + 1)
-			wprintf(writer, "[Value] %v\n", element.value)
+
+		for value in element.value {
+			switch v in value {
+			case string:
+				tab(writer, indent + 1)
+				wprintf(writer, "[Value] %v\n", v)
+			case Element_ID:
+				print_element(writer, doc, v, indent + 1)
+			}
 		}
 
 		for attr in element.attribs {
 			tab(writer, indent + 1)
 			wprintf(writer, "[Attr] %v: %v\n", attr.key, attr.val)
 		}
-
-		for child in element.children {
-			print_element(writer, doc, child, indent + 1)
-		}
 	} else if element.kind == .Comment {
 		wprintf(writer, "[COMMENT] %v\n", element.value)
 	}
diff --git a/core/encoding/xml/example/xml_example.odin b/core/encoding/xml/example/xml_example.odin
index 887b40764..aebb8d0ea 100644
--- a/core/encoding/xml/example/xml_example.odin
+++ b/core/encoding/xml/example/xml_example.odin
@@ -72,10 +72,10 @@ example :: proc() {
 	 	return
 	}
 
-	printf("Found `<charlist>` with %v children, %v elements total\n", len(docs[0].elements[charlist].children), docs[0].element_count)
+	printf("Found `<charlist>` with %v children, %v elements total\n", len(docs[0].elements[charlist].value), docs[0].element_count)
 
-	crc32 := doc_hash(docs[0])
-	printf("[%v] CRC32: 0x%08x\n", "🎉" if crc32 == 0xcaa042b9 else "🤬", crc32)
+	crc32 := doc_hash(docs[0], false)
+	printf("[%v] CRC32: 0x%08x\n", "🎉" if crc32 == 0x420dbac5 else "🤬", crc32)
 
 	for round in 0..<N {
 		defer xml.destroy(docs[round])
diff --git a/core/encoding/xml/helpers.odin b/core/encoding/xml/helpers.odin
index 48f058334..200c5c1de 100644
--- a/core/encoding/xml/helpers.odin
+++ b/core/encoding/xml/helpers.odin
@@ -13,20 +13,25 @@ find_child_by_ident :: proc(doc: ^Document, parent_id: Element_ID, ident: string
 	tag := doc.elements[parent_id]
 
 	count := 0
-	for child_id in tag.children {
-		child := doc.elements[child_id]
-		/*
-			Skip commments. They have no name.
-		*/
-		if child.kind  != .Element                { continue }
+	for v in tag.value {
+		switch child_id in v {
+		case string: continue
+		case Element_ID:
+			child := doc.elements[child_id]
+			/*
+				Skip commments. They have no name.
+			*/
+			if child.kind  != .Element                { continue }
 
-		/*
-			If the ident matches and it's the nth such child, return it.
-		*/
-		if child.ident == ident {
-			if count == nth                       { return child_id, true }
-			count += 1
+			/*
+				If the ident matches and it's the nth such child, return it.
+			*/
+			if child.ident == ident {
+				if count == nth                       { return child_id, true }
+				count += 1
+			}
 		}
+
 	}
 	return 0, false
 }
diff --git a/core/encoding/xml/xml_reader.odin b/core/encoding/xml/xml_reader.odin
index f5523c299..f4f8a4b05 100644
--- a/core/encoding/xml/xml_reader.odin
+++ b/core/encoding/xml/xml_reader.odin
@@ -125,16 +125,19 @@ Document :: struct {
 
 Element :: struct {
 	ident:   string,
-	value:   string,
+	value:   [dynamic]Value,
 	attribs: Attributes,
 
 	kind: enum {
 		Element = 0,
 		Comment,
 	},
-
 	parent:   Element_ID,
-	children: [dynamic]Element_ID,
+}
+
+Value :: union {
+	string,
+	Element_ID,
 }
 
 Attribute :: struct {
@@ -247,9 +250,6 @@ parse_bytes :: proc(data: []u8, options := DEFAULT_OPTIONS, path := "", error_ha
 
 	err =            .Unexpected_Token
 	element, parent: Element_ID
-
-	tag_is_open   := false
-	first_element := true
 	open: Token
 
 	/*
@@ -275,16 +275,10 @@ parse_bytes :: proc(data: []u8, options := DEFAULT_OPTIONS, path := "", error_ha
 					e.g. <odin - Start of new element.
 				*/
 				element = new_element(doc)
-				tag_is_open = true
-
-				if first_element {
-					/*
-						First element.
-					*/
-					parent   = element
-					first_element = false
+				if element == 0 { // First Element
+					parent = element
 				} else {
-					append(&doc.elements[parent].children, element)
+					append(&doc.elements[parent].value, element)
 				}
 
 				doc.elements[element].parent = parent
@@ -324,7 +318,6 @@ parse_bytes :: proc(data: []u8, options := DEFAULT_OPTIONS, path := "", error_ha
 					expect(t, .Gt) or_return
 					parent      = doc.elements[element].parent
 					element     = parent
-					tag_is_open = false
 
 				case:
 					error(t, t.offset, "Expected close tag, got: %#v\n", end_token)
@@ -344,7 +337,6 @@ parse_bytes :: proc(data: []u8, options := DEFAULT_OPTIONS, path := "", error_ha
 				}
 				parent      = doc.elements[element].parent
 				element     = parent
-				tag_is_open = false
 
 			} else if open.kind == .Exclaim {
 				/*
@@ -392,8 +384,8 @@ parse_bytes :: proc(data: []u8, options := DEFAULT_OPTIONS, path := "", error_ha
 							el := new_element(doc)
 							doc.elements[el].parent = element
 							doc.elements[el].kind   = .Comment
-							doc.elements[el].value  = comment
-							append(&doc.elements[element].children, el)
+							append(&doc.elements[el].value, comment)
+							append(&doc.elements[element].value, el)
 						}
 					}
 
@@ -436,9 +428,6 @@ parse_bytes :: proc(data: []u8, options := DEFAULT_OPTIONS, path := "", error_ha
 			/*
 				End of file.
 			*/
-			if tag_is_open {
-				return doc, .Premature_EOF
-			}
 			break loop
 
 		case:
@@ -450,7 +439,7 @@ parse_bytes :: proc(data: []u8, options := DEFAULT_OPTIONS, path := "", error_ha
 			needs_processing |= .Decode_SGML_Entities in opts.flags
 
 			if !needs_processing {
-				doc.elements[element].value = body_text
+				append(&doc.elements[element].value, body_text)
 				continue
 			}
 
@@ -472,10 +461,10 @@ parse_bytes :: proc(data: []u8, options := DEFAULT_OPTIONS, path := "", error_ha
 
 			decoded, decode_err := entity.decode_xml(body_text, decode_opts)
 			if decode_err == .None {
-				doc.elements[element].value = decoded
+				append(&doc.elements[element].value, decoded)
 				append(&doc.strings_to_free, decoded)
 			} else {
-				doc.elements[element].value = body_text
+				append(&doc.elements[element].value, body_text)
 			}
 		}
 	}
@@ -518,7 +507,7 @@ destroy :: proc(doc: ^Document) {
 
 	for el in doc.elements {
 		delete(el.attribs)
-		delete(el.children)
+		delete(el.value)
 	}
 	delete(doc.elements)
 
@@ -710,6 +699,5 @@ new_element :: proc(doc: ^Document) -> (id: Element_ID) {
 
 	cur := doc.element_count
 	doc.element_count += 1
-
 	return cur
 }
\ No newline at end of file
-- 
cgit v1.2.3