From e20f4d7a656e47553f9da9d594e299e2fa2dbe41 Mon Sep 17 00:00:00 2001
From: Nick Wellnhofer <wellnhofer@aevum.de>
Date: Mon, 13 Feb 2023 14:38:05 +0100
Subject: [PATCH] xinclude: Fix quadratic behavior in xmlXIncludeLoadTxt

Also make text inclusions work with memory buffers, for example when
using a custom entity loader, and fix a memory leak in case of invalid
characters.

Fixes #483.

Reference:https://github.com/GNOME/libxml2/commit/e20f4d7a656e47553f9da9d594e299e2fa2dbe41
Conflict:NA

---
 result/XInclude/invalid_char.xml.err |  2 +
 result/XInclude/invalid_char.xml.rdr |  7 ++++
 test/XInclude/docs/invalid_char.xml  |  3 ++
 test/XInclude/ents/invalid_char.txt  |  1 +
 xinclude.c                           | 61 ++++++++++++----------------
 5 files changed, 39 insertions(+), 35 deletions(-)
 create mode 100644 result/XInclude/invalid_char.xml.err
 create mode 100644 result/XInclude/invalid_char.xml.rdr
 create mode 100644 test/XInclude/docs/invalid_char.xml
 create mode 100644 test/XInclude/ents/invalid_char.txt

diff --git a/result/XInclude/invalid_char.xml.err b/result/XInclude/invalid_char.xml.err
new file mode 100644
index 0000000..c28c109
--- /dev/null
+++ b/result/XInclude/invalid_char.xml.err
@@ -0,0 +1,2 @@
+./test/XInclude/docs/invalid_char.xml:2: element include: XInclude error : test/XInclude/ents/invalid_char.txt contains invalid char
+./test/XInclude/docs/invalid_char.xml:2: element include: XInclude error : could not load test/XInclude/ents/invalid_char.txt, and no fallback was found
diff --git a/result/XInclude/invalid_char.xml.rdr b/result/XInclude/invalid_char.xml.rdr
new file mode 100644
index 0000000..1fb5774
--- /dev/null
+++ b/result/XInclude/invalid_char.xml.rdr
@@ -0,0 +1,7 @@
+0 1 x 0 0
+1 14 #text 0 1 
+   
+1 1 xinclude:include 1 0
+1 14 #text 0 1 
+
+0 15 x 0 0
diff --git a/test/XInclude/docs/invalid_char.xml b/test/XInclude/docs/invalid_char.xml
new file mode 100644
index 0000000..28e5a48
--- /dev/null
+++ b/test/XInclude/docs/invalid_char.xml
@@ -0,0 +1,3 @@
+<x xmlns:xinclude="http://www.w3.org/2001/XInclude">
+   <xinclude:include href="../ents/invalid_char.txt" parse="text"/>
+</x>
diff --git a/test/XInclude/ents/invalid_char.txt b/test/XInclude/ents/invalid_char.txt
new file mode 100644
index 0000000..ae06618
--- /dev/null
+++ b/test/XInclude/ents/invalid_char.txt
@@ -0,0 +1 @@
+invalid: ÿ
\ No newline at end of file
diff --git a/xinclude.c b/xinclude.c
index cc486f5..6e5b61d 100644
--- a/xinclude.c
+++ b/xinclude.c
@@ -1798,7 +1798,9 @@ xmlXIncludeLoadTxt(xmlXIncludeCtxtPtr ctxt, const xmlChar *url, int nr) {
     xmlCharEncoding enc = (xmlCharEncoding) 0;
     xmlParserCtxtPtr pctxt;
     xmlParserInputPtr inputStream;
-    int xinclude_multibyte_fallback_used = 0;
+    int len;
+    const xmlChar *content;
+
 
     /* Don't read from stdin. */
     if (xmlStrcmp(url, BAD_CAST "-") == 0)
@@ -1905,41 +1907,30 @@ xmlXIncludeLoadTxt(xmlXIncludeCtxtPtr ctxt, const xmlChar *url, int nr) {
     /*
      * Scan all chars from the resource and add the to the node
      */
-xinclude_multibyte_fallback:
-    while (xmlParserInputBufferRead(buf, 128) > 0) {
-	int len;
-	const xmlChar *content;
-
-	content = xmlBufContent(buf->buffer);
-	len = xmlBufLength(buf->buffer);
-	for (i = 0;i < len;) {
-	    int cur;
-	    int l;
-
-	    cur = xmlStringCurrentChar(NULL, &content[i], &l);
-	    if (!IS_CHAR(cur)) {
-		/* Handle split multibyte char at buffer boundary */
-		if (((len - i) < 4) && (!xinclude_multibyte_fallback_used)) {
-		    xinclude_multibyte_fallback_used = 1;
-		    xmlBufShrink(buf->buffer, i);
-		    goto xinclude_multibyte_fallback;
-		} else {
-		    xmlXIncludeErr(ctxt, ctxt->incTab[nr]->ref,
-				   XML_XINCLUDE_INVALID_CHAR,
-				   "%s contains invalid char\n", URL);
-		    xmlFreeParserCtxt(pctxt);
-		    xmlFreeParserInputBuffer(buf);
-		    xmlFree(URL);
-		    return(-1);
-		}
-	    } else {
-		xinclude_multibyte_fallback_used = 0;
-		xmlNodeAddContentLen(node, &content[i], l);
-	    }
-	    i += l;
-	}
-	xmlBufShrink(buf->buffer, len);
+    while (xmlParserInputBufferRead(buf, 4096) > 0)
+        ;
+
+    content = xmlBufContent(buf->buffer);
+    len = xmlBufLength(buf->buffer);
+    for (i = 0; i < len;) {
+        int cur;
+        int l;
+
+        cur = xmlStringCurrentChar(NULL, &content[i], &l);
+        if (!IS_CHAR(cur)) {
+            xmlXIncludeErr(ctxt, ctxt->incTab[nr]->ref, XML_XINCLUDE_INVALID_CHAR,
+                           "%s contains invalid char\n", URL);
+            xmlFreeNode(node);
+            xmlFreeInputStream(inputStream);
+            xmlFreeParserCtxt(pctxt);
+            xmlFree(URL);
+            return(-1);
+        }
+
+        i += l;
     }
+
+    xmlNodeAddContentLen(node, content, len);
     xmlFreeParserCtxt(pctxt);
     xmlXIncludeAddTxt(ctxt, node->content, URL);
     xmlFreeInputStream(inputStream);
-- 
2.27.0