From bd9eed4694fe56e8b26405cefb3d22637b21a4d4 Mon Sep 17 00:00:00 2001
From: Nick Wellnhofer <wellnhofer@aevum.de>
Date: Mon, 2 Sep 2024 18:37:41 +0200
Subject: [PATCH] parser: Make unsupported encodings an error in declarations

This was changed in 45157261, but in encoding declarations, unsupported
encodings should raise a fatal error.

Fixes #794.
---
 parserInternals.c                          | 40 ++++++++++++++++------
 result/errors/unsupported-encoding.xml.ent |  3 ++
 result/errors/unsupported-encoding.xml.err |  3 ++
 result/errors/unsupported-encoding.xml.str |  4 +++
 test/errors/unsupported-encoding.xml       |  2 ++
 5 files changed, 42 insertions(+), 10 deletions(-)
 create mode 100644 result/errors/unsupported-encoding.xml.ent
 create mode 100644 result/errors/unsupported-encoding.xml.err
 create mode 100644 result/errors/unsupported-encoding.xml.str
 create mode 100644 test/errors/unsupported-encoding.xml

diff --git a/parserInternals.c b/parserInternals.c
index eb13be7e..a6ddf502 100644
--- a/parserInternals.c
+++ b/parserInternals.c
@@ -475,20 +475,14 @@ void
 xmlFatalErr(xmlParserCtxtPtr ctxt, xmlParserErrors code, const char *info)
 {
     const char *errmsg;
-    xmlErrorLevel level;
-
-    if (code == XML_ERR_UNSUPPORTED_ENCODING)
-        level = XML_ERR_WARNING;
-    else
-        level = XML_ERR_FATAL;
 
     errmsg = xmlErrString(code);
 
     if (info == NULL) {
-        xmlCtxtErr(ctxt, NULL, XML_FROM_PARSER, code, level,
+        xmlCtxtErr(ctxt, NULL, XML_FROM_PARSER, code, XML_ERR_FATAL,
                    NULL, NULL, NULL, 0, "%s\n", errmsg);
     } else {
-        xmlCtxtErr(ctxt, NULL, XML_FROM_PARSER, code, level,
+        xmlCtxtErr(ctxt, NULL, XML_FROM_PARSER, code, XML_ERR_FATAL,
                    (const xmlChar *) info, NULL, NULL, 0,
                    "%s: %s\n", errmsg, info);
     }
@@ -1223,7 +1217,11 @@ xmlSwitchInputEncodingName(xmlParserCtxtPtr ctxt, xmlParserInputPtr input,
 
     res = xmlCreateCharEncodingHandler(encoding, /* output */ 0,
             ctxt->convImpl, ctxt->convCtxt, &handler);
-    if (res != XML_ERR_OK) {
+    if (res == XML_ERR_UNSUPPORTED_ENCODING) {
+        xmlWarningMsg(ctxt, XML_ERR_UNSUPPORTED_ENCODING,
+                      "Unsupported encoding: %s\n", BAD_CAST encoding, NULL);
+        return(-1);
+    } else if (res != XML_ERR_OK) {
         xmlFatalErr(ctxt, res, encoding);
         return(-1);
     }
@@ -1535,7 +1533,29 @@ void
 xmlSetDeclaredEncoding(xmlParserCtxtPtr ctxt, xmlChar *encoding) {
     if (((ctxt->input->flags & XML_INPUT_HAS_ENCODING) == 0) &&
         ((ctxt->options & XML_PARSE_IGNORE_ENC) == 0)) {
-        xmlSwitchEncodingName(ctxt, (const char *) encoding);
+        xmlCharEncodingHandlerPtr handler;
+        int res;
+
+        /*
+         * xmlSwitchEncodingName treats unsupported encodings as
+         * warnings, but we want it to be an error in an encoding
+         * declaration.
+         */
+        res = xmlCreateCharEncodingHandler((const char *) encoding,
+                /* output */ 0, ctxt->convImpl, ctxt->convCtxt, &handler);
+        if (res != XML_ERR_OK) {
+            xmlFatalErr(ctxt, res, (const char *) encoding);
+            xmlFree(encoding);
+            return;
+        }
+
+        res  = xmlInputSetEncodingHandler(ctxt->input, handler);
+        if (res != XML_ERR_OK) {
+            xmlCtxtErrIO(ctxt, res, NULL);
+            xmlFree(encoding);
+            return;
+        }
+
         ctxt->input->flags |= XML_INPUT_USES_ENC_DECL;
     } else if (ctxt->input->flags & XML_INPUT_AUTO_ENCODING) {
         static const char *allowedUTF8[] = {
diff --git a/result/errors/unsupported-encoding.xml.ent b/result/errors/unsupported-encoding.xml.ent
new file mode 100644
index 00000000..a275ec19
--- /dev/null
+++ b/result/errors/unsupported-encoding.xml.ent
@@ -0,0 +1,3 @@
+./test/errors/unsupported-encoding.xml:1: parser error : Unsupported encoding: unsupported-encoding
+<?xml version="1.0" encoding="unsupported-encoding"?>
+                                                   ^
diff --git a/result/errors/unsupported-encoding.xml.err b/result/errors/unsupported-encoding.xml.err
new file mode 100644
index 00000000..a275ec19
--- /dev/null
+++ b/result/errors/unsupported-encoding.xml.err
@@ -0,0 +1,3 @@
+./test/errors/unsupported-encoding.xml:1: parser error : Unsupported encoding: unsupported-encoding
+<?xml version="1.0" encoding="unsupported-encoding"?>
+                                                   ^
diff --git a/result/errors/unsupported-encoding.xml.str b/result/errors/unsupported-encoding.xml.str
new file mode 100644
index 00000000..1f06d608
--- /dev/null
+++ b/result/errors/unsupported-encoding.xml.str
@@ -0,0 +1,4 @@
+./test/errors/unsupported-encoding.xml:1: parser error : Unsupported encoding: unsupported-encoding
+<?xml version="1.0" encoding="unsupported-encoding"?>
+                                                   ^
+./test/errors/unsupported-encoding.xml : failed to parse
diff --git a/test/errors/unsupported-encoding.xml b/test/errors/unsupported-encoding.xml
new file mode 100644
index 00000000..8aba00e4
--- /dev/null
+++ b/test/errors/unsupported-encoding.xml
@@ -0,0 +1,2 @@
+<?xml version="1.0" encoding="unsupported-encoding"?>
+<doc/>