Bug 758605: Heap-based buffer overread in xmlDictAddString <https://bugzilla.gnome.org/show_bug.cgi?id=758605>
Reviewed by David Kilzer.
* HTMLparser.c:
(htmlParseName): Add bounds check.
(htmlParseNameComplex): Ditto.
* result/HTML/758605.html: Added.
* result/HTML/758605.html.err: Added.
* result/HTML/758605.html.sax: Added.
* runtest.c:
(pushParseTest): The input for the new test case was so small
(4 bytes) that htmlParseChunk() was never called after
htmlCreatePushParserCtxt(), thereby creating a false positive
test failure. Fixed by using a do-while loop so we always call
htmlParseChunk() at least once.
* test/HTML/758605.html: Added.
diff --git a/HTMLparser.c b/HTMLparser.c
index 69eed2b..1c112cc 100644
--- a/HTMLparser.c
+++ b/HTMLparser.c
@@ -2471,6 +2471,10 @@
(*in == '_') || (*in == '-') ||
(*in == ':') || (*in == '.'))
in++;
+
+ if (in == ctxt->input->end)
+ return(NULL);
+
if ((*in > 0) && (*in < 0x80)) {
count = in - ctxt->input->cur;
ret = xmlDictLookup(ctxt->dict, ctxt->input->cur, count);
@@ -2514,6 +2518,10 @@
NEXTL(l);
c = CUR_CHAR(l);
}
+
+ if (ctxt->input->base > ctxt->input->cur - len)
+ return(NULL);
+
return(xmlDictLookup(ctxt->dict, ctxt->input->cur - len, len));
}
diff --git a/result/HTML/758605.html b/result/HTML/758605.html
new file mode 100644
index 0000000..a085cce
--- /dev/null
+++ b/result/HTML/758605.html
@@ -0,0 +1,3 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
+<html><body><p>&
+</p></body></html>
diff --git a/result/HTML/758605.html.err b/result/HTML/758605.html.err
new file mode 100644
index 0000000..2b82be6
--- /dev/null
+++ b/result/HTML/758605.html.err
@@ -0,0 +1,3 @@
+./test/HTML/758605.html:1: HTML parser error : htmlParseEntityRef: no name
+ê
+ ^
diff --git a/result/HTML/758605.html.sax b/result/HTML/758605.html.sax
new file mode 100644
index 0000000..1f5cd32
--- /dev/null
+++ b/result/HTML/758605.html.sax
@@ -0,0 +1,13 @@
+SAX.setDocumentLocator()
+SAX.startDocument()
+SAX.error: htmlParseEntityRef: no name
+SAX.startElement(html)
+SAX.startElement(body)
+SAX.startElement(p)
+SAX.characters(&, 1)
+SAX.ignorableWhitespace(
+, 1)
+SAX.endElement(p)
+SAX.endElement(body)
+SAX.endElement(html)
+SAX.endDocument()
diff --git a/runtest.c b/runtest.c
index 36fbe5a..bb74d2a 100644
--- a/runtest.c
+++ b/runtest.c
@@ -1873,7 +1873,7 @@
ctxt = xmlCreatePushParserCtxt(NULL, NULL, base + cur, 4, filename);
xmlCtxtUseOptions(ctxt, options);
cur += 4;
- while (cur < size) {
+ do {
if (cur + 1024 >= size) {
#ifdef LIBXML_HTML_ENABLED
if (options & XML_PARSE_HTML)
@@ -1891,7 +1891,7 @@
xmlParseChunk(ctxt, base + cur, 1024, 0);
cur += 1024;
}
- }
+ } while (cur < size);
doc = ctxt->myDoc;
#ifdef LIBXML_HTML_ENABLED
if (options & XML_PARSE_HTML)
diff --git a/test/HTML/758605.html b/test/HTML/758605.html
new file mode 100644
index 0000000..9b1b3c2
--- /dev/null
+++ b/test/HTML/758605.html
@@ -0,0 +1 @@
+&:ê
