From 624b5702d75d63dde56fbbc89680358241035aa7 Mon Sep 17 00:00:00 2001
From: "arseny.kapoulkine"
 <arseny.kapoulkine@99668b35-9821-0410-8761-19e4c4f06640>
Date: Fri, 11 Jun 2010 20:39:57 +0000
Subject: Rewritten numeric character reference parsing (fixed &#; and &#x;
 parsing), added more character reference tests

git-svn-id: http://pugixml.googlecode.com/svn/trunk@512 99668b35-9821-0410-8761-19e4c4f06640
---
 src/pugixml.cpp      | 51 +++++++++++++++++++++++++++++++--------------------
 tests/test_parse.cpp | 48 ++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 79 insertions(+), 20 deletions(-)

diff --git a/src/pugixml.cpp b/src/pugixml.cpp
index 242202e..68a68d0 100644
--- a/src/pugixml.cpp
+++ b/src/pugixml.cpp
@@ -1374,37 +1374,48 @@ namespace
 			{
 				unsigned int ucsc = 0;
 
-				++stre;
-
-				if (*stre == 'x') // &#x... (hex code)
+				if (stre[1] == 'x') // &#x... (hex code)
 				{
-					++stre;
-					
-					while (*stre)
+					stre += 2;
+
+					char_t ch = *stre;
+
+					if (ch == ';') return stre;
+
+					for (;;)
 					{
-						if (*stre >= '0' && *stre <= '9')
-							ucsc = 16 * ucsc + (*stre++ - '0');
-						else if (*stre >= 'A' && *stre <= 'F')
-							ucsc = 16 * ucsc + (*stre++ - 'A' + 10);
-						else if (*stre >= 'a' && *stre <= 'f')
-							ucsc = 16 * ucsc + (*stre++ - 'a' + 10);
-						else if (*stre == ';')
+						if (static_cast<unsigned int>(ch - '0') <= 9)
+							ucsc = 16 * ucsc + (ch - '0');
+						else if (static_cast<unsigned int>((ch | ' ') - 'a') <= 5)
+							ucsc = 16 * ucsc + ((ch | ' ') - 'a' + 10);
+						else if (ch == ';')
 							break;
 						else // cancel
 							return stre;
-					}
 
-					if (*stre != ';') return stre;
-						
+						ch = *++stre;
+					}
+					
 					++stre;
 				}
 				else	// &#... (dec code)
 				{
-					while (*stre >= '0' && *stre <= '9')
-						ucsc = 10 * ucsc + (*stre++ - '0');
+					char_t ch = *++stre;
 
-					if (*stre != ';') return stre;
-						
+					if (ch == ';') return stre;
+
+					for (;;)
+					{
+						if (static_cast<unsigned int>(ch - '0') <= 9)
+							ucsc = 10 * ucsc + (ch - '0');
+						else if (ch == ';')
+							break;
+						else // cancel
+							return stre;
+
+						ch = *++stre;
+					}
+					
 					++stre;
 				}
 
diff --git a/tests/test_parse.cpp b/tests/test_parse.cpp
index c2f56e5..50f8867 100644
--- a/tests/test_parse.cpp
+++ b/tests/test_parse.cpp
@@ -281,6 +281,47 @@ TEST(parse_escapes_code)
 	CHECK_STRING(doc.child_value(STR("node")), STR("\01  "));
 }
 
+TEST(parse_escapes_code_exhaustive_dec)
+{
+	xml_document doc;
+	CHECK(doc.load(STR("<node>&#/;&#01;&#2;&#3;&#4;&#5;&#6;&#7;&#8;&#9;&#:;&#a;&#A;&#XA;</node>"), parse_minimal | parse_escapes));
+	CHECK_STRING(doc.child_value(STR("node")), STR("&#/;\x1\x2\x3\x4\x5\x6\x7\x8\x9&#:;&#a;&#A;&#XA;"));
+}
+
+TEST(parse_escapes_code_exhaustive_hex)
+{
+	xml_document doc;
+	CHECK(doc.load(STR("<node>&#x/;&#x01;&#x2;&#x3;&#x4;&#x5;&#x6;&#x7;&#x8;&#x9;&#x:;&#x@;&#xA;&#xB;&#xC;&#xD;&#xE;&#xF;&#xG;&#x`;&#xa;&#xb;&#xc;&#xd;&#xe;&#xf;&#xg;</node>"), parse_minimal | parse_escapes));
+	CHECK_STRING(doc.child_value(STR("node")), STR("&#x/;\x1\x2\x3\x4\x5\x6\x7\x8\x9&#x:;&#x@;\xa\xb\xc\xd\xe\xf&#xG;&#x`;\xa\xb\xc\xd\xe\xf&#xg;"));
+}
+
+TEST(parse_escapes_code_restore)
+{
+	xml_document doc;
+	CHECK(doc.load(STR("<node>&#1&#32;&#x1&#32;&#1-&#32;&#x1-&#32;</node>"), parse_minimal | parse_escapes));
+	CHECK_STRING(doc.child_value(STR("node")), STR("&#1 &#x1 &#1- &#x1- "));
+}
+
+TEST(parse_escapes_char_restore)
+{
+	xml_document doc;
+
+	CHECK(doc.load(STR("<node>&q&#32;&qu&#32;&quo&#32;&quot&#32;</node>"), parse_minimal | parse_escapes));
+	CHECK_STRING(doc.child_value(STR("node")), STR("&q &qu &quo &quot "));
+
+	CHECK(doc.load(STR("<node>&a&#32;&ap&#32;&apo&#32;&apos&#32;</node>"), parse_minimal | parse_escapes));
+	CHECK_STRING(doc.child_value(STR("node")), STR("&a &ap &apo &apos "));
+
+	CHECK(doc.load(STR("<node>&a&#32;&am&#32;&amp&#32;</node>"), parse_minimal | parse_escapes));
+	CHECK_STRING(doc.child_value(STR("node")), STR("&a &am &amp "));
+
+	CHECK(doc.load(STR("<node>&l&#32;&lt&#32;</node>"), parse_minimal | parse_escapes));
+	CHECK_STRING(doc.child_value(STR("node")), STR("&l &lt "));
+
+	CHECK(doc.load(STR("<node>&g&#32;&gt&#32;</node>"), parse_minimal | parse_escapes));
+	CHECK_STRING(doc.child_value(STR("node")), STR("&g &gt "));
+}
+
 TEST(parse_escapes_unicode)
 {
 	xml_document doc;
@@ -314,6 +355,13 @@ TEST(parse_escapes_error)
 	CHECK(!doc.load(STR("<node id='&apos")));
 }
 
+TEST(parse_escapes_code_invalid)
+{
+	xml_document doc;
+	CHECK(doc.load(STR("<node>&#;&#x;&;&#x-;&#-;</node>"), parse_minimal | parse_escapes));
+	CHECK_STRING(doc.child_value(STR("node")), STR("&#;&#x;&;&#x-;&#-;"));
+}
+
 TEST(parse_attribute_spaces)
 {
 	xml_document doc;
-- 
cgit v1.2.3