Refactor various hex escaping code.

src/Desc.cc

@@ -216,18 +216,32 @@ void ODesc::Indent()
 		}
 	}
 
-static const char hex_chars[] = "0123456789abcdef";
-
-static const char* find_first_unprintable(ODesc* d, const char* bytes, unsigned int n)
+static bool starts_with(const char* str1, const char* str2, size_t len)
 	{
-	if ( d->IsBinary() )
+	for ( size_t i = 0; i < len; ++i )
+		if ( str1[i] != str2[i] )
+			return false;
+
+	return true;
+	}
+
+size_t ODesc::StartsWithEscapeSequence(const char* start, const char* end)
+	{
+	if ( escape_sequences.empty() )
 		return 0;
 
-	while ( n-- )
+	escape_set::const_iterator it;
+
+	for ( it = escape_sequences.begin(); it != escape_sequences.end(); ++it )
 		{
-		if ( ! isprint(*bytes) )
-			return bytes;
-		++bytes;
+		const string& esc_str = *it;
+		size_t esc_len = esc_str.length();
+
+		if ( start + esc_len > end )
+			continue;
+
+		if ( starts_with(start, esc_str.c_str(), esc_len) )
+			return esc_len;
 		}
 
 	return 0;
@@ -235,21 +249,23 @@ static const char* find_first_unprintable(ODesc* d, const char* bytes, unsigned
 
 pair<const char*, size_t> ODesc::FirstEscapeLoc(const char* bytes, size_t n)
 	{
-	pair<const char*, size_t> p(find_first_unprintable(this, bytes, n), 1);
+	typedef pair<const char*, size_t> escape_pos;
 
-	string str(bytes, n);
-	list<string>::const_iterator it;
-	for ( it = escape_sequences.begin(); it != escape_sequences.end(); ++it )
+	if ( IsBinary() )
+		return escape_pos(0, 0);
+
+	for ( size_t i = 0; i < n; ++i )
 		{
-		size_t pos = str.find(*it);
-		if ( pos != string::npos && (p.first == 0 || bytes + pos < p.first) )
-			{
-			p.first = bytes + pos;
-			p.second = it->size();
-			}
+		if ( ! isprint(bytes[i]) )
+			return escape_pos(bytes + i, 1);
+
+		size_t len = StartsWithEscapeSequence(bytes + i, bytes + n);
+
+		if ( len )
+			return escape_pos(bytes + i, len);
 		}
 
-	return p;
+	return escape_pos(0, 0);
 	}
 
 void ODesc::AddBytes(const void* bytes, unsigned int n)
@@ -266,21 +282,11 @@ void ODesc::AddBytes(const void* bytes, unsigned int n)
 	while ( s < e )
 		{
 		pair<const char*, size_t> p = FirstEscapeLoc(s, e - s);
+
 		if ( p.first )
 			{
 			AddBytesRaw(s, p.first - s);
-			if ( p.second == 1 )
-				{
-				char hex[6] = "\\x00";
-				hex[2] = hex_chars[((*p.first) & 0xf0) >> 4];
-				hex[3] = hex_chars[(*p.first) & 0x0f];
-				AddBytesRaw(hex, 4);
-				}
-			else
-				{
-				string esc_str = get_escaped_string(string(p.first, p.second), true);
-				AddBytesRaw(esc_str.c_str(), esc_str.size());
-				}
+			get_escaped_string(this, p.first, p.second, true);
 			s = p.first + p.second;
 			}
 		else

src/Desc.h

@@ -4,7 +4,7 @@
 #define descriptor_h
 
 #include <stdio.h>
-#include <list>
+#include <set>
 #include <utility>
 
 #include "BroString.h"
@@ -54,16 +54,16 @@ public:
 	void SetFlush(int arg_do_flush)	{ do_flush = arg_do_flush; }
 
 	void EnableEscaping();
-	void AddEscapeSequence(const char* s) { escape_sequences.push_back(s); }
+	void AddEscapeSequence(const char* s) { escape_sequences.insert(s); }
 	void AddEscapeSequence(const char* s, size_t n)
-	    { escape_sequences.push_back(string(s, n)); }
+	    { escape_sequences.insert(string(s, n)); }
 	void AddEscapeSequence(const string & s)
-	    { escape_sequences.push_back(s); }
-	void RemoveEscapeSequence(const char* s) { escape_sequences.remove(s); }
+	    { escape_sequences.insert(s); }
+	void RemoveEscapeSequence(const char* s) { escape_sequences.erase(s); }
 	void RemoveEscapeSequence(const char* s, size_t n)
-	    { escape_sequences.remove(string(s, n)); }
+	    { escape_sequences.erase(string(s, n)); }
 	void RemoveEscapeSequence(const string & s)
-	    { escape_sequences.remove(s); }
+	    { escape_sequences.erase(s); }
 
 	void PushIndent();
 	void PopIndent();
@@ -163,6 +163,15 @@ protected:
 	 */
 	pair<const char*, size_t> FirstEscapeLoc(const char* bytes, size_t n);
 
+	/**
+	 * @param start start of string to check for starting with an espace
+	 *              sequence.
+	 * @param end one byte past the last character in the string.
+	 * @return The number of bytes in the escape sequence that the string
+	 *         starts with.
+	 */
+	size_t StartsWithEscapeSequence(const char* start, const char* end);
+
 	desc_type type;
 	desc_style style;
 
@@ -171,7 +180,8 @@ protected:
 	unsigned int size;	// size of buffer in bytes
 
 	bool escape;	// escape unprintable characters in output?
-	list<string> escape_sequences; // additional sequences of chars to escape
+	typedef set<string> escape_set;
+	escape_set escape_sequences; // additional sequences of chars to escape
 
 	BroFile* f;	// or the file we're using.
 

src/logging/writers/Ascii.cc

@@ -335,10 +335,10 @@ bool Ascii::DoWrite(int num_fields, const Field* const * fields,
 	if ( strncmp(bytes, meta_prefix.data(), meta_prefix.size()) == 0 )
 		{
 		// It would so escape the first character.
-		char buf[16];
-		snprintf(buf, sizeof(buf), "\\x%02x", bytes[0]);
+		char hex[4] = {'\\', 'x', '0', '0'};
+		bytetohex(bytes[0], hex + 2);
 
-		if ( ! safe_write(fd, buf, strlen(buf)) )
+		if ( ! safe_write(fd, hex, 4) )
 			goto write_error;
 
 		++bytes;

src/threading/formatters/Ascii.cc

@@ -122,10 +122,8 @@ bool Ascii::Describe(ODesc* desc, threading::Value* val, const string& name) con
 			// place-holder we use for unset optional fields. We
 			// escape the first character so that the output
 			// won't be ambigious.
-			static const char hex_chars[] = "0123456789abcdef";
-			char hex[6] = "\\x00";
-			hex[2] = hex_chars[((*data) & 0xf0) >> 4];
-			hex[3] = hex_chars[(*data) & 0x0f];
+			char hex[4] = {'\\', 'x', '0', '0'};
+			bytetohex(*data, hex + 2);
 			desc->AddRaw(hex, 4);
 
 			++data;

src/threading/formatters/JSON.cc

@@ -160,10 +160,11 @@ bool JSON::Describe(ODesc* desc, Value* val, const string& name) const
 				// 2byte Unicode escape special characters.
 				if ( c < 32 || c > 126 || c == '\n' || c == '"' || c == '\'' || c == '\\' || c == '&' )
 					{
-					static const char hex_chars[] = "0123456789abcdef";
 					desc->AddRaw("\\u00", 4);
-					desc->AddRaw(&hex_chars[(c & 0xf0) >> 4], 1);
-					desc->AddRaw(&hex_chars[c & 0x0f], 1);
+					char hex[2] = {'0', '0'};
+					bytetohex(c, hex);
+					desc->AddRaw(hex, 1);
+					desc->AddRaw(hex + 1, 1);
 					}
 				else
 					desc->AddRaw(&c, 1);

src/util.cc

@@ -120,31 +120,41 @@ std::string get_unescaped_string(const std::string& arg_str)
  * Takes a string, escapes characters into equivalent hex codes (\x##), and
  * returns a string containing all escaped values.
  *
+ * @param d an ODesc object to store the escaped hex version of the string,
+ *          if null one will be allocated and returned from the function.
  * @param str string to escape
  * @param escape_all If true, all characters are escaped. If false, only
  * characters are escaped that are either whitespace or not printable in
  * ASCII.
- * @return A std::string containing a list of escaped hex values of the form
- * \x## */
-std::string get_escaped_string(const std::string& str, bool escape_all)
+ * @return A ODesc object containing a list of escaped hex values of the form
+ *         \x##, which may be newly allocated if \a d was a null pointer. */
+ODesc* get_escaped_string(ODesc* d, const char* str, size_t len,
+                          bool escape_all)
 	{
-	char tbuf[16];
-	string esc = "";
+	if ( ! d )
+		d = new ODesc();
 
-	for ( size_t i = 0; i < str.length(); ++i )
+	for ( size_t i = 0; i < len; ++i )
 		{
 		char c = str[i];
 
 		if ( escape_all || isspace(c) || ! isascii(c) || ! isprint(c) )
 			{
-			snprintf(tbuf, sizeof(tbuf), "\\x%02x", str[i]);
-			esc += tbuf;
+			char hex[4] = {'\\', 'x', '0', '0' };
+			bytetohex(c, hex + 2);
+			d->AddRaw(hex, 4);
 			}
 		else
-			esc += c;
+			d->AddRaw(&c, 1);
 		}
 
-	return esc;
+	return d;
+	}
+
+std::string get_escaped_string(const char* str, size_t len, bool escape_all)
+	{
+	ODesc d;
+	return get_escaped_string(&d, str, len, escape_all)->Description();
 	}
 
 char* copy_string(const char* s)

src/util.h

@@ -102,8 +102,25 @@ void delete_each(T* t)
 std::string extract_ip(const std::string& i);
 std::string extract_ip_and_len(const std::string& i, int* len);
 
+inline void bytetohex(unsigned char byte, char* hex_out)
+	{
+	static const char hex_chars[] = "0123456789abcdef";
+	hex_out[0] = hex_chars[(byte & 0xf0) >> 4];
+	hex_out[1] = hex_chars[byte & 0x0f];
+	}
+
 std::string get_unescaped_string(const std::string& str);
-std::string get_escaped_string(const std::string& str, bool escape_all);
+
+class ODesc;
+
+ODesc* get_escaped_string(ODesc* d, const char* str, size_t len,
+                          bool escape_all);
+std::string get_escaped_string(const char* str, size_t len, bool escape_all);
+
+inline std::string get_escaped_string(const std::string& str, bool escape_all)
+	{
+	return get_escaped_string(str.data(), str.length(), escape_all);
+	}
 
 std::vector<std::string>* tokenize_string(std::string input,
 					  const std::string& delim,