Switched to reading the entire file into a buffer at the start.\nThis…

… speeds it up a TON (like 100x).
larics · Feb 1, 2009 · f9c0725 · f9c0725
1 parent a6d5902
commit f9c0725
Show file tree

Hide file tree

Showing 9 changed files with 72 additions and 75 deletions.
diff --git a/src/exp.h b/src/exp.h
@@ -3,7 +3,6 @@
 #include "regex.h"
 #include <string>
 #include <ios>
-#include <cstdio>
 #include "stream.h"
 
 namespace YAML
@@ -24,10 +23,10 @@ namespace YAML
 
 		// actual tags
 
-		const RegEx DocStart = RegEx("---") + (BlankOrBreak || RegEx(EOF) || RegEx());
-		const RegEx DocEnd = RegEx("...") + (BlankOrBreak || RegEx(EOF) || RegEx());
+		const RegEx DocStart = RegEx("---") + (BlankOrBreak || RegEx());
+		const RegEx DocEnd = RegEx("...") + (BlankOrBreak || RegEx());
 		const RegEx DocIndicator = DocStart || DocEnd;
-		const RegEx BlockEntry = RegEx('-') + (BlankOrBreak || RegEx(EOF));
+		const RegEx BlockEntry = RegEx('-') + BlankOrBreak;
 		const RegEx Key = RegEx('?'),
 		            KeyInFlow = RegEx('?') + BlankOrBreak;
 		const RegEx Value = RegEx(':') + BlankOrBreak,

diff --git a/src/regex.cpp b/src/regex.cpp
@@ -90,12 +90,12 @@ namespace YAML
 		return Match(str) >= 0;
 	}
 
-	bool RegEx::Matches(std::istream& in) const
+	bool RegEx::Matches(const char *buffer) const
 	{
-		return Match(in) >= 0;
+		return Match(buffer) >= 0;
 	}
 
-	bool RegEx::Matches(Stream& in) const
+	bool RegEx::Matches(const Stream& in) const
 	{
 		return Match(in) >= 0;
 	}
@@ -115,9 +115,9 @@ namespace YAML
 	}
 
 	// Match
-	int RegEx::Match(Stream& in) const
+	int RegEx::Match(const Stream& in) const
 	{
-		return Match(in.stream());
+		return Match(in.current());
 	}
 
 	// Match
@@ -126,19 +126,12 @@ namespace YAML
 	// . Note: the istream is not a const reference, but we guarantee
 	//   that the pointer will be in the same spot, and we'll clear its
 	//   flags before we end.
-	int RegEx::Match(std::istream& in) const
+	int RegEx::Match(const char *buffer) const
 	{
 		if(!m_pOp)
 			return -1;
 
-		int pos = in.tellg();
-		int ret = m_pOp->Match(in, *this);
-
-		// reset input stream!
-		in.clear();
-		in.seekg(pos);
-
-		return ret;
+		return m_pOp->Match(buffer, *this);
 	}
 
 	RegEx operator ! (const RegEx& ex)
@@ -184,9 +177,9 @@ namespace YAML
 	}
 
 
-	int RegEx::MatchOperator::Match(std::istream& in, const RegEx& regex) const
+	int RegEx::MatchOperator::Match(const char *buffer, const RegEx& regex) const
 	{
-		if(!in || in.peek() != regex.m_a)
+		if(buffer[0] != regex.m_a)
 			return -1;
 		return 1;
 	}
@@ -199,9 +192,9 @@ namespace YAML
 		return 1;
 	}
 
-	int RegEx::RangeOperator::Match(std::istream& in, const RegEx& regex) const
+	int RegEx::RangeOperator::Match(const char *buffer, const RegEx& regex) const
 	{
-		if(!in || regex.m_a > in.peek() || regex.m_z < in.peek())
+		if(regex.m_a > buffer[0] || regex.m_z < buffer[0])
 			return -1;
 		return 1;
 	}
@@ -217,10 +210,10 @@ namespace YAML
 		return -1;
 	}
 
-	int RegEx::OrOperator::Match(std::istream& in, const RegEx& regex) const
+	int RegEx::OrOperator::Match(const char *buffer, const RegEx& regex) const
 	{
 		for(unsigned i=0;i<regex.m_params.size();i++) {
-			int n = regex.m_params[i].Match(in);
+			int n = regex.m_params[i].Match(buffer);
 			if(n >= 0)
 				return n;
 		}
@@ -244,11 +237,11 @@ namespace YAML
 		return first;
 	}
 
-	int RegEx::AndOperator::Match(std::istream& in, const RegEx& regex) const
+	int RegEx::AndOperator::Match(const char *buffer, const RegEx& regex) const
 	{
 		int first = -1;
 		for(unsigned i=0;i<regex.m_params.size();i++) {
-			int n = regex.m_params[i].Match(in);
+			int n = regex.m_params[i].Match(buffer);
 			if(n == -1)
 				return -1;
 			if(i == 0)
@@ -267,11 +260,11 @@ namespace YAML
 		return 1;
 	}
 
-	int RegEx::NotOperator::Match(std::istream& in, const RegEx& regex) const
+	int RegEx::NotOperator::Match(const char *buffer, const RegEx& regex) const
 	{
 		if(regex.m_params.empty())
 			return -1;
-		if(regex.m_params[0].Match(in) >= 0)
+		if(regex.m_params[0].Match(buffer) >= 0)
 			return -1;
 		return 1;
 	}
@@ -289,16 +282,15 @@ namespace YAML
 		return offset;
 	}
 
-	int RegEx::SeqOperator::Match(std::istream& in, const RegEx& regex) const
+	int RegEx::SeqOperator::Match(const char *buffer, const RegEx& regex) const
 	{
 		int offset = 0;
 		for(unsigned i=0;i<regex.m_params.size();i++) {
-			int n = regex.m_params[i].Match(in);
+			int n = regex.m_params[i].Match(buffer + offset);
 			if(n == -1)
 				return -1;
 
 			offset += n;
-			in.ignore(n);
 		}
 
 		return offset;

diff --git a/src/regex.h b/src/regex.h
@@ -2,7 +2,6 @@
 
 #include <vector>
 #include <string>
-#include <ios>
 
 namespace YAML
 {
@@ -19,37 +18,37 @@ namespace YAML
 		struct Operator {
 			virtual ~Operator() {}
 			virtual int Match(const std::string& str, const RegEx& regex) const = 0;
-			virtual int Match(std::istream& in, const RegEx& regex) const = 0;
+			virtual int Match(const char *buffer, const RegEx& regex) const = 0;
 		};
 
 		struct MatchOperator: public Operator {
 			virtual int Match(const std::string& str, const RegEx& regex) const;
-			virtual int Match(std::istream& in, const RegEx& regex) const;
+			virtual int Match(const char *buffer, const RegEx& regex) const;
 		};
 
 		struct RangeOperator: public Operator {
 			virtual int Match(const std::string& str, const RegEx& regex) const;
-			virtual int Match(std::istream& in, const RegEx& regex) const;
+			virtual int Match(const char *buffer, const RegEx& regex) const;
 		};
 
 		struct OrOperator: public Operator {
 			virtual int Match(const std::string& str, const RegEx& regex) const;
-			virtual int Match(std::istream& in, const RegEx& regex) const;
+			virtual int Match(const char *buffer, const RegEx& regex) const;
 		};
 
 		struct AndOperator: public Operator {
 			virtual int Match(const std::string& str, const RegEx& regex) const;
-			virtual int Match(std::istream& in, const RegEx& regex) const;
+			virtual int Match(const char *buffer, const RegEx& regex) const;
 		};
 
 		struct NotOperator: public Operator {
 			virtual int Match(const std::string& str, const RegEx& regex) const;
-			virtual int Match(std::istream& in, const RegEx& regex) const;
+			virtual int Match(const char *buffer, const RegEx& regex) const;
 		};
 
 		struct SeqOperator: public Operator {
 			virtual int Match(const std::string& str, const RegEx& regex) const;
-			virtual int Match(std::istream& in, const RegEx& regex) const;
+			virtual int Match(const char *buffer, const RegEx& regex) const;
 		};
 
 	public:
@@ -66,11 +65,11 @@ namespace YAML
 
 		bool Matches(char ch) const;
 		bool Matches(const std::string& str) const;
-		bool Matches(std::istream& in) const;
-		bool Matches(Stream& in) const;
+		bool Matches(const char *buffer) const;
+		bool Matches(const Stream& in) const;
 		int Match(const std::string& str) const;
-		int Match(std::istream& in) const;
-		int Match(Stream& in) const;
+		int Match(const char *buffer) const;
+		int Match(const Stream& in) const;
 
 		friend RegEx operator ! (const RegEx& ex);
 		friend RegEx operator || (const RegEx& ex1, const RegEx& ex2);

diff --git a/src/scanner.cpp b/src/scanner.cpp
@@ -100,7 +100,7 @@ namespace YAML
 		// *****
 
 		// end of stream
-		if(INPUT.peek() == EOF)
+		if(!INPUT)
 			return EndStream();
 
 		if(INPUT.column == 0 && INPUT.peek() == Keys::Directive)

diff --git a/src/simplekey.cpp b/src/simplekey.cpp
@@ -32,7 +32,7 @@ namespace YAML
 	//   and saves it on a stack.
 	void Scanner::InsertSimpleKey()
 	{
-		SimpleKey key(INPUT.pos(), INPUT.line, INPUT.column, m_flowLevel);
+		SimpleKey key(INPUT.pos, INPUT.line, INPUT.column, m_flowLevel);
 
 		// first add a map start, if necessary
 		key.pMapStart = PushIndentTo(INPUT.column, false);
@@ -78,7 +78,7 @@ namespace YAML
 			isValid = false;
 
 		// also needs to be less than 1024 characters and inline
-		if(INPUT.line != key.line || INPUT.pos() - key.pos > 1024)
+		if(INPUT.line != key.line || INPUT.pos - key.pos > 1024)
 			isValid = false;
 
 		// invalidate key

diff --git a/src/stream.cpp b/src/stream.cpp
@@ -4,26 +4,39 @@
 
 namespace YAML
 {
-	int Stream::pos() const
+	Stream::Stream(std::istream& input): buffer(0), pos(0), line(0), column(0), size(0)
 	{
-		return input.tellg();
+		std::streambuf *pBuf = input.rdbuf();
+
+		// store entire file in buffer
+		size = pBuf->pubseekoff(0, std::ios::end, std::ios::in);
+		pBuf->pubseekpos(0, std::ios::in);
+		buffer = new char[size];
+		pBuf->sgetn(buffer, size);
 	}
-
+
+	Stream::~Stream()
+	{
+		delete [] buffer;
+	}
+
+
 	char Stream::peek()
 	{
-		return input.peek();
+		return buffer[pos];
 	}
 
-	Stream::operator bool()
+	Stream::operator bool() const
 	{
-		return input.good();
+		return pos < size;
 	}
 
 	// get
 	// . Extracts a character from the stream and updates our position
 	char Stream::get()
 	{
-		char ch = input.get();
+		char ch = buffer[pos];
+		pos++;
 		column++;
 		if(ch == '\n') {
 			column = 0;

diff --git a/src/stream.h b/src/stream.h
@@ -5,21 +5,24 @@
 
 namespace YAML
 {
-	struct Stream
+	class Stream
 	{
-		Stream(std::istream& input_): input(input_), line(0), column(0) {}
+	public:
+		Stream(std::istream& input);
+		~Stream();
 
-		int pos() const;
-		operator bool();
-		bool operator !() { return !(*this); }
+		operator bool() const;
+		bool operator !() const { return !static_cast <bool>(*this); }
 
-		std::istream& stream() const { return input; }
+		const char *current() const { return buffer + pos; }
 		char peek();
 		char get();
 		std::string get(int n);
 		void eat(int n = 1);
 
-		std::istream& input;
-		int line, column;
+		int pos, line, column, size;
+
+	private:
+		char *buffer;
 	};
 }
diff --git a/yaml-reader/main.cpp b/yaml-reader/main.cpp
@@ -7,20 +7,13 @@
 void run()
 {
 	std::ifstream fin("tests/test.yaml");
+	YAML::Parser parser(fin);
 
-	try {
-		YAML::Parser parser(fin);
+	while(parser)
+	{
 		YAML::Node doc;
 		parser.GetNextDocument(doc);
-
-		std::cout << "name: " << doc["name"] << "\n";
-		std::cout << "age: " << doc["age"] << "\n";
-	} catch(YAML::TypedKeyNotFound <std::string>& e) {
-		std::cout << "Key '" << e.key << "' not found at line " << e.line+1 << ", col " << e.column+1 << "\n";
-	} catch(YAML::KeyNotFound& e) {
-		std::cout << "Key not found at line " << e.line+1 << ", col " << e.column+1 << "\n";
-	} catch(YAML::Exception& e) {
-		std::cout << "Error at line " << e.line+1 << ", col " << e.column+1 << ": " << e.msg << "\n";
+		std::cout << doc;
 	}
 }
 

diff --git a/yaml-reader/tests/test.yaml b/yaml-reader/tests/test.yaml
@@ -1,3 +1 @@
-name: Brett Favre
-position: QB
-teams: [ Falcons, Packers, Jets ]
+- test