Remove CharStream from SourceLocation.

Author: chriseth

libevmasm/Assembly.cpp

@@ -76,15 +76,15 @@ namespace
 string locationFromSources(StringMap const& _sourceCodes, SourceLocation const& _location)
 {
 	if (!_location.hasText() || _sourceCodes.empty())
-		return "";
+		return {};
 
-	auto it = _sourceCodes.find(_location.source->name());
+	auto it = _sourceCodes.find(*_location.sourceName);
 	if (it == _sourceCodes.end())
-		return "";
+		return {};
 
 	string const& source = it->second;
 	if (static_cast<size_t>(_location.start) >= source.size())
-		return "";
+		return {};
 
 	string cut = source.substr(static_cast<size_t>(_location.start), static_cast<size_t>(_location.end - _location.start));
 	auto newLinePos = cut.find_first_of("\n");
@@ -152,8 +152,8 @@ class Functionalizer
 		if (!m_location.isValid())
 			return;
 		m_out << m_prefix << "    /*";
-		if (m_location.source)
-			m_out << " \"" + m_location.source->name() + "\"";
+		if (m_location.sourceName)
+			m_out << " " + escapeAndQuoteYulString(*m_location.sourceName);
 		if (m_location.hasText())
 			m_out << ":" << to_string(m_location.start) + ":" + to_string(m_location.end);
 		m_out << "  " << locationFromSources(m_sourceCodes, m_location);
@@ -235,9 +235,9 @@ Json::Value Assembly::assemblyJSON(map<string, unsigned> const& _sourceIndices)
 	for (AssemblyItem const& i: m_items)
 	{
 		int sourceIndex = -1;
-		if (i.location().source)
+		if (i.location().sourceName)
 		{
-			auto iter = _sourceIndices.find(i.location().source->name());
+			auto iter = _sourceIndices.find(*i.location().sourceName);
 			if (iter != _sourceIndices.end())
 				sourceIndex = static_cast<int>(iter->second);
 		}

libevmasm/AssemblyItem.cpp

@@ -350,8 +350,8 @@ std::string AssemblyItem::computeSourceMapping(
 		SourceLocation const& location = item.location();
 		int length = location.start != -1 && location.end != -1 ? location.end - location.start : -1;
 		int sourceIndex =
-			location.source && _sourceIndicesMap.count(location.source->name()) ?
-			static_cast<int>(_sourceIndicesMap.at(location.source->name())) :
+			(location.sourceName && _sourceIndicesMap.count(*location.sourceName)) ?
+			static_cast<int>(_sourceIndicesMap.at(*location.sourceName)) :
 			-1;
 		char jump = '-';
 		if (item.getJumpType() == evmasm::AssemblyItem::JumpType::IntoFunction)

liblangutil/CMakeLists.txt

@@ -13,6 +13,7 @@ set(sources
 	ParserBase.h
 	Scanner.cpp
 	Scanner.h
+	CharStreamProvider.h
 	SemVerHandler.cpp
 	SemVerHandler.h
 	SourceLocation.h

liblangutil/CharStream.cpp

@@ -45,9 +45,7 @@
  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */
 /**
- * @author Christian <[email protected]>
- * @date 2014
- * Solidity scanner.
+ * Character stream / input file.
  */
 
 #include <liblangutil/CharStream.h>
@@ -118,3 +116,15 @@ tuple<int, int> CharStream::translatePositionToLineColumn(int _position) const
 	}
 	return tuple<int, int>(lineNumber, searchPosition - lineStart);
 }
+
+string_view CharStream::text(SourceLocation const& _location) const
+{
+	if (!_location.hasText())
+		return {};
+	solAssert(_location.sourceName && *_location.sourceName == m_name, "");
+	solAssert(static_cast<size_t>(_location.end) <= m_source.size(), "");
+	return string_view{m_source}.substr(
+		static_cast<size_t>(_location.start),
+		static_cast<size_t>(_location.end - _location.start)
+	);
+}

liblangutil/CharStream.h

@@ -45,9 +45,7 @@
  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */
 /**
- * @author Christian <[email protected]>
- * @date 2014
- * Solidity scanner.
+ * Character stream / input file.
  */
 
 #pragma once
@@ -60,6 +58,8 @@
 namespace solidity::langutil
 {
 
+struct SourceLocation;
+
 /**
  * Bidirectional stream of characters.
  *
@@ -69,8 +69,8 @@ class CharStream
 {
 public:
 	CharStream() = default;
-	explicit CharStream(std::string  _source, std::string  name):
-		m_source(std::move(_source)), m_name(std::move(name)) {}
+	CharStream(std::string _source, std::string _name):
+		m_source(std::move(_source)), m_name(std::move(_name)) {}
 
 	size_t position() const { return m_position; }
 	bool isPastEndOfInput(size_t _charsForward = 0) const { return (m_position + _charsForward) >= m_source.size(); }
@@ -90,6 +90,8 @@ class CharStream
 	std::string const& source() const noexcept { return m_source; }
 	std::string const& name() const noexcept { return m_name; }
 
+	size_t size() const { return m_source.size(); }
+
 	///@{
 	///@name Error printing helper functions
 	/// Functions that help pretty-printing parse errors
@@ -112,6 +114,10 @@ class CharStream
 		return true;
 	}
 
+	/// @returns the substring of the source that the source location references.
+	/// Returns an empty string view if the source location does not `hasText()`.
+	std::string_view text(SourceLocation const& _location) const;
+
 private:
 	std::string m_source;
 	std::string m_name;

liblangutil/CharStreamProvider.h

@@ -0,0 +1,57 @@
+/*
+	This file is part of solidity.
+
+	solidity is free software: you can redistribute it and/or modify
+	it under the terms of the GNU General Public License as published by
+	the Free Software Foundation, either version 3 of the License, or
+	(at your option) any later version.
+
+	solidity is distributed in the hope that it will be useful,
+	but WITHOUT ANY WARRANTY; without even the implied warranty of
+	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+	GNU General Public License for more details.
+
+	You should have received a copy of the GNU General Public License
+	along with solidity.  If not, see <http://www.gnu.org/licenses/>.
+*/
+// SPDX-License-Identifier: GPL-3.0
+/**
+ * Interface to retrieve the character stream by a source name.
+ */
+
+#pragma once
+
+#include <liblangutil/CharStream.h>
+#include <liblangutil/Exceptions.h>
+
+#include <string>
+
+namespace solidity::langutil
+{
+
+/**
+ * Interface to retrieve a CharStream (source) from a source name.
+ * Used especially for printing error information.
+ */
+class CharStreamProvider
+{
+public:
+	virtual ~CharStreamProvider() = default;
+	virtual CharStream const& charStream(std::string const& _sourceName) const = 0;
+};
+
+class SingletonCharStreamProvider: public CharStreamProvider
+{
+public:
+	explicit SingletonCharStreamProvider(CharStream const& _charStream):
+		m_charStream(_charStream) {}
+	CharStream const& charStream(std::string const& _sourceName) const override
+	{
+		solAssert(m_charStream.name() == _sourceName, "");
+		return m_charStream;
+	}
+private:
+	CharStream const& m_charStream;
+};
+
+}

liblangutil/Scanner.cpp

@@ -138,13 +138,15 @@ class LiteralScope
 void Scanner::reset(CharStream _source)
 {
 	m_source = make_shared<CharStream>(std::move(_source));
+	m_sourceName = make_shared<string>(m_source->name());
 	reset();
 }
 
 void Scanner::reset(shared_ptr<CharStream> _source)
 {
 	solAssert(_source.get() != nullptr, "You MUST provide a CharStream when resetting.");
 	m_source = std::move(_source);
+	m_sourceName = make_shared<string>(m_source->name());
 	reset();
 }
 
@@ -497,7 +499,7 @@ Token Scanner::scanSlash()
 				return skipSingleLineComment();
 			// doxygen style /// comment
 			m_skippedComments[NextNext].location.start = firstSlashPosition;
-			m_skippedComments[NextNext].location.source = m_source;
+			m_skippedComments[NextNext].location.sourceName = m_sourceName;
 			m_skippedComments[NextNext].token = Token::CommentLiteral;
 			m_skippedComments[NextNext].location.end = static_cast<int>(scanSingleLineDocComment());
 			return Token::Whitespace;
@@ -526,7 +528,7 @@ Token Scanner::scanSlash()
 				return skipMultiLineComment();
 			// we actually have a multiline documentation comment
 			m_skippedComments[NextNext].location.start = firstSlashPosition;
-			m_skippedComments[NextNext].location.source = m_source;
+			m_skippedComments[NextNext].location.sourceName = m_sourceName;
 			Token comment = scanMultiLineDocComment();
 			m_skippedComments[NextNext].location.end = static_cast<int>(sourcePos());
 			m_skippedComments[NextNext].token = comment;
@@ -766,7 +768,7 @@ void Scanner::scanToken()
 	}
 	while (token == Token::Whitespace);
 	m_tokens[NextNext].location.end = static_cast<int>(sourcePos());
-	m_tokens[NextNext].location.source = m_source;
+	m_tokens[NextNext].location.sourceName = m_sourceName;
 	m_tokens[NextNext].token = token;
 	m_tokens[NextNext].extendedTokenInfo = make_tuple(m, n);
 }

liblangutil/Scanner.h

@@ -177,14 +177,6 @@ class Scanner
 	Token peekNextNextToken() const { return m_tokens[NextNext].token; }
 	///@}
 
-	///@{
-	///@name Error printing helper functions
-	/// Functions that help pretty-printing parse errors
-	/// Do only use in error cases, they are quite expensive.
-	std::string lineAtPosition(int _position) const { return m_source->lineAtPosition(_position); }
-	std::tuple<int, int> translatePositionToLineColumn(int _position) const { return m_source->translatePositionToLineColumn(_position); }
-	///@}
-
 private:
 
 	inline Token setError(ScannerError _error) noexcept
@@ -270,6 +262,7 @@ class Scanner
 	TokenDesc m_tokens[3] = {}; // desc for the current, next and nextnext token
 
 	std::shared_ptr<CharStream> m_source;
+	std::shared_ptr<std::string const> m_sourceName;
 
 	ScannerKind m_kind = ScannerKind::Solidity;
 

liblangutil/SourceLocation.cpp

@@ -34,6 +34,7 @@ SourceLocation const parseSourceLocation(std::string const& _input, std::string
 
 	boost::algorithm::split(pos, _input, boost::is_any_of(":"));
 
+	// TODO What to do with sourceIndex?
 	solAssert(pos.size() == 3, "SourceLocation string must have 3 colon separated numeric fields.");
 	auto const sourceIndex = stoi(pos[Index]);
 
@@ -45,12 +46,7 @@ SourceLocation const parseSourceLocation(std::string const& _input, std::string
 	int start = stoi(pos[Start]);
 	int end = start + stoi(pos[Length]);
 
-	// ASSUMPTION: only the name of source is used from here on, the m_source of the CharStream-Object can be empty
-	std::shared_ptr<langutil::CharStream> source;
-	if (sourceIndex != -1)
-		source = std::make_shared<langutil::CharStream>("", _sourceName);
-
-	return SourceLocation{start, end, source};
+	return SourceLocation{start, end, std::make_shared<std::string>(_sourceName)};
 }
 
 }

liblangutil/SourceLocation.h

@@ -26,8 +26,6 @@
 #include <libsolutil/Assertions.h>
 #include <libsolutil/Exceptions.h>
 
-#include <liblangutil/CharStream.h>
-
 #include <limits>
 #include <memory>
 #include <string>
@@ -44,51 +42,44 @@ struct SourceLocation
 {
 	bool operator==(SourceLocation const& _other) const
 	{
-		return source.get() == _other.source.get() && start == _other.start && end == _other.end;
+		return start == _other.start && end == _other.end && equalSources(_other);
 	}
 	bool operator!=(SourceLocation const& _other) const { return !operator==(_other); }
 
-	inline bool operator<(SourceLocation const& _other) const
+	bool operator<(SourceLocation const& _other) const
 	{
-		if (!source|| !_other.source)
-			return std::make_tuple(int(!!source), start, end) < std::make_tuple(int(!!_other.source), _other.start, _other.end);
+		if (!sourceName || !_other.sourceName)
+			return std::make_tuple(int(!!sourceName), start, end) < std::make_tuple(int(!!_other.sourceName), _other.start, _other.end);
 		else
-			return std::make_tuple(source->name(), start, end) < std::make_tuple(_other.source->name(), _other.start, _other.end);
+			return std::make_tuple(*sourceName, start, end) < std::make_tuple(*_other.sourceName, _other.start, _other.end);
 	}
 
-	inline bool contains(SourceLocation const& _other) const
+	bool contains(SourceLocation const& _other) const
 	{
-		if (!hasText() || !_other.hasText() || source.get() != _other.source.get())
+		if (!hasText() || !_other.hasText() || !equalSources(_other))
 			return false;
 		return start <= _other.start && _other.end <= end;
 	}
 
-	inline bool intersects(SourceLocation const& _other) const
+	bool intersects(SourceLocation const& _other) const
 	{
-		if (!hasText() || !_other.hasText() || source.get() != _other.source.get())
+		if (!hasText() || !_other.hasText() || !equalSources(_other))
 			return false;
 		return _other.start < end && start < _other.end;
 	}
 
-	bool isValid() const { return source || start != -1 || end != -1; }
-
-	bool hasText() const
+	bool equalSources(SourceLocation const& _other) const
 	{
-		return
-			source &&
-			0 <= start &&
-			start <= end &&
-			end <= int(source->source().length());
+		if (!!sourceName != !!_other.sourceName)
+			return false;
+		if (sourceName && *sourceName != *_other.sourceName)
+			return false;
+		return true;
 	}
 
-	std::string text() const
-	{
-		assertThrow(source, SourceLocationError, "Requested text from null source.");
-		assertThrow(0 <= start, SourceLocationError, "Invalid source location.");
-		assertThrow(start <= end, SourceLocationError, "Invalid source location.");
-		assertThrow(end <= int(source->source().length()), SourceLocationError, "Invalid source location.");
-		return source->source().substr(size_t(start), size_t(end - start));
-	}
+	bool isValid() const { return sourceName || start != -1 || end != -1; }
+
+	bool hasText() const { return sourceName && 0 <= start && start <= end; }
 
 	/// @returns the smallest SourceLocation that contains both @param _a and @param _b.
 	/// Assumes that @param _a and @param _b refer to the same source (exception: if the source of either one
@@ -97,8 +88,8 @@ struct SourceLocation
 	/// @param _b, then start resp. end of the result will be -1 as well).
 	static SourceLocation smallestCovering(SourceLocation _a, SourceLocation const& _b)
 	{
-		if (!_a.source)
-			_a.source = _b.source;
+		if (!_a.sourceName)
+			_a.sourceName = _b.sourceName;
 
 		if (_a.start < 0)
 			_a.start = _b.start;
@@ -112,7 +103,7 @@ struct SourceLocation
 
 	int start = -1;
 	int end = -1;
-	std::shared_ptr<CharStream> source;
+	std::shared_ptr<std::string const> sourceName;
 };
 
 SourceLocation const parseSourceLocation(
@@ -127,8 +118,8 @@ inline std::ostream& operator<<(std::ostream& _out, SourceLocation const& _locat
 	if (!_location.isValid())
 		return _out << "NO_LOCATION_SPECIFIED";
 
-	if (_location.source)
-		_out << _location.source->name();
+	if (_location.sourceName)
+		_out << *_location.sourceName;
 
 	_out << "[" << _location.start << "," << _location.end << "]";