19 files changed, 86 insertions, 150 deletions
diff --git a/libevmasm/Assembly.cpp b/libevmasm/Assembly.cpp
index e63194a0..29d9846d 100644
--- a/libevmasm/Assembly.cpp
+++ b/libevmasm/Assembly.cpp
@@ -35,6 +35,7 @@
 using namespace std;
 using namespace dev;
 using namespace dev::eth;
+using namespace langutil;
 
 void Assembly::append(Assembly const& _a)
 {
@@ -113,10 +114,10 @@ namespace
 
 string locationFromSources(StringMap const& _sourceCodes, SourceLocation const& _location)
 {
-	if (_location.isEmpty() || _sourceCodes.empty() || _location.start >= _location.end || _location.start < 0)
+	if (_location.isEmpty() || !_location.source.get() || _sourceCodes.empty() || _location.start >= _location.end || _location.start < 0)
 		return "";
 
-	auto it = _sourceCodes.find(*_location.sourceName);
+	auto it = _sourceCodes.find(_location.source->name());
 	if (it == _sourceCodes.end())
 		return "";
 
@@ -185,11 +186,11 @@ public:
 
 	void printLocation()
 	{
-		if (!m_location.sourceName && m_location.isEmpty())
+		if (!m_location.source && m_location.isEmpty())
 			return;
 		m_out << m_prefix << "    /*";
-		if (m_location.sourceName)
-			m_out << " \"" + *m_location.sourceName + "\"";
+		if (m_location.source)
+			m_out << " \"" + m_location.source->name() + "\"";
 		if (!m_location.isEmpty())
 			m_out << ":" << to_string(m_location.start) + ":" + to_string(m_location.end);
 		m_out << "  " << locationFromSources(m_sourceCodes, m_location);
diff --git a/libevmasm/Assembly.h b/libevmasm/Assembly.h
index 8ef36923..d846b475 100644
--- a/libevmasm/Assembly.h
+++ b/libevmasm/Assembly.h
@@ -18,12 +18,12 @@
 #pragma once
 
 #include <libevmasm/Instruction.h>
-#include <libevmasm/SourceLocation.h>
+#include <liblangutil/SourceLocation.h>
 #include <libevmasm/AssemblyItem.h>
 #include <libevmasm/LinkerObject.h>
 #include <libevmasm/Exceptions.h>
 
-#include <libsolidity/interface/EVMVersion.h>
+#include <liblangutil/EVMVersion.h>
 
 #include <libdevcore/Common.h>
 #include <libdevcore/Assertions.h>
@@ -93,7 +93,7 @@ public:
 	void setDeposit(int _deposit) { m_deposit = _deposit; assertThrow(m_deposit >= 0, InvalidDeposit, ""); }
 
 	/// Changes the source location used for each appended item.
-	void setSourceLocation(SourceLocation const& _location) { m_currentSourceLocation = _location; }
+	void setSourceLocation(langutil::SourceLocation const& _location) { m_currentSourceLocation = _location; }
 
 	/// Assembles the assembly into bytecode. The assembly should not be modified after this call, since the assembled version is cached.
 	LinkerObject const& assemble() const;
@@ -178,7 +178,7 @@ protected:
 
 	int m_deposit = 0;
 
-	SourceLocation m_currentSourceLocation;
+	langutil::SourceLocation m_currentSourceLocation;
 };
 
 inline std::ostream& operator<<(std::ostream& _out, Assembly const& _a)
diff --git a/libevmasm/AssemblyItem.cpp b/libevmasm/AssemblyItem.cpp
index a3a4d6b6..52f246d1 100644
--- a/libevmasm/AssemblyItem.cpp
+++ b/libevmasm/AssemblyItem.cpp
@@ -117,6 +117,8 @@ int AssemblyItem::returnValues() const
 
 bool AssemblyItem::canBeFunctional() const
 {
+	if (m_jumpType != JumpType::Ordinary)
+		return false;
 	switch (m_type)
 	{
 	case Operation:
diff --git a/libevmasm/AssemblyItem.h b/libevmasm/AssemblyItem.h
index 6187e18f..a7875171 100644
--- a/libevmasm/AssemblyItem.h
+++ b/libevmasm/AssemblyItem.h
@@ -26,7 +26,7 @@
 #include <libdevcore/Common.h>
 #include <libdevcore/Assertions.h>
 #include <libevmasm/Instruction.h>
-#include <libevmasm/SourceLocation.h>
+#include <liblangutil/SourceLocation.h>
 #include "Exceptions.h"
 using namespace dev::solidity;
 
@@ -57,14 +57,14 @@ class AssemblyItem
 public:
 	enum class JumpType { Ordinary, IntoFunction, OutOfFunction };
 
-	AssemblyItem(u256 _push, SourceLocation const& _location = SourceLocation()):
+	AssemblyItem(u256 _push, langutil::SourceLocation const& _location = langutil::SourceLocation()):
 		AssemblyItem(Push, _push, _location) { }
-	AssemblyItem(solidity::Instruction _i, SourceLocation const& _location = SourceLocation()):
+	AssemblyItem(solidity::Instruction _i, langutil::SourceLocation const& _location = langutil::SourceLocation()):
 		m_type(Operation),
 		m_instruction(_i),
 		m_location(_location)
 	{}
-	AssemblyItem(AssemblyItemType _type, u256 _data = 0, SourceLocation const& _location = SourceLocation()):
+	AssemblyItem(AssemblyItemType _type, u256 _data = 0, langutil::SourceLocation const& _location = langutil::SourceLocation()):
 		m_type(_type),
 		m_location(_location)
 	{
@@ -124,8 +124,8 @@ public:
 	/// @returns true if the assembly item can be used in a functional context.
 	bool canBeFunctional() const;
 
-	void setLocation(SourceLocation const& _location) { m_location = _location; }
-	SourceLocation const& location() const { return m_location; }
+	void setLocation(langutil::SourceLocation const& _location) { m_location = _location; }
+	langutil::SourceLocation const& location() const { return m_location; }
 
 	void setJumpType(JumpType _jumpType) { m_jumpType = _jumpType; }
 	JumpType getJumpType() const { return m_jumpType; }
@@ -140,7 +140,7 @@ private:
 	AssemblyItemType m_type;
 	Instruction m_instruction; ///< Only valid if m_type == Operation
 	std::shared_ptr<u256> m_data; ///< Only valid if m_type != Operation
-	SourceLocation m_location;
+	langutil::SourceLocation m_location;
 	JumpType m_jumpType = JumpType::Ordinary;
 	/// Pushed value for operations with data to be determined during assembly stage,
 	/// e.g. PushSubSize, PushTag, PushSub, etc.
diff --git a/libevmasm/CMakeLists.txt b/libevmasm/CMakeLists.txt
index 86192c1b..e0e3389a 100644
--- a/libevmasm/CMakeLists.txt
+++ b/libevmasm/CMakeLists.txt
@@ -1,5 +1,21 @@
-file(GLOB sources "*.cpp")
-file(GLOB headers "*.h")
+set(sources
+	Assembly.cpp
+	AssemblyItem.cpp
+	BlockDeduplicator.cpp
+	CommonSubexpressionEliminator.cpp
+	ConstantOptimiser.cpp
+	ControlFlowGraph.cpp
+	ExpressionClasses.cpp
+	GasMeter.cpp
+	Instruction.cpp
+	JumpdestRemover.cpp
+	KnownState.cpp
+	LinkerObject.cpp
+	PathGasMeter.cpp
+	PeepholeOptimiser.cpp
+	SemanticInformation.cpp
+	SimplificationRules.cpp
+)
 
-add_library(evmasm ${sources} ${headers})
+add_library(evmasm ${sources})
 target_link_libraries(evmasm PUBLIC devcore)
diff --git a/libevmasm/CommonSubexpressionEliminator.cpp b/libevmasm/CommonSubexpressionEliminator.cpp
index 04926986..949d2c75 100644
--- a/libevmasm/CommonSubexpressionEliminator.cpp
+++ b/libevmasm/CommonSubexpressionEliminator.cpp
@@ -30,6 +30,7 @@
 using namespace std;
 using namespace dev;
 using namespace dev::eth;
+using namespace langutil;
 
 vector<AssemblyItem> CommonSubexpressionEliminator::getOptimizedItems()
 {
diff --git a/libevmasm/CommonSubexpressionEliminator.h b/libevmasm/CommonSubexpressionEliminator.h
index b20de246..eba25db0 100644
--- a/libevmasm/CommonSubexpressionEliminator.h
+++ b/libevmasm/CommonSubexpressionEliminator.h
@@ -34,6 +34,11 @@
 #include <libevmasm/SemanticInformation.h>
 #include <libevmasm/KnownState.h>
 
+namespace langutil
+{
+struct SourceLocation;
+}
+
 namespace dev
 {
 namespace eth
@@ -137,10 +142,10 @@ private:
 	bool removeStackTopIfPossible();
 
 	/// Appends a dup instruction to m_generatedItems to retrieve the element at the given stack position.
-	void appendDup(int _fromPosition, SourceLocation const& _location);
+	void appendDup(int _fromPosition, langutil::SourceLocation const& _location);
 	/// Appends a swap instruction to m_generatedItems to retrieve the element at the given stack position.
 	/// @note this might also remove the last item if it exactly the same swap instruction.
-	void appendOrRemoveSwap(int _fromPosition, SourceLocation const& _location);
+	void appendOrRemoveSwap(int _fromPosition, langutil::SourceLocation const& _location);
 	/// Appends the given assembly item.
 	void appendItem(AssemblyItem const& _item);
 
diff --git a/libevmasm/ConstantOptimiser.cpp b/libevmasm/ConstantOptimiser.cpp
index 9844ba3a..ae1a5643 100644
--- a/libevmasm/ConstantOptimiser.cpp
+++ b/libevmasm/ConstantOptimiser.cpp
@@ -193,7 +193,7 @@ AssemblyItems ComputeMethod::findRepresentation(u256 const& _value)
 		bigint bestGas = gasNeeded(routine);
 		for (unsigned bits = 255; bits > 8 && m_maxSteps > 0; --bits)
 		{
-			unsigned gapDetector = unsigned(_value >> (bits - 8)) & 0x1ff;
+			unsigned gapDetector = unsigned((_value >> (bits - 8)) & 0x1ff);
 			if (gapDetector != 0xff && gapDetector != 0x100)
 				continue;
 
diff --git a/libevmasm/ConstantOptimiser.h b/libevmasm/ConstantOptimiser.h
index 2c753fa8..04c43c5d 100644
--- a/libevmasm/ConstantOptimiser.h
+++ b/libevmasm/ConstantOptimiser.h
@@ -23,7 +23,7 @@
 
 #include <libevmasm/Exceptions.h>
 
-#include <libsolidity/interface/EVMVersion.h>
+#include <liblangutil/EVMVersion.h>
 
 #include <libdevcore/Assertions.h>
 #include <libdevcore/CommonData.h>
@@ -109,8 +109,8 @@ class LiteralMethod: public ConstantOptimisationMethod
 public:
 	explicit LiteralMethod(Params const& _params, u256 const& _value):
 		ConstantOptimisationMethod(_params, _value) {}
-	virtual bigint gasNeeded() const override;
-	virtual AssemblyItems execute(Assembly&) const override { return AssemblyItems{}; }
+	bigint gasNeeded() const override;
+	AssemblyItems execute(Assembly&) const override { return AssemblyItems{}; }
 };
 
 /**
@@ -120,8 +120,8 @@ class CodeCopyMethod: public ConstantOptimisationMethod
 {
 public:
 	explicit CodeCopyMethod(Params const& _params, u256 const& _value);
-	virtual bigint gasNeeded() const override;
-	virtual AssemblyItems execute(Assembly& _assembly) const override;
+	bigint gasNeeded() const override;
+	AssemblyItems execute(Assembly& _assembly) const override;
 
 protected:
 	static AssemblyItems const& copyRoutine();
@@ -144,8 +144,8 @@ public:
 		);
 	}
 
-	virtual bigint gasNeeded() const override { return gasNeeded(m_routine); }
-	virtual AssemblyItems execute(Assembly&) const override
+	bigint gasNeeded() const override { return gasNeeded(m_routine); }
+	AssemblyItems execute(Assembly&) const override
 	{
 		return m_routine;
 	}
diff --git a/libevmasm/ControlFlowGraph.cpp b/libevmasm/ControlFlowGraph.cpp
index 86f16d48..d62f5436 100644
--- a/libevmasm/ControlFlowGraph.cpp
+++ b/libevmasm/ControlFlowGraph.cpp
@@ -275,7 +275,7 @@ void ControlFlowGraph::gatherKnowledge()
 			//@todo in the case of JUMPI, add knowledge about the condition to the state
 			// (for both values of the condition)
 			set<u256> tags = state->tagsInExpression(
-				state->stackElement(state->stackHeight(), SourceLocation())
+				state->stackElement(state->stackHeight(), langutil::SourceLocation{})
 			);
 			state->feedItem(m_items.at(pc++));
 
diff --git a/libevmasm/ExpressionClasses.cpp b/libevmasm/ExpressionClasses.cpp
index 42a1819a..abbbbc2c 100644
--- a/libevmasm/ExpressionClasses.cpp
+++ b/libevmasm/ExpressionClasses.cpp
@@ -34,7 +34,7 @@
 using namespace std;
 using namespace dev;
 using namespace dev::eth;
-
+using namespace langutil;
 
 bool ExpressionClasses::Expression::operator<(ExpressionClasses::Expression const& _other) const
 {
diff --git a/libevmasm/ExpressionClasses.h b/libevmasm/ExpressionClasses.h
index df8082f9..a34844c5 100644
--- a/libevmasm/ExpressionClasses.h
+++ b/libevmasm/ExpressionClasses.h
@@ -31,6 +31,11 @@
 #include <memory>
 #include <set>
 
+namespace langutil
+{
+struct SourceLocation;
+}
+
 namespace dev
 {
 namespace eth
@@ -82,7 +87,7 @@ public:
 	void forceEqual(Id _id, AssemblyItem const& _item, Ids const& _arguments, bool _copyItem = true);
 
 	/// @returns the id of a new class which is different to all other classes.
-	Id newClass(SourceLocation const& _location);
+	Id newClass(langutil::SourceLocation const& _location);
 
 	/// @returns true if the values of the given classes are known to be different (on every input).
 	/// @note that this function might still return false for some different inputs.
diff --git a/libevmasm/GasMeter.h b/libevmasm/GasMeter.h
index da90b028..e3909e86 100644
--- a/libevmasm/GasMeter.h
+++ b/libevmasm/GasMeter.h
@@ -24,7 +24,7 @@
 #include <libevmasm/ExpressionClasses.h>
 #include <libevmasm/AssemblyItem.h>
 
-#include <libsolidity/interface/EVMVersion.h>
+#include <liblangutil/EVMVersion.h>
 
 #include <ostream>
 #include <tuple>
diff --git a/libevmasm/KnownState.cpp b/libevmasm/KnownState.cpp
index b6c1bcc9..a5546e61 100644
--- a/libevmasm/KnownState.cpp
+++ b/libevmasm/KnownState.cpp
@@ -29,6 +29,7 @@
 using namespace std;
 using namespace dev;
 using namespace dev::eth;
+using namespace langutil;
 
 ostream& KnownState::stream(ostream& _out) const
 {
diff --git a/libevmasm/KnownState.h b/libevmasm/KnownState.h
index cd50550e..3ab1c4b1 100644
--- a/libevmasm/KnownState.h
+++ b/libevmasm/KnownState.h
@@ -46,6 +46,11 @@
 #include <libevmasm/ExpressionClasses.h>
 #include <libevmasm/SemanticInformation.h>
 
+namespace langutil
+{
+struct SourceLocation;
+}
+
 namespace dev
 {
 namespace eth
@@ -121,9 +126,9 @@ public:
 
 	/// Retrieves the current equivalence class fo the given stack element (or generates a new
 	/// one if it does not exist yet).
-	Id stackElement(int _stackHeight, SourceLocation const& _location);
+	Id stackElement(int _stackHeight, langutil::SourceLocation const& _location);
 	/// @returns the stackElement relative to the current stack height.
-	Id relativeStackElement(int _stackOffset, SourceLocation const& _location = SourceLocation());
+	Id relativeStackElement(int _stackOffset, langutil::SourceLocation const& _location = {});
 
 	/// @returns its set of tags if the given expression class is a known tag union; returns a set
 	/// containing the tag if it is a PushTag expression and the empty set otherwise.
@@ -142,22 +147,22 @@ private:
 	/// Assigns a new equivalence class to the next sequence number of the given stack element.
 	void setStackElement(int _stackHeight, Id _class);
 	/// Swaps the given stack elements in their next sequence number.
-	void swapStackElements(int _stackHeightA, int _stackHeightB, SourceLocation const& _location);
+	void swapStackElements(int _stackHeightA, int _stackHeightB, langutil::SourceLocation const& _location);
 
 	/// Increments the sequence number, deletes all storage information that might be overwritten
 	/// and stores the new value at the given slot.
 	/// @returns the store operation, which might be invalid if storage was not modified
-	StoreOperation storeInStorage(Id _slot, Id _value, SourceLocation const& _location);
+	StoreOperation storeInStorage(Id _slot, Id _value, langutil::SourceLocation const& _location);
 	/// Retrieves the current value at the given slot in storage or creates a new special sload class.
-	Id loadFromStorage(Id _slot, SourceLocation const& _location);
+	Id loadFromStorage(Id _slot, langutil::SourceLocation const& _location);
 	/// Increments the sequence number, deletes all memory information that might be overwritten
 	/// and stores the new value at the given slot.
 	/// @returns the store operation, which might be invalid if memory was not modified
-	StoreOperation storeInMemory(Id _slot, Id _value, SourceLocation const& _location);
+	StoreOperation storeInMemory(Id _slot, Id _value, langutil::SourceLocation const& _location);
 	/// Retrieves the current value at the given slot in memory or creates a new special mload class.
-	Id loadFromMemory(Id _slot, SourceLocation const& _location);
+	Id loadFromMemory(Id _slot, langutil::SourceLocation const& _location);
 	/// Finds or creates a new expression that applies the Keccak-256 hash function to the contents in memory.
-	Id applyKeccak256(Id _start, Id _length, SourceLocation const& _location);
+	Id applyKeccak256(Id _start, Id _length, langutil::SourceLocation const& _location);
 
 	/// @returns a new or already used Id representing the given set of tags.
 	Id tagUnion(std::set<u256> _tags);
diff --git a/libevmasm/PathGasMeter.h b/libevmasm/PathGasMeter.h
index fb821684..772df484 100644
--- a/libevmasm/PathGasMeter.h
+++ b/libevmasm/PathGasMeter.h
@@ -23,7 +23,7 @@
 
 #include <libevmasm/GasMeter.h>
 
-#include <libsolidity/interface/EVMVersion.h>
+#include <liblangutil/EVMVersion.h>
 
 #include <set>
 #include <vector>
diff --git a/libevmasm/SimplificationRules.cpp b/libevmasm/SimplificationRules.cpp
index 120d1787..1dce5f1e 100644
--- a/libevmasm/SimplificationRules.cpp
+++ b/libevmasm/SimplificationRules.cpp
@@ -38,7 +38,7 @@
 using namespace std;
 using namespace dev;
 using namespace dev::eth;
-
+using namespace langutil;
 
 SimplificationRule<Pattern> const* Rules::findFirstMatch(
 	Expression const& _expr,
diff --git a/libevmasm/SimplificationRules.h b/libevmasm/SimplificationRules.h
index fbe5a2b0..fc45a46c 100644
--- a/libevmasm/SimplificationRules.h
+++ b/libevmasm/SimplificationRules.h
@@ -31,6 +31,11 @@
 #include <functional>
 #include <vector>
 
+namespace langutil
+{
+struct SourceLocation;
+}
+
 namespace dev
 {
 namespace eth
@@ -97,7 +102,7 @@ public:
 	unsigned matchGroup() const { return m_matchGroup; }
 	bool matches(Expression const& _expr, ExpressionClasses const& _classes) const;
 
-	AssemblyItem toAssemblyItem(SourceLocation const& _location) const;
+	AssemblyItem toAssemblyItem(langutil::SourceLocation const& _location) const;
 	std::vector<Pattern> arguments() const { return m_arguments; }
 
 	/// @returns the id of the matched expression if this pattern is part of a match group.
@@ -135,7 +140,7 @@ struct ExpressionTemplate
 {
 	using Expression = ExpressionClasses::Expression;
 	using Id = ExpressionClasses::Id;
-	explicit ExpressionTemplate(Pattern const& _pattern, SourceLocation const& _location);
+	explicit ExpressionTemplate(Pattern const& _pattern, langutil::SourceLocation const& _location);
 	std::string toString() const;
 	bool hasId = false;
 	/// Id of the matched expression, if available.
diff --git a/libevmasm/SourceLocation.h b/libevmasm/SourceLocation.h
deleted file mode 100644
index b42c3aa9..00000000
--- a/libevmasm/SourceLocation.h
+++ /dev/null
@@ -1,105 +0,0 @@
-/*
-    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/>.
-*/
-/**
- * @author Lefteris Karapetsas <lefteris@ethdev.com>
- * @date 2015
- * Represents a location in a source file
- */
-
-#pragma once
-
-#include <memory>
-#include <string>
-#include <ostream>
-#include <tuple>
-#include <libdevcore/Common.h> // defines noexcept macro for MSVC
-
-namespace dev
-{
-
-/**
- * Representation of an interval of source positions.
- * The interval includes start and excludes end.
- */
-struct SourceLocation
-{
-	SourceLocation(): start(-1), end(-1) { }
-	SourceLocation(int _start, int _end, std::shared_ptr<std::string const> _sourceName):
-		start(_start), end(_end), sourceName(_sourceName) { }
-	SourceLocation(SourceLocation&& _other) noexcept:
-		start(_other.start),
-		end(_other.end),
-		sourceName(std::move(_other.sourceName))
-	{}
-	SourceLocation(SourceLocation const&) = default;
-	SourceLocation& operator=(SourceLocation const&) = default;
-	SourceLocation& operator=(SourceLocation&& _other) noexcept
-	{
-		start = _other.start;
-		end = _other.end;
-		sourceName = std::move(_other.sourceName);
-		return *this;
-	}
-
-	bool operator==(SourceLocation const& _other) const
-	{
-		return start == _other.start && end == _other.end &&
-			((!sourceName && !_other.sourceName) || (sourceName && _other.sourceName && *sourceName == *_other.sourceName));
-	}
-	bool operator!=(SourceLocation const& _other) const { return !operator==(_other); }
-	inline bool operator<(SourceLocation const& _other) const;
-	inline bool contains(SourceLocation const& _other) const;
-	inline bool intersects(SourceLocation const& _other) const;
-
-	bool isEmpty() const { return start == -1 && end == -1; }
-
-	int start;
-	int end;
-	std::shared_ptr<std::string const> sourceName;
-};
-
-/// Stream output for Location (used e.g. in boost exceptions).
-inline std::ostream& operator<<(std::ostream& _out, SourceLocation const& _location)
-{
-	if (_location.isEmpty())
-		return _out << "NO_LOCATION_SPECIFIED";
-	return _out << *_location.sourceName << "[" << _location.start << "," << _location.end << ")";
-}
-
-bool SourceLocation::operator<(SourceLocation const& _other) const
-{
-	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(*sourceName, start, end) < std::make_tuple(*_other.sourceName, _other.start, _other.end);
-}
-
-bool SourceLocation::contains(SourceLocation const& _other) const
-{
-	if (isEmpty() || _other.isEmpty() || ((!sourceName || !_other.sourceName || *sourceName != *_other.sourceName) && (sourceName || _other.sourceName)))
-		return false;
-	return start <= _other.start && _other.end <= end;
-}
-
-bool SourceLocation::intersects(SourceLocation const& _other) const
-{
-	if (isEmpty() || _other.isEmpty() || ((!sourceName || !_other.sourceName || *sourceName != *_other.sourceName) && (sourceName || _other.sourceName)))
-		return false;
-	return _other.start < end && start < _other.end;
-}
-
-}