Merge pull request #10258 from ethereum/develop

Merge develop into breaking.

.editorconfig

@@ -14,7 +14,7 @@ indent_size = 4
 indent_style = space
 indent_size = 4
 
-[*.sol]
+[*.{sol,yul}]
 indent_style = space
 indent_size = 4
 

libyul/backends/wasm/TextTransform.cpp

@@ -19,10 +19,13 @@
  * Component that transforms internal Wasm representation to text.
  */
 
+#include <libyul/backends/wasm/BinaryTransform.h>
 #include <libyul/backends/wasm/TextTransform.h>
 
 #include <libyul/Exceptions.h>
 
+#include <libsolutil/CommonData.h>
+#include <libsolutil/Keccak256.h>
 #include <libsolutil/StringUtils.h>
 #include <libsolutil/Visitor.h>
 
@@ -39,16 +42,27 @@ using namespace solidity::util;
 string TextTransform::run(wasm::Module const& _module)
 {
 	string ret = "(module\n";
-	for (auto const& sub: _module.subModules)
+	for (auto const& [name, module]: _module.subModules)
 		ret +=
-			"    ;; sub-module \"" +
+			"    ;; custom section for sub-module\n"
-			sub.first +
+			"    ;; The Keccak-256 hash of the text representation of \"" +
-			"\" will be encoded as custom section in binary here, but is skipped in text mode.\n";
+			name +
-	for (auto const& data: _module.customSections)
+			"\": " +
+			toHex(keccak256(run(module))) +
+			"\n"
+			"    ;; (@custom \"" +
+			name +
+			"\" \"" +
+			toHex(BinaryTransform::run(module)) +
+			"\")\n";
+	for (auto const& [name, data]: _module.customSections)
 		ret +=
-			"    ;; custom-section \"" +
+			"    ;; custom section for data\n"
-			data.first +
+			"    ;; (@custom \"" +
-			"\" will be encoded as custom section in binary here, but is skipped in text mode.\n";
+			name +
+			"\" \"" +
+			toHex(data) +
+			"\")\n";
 	for (wasm::FunctionImport const& imp: _module.imports)
 	{
 		ret += "    (import \"" + imp.module + "\" \"" + imp.externalName + "\" (func $" + imp.internalName;

test/cmdlineTests.sh

@@ -261,12 +261,6 @@ printTask "Running general commandline tests..."
 
         inputFiles="$(find "${tdir}" -name 'input.*' -type f -exec printf "%s\n" "{}" \;)"
         inputCount="$(echo "${inputFiles}" | wc -l)"
-        if (( ${inputCount} == 0 ))
-        then
-            printError "No input files found."
-            exit 1
-        fi
-
         if (( ${inputCount} > 1 ))
         then
             printError "Ambiguous input. Found input files in multiple formats:"
@@ -277,6 +271,11 @@ printTask "Running general commandline tests..."
         # Use printf to get rid of the trailing newline
         inputFile=$(printf "%s" "${inputFiles}")
 
+        # If no files specified, assume input.sol as the default
+        if [ -z "${inputFile}" ]; then
+            inputFile="${tdir}/input.sol"
+        fi
+
         if [ "${inputFile}" = "${tdir}/input.json" ]
         then
             stdin="${inputFile}"

test/cmdlineTests/standard_eWasm_requested/output.json

@@ -1,5 +1,7 @@
 {"contracts":{"A":{"C":{"ewasm":{"wast":"(module
-    ;; sub-module \"C_2_deployed\" will be encoded as custom section in binary here, but is skipped in text mode.
+    ;; custom section for sub-module
+    ;; The Keccak-256 hash of the text representation of \"C_2_deployed\": f03f5b9154b9eb6803a947177e38e92e2860de95e90ba0e75eb71a58f18ed589
+    ;; (@custom \"C_2_deployed\" \"0061736d0100000001160460000060017e017e60047e7e7e7e017f60027f7f0002130108657468657265756d067265766572740003030504000201010503010001060100071102066d656d6f72790200046d61696e00010ab60204ca0104017e027f057e037f02404200210020002000200042c00010022101200141c0006a210220022001490440000b20001003421086210320032000421088100384422086210420042000422088100484210520022005370000200241086a2005370000200241106a20053700004280011003421086210620064280014210881003844220862107200241186a2007428001422088100484370000200020002000200010022108200020002000200010022109200941c0006a210a200a2009490440000b200a200810000b0b2901017f024042002000200184200284520440000b42002003422088520440000b2003a721040b20040b1f01017e024020004208864280fe0383200042088842ff01838421010b20010b1e01027e02402000100342108621022002200042108810038421010b20010b\")
     (import \"ethereum\" \"codeCopy\" (func $eth.codeCopy (param i32 i32 i32)))
     (import \"ethereum\" \"revert\" (func $eth.revert (param i32 i32)))
     (import \"ethereum\" \"getCallValue\" (func $eth.getCallValue (param i32)))

test/libyul/ObjectCompilerTest.cpp

@@ -44,6 +44,7 @@ ObjectCompilerTest::ObjectCompilerTest(string const& _filename):
 {
 	m_source = m_reader.source();
 	m_optimize = m_reader.boolSetting("optimize", false);
+	m_wasm = m_reader.boolSetting("wasm", false);
 	m_expectation = m_reader.simpleExpectations();
 }
 
@@ -51,7 +52,7 @@ TestCase::TestResult ObjectCompilerTest::run(ostream& _stream, string const& _li
 {
 	AssemblyStack stack(
 		EVMVersion(),
-		AssemblyStack::Language::StrictAssembly,
+		m_wasm ? AssemblyStack::Language::Ewasm : AssemblyStack::Language::StrictAssembly,
 		m_optimize ? OptimiserSettings::full() : OptimiserSettings::minimal()
 	);
 	if (!stack.parseAndAnalyze("source", m_source))
@@ -62,22 +63,33 @@ TestCase::TestResult ObjectCompilerTest::run(ostream& _stream, string const& _li
 	}
 	stack.optimize();
 
-	MachineAssemblyObject obj = stack.assemble(AssemblyStack::Machine::EVM);
+	if (m_wasm)
-	solAssert(obj.bytecode, "");
+	{
-	solAssert(obj.sourceMappings, "");
+		MachineAssemblyObject obj = stack.assemble(AssemblyStack::Machine::Ewasm);
+		solAssert(obj.bytecode, "");
 
-	m_obtainedResult = "Assembly:\n" + obj.assembly;
+		m_obtainedResult = "Text:\n" + obj.assembly + "\n";
-	if (obj.bytecode->bytecode.empty())
+		m_obtainedResult += "Binary:\n" + toHex(obj.bytecode->bytecode) + "\n";
-		m_obtainedResult += "-- empty bytecode --\n";
+	}
 	else
-		m_obtainedResult +=
+	{
-			"Bytecode: " +
+		MachineAssemblyObject obj = stack.assemble(AssemblyStack::Machine::EVM);
-			toHex(obj.bytecode->bytecode) +
+		solAssert(obj.bytecode, "");
-			"\nOpcodes: " +
+		solAssert(obj.sourceMappings, "");
-			boost::trim_copy(evmasm::disassemble(obj.bytecode->bytecode)) +
+
-			"\nSourceMappings:" +
+		m_obtainedResult = "Assembly:\n" + obj.assembly;
-			(obj.sourceMappings->empty() ? "" : " " + *obj.sourceMappings) +
+		if (obj.bytecode->bytecode.empty())
-			"\n";
+			m_obtainedResult += "-- empty bytecode --\n";
+		else
+			m_obtainedResult +=
+				"Bytecode: " +
+				toHex(obj.bytecode->bytecode) +
+				"\nOpcodes: " +
+				boost::trim_copy(evmasm::disassemble(obj.bytecode->bytecode)) +
+				"\nSourceMappings:" +
+				(obj.sourceMappings->empty() ? "" : " " + *obj.sourceMappings) +
+				"\n";
+	}
 
 	return checkResult(_stream, _linePrefix, _formatted);
 }

test/libyul/ObjectCompilerTest.h

@@ -55,6 +55,7 @@ private:
 	static void printErrors(std::ostream& _stream, langutil::ErrorList const& _errors);
 
 	bool m_optimize = false;
+	bool m_wasm = false;
 };
 
 }

test/libyul/objectCompiler/wasm/no_main_function.yul

@@ -0,0 +1,22 @@
+{
+    function not_main() {
+        i64.drop(i64.add(0, 1))
+    }
+}
+// ====
+// wasm: true
+// ----
+// Text:
+// (module
+//     (memory $memory (export "memory") 1)
+//
+// (func $not_main
+//     (block $label_
+//         (drop (i64.add (i64.const 0) (i64.const 1)))
+//     )
+// )
+//
+// )
+//
+// Binary:
+// 0061736d01000000010401600000020100030201000503010001060100070a01066d656d6f727902000a0801060002401a0b0b

test/libyul/objectCompiler/wasm/simple.yul

@@ -0,0 +1,23 @@
+{
+    function main() {
+        i64.drop(i64.add(0, 1))
+    }
+}
+// ====
+// wasm: true
+// ----
+// Text:
+// (module
+//     (memory $memory (export "memory") 1)
+//     (export "main" (func $main))
+//
+// (func $main
+//     (block $label_
+//         (drop (i64.add (i64.const 0) (i64.const 1)))
+//     )
+// )
+//
+// )
+//
+// Binary:
+// 0061736d01000000010401600000020100030201000503010001060100071102066d656d6f72790200046d61696e00000a0801060002401a0b0b

test/libyul/objectCompiler/wasm/subObject.yul

@@ -0,0 +1,22 @@
+object "a" {
+    code {}
+    // Unreferenced data is not added to the assembled bytecode.
+    data "str" "Hello, World!"
+    object "sub" { code { function main() { i64.drop(11) } } }
+}
+// ====
+// wasm: true
+// ----
+// Text:
+// (module
+//     ;; custom section for sub-module
+//     ;; The Keccak-256 hash of the text representation of "sub": 78ac3419d75c8d6f42f663717b8e964eeb994d77ff175145133084422dbd23d7
+//     ;; (@custom "sub" "0061736d01000000010401600000020100030201000503010001060100071102066d656d6f72790200046d61696e00000a0801060002401a0b0b")
+//     ;; custom section for data
+//     ;; (@custom "str" "48656c6c6f2c20576f726c6421")
+//     (memory $memory (export "memory") 1)
+//
+// )
+//
+// Binary:
+// 0061736d010000000101000201000301000503010001060100070a01066d656d6f72790200003e037375620061736d01000000010401600000020100030201000503010001060100071102066d656d6f72790200046d61696e00000a0801060002401a0b0b00110373747248656c6c6f2c20576f726c64210a0100

test/libyul/objectCompiler/wasm/subObjectAccess.yul

@@ -0,0 +1,79 @@
+object "A" {
+    code {
+        function main() {
+            // TODO: support this
+            // i64.drop(dataoffset("A"))
+            // i64.drop(datasize("A"))
+            i64.drop(dataoffset("B"))
+            i64.drop(datasize("B"))
+            // TODO: support sub-subobjects
+            // i64.drop(dataoffset("B.C"))
+            // i64.drop(datasize("B.C"))
+            // i64.drop(dataoffset("B.E"))
+            // i64.drop(datasize("B.E"))
+            // i64.drop(dataoffset("B.C.D"))
+            // i64.drop(datasize("B.C.D"))
+        }
+    }
+
+    data "data1" "Hello, World!"
+
+    object "B" {
+        code {
+            function main() {
+                i64.drop(dataoffset("C"))
+                i64.drop(datasize("C"))
+                i64.drop(dataoffset("E"))
+                i64.drop(datasize("E"))
+                // i64.drop(dataoffset("C.D"))
+                // i64.drop(datasize("C.D"))
+            }
+        }
+        object "C" {
+            code {
+                function main() {
+                    i64.drop(dataoffset("D"))
+                    i64.drop(datasize("D"))
+                }
+            }
+            object "D" {
+                code {
+                    function main() {
+                        unreachable()
+                    }
+                }
+            }
+        }
+        object "E" {
+            code {
+                function main() {
+                    unreachable()
+                }
+            }
+        }
+    }
+}
+// ====
+// wasm: true
+// ----
+// Text:
+// (module
+//     ;; custom section for sub-module
+//     ;; The Keccak-256 hash of the text representation of "B": 1eeffe5bc8d8819350ead60cc71ccd92c223cf52a908330db53461eb9ac89b62
+//     ;; (@custom "B" "0061736d01000000010401600000020100030201000503010001060100071102066d656d6f72790200046d61696e0000007b01430061736d01000000010401600000020100030201000503010001060100071102066d656d6f72790200046d61696e0000003c01440061736d01000000010401600000020100030201000503010001060100071102066d656d6f72790200046d61696e00000a080106000240000b0b0a0901070002401a1a0b0b003c01450061736d01000000010401600000020100030201000503010001060100071102066d656d6f72790200046d61696e00000a080106000240000b0b0a0b01090002401a1a1a1a0b0b")
+//     ;; custom section for data
+//     ;; (@custom "data1" "48656c6c6f2c20576f726c6421")
+//     (memory $memory (export "memory") 1)
+//     (export "main" (func $main))
+//
+// (func $main
+//     (block $label_
+//         (drop (dataoffset "B"))
+//         (drop (datasize "B"))
+//     )
+// )
+//
+// )
+//
+// Binary:
+// 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