prepare_report.py: Add support for switching between CLI and Standard JSON compiler interfaces

Author: cameel

scripts/bytecodecompare/prepare_report.py

@@ -3,13 +3,26 @@
 import sys
 import subprocess
 import json
+import re
 from argparse import ArgumentParser
 from dataclasses import dataclass
+from enum import Enum
 from glob import glob
 from pathlib import Path
+from tempfile import TemporaryDirectory
 from typing import List, Optional, Tuple, Union
 
 
+CONTRACT_SEPARATOR_PATTERN = re.compile(r'^======= (?P<file_name>.+):(?P<contract_name>[^:]+) =======$', re.MULTILINE)
+BYTECODE_REGEX = re.compile(r'^Binary:\n(?P<bytecode>.*)$', re.MULTILINE)
+METADATA_REGEX = re.compile(r'^Metadata:\n(?P<metadata>\{.*\})$', re.MULTILINE)
+
+
+class CompilerInterface(Enum):
+    CLI = 'cli'
+    STANDARD_JSON = 'standard-json'
+
+
 @dataclass(frozen=True)
 class ContractReport:
     contract_name: str
@@ -74,62 +87,145 @@ def parse_standard_json_output(source_file_name: Path, standard_json_output: str
     return file_report
 
 
-def prepare_compiler_input(compiler_path: Path, source_file_name: Path, optimize: bool) -> Tuple[List[str], str]:
-    json_input: dict = {
-        'language': 'Solidity',
-        'sources': {
-            str(source_file_name): {'content': load_source(source_file_name)}
-        },
-        'settings': {
-            'optimizer': {'enabled': optimize},
-            'outputSelection': {'*': {'*': ['evm.bytecode.object', 'metadata']}},
-            'modelChecker': {'engine': 'none'},
-        }
-    }
+def parse_cli_output(source_file_name: Path, cli_output: str) -> FileReport:
+    # re.split() returns a list containing the text between pattern occurrences but also inserts the
+    # content of matched groups in between. It also never omits the empty elements so the number of
+    # list items is predictable (3 per match + the text before the first match)
+    output_segments = re.split(CONTRACT_SEPARATOR_PATTERN, cli_output)
+    assert len(output_segments) % 3 == 1
 
-    command_line = [str(compiler_path), '--standard-json']
-    compiler_input = json.dumps(json_input)
+    if len(output_segments) == 1:
+        return FileReport(file_name=source_file_name, contract_reports=None)
 
-    return (command_line, compiler_input)
+    file_report = FileReport(file_name=source_file_name, contract_reports=[])
+    for file_name, contract_name, contract_output in zip(output_segments[1::3], output_segments[2::3], output_segments[3::3]):
+        bytecode_match = re.search(BYTECODE_REGEX, contract_output)
+        metadata_match = re.search(METADATA_REGEX, contract_output)
+
+        assert file_report.contract_reports is not None
+        file_report.contract_reports.append(ContractReport(
+            contract_name=contract_name,
+            file_name=Path(file_name),
+            bytecode=bytecode_match['bytecode'] if bytecode_match is not None else None,
+            metadata=metadata_match['metadata'] if metadata_match is not None else None,
+        ))
 
+    return file_report
 
-def run_compiler(compiler_path: Path, source_file_name: Path, optimize: bool) -> FileReport:
-    (command_line, compiler_input) = prepare_compiler_input(compiler_path, Path(Path(source_file_name).name), optimize)
 
-    process = subprocess.run(
-        command_line,
-        input=compiler_input,
-        encoding='utf8',
-        capture_output=True,
-        check=False,
-    )
+def prepare_compiler_input(
+    compiler_path: Path,
+    source_file_name: Path,
+    optimize: bool,
+    interface: CompilerInterface
+) -> Tuple[List[str], str]:
+
+    if interface == CompilerInterface.STANDARD_JSON:
+        json_input: dict = {
+            'language': 'Solidity',
+            'sources': {
+                str(source_file_name): {'content': load_source(source_file_name)}
+            },
+            'settings': {
+                'optimizer': {'enabled': optimize},
+                'outputSelection': {'*': {'*': ['evm.bytecode.object', 'metadata']}},
+                'modelChecker': {'engine': 'none'},
+            }
+        }
+
+        command_line = [str(compiler_path), '--standard-json']
+        compiler_input = json.dumps(json_input)
+    else:
+        assert interface == CompilerInterface.CLI
 
-    return parse_standard_json_output(Path(source_file_name), process.stdout)
+        compiler_options = [str(source_file_name), '--bin', '--metadata', '--model-checker-engine', 'none']
+        if optimize:
+            compiler_options.append('--optimize')
 
+        command_line = [str(compiler_path)] + compiler_options
+        compiler_input = load_source(source_file_name)
 
-def generate_report(source_file_names: List[str], compiler_path: Path):
+    return (command_line, compiler_input)
+
+
+def run_compiler(
+    compiler_path: Path,
+    source_file_name: Path,
+    optimize: bool,
+    interface: CompilerInterface,
+    tmp_dir: Path,
+) -> FileReport:
+
+    if interface == CompilerInterface.STANDARD_JSON:
+        (command_line, compiler_input) = prepare_compiler_input(
+            compiler_path,
+            Path(Path(source_file_name).name),
+            optimize,
+            interface
+        )
+
+        process = subprocess.run(
+            command_line,
+            input=compiler_input,
+            encoding='utf8',
+            capture_output=True,
+            check=False,
+        )
+
+        return parse_standard_json_output(Path(source_file_name), process.stdout)
+    else:
+        assert interface == CompilerInterface.CLI
+        assert tmp_dir is not None
+
+        (command_line, compiler_input) = prepare_compiler_input(
+            compiler_path.absolute(),
+            Path(source_file_name.name),
+            optimize,
+            interface
+        )
+
+        # Create a copy that we can use directly with the CLI interface
+        modified_source_path = tmp_dir / source_file_name.name
+        # NOTE: newline='' disables newline conversion.
+        # We want the file exactly as is because changing even a single byte in the source affects metadata.
+        with open(modified_source_path, 'w', encoding='utf8', newline='') as modified_source_file:
+            modified_source_file.write(compiler_input)
+
+        process = subprocess.run(
+            command_line,
+            cwd=tmp_dir,
+            encoding='utf8',
+            capture_output=True,
+            check=False,
+        )
+
+        return parse_cli_output(Path(source_file_name), process.stdout)
+
+
+def generate_report(source_file_names: List[str], compiler_path: Path, interface: CompilerInterface):
     with open('report.txt', mode='w', encoding='utf8', newline='\n') as report_file:
         for optimize in [False, True]:
-            for source_file_name in sorted(source_file_names):
-                try:
-                    report = run_compiler(Path(compiler_path), Path(source_file_name), optimize)
-                    report_file.write(report.format_report())
-                except subprocess.CalledProcessError as exception:
-                    print(
-                        f"\n\nInterrupted by an exception while processing file "
-                        f"'{source_file_name}' with optimize={optimize}\n\n"
-                        f"COMPILER STDOUT:\n{exception.stdout}\n"
-                        f"COMPILER STDERR:\n{exception.stderr}\n",
-                        file=sys.stderr
-                    )
-                    raise
-                except:
-                    print(
-                        f"\n\nInterrupted by an exception while processing file "
-                        f"'{source_file_name}' with optimize={optimize}\n",
-                        file=sys.stderr
-                    )
-                    raise
+            with TemporaryDirectory(prefix='prepare_report-') as tmp_dir:
+                for source_file_name in sorted(source_file_names):
+                    try:
+                        report = run_compiler(compiler_path, Path(source_file_name), optimize, interface, Path(tmp_dir))
+                        report_file.write(report.format_report())
+                    except subprocess.CalledProcessError as exception:
+                        print(
+                            f"\n\nInterrupted by an exception while processing file "
+                            f"'{source_file_name}' with optimize={optimize}\n\n"
+                            f"COMPILER STDOUT:\n{exception.stdout}\n"
+                            f"COMPILER STDERR:\n{exception.stderr}\n",
+                            file=sys.stderr
+                        )
+                        raise
+                    except:
+                        print(
+                            f"\n\nInterrupted by an exception while processing file "
+                            f"'{source_file_name}' with optimize={optimize}\n",
+                            file=sys.stderr
+                        )
+                        raise
 
 
 def commandline_parser() -> ArgumentParser:
@@ -140,6 +236,13 @@ def commandline_parser() -> ArgumentParser:
 
     parser = ArgumentParser(description=script_description)
     parser.add_argument(dest='compiler_path', help="Solidity compiler executable")
+    parser.add_argument(
+        '--interface',
+        dest='interface',
+        default=CompilerInterface.STANDARD_JSON.value,
+        choices=[c.value for c in CompilerInterface],
+        help="Compiler interface to use."
+    )
     return parser;
 
 
@@ -148,4 +251,5 @@ def commandline_parser() -> ArgumentParser:
     generate_report(
         glob("*.sol"),
         Path(options.compiler_path),
+        CompilerInterface(options.interface),
     )

test/scripts/fixtures/library_inherited2_sol_cli_output.txt

@@ -0,0 +1,24 @@
+Warning: SPDX license identifier not provided in source file. Before publishing, consider adding a comment containing "SPDX-License-Identifier: <SPDX-License>" to each source file. Use "SPDX-License-Identifier: UNLICENSED" for non-open-source code. Please see https://spdx.org for more information.
+--> syntaxTests/scoping/library_inherited2.sol
+
+Warning: Source file does not specify required compiler version! Consider adding "pragma solidity ^0.8.0;"
+--> syntaxTests/scoping/library_inherited2.sol
+
+
+======= syntaxTests/scoping/library_inherited2.sol:A =======
+Binary:
+6080604052348015600f57600080fd5b50603f80601d6000396000f3fe6080604052600080fdfea264697066735822122086e727f29d40b264a19bbfcad38d64493dca4bab5dbba8c82ffdaae389d2bba064736f6c63430008000033
+Metadata:
+{"compiler":{"version":"0.8.0+commit.c7dfd78e"},"language":"Solidity","output":{"abi":[],"devdoc":{"kind":"dev","methods":{},"version":1},"userdoc":{"kind":"user","methods":{},"version":1}},"settings":{"compilationTarget":{"syntaxTests/scoping/library_inherited2.sol":"A"},"evmVersion":"istanbul","libraries":{},"metadata":{"bytecodeHash":"ipfs"},"optimizer":{"enabled":false,"runs":200},"remappings":[]},"sources":{"syntaxTests/scoping/library_inherited2.sol":{"keccak256":"0xd0619f00638fdfea187368965615dbd599fead93dd14b6558725e85ec7011d96","urls":["bzz-raw://ec7af066be66a223f0d25ba3bf9ba6dc103e1a57531a66a38a5ca2b6ce172f55","dweb:/ipfs/QmW1NrqQNhnY1Tkgr3Z9oM8buCGLUJCJVCDTVejJTT5Vet"]}},"version":1}
+
+======= syntaxTests/scoping/library_inherited2.sol:B =======
+Binary:
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
+Metadata:
+{"compiler":{"version":"0.8.0+commit.c7dfd78e"},"language":"Solidity","output":{"abi":[{"inputs":[{"internalType":"uint256","name":"value","type":"uint256"}],"name":"bar","outputs":[{"internalType":"uint256","name":"","type":"uint256"}],"stateMutability":"pure","type":"function"}],"devdoc":{"kind":"dev","methods":{},"version":1},"userdoc":{"kind":"user","methods":{},"version":1}},"settings":{"compilationTarget":{"syntaxTests/scoping/library_inherited2.sol":"B"},"evmVersion":"istanbul","libraries":{},"metadata":{"bytecodeHash":"ipfs"},"optimizer":{"enabled":false,"runs":200},"remappings":[]},"sources":{"syntaxTests/scoping/library_inherited2.sol":{"keccak256":"0xd0619f00638fdfea187368965615dbd599fead93dd14b6558725e85ec7011d96","urls":["bzz-raw://ec7af066be66a223f0d25ba3bf9ba6dc103e1a57531a66a38a5ca2b6ce172f55","dweb:/ipfs/QmW1NrqQNhnY1Tkgr3Z9oM8buCGLUJCJVCDTVejJTT5Vet"]}},"version":1}
+
+======= syntaxTests/scoping/library_inherited2.sol:Lib =======
+Binary:
+60566050600b82828239805160001a6073146043577f4e487b7100000000000000000000000000000000000000000000000000000000600052600060045260246000fd5b30600052607381538281f3fe73000000000000000000000000000000000000000030146080604052600080fdfea26469706673582212207f9515e2263fa71a7984707e2aefd82241fac15c497386ca798b526f14f8ba6664736f6c63430008000033
+Metadata:
+{"compiler":{"version":"0.8.0+commit.c7dfd78e"},"language":"Solidity","output":{"abi":[],"devdoc":{"kind":"dev","methods":{},"version":1},"userdoc":{"kind":"user","methods":{},"version":1}},"settings":{"compilationTarget":{"syntaxTests/scoping/library_inherited2.sol":"Lib"},"evmVersion":"istanbul","libraries":{},"metadata":{"bytecodeHash":"ipfs"},"optimizer":{"enabled":false,"runs":200},"remappings":[]},"sources":{"syntaxTests/scoping/library_inherited2.sol":{"keccak256":"0xd0619f00638fdfea187368965615dbd599fead93dd14b6558725e85ec7011d96","urls":["bzz-raw://ec7af066be66a223f0d25ba3bf9ba6dc103e1a57531a66a38a5ca2b6ce172f55","dweb:/ipfs/QmW1NrqQNhnY1Tkgr3Z9oM8buCGLUJCJVCDTVejJTT5Vet"]}},"version":1}

test/scripts/fixtures/unknown_pragma_sol_cli_output.txt

@@ -0,0 +1,11 @@
+Warning: SPDX license identifier not provided in source file. Before publishing, consider adding a comment containing "SPDX-License-Identifier: <SPDX-License>" to each source file. Use "SPDX-License-Identifier: UNLICENSED" for non-open-source code. Please see https://spdx.org for more information.
+--> syntaxTests/pragma/unknown_pragma.sol
+
+Error: Unknown pragma "thisdoesntexist"
+--> syntaxTests/pragma/unknown_pragma.sol:1:1:
+|
+1 | pragma thisdoesntexist;
+| ^^^^^^^^^^^^^^^^^^^^^^^
+
+Warning: Source file does not specify required compiler version! Consider adding "pragma solidity ^0.8.0;"
+--> syntaxTests/pragma/unknown_pragma.sol