Add bytecode test for multiple sources compiler at same time

Co-authored-by: Kamil Śliwak <[email protected]>

Author: r0qs

scripts/common/cmdline_helpers.py

@@ -0,0 +1,67 @@
+import os
+import subprocess
+from pathlib import Path
+from shutil import rmtree
+from tempfile import mkdtemp
+from textwrap import dedent
+from typing import List
+from typing import Optional
+
+from bytecodecompare.prepare_report import FileReport
+from bytecodecompare.prepare_report import parse_cli_output
+
+
+DEFAULT_PREAMBLE = dedent("""
+    // SPDX-License-Identifier: UNLICENSED
+    pragma solidity >=0.0;
+""")
+
+
+def inside_temporary_dir(prefix):
+    """
+    Creates a temporary directory, enters the directory and executes the function inside it.
+    Restores the previous working directory after executing the function.
+    """
+    def tmp_dir_decorator(fn):
+        previous_dir = os.getcwd()
+        def f(*args, **kwargs):
+            try:
+                tmp_dir = mkdtemp(prefix=prefix)
+                os.chdir(tmp_dir)
+                result = fn(*args, **kwargs)
+                rmtree(tmp_dir)
+                return result
+            finally:
+                os.chdir(previous_dir)
+        return f
+    return tmp_dir_decorator
+
+
+def solc_bin_report(solc_binary: str, input_files: List[Path], via_ir: bool) -> FileReport:
+    """
+    Runs the solidity compiler binary
+    """
+
+    output = subprocess.check_output(
+        [solc_binary, '--bin'] +
+        input_files +
+        (['--via-ir'] if via_ir else []),
+        encoding='utf8',
+    )
+    return parse_cli_output('', output)
+
+
+def save_bytecode(bytecode_path: Path, reports: FileReport, contract: Optional[str] = None):
+    with open(bytecode_path, 'w', encoding='utf8') as f:
+        for report in reports.contract_reports:
+            if contract is None or report.contract_name == contract:
+                bytecode = report.bytecode if report.bytecode is not None else '<NO BYTECODE>'
+                f.write(f'{report.contract_name}: {bytecode}\n')
+
+
+def add_preamble(source_path: Path, preamble: str = DEFAULT_PREAMBLE):
+    for source in source_path.glob('**/*.sol'):
+        with open(source, 'r+', encoding='utf8') as f:
+            content = f.read()
+            f.seek(0, 0)
+            f.write(preamble + content)

scripts/common/git_helpers.py

@@ -1,4 +1,6 @@
 import subprocess
+from pathlib import Path
+from shutil import which
 
 
 def run_git_command(command):
@@ -17,3 +19,21 @@ def git_current_branch():
 
 def git_commit_hash(ref: str = 'HEAD'):
     return run_git_command(['git', 'rev-parse', '--verify', ref])
+
+
+def git_diff(file_a: Path, file_b: Path) -> int:
+    if which('git') is None:
+        raise RuntimeError('git not found.')
+
+    return subprocess.run([
+        'git',
+        'diff',
+        '--color',
+        '--word-diff=plain',
+        '--word-diff-regex=.',
+        '--ignore-space-change',
+        '--ignore-blank-lines',
+        '--exit-code',
+        file_a,
+        file_b,
+    ], encoding='utf8', check=False).returncode

scripts/splitSources.py

@@ -13,9 +13,6 @@
 import os
 import traceback
 
-hasMultipleSources = False
-createdSources = []
-
 
 def uncaught_exception_hook(exc_type, exc_value, exc_traceback):
     # The script `scripts/ASTImportTest.sh` will interpret return code 3
@@ -43,41 +40,39 @@ def writeSourceToFile(lines):
     if filePath:
         os.system("mkdir -p " + filePath)
     with open(srcName, mode='a+', encoding='utf8', newline='') as f:
-        createdSources.append(srcName)
         for idx, line in enumerate(lines[1:]):
             # write to file
             if line[:12] != "==== Source:":
                 f.write(line + '\n')
 
             # recursive call if there is another source
             else:
-                writeSourceToFile(lines[1+idx:])
-                break
+                return [srcName] + writeSourceToFile(lines[1+idx:])
 
+    return [srcName]
 
-if __name__ == '__main__':
-    filePath = sys.argv[1]
+def split_sources(filePath, suppress_output = False):
     sys.excepthook = uncaught_exception_hook
 
     try:
         # decide if file has multiple sources
         with open(filePath, mode='r', encoding='utf8', newline='') as f:
             lines = f.read().splitlines()
         if len(lines) >= 1 and lines[0][:12] == "==== Source:":
-            hasMultipleSources = True
-            writeSourceToFile(lines)
-
-        if hasMultipleSources:
             srcString = ""
-            for src in createdSources:
+            for src in writeSourceToFile(lines):
                 srcString += src + ' '
-            print(srcString)
-            sys.exit(0)
-        else:
-            sys.exit(1)
+            if not suppress_output:
+                print(srcString)
+            return 0
+        return 1
 
     except UnicodeDecodeError as ude:
         print("UnicodeDecodeError in '" + filePath + "': " + str(ude))
         print("This is expected for some tests containing invalid utf8 sequences. "
               "Exception will be ignored.")
-        sys.exit(2)
+        return 2
+
+
+if __name__ == '__main__':
+    sys.exit(split_sources(sys.argv[1]))

test/cmdlineTests/~bytecode_equivalence_independent_of_import_discovery_order/inputs.sol

@@ -0,0 +1,29 @@
+==== Source: A.sol ====
+import "@/D.sol";
+import "B.sol";
+
+contract A is B {
+    function a() public pure {
+        e();
+    }
+}
+
+==== Source: B.sol ====
+import "C.sol";
+
+abstract contract B is C {}
+
+==== Source: C.sol ====
+abstract contract C {
+    function c() public pure returns (uint) {
+        return 0;
+    }
+}
+
+==== Source: @/D.sol ====
+import "@/E.sol";
+
+==== Source: @/E.sol ====
+function e() pure returns (bytes memory returndata) {
+    return "";
+}

test/cmdlineTests/~bytecode_equivalence_independent_of_import_discovery_order/test.py

@@ -0,0 +1,43 @@
+#!/usr/bin/env python3
+
+import os
+import sys
+from pathlib import Path
+from textwrap import dedent
+
+# pylint: disable=wrong-import-position
+PROJECT_ROOT = Path(__file__).parents[3]
+sys.path.insert(0, str(PROJECT_ROOT / 'scripts'))
+
+from common.cmdline_helpers import add_preamble
+from common.cmdline_helpers import inside_temporary_dir
+from common.cmdline_helpers import save_bytecode
+from common.cmdline_helpers import solc_bin_report
+from common.git_helpers import git_diff
+from splitSources import split_sources
+
+
+@inside_temporary_dir(Path(__file__).parent.name)
+def test_bytecode_equivalence():
+    source_file_path = Path(__file__).parent / 'inputs.sol'
+    split_sources(source_file_path, suppress_output=True)
+    add_preamble(Path.cwd())
+
+    solc_binary = os.environ.get('SOLC')
+    if solc_binary is None:
+        raise RuntimeError(dedent("""\
+            `solc` compiler not found.
+            Please ensure you set the SOLC environment variable
+            with the correct path to the compiler's binary.
+        """))
+
+    # Whether a file is passed to the compiler explicitly or only discovered when traversing imports
+    # may affect the order in which files are processed and result in different AST IDs.
+    # This, however, must not result in different bytecode being generated.
+    save_bytecode(Path('A.bin'), solc_bin_report(solc_binary, [Path('A.sol')], via_ir=True))
+    save_bytecode(Path('AB.bin'), solc_bin_report(solc_binary, [Path('A.sol'), Path('B.sol')], via_ir=True))
+    return git_diff(Path('A.bin'), Path('AB.bin'))
+
+
+if __name__ == '__main__':
+    sys.exit(test_bytecode_equivalence())