#!/usr/bin/env python3
# Copyright 2011 The Chromium Authors
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.
"""Implements a simple "negative compile" test for C++ on linux.

Sometimes a C++ API needs to ensure that various usages cannot compile. To
enable unittesting of these assertions, we use this python script to
invoke the compiler on a source file and assert that compilation fails.

For more info, see:
  http://dev.chromium.org/developers/testing/no-compile-tests
"""

import argparse
import ast
import concurrent.futures
import functools
import io
import os
import re
import subprocess
import sys
import tempfile
import time

from typing import Any
from typing import IO
from typing import Optional
from typing import Sequence
from typing import Set
from typing import Tuple
from typing import TypedDict

sys.path.append(
    os.path.join(os.path.dirname(__file__), os.pardir, os.pardir, 'build'))
import action_helpers

# Matches lines that start with #if and have the substring TEST in the
# conditional. Also extracts the comment.  This allows us to search for
# lines like the following:
#
#   #ifdef NCTEST_NAME_OF_TEST  // [r'expected output']
#   #if defined(NCTEST_NAME_OF_TEST)  // [r'expected output']
#   #if NCTEST_NAME_OF_TEST  // [r'expected output']
#   #elif NCTEST_NAME_OF_TEST  // [r'expected output']
#   #elif DISABLED_NCTEST_NAME_OF_TEST  // [r'expected output']
#
# inside the unittest file.
NCTEST_CONFIG_RE = re.compile(r'^#(?:el)?if.*\s+(\S*NCTEST\S*)\s*(//.*)?')

# Matches and removes the defined() preprocesor predicate. This is useful
# for test cases that use the preprocessor if-statement form:
#
#   #if defined(NCTEST_NAME_OF_TEST)
#
# Should be used to post-process the results found by NCTEST_CONFIG_RE.
STRIP_DEFINED_RE = re.compile(r'defined\((.*)\)')

# Used to grab the expectation from comment at the end of an #ifdef.  See
# NCTEST_CONFIG_RE's comment for examples of what the format should look like.
#
# The extracted substring should be a python array of regular expressions.
EXTRACT_EXPECTATION_RE = re.compile(r'//\s*(\[.*\])')

# The header for the result file so that it can be compiled.
RESULT_FILE_HEADER = """
// This file is generated by the no compile test from:
//   %s

#include "base/logging.h"
#include "testing/gtest/include/gtest/gtest.h"

"""

# The log message on a test completion.
LOG_TEMPLATE = """
TEST(%s, %s) took %f secs. Started at %f, ended at %f.
"""

# The GUnit test function to output for a successful or disabled test.
GUNIT_TEMPLATE = """
TEST(%s, %s) { }
"""

# How long a nocompile test should be able to run for before timing out.
NCTEST_TERMINATE_TIMEOUT_SEC = 120


class TestResult(TypedDict):
  """Represents the result of one nocompile test.

  Attributes:
    cmdline: The executed command line.
    stdout: A temporary file object containing stdout.
    name: The name of the test.
    suite_name: The suite name to use when generating the gunit test result.
    started_at: A timestamp in seconds since the epoch for when this test was
      started.
    aborted_at: A timestamp in seconds since the epoch for when this test was
      aborted. If the test completed successfully, this value is 0.
    finished_at: A timestamp in seconds since the epoch for when this test was
      successfully complete.  If the test is aborted, this value is 0.
    expectations: A dictionary with the test expectations. See
      ParseExpectation() for the structure.
    return_code: The return code of the test process, if not aborted.
  """
  cmdline: str
  stdout: IO[str]
  stderr: IO[str]
  name: str
  suite_name: str
  started_at: float
  aborted_at: float
  finished_at: float
  expectations: Sequence[re.Pattern]
  returncode: int


def ValidateInput(compiler, parallelism, sourcefile_path, cflags,
                  resultfile_path):
  """Make sure the arguments being passed in are sane."""
  assert os.path.isfile(compiler)
  assert parallelism >= 1
  assert type(sourcefile_path) is str
  assert type(cflags) is list
  for flag in cflags:
    assert type(flag) is str
  assert type(resultfile_path) is str


def ParseExpectation(expectation_string) -> Sequence[re.Pattern]:
  """Extracts expectation definition from the trailing comment on the ifdef.

  See the comment on NCTEST_CONFIG_RE for examples of the format we are parsing.

  Args:
    expectation_string: A string like "// [r'some_regex']"

  Returns:
    A list of compiled regular expressions indicating all possible valid
    compiler outputs.  If the list is empty, all outputs are considered valid.
  """
  assert expectation_string is not None

  match = EXTRACT_EXPECTATION_RE.match(expectation_string)
  assert match

  raw_expectation = ast.literal_eval(match.group(1))
  assert type(raw_expectation) is list

  expectation = []
  for regex_str in raw_expectation:
    assert type(regex_str) is str
    expectation.append(re.compile(regex_str))
  return expectation


def ExtractTestConfigs(sourcefile_path, suite_name, resultfile, resultlog):
  """Parses the source file for test configurations.

  Each no-compile test in the file is separated by an ifdef macro.  We scan
  the source file with the NCTEST_CONFIG_RE to find all ifdefs that look like
  they demark one no-compile test and try to extract the test configuration
  from that.

  Args:
    sourcefile_path: The path to the source file.
    suite_name: The name of the test suite.
    resultfile: File object for .cc file that results are written to.
    resultlog: File object for the log file.

  Returns:
    A list of test configurations, excluding tests prefixed with DISABLED_. Each
    test configuration is a dictionary of the form:

      { name: 'NCTEST_NAME'
        suite_name: 'SOURCE_FILE_NAME'
        expectations: [re.Pattern, re.Pattern] }

    The |suite_name| is used to generate a pretty gtest output on successful
    completion of the no compile test.

    The compiled regexps in |expectations| define the valid outputs of the
    compiler.  If any one of the listed patterns matches either the stderr or
    stdout from the compilation, and the compilation failed, then the test is
    considered to have succeeded.  If the list is empty, than we ignore the
    compiler output and just check for failed compilation. If |expectations|
    is actually None, then this specifies a compiler sanity check test, which
    should expect a SUCCESSFUL compilation.
  """
  with open(sourcefile_path, 'r', encoding='utf-8') as sourcefile:
    # Start with a compiler smoke test. This is important to show that compiler
    # flags and configuration are not just wrong. Otherwise, having a
    # misconfigured compiler, or an error in the shared portions of the .nc file
    # would cause all tests to erroneously pass.
    test_configs = [{
        'name': 'NCTEST_SMOKE',
        'suite_name': suite_name,
        'expectations': None,
    }]

    for line in sourcefile:
      match_result = NCTEST_CONFIG_RE.match(line)
      if not match_result:
        continue

      groups = match_result.groups()

      # Grab the name and remove the defined() predicate if there is one.
      name = groups[0]
      strip_result = STRIP_DEFINED_RE.match(name)
      if strip_result:
        name = strip_result.group(1)

      config = {
          'name': name,
          'suite_name': suite_name,
          'expectations': ParseExpectation(groups[1])
      }

      if config['name'].startswith('DISABLED_'):
        PassTest(resultfile, resultlog, config)
        continue

      test_configs.append(config)
  return test_configs


def RunTest(compiler, tempfile_dir, cflags, config) -> TestResult:
  """Runs one negative compile test.

  Args:
    compiler: The path to the compiler.
    tempfile_dir: A directory to store temporary data from tests.
    cflags: An array of strings with all the CFLAGS to give to gcc.
    config: A dictionary describing the test.  See ExtractTestConfigs
      for a description of the config format.

  Returns:
    A TestResult containing all the information about the started test.
  """
  cmdline = [compiler]
  cmdline.extend(cflags)
  name = config['name']
  expectations = config['expectations']
  if expectations is not None:
    cmdline.append('-D%s' % name)
  test_stdout = tempfile.TemporaryFile(dir=tempfile_dir,
                                       mode='w+',
                                       encoding='utf-8')
  test_stderr = tempfile.TemporaryFile(dir=tempfile_dir,
                                       mode='w+',
                                       encoding='utf-8')

  # Note: this could use pipes, but having temp files might make for somewhat
  # better debuggability.
  try:
    started_at = time.time()
    returncode = subprocess.run(cmdline,
                                stdout=test_stdout,
                                stderr=test_stderr,
                                timeout=NCTEST_TERMINATE_TIMEOUT_SEC).returncode
    aborted_at = 0
    finished_at = time.time()
  except subprocess.CalledProcessError:
    returncode = -1
    aborted_at = time.time()
    finished_at = 0
  return TestResult(cmdline=' '.join(cmdline),
                    stdout=test_stdout,
                    stderr=test_stderr,
                    name=name,
                    suite_name=config['suite_name'],
                    started_at=started_at,
                    aborted_at=aborted_at,
                    finished_at=finished_at,
                    expectations=expectations,
                    returncode=returncode)


def PassTest(resultfile, resultlog, test):
  """Logs the result of a test run with RunTest(), or a disabled test
  configuration.

  Args:
    resultfile: File object for .cc file that results are written to.
    resultlog: File object for the log file.
    test: An instance of the dictionary returned by RunTest(), a
          configuration from ExtractTestConfigs().
  """
  resultfile.write(GUNIT_TEMPLATE % (test['suite_name'], test['name']))

  # The 'started_at' key is only added if a test has been started.
  if 'started_at' in test:
    resultlog.write(
        LOG_TEMPLATE %
        (test['suite_name'], test['name'], test['finished_at'] -
         test['started_at'], test['started_at'], test['finished_at']))


def FailTest(resultfile, test, error, stdout=None, stderr=None):
  """Logs the result of a test run with by RunTest()

  Args:
    resultfile: File object for .cc file that results are written to.
    test: An instance of the dictionary returned by StartTest()
    error: The printable reason for the failure.
    stdout: The test's output to stdout.
    stderr: The test's output to stderr.
  """
  resultfile.write('\n')
  resultfile.write('#error %s Failed: %s\n' % (test['name'], error))
  resultfile.write('#error compile line: %s\n' % test['cmdline'])
  if stdout and len(stdout) != 0:
    resultfile.write('#error %s stdout:\n' % test['name'])
    for line in stdout.split('\n'):
      resultfile.write('#error "  %s:"\n' % line)

  if stderr and len(stderr) != 0:
    resultfile.write('#error %s stderr:"\n' % test['name'])
    for line in stderr.split('\n'):
      resultfile.write('#error "  %s"\n' % line)


def WriteStats(resultlog, suite_name, timings):
  """Logs the peformance timings for each stage of the script.

  Args:
    resultlog: File object for the log file.
    suite_name: The name of the GUnit suite this test belongs to.
    timings: Dictionary with timestamps for each stage of the script run.
  """
  stats_template = """
TEST(%s): Started %f, Ended %f, Total %fs, Extract %fs, Compile %fs, Process %fs
"""
  total_secs = timings['results_processed'] - timings['started']
  extract_secs = timings['extract_done'] - timings['started']
  compile_secs = timings['compile_done'] - timings['extract_done']
  process_secs = timings['results_processed'] - timings['compile_done']
  resultlog.write(stats_template %
                  (suite_name, timings['started'], timings['results_processed'],
                   total_secs, extract_secs, compile_secs, process_secs))


def ExtractTestOutputAndCleanup(test: TestResult) -> Tuple[str, str]:
  """Test output is in temp files. Read those and delete them.
  Returns: A tuple (stderr, stdout).
  """
  def ReadStreamAndClose(stream: IO[str]) -> str:
    with stream:
      stream.seek(0)
      return stream.read()

  return (ReadStreamAndClose(test['stdout']),
          ReadStreamAndClose(test['stderr']))


def ProcessTestResult(sourcefile_path: str,
                      resultfile: IO[str],
                      resultlog: IO[str],
                      test: TestResult,
                      includes: Optional[Set[str]] = None) -> None:
  """Interprets and logs the result of a test run by RunTest()

  Args:
    sourcefile_path: Path to the source .cc file derived from the .nc file.
    resultfile: File object for .cc file that results are written to.
    resultlog: File object for the log file.
    test: The dictionary from RunTest() to process.
    includes: Must either be a set or None. If a set, the driver will scrape
              stdout for the /showIncludes format and insert any headers found
              into `includes`.
  """
  (stdout, stderr) = ExtractTestOutputAndCleanup(test)

  if includes is not None:
    # /showIncludes format:
    # Note: including file: third_party/libc++/src/include/stdio.h
    # Note: including file:  third_party/libc++/src/include/__config
    # Note: including file:   buildtools/third_party/libc++/__config_site
    # Note: including file: third_party/libc++/src/include/stdint.h
    INCLUDE_PREFIX = 'Note: including file: '
    includes.update(
        map(
            lambda x: os.path.relpath(x[len(INCLUDE_PREFIX):].strip()),
            filter(
                lambda x: x.startswith(INCLUDE_PREFIX),
                stdout.splitlines(),
            ),
        ))

  if test['aborted_at'] != 0:
    FailTest(
        resultfile, test, "Compile timed out. Started %f ended %f." %
        (test['started_at'], test['aborted_at']))
    return

  if test['returncode'] == 0:
    # Handle failure due to successful compile.
    FailTest(resultfile, test, 'Unexpected successful compilation.', stdout,
             stderr)
    return
  else:
    # Check the output has the right expectations.  If there are no
    # expectations, then we just consider the output "matched" by default.
    if len(test['expectations']) == 0:
      PassTest(resultfile, resultlog, test)
      return

    # Otherwise test against all expectations.
    for regexp in test['expectations']:
      if (regexp.search(stdout) is not None
          or regexp.search(stderr) is not None):
        PassTest(resultfile, resultlog, test)
        return
    expectation_str = ', '.join(
        ["r'%s'" % regexp.pattern for regexp in test['expectations']])
    FailTest(resultfile, test,
             'Expectations [%s] did not match output.' % expectation_str,
             stdout, stderr)
    return


def main():
  parser = argparse.ArgumentParser(prog=sys.argv[0])
  parser.add_argument('compiler')
  parser.add_argument('parallelism', type=int)
  parser.add_argument('sourcefile')
  parser.add_argument('resultfile')
  parser.add_argument('--depfile', default='')
  parser.add_argument('compiler_options', nargs=argparse.REMAINDER)

  # Force us into the "C" locale so the compiler doesn't localize its output.
  # In particular, this stops gcc from using smart quotes when in english UTF-8
  # locales.  This makes the expectation writing much easier.
  os.environ['LC_ALL'] = 'C'

  args = parser.parse_args()
  compiler = args.compiler
  parallelism = args.parallelism
  sourcefile_path = args.sourcefile
  resultfile_path = args.resultfile
  cflags = args.compiler_options

  timings = {'started': time.time()}

  ValidateInput(compiler, parallelism, sourcefile_path, cflags, resultfile_path)

  # Convert filename from underscores to CamelCase.
  words = os.path.splitext(os.path.basename(sourcefile_path))[0].split('_')
  words = [w.capitalize() for w in words]
  suite_name = 'NoCompile' + ''.join(words)

  with io.StringIO() as resultfile, io.StringIO() as resultlog:
    resultfile.write(RESULT_FILE_HEADER % sourcefile_path)

    test_configs = ExtractTestConfigs(sourcefile_path, suite_name, resultfile,
                                      resultlog)
    timings['extract_done'] = time.time()

    # Run the no-compile tests, but ensure we do not run more than |parallelism|
    # tests at once.
    timings['header_written'] = time.time()
    finished_tests = []

    includes = set() if args.depfile else None

    with concurrent.futures.ThreadPoolExecutor(
        max_workers=parallelism) as executor:
      finished_tests = executor.map(
          functools.partial(RunTest, compiler, os.path.dirname(resultfile_path),
                            cflags), test_configs)

      timings['compile_done'] = time.time()

      finished_tests = sorted(finished_tests, key=lambda test: test['name'])
      for test in finished_tests:
        if test['name'] == 'NCTEST_SMOKE':
          (stdout, stderr) = ExtractTestOutputAndCleanup(test)
          return_code = test['returncode']
          if return_code != 0:
            sys.stdout.write(stdout)
            sys.stderr.write(stderr)
          continue
        ProcessTestResult(sourcefile_path, resultfile, resultlog, test,
                          includes)
      timings['results_processed'] = time.time()

    WriteStats(resultlog, suite_name, timings)

    with open(resultfile_path + '.log', 'w') as fd:
      fd.write(resultlog.getvalue())
    if return_code == 0:
      with open(resultfile_path, 'w') as fd:
        fd.write(resultfile.getvalue())

      # Even if includes is empty, write a depfile if it was requested.
      if args.depfile:
        action_helpers.write_depfile(args.depfile, resultfile_path, includes)

    if return_code != 0:
      print("No-compile driver failure with return_code %d. Result log:" %
            return_code)
      print(resultlog.getvalue())

    sys.exit(return_code)


if __name__ == '__main__':
  main()