SERVER-18575: clang_format.py

3 files changed, 627 insertions, 1 deletions

diff --git a/buildscripts/clang_format.py b/buildscripts/clang_format.py
new file mode 100755
index 00000000000..f0473c61c01
--- /dev/null
+++ b/buildscripts/clang_format.py
@@ -0,0 +1,592 @@
+#! /usr/bin/env python
+"""
+A script that provides:
+1. Ability to grab binaries where possible from LLVM.
+2. Ability to download binaries from MongoDB cache for clang-format.
+3. Validates clang-format is the right version.
+4. Has support for checking which files are to be checked.
+5. Supports validating and updating a set of files to the right coding style.
+"""
+from __future__ import print_function, absolute_import
+
+import Queue
+import difflib
+import itertools
+import os
+import re
+import shutil
+import string
+import subprocess
+import sys
+import tempfile
+import threading
+import time
+import urllib
+from distutils import spawn
+from optparse import OptionParser
+from multiprocessing import cpu_count
+
+# Get relative imports to work when the package is not installed on the PYTHONPATH.
+if __name__ == "__main__" and __package__ is None:
+    sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from buildscripts.resmokelib.utils import globstar
+from buildscripts import moduleconfig
+
+
+##############################################################################
+#
+# Constants for clang-format
+#
+#
+
+# Expected version of clang-format
+CLANG_FORMAT_VERSION = "3.6.0"
+
+# Name of clang-format as a binary
+CLANG_FORMAT_PROGNAME = "clang-format"
+
+# URL location of the "cached" copy of clang-format to download
+# for users which do not have clang-format installed
+CLANG_FORMAT_HTTP_LINUX_CACHE = "https://s3.amazonaws.com/boxes.10gen.com/build/clang-format-rhel55.tar.gz"
+
+# URL on LLVM's website to download the clang tarball
+CLANG_FORMAT_SOURCE_URL_BASE = string.Template("http://llvm.org/releases/$version/clang+llvm-$version-$llvm_distro.tar.xz")
+
+# Path in the tarball to the clang-format binary
+CLANG_FORMAT_SOURCE_TAR_BASE = string.Template("clang+llvm-$version-$tar_path/bin/" + CLANG_FORMAT_PROGNAME)
+
+# Path to the modules in the mongodb source tree
+# Has to match the string in SConstruct
+MODULE_DIR = "src/mongo/db/modules"
+
+# Copied from python 2.7 version of subprocess.py
+# Exception classes used by this module.
+class CalledProcessError(Exception):
+    """This exception is raised when a process run by check_call() or
+    check_output() returns a non-zero exit status.
+    The exit status will be stored in the returncode attribute;
+    check_output() will also store the output in the output attribute.
+    """
+    def __init__(self, returncode, cmd, output=None):
+        self.returncode = returncode
+        self.cmd = cmd
+        self.output = output
+    def __str__(self):
+        return ("Command '%s' returned non-zero exit status %d with output %s" %
+            (self.cmd, self.returncode, self.output))
+
+
+# Copied from python 2.7 version of subprocess.py
+def check_output(*popenargs, **kwargs):
+    r"""Run command with arguments and return its output as a byte string.
+
+    If the exit code was non-zero it raises a CalledProcessError.  The
+    CalledProcessError object will have the return code in the returncode
+    attribute and output in the output attribute.
+
+    The arguments are the same as for the Popen constructor.  Example:
+
+    >>> check_output(["ls", "-l", "/dev/null"])
+    'crw-rw-rw- 1 root root 1, 3 Oct 18  2007 /dev/null\n'
+
+    The stdout argument is not allowed as it is used internally.
+    To capture standard error in the result, use stderr=STDOUT.
+
+    >>> check_output(["/bin/sh", "-c",
+    ...               "ls -l non_existent_file ; exit 0"],
+    ...              stderr=STDOUT)
+    'ls: non_existent_file: No such file or directory\n'
+    """
+    if 'stdout' in kwargs:
+        raise ValueError('stdout argument not allowed, it will be overridden.')
+    process = subprocess.Popen(stdout=subprocess.PIPE, *popenargs, **kwargs)
+    output, unused_err = process.communicate()
+    retcode = process.poll()
+    if retcode:
+        cmd = kwargs.get("args")
+        if cmd is None:
+            cmd = popenargs[0]
+        raise CalledProcessError(retcode, cmd, output)
+    return output
+
+def callo(args):
+    """Call a program, and capture its output
+    """
+    return check_output(args)
+
+def get_llvm_url(version, llvm_distro):
+    """Get the url to download clang-format from llvm.org
+    """
+    return CLANG_FORMAT_SOURCE_URL_BASE.substitute(
+        version=version,
+        llvm_distro=llvm_distro)
+
+def get_tar_path(version, tar_path):
+    """ Get the path to clang-format in the llvm tarball
+    """
+    return CLANG_FORMAT_SOURCE_TAR_BASE.substitute(
+        version=version,
+        tar_path=tar_path)
+
+def get_clang_format_from_llvm(llvm_distro, tar_path, dest_file):
+    """Download clang-format from llvm.org, unpack the tarball,
+    and put clang-format in the specified place
+    """
+    # Build URL
+    url = get_llvm_url(CLANG_FORMAT_VERSION, llvm_distro)
+
+    dest_dir = tempfile.gettempdir()
+    temp_tar_file = os.path.join(dest_dir, "temp.tar.xz")
+
+    # Download from LLVM
+    print("Downloading clang-format %s from %s, saving to %s" % (CLANG_FORMAT_VERSION,
+            url, temp_tar_file))
+    urllib.urlretrieve(url, temp_tar_file)
+
+    # Extract just clang format file
+    subprocess.call(['tar', 'zxvf', temp_tar_file, '*clang-format*'])
+
+    # Destination Path
+    shutil.move(get_tar_path(CLANG_FORMAT_VERSION, tar_path), dest_file)
+
+def get_clang_format_from_linux_cache(dest_file):
+    """Get clang-format from mongodb's cache
+    """
+    # Get URL
+    url = CLANG_FORMAT_HTTP_LINUX_CACHE
+
+    dest_dir = tempfile.gettempdir()
+    temp_tar_file = os.path.join(dest_dir, "temp.tar.xz")
+
+    # Download the file
+    print("Downloading clang-format %s from %s, saving to %s" % (CLANG_FORMAT_VERSION,
+            url, temp_tar_file))
+    urllib.urlretrieve(url, temp_tar_file)
+
+    # Extract just clang format file
+    subprocess.call(['tar', 'zxvf', temp_tar_file, '*clang-format*'])
+
+    # Destination Path
+    shutil.move("llvm/Release/bin/clang-format", dest_file)
+
+
+class ClangFormat(object):
+    """Class encapsulates finding a suitable copy of clang-format,
+    and linting/formating an individual file
+    """
+    def __init__(self, path, cache_dir):
+        clang_format_progname = CLANG_FORMAT_PROGNAME
+
+        # Initialize clang-format configuration information
+        if sys.platform.startswith("linux"):
+              #"3.6.0/clang+llvm-3.6.0-x86_64-linux-gnu-ubuntu-14.04.tar.xz
+            self.platform = "linux_x64"
+            self.llvm_distro = "x86_64-linux-gnu-ubuntu"
+            self.tar_path = "x86_64-linux-gnu"
+        elif sys.platform == "win32":
+            self.platform = "windows_x64"
+            self.llvm_distro = "windows_x64"
+            self.tar_path = None
+            clang_format_progname += ".exe"
+        elif sys.platform == "darwin":
+             #"3.6.0/clang+llvm-3.6.0-x86_64-apple-darwin.tar.xz
+            self.platform = "darwin_x64"
+            self.llvm_distro = "x86_64-apple-darwin"
+            self.tar_path = "x86_64-apple-darwin"
+
+        self.path = None
+
+        # Find Clang-Format now
+        if path is not None:
+            if os.path.isfile(path):
+                self.path = path
+            else:
+                print("WARNING: Could not find clang-format %s" % (path))
+
+        # Check the envionrment variable
+        if "MONGO_CLANG_FORMAT" in os.environ:
+            self.path = os.environ["MONGO_CLANG_FORMAT"]
+
+            if self.path and not self._validate_version(warn=True):
+                self.path = None
+
+        # Check the users' PATH environment variable now
+        if self.path is None:
+            self.path = spawn.find_executable(clang_format_progname)
+
+            if self.path and not self._validate_version(warn=True):
+                self.path = None
+
+        # If Windows, try to grab it from Program Files
+        if sys.platform == "win32":
+            win32bin = os.path.join(os.environ["ProgramFiles(x86)"], "LLVM\\bin\\clang-format.exe")
+            if os.path.exists(win32bin):
+                self.path = win32bin
+
+        # Have not found it yet, download it from the web
+        if self.path is None:
+            if not os.path.isdir(cache_dir):
+                os.makedirs(cache_dir)
+
+            self.path = os.path.join(cache_dir, clang_format_progname)
+
+            if not os.path.isfile(self.path):
+                if sys.platform.startswith("linux"):
+                    get_clang_format_from_linux_cache(self.path)
+                elif sys.platform == "darwin":
+                    get_clang_format_from_llvm(self.llvm_distro, self.tar_path, self.path)
+                else:
+                    print("ERROR: clang-format.py does not support downloading clang-format " +
+                        " on this platform, please install clang-format " + CLANG_FORMAT_VERSION)
+
+        # Validate we have the correct version
+        self._validate_version()
+
+        self.print_lock = threading.Lock()
+
+    def _validate_version(self, warn=False):
+        """Validate clang-format is the expected version
+        """
+        cf_version = callo([self.path, "--version"])
+
+        if CLANG_FORMAT_VERSION in cf_version:
+            return True
+
+        if warn:
+            print("WARNING: clang-format found in path, but incorrect version found at " +
+                    self.path + " with version: " + cf_version)
+
+        return False
+
+    def lint(self, file_name):
+        """Check the specified file has the correct format
+        """
+        with open(file_name, 'r') as original_text:
+            original_file = original_text.read()
+
+        # Get formatted file as clang-format would format the file
+        formatted_file = callo([self.path, "--style=file", file_name])
+
+        if original_file != formatted_file:
+
+            original_lines = original_file.splitlines()
+            formatted_lines = formatted_file.splitlines()
+            result = difflib.unified_diff(original_lines, formatted_lines)
+
+            # Take a lock to ensure diffs do not get mixed
+            with self.print_lock:
+                print("ERROR: Found diff for " + file_name)
+                print("To fix formatting errors, run %s --style=file -i %s" %
+                        (self.path, file_name))
+                for line in result:
+                    print(line.rstrip())
+
+            return False
+
+        return True
+
+    def format(self, file_name):
+        """Update the format of the specified file
+        """
+
+        # Update the file with clang-format
+        return not subprocess.call([self.path, "--style=file", "-i", file_name])
+
+
+def parallel_process(items, func):
+    """Run a set of work items to completion
+    """
+    try:
+        cpus = cpu_count()
+    except NotImplementedError:
+        cpus = 1
+
+    print("Running across %d cpus" % (cpus))
+
+    task_queue = Queue.Queue()
+
+    # Use a list so that worker function will capture this variable
+    pp_event = threading.Event()
+    pp_result = [True]
+    pp_lock = threading.Lock()
+
+    def worker():
+        """Worker thread to process work items in parallel
+        """
+        while not pp_event.is_set():
+            try:
+                item = task_queue.get_nowait()
+            except Queue.Empty:
+                # if the queue is empty, exit the worker thread
+                pp_event.set()
+                return
+
+            try:
+                ret = func(item)
+            finally:
+                # Tell the queue we finished with the item
+                task_queue.task_done()
+
+            # Return early if we fail, and signal we are done
+            if not ret:
+                with pp_lock:
+                    pp_result[0] = False
+
+                pp_event.set()
+                return
+
+    # Enqueue all the work we want to process
+    for item in items:
+        task_queue.put(item)
+
+    # Process all the work
+    threads = []
+    for cpu in range(cpus):
+        thread = threading.Thread(target=worker)
+
+        thread.daemon = True
+        thread.start()
+        threads.append(thread)
+
+    # Wait for the threads to finish
+    # Loop with a timeout so that we can process Ctrl-C interrupts
+    # Note: On Python 2.6 wait always returns None so we check is_set also,
+    #  This works because we only set the event once, and never reset it
+    while not pp_event.wait(1) and not pp_event.is_set():
+        time.sleep(1)
+
+    for thread in threads:
+        thread.join()
+
+    return pp_result[0]
+
+def get_base_dir():
+    """Get the base directory for mongo repo.
+        This script assumes that it is running in buildscripts/, and uses
+        that to find the base directory.
+    """
+    script_path = os.path.dirname(os.path.realpath(__file__))
+
+    return os.path.dirname(script_path)
+
+def get_repos():
+    """Get a list of Repos to check clang-format for
+    """
+    base_dir = get_base_dir()
+
+    # Get a list of modules
+    # TODO: how do we filter rocks, does it matter?
+    mongo_modules = moduleconfig.discover_module_directories(
+                        os.path.join(base_dir, MODULE_DIR), None)
+
+    paths = [os.path.join(base_dir, MODULE_DIR, m) for m in mongo_modules]
+
+    paths.append(base_dir)
+
+    return [Repo(p) for p in paths]
+
+
+class Repo(object):
+    """Class encapsulates all knowledge about a git repository, and its metadata
+        to run clang-format.
+    """
+    def __init__(self, path):
+        self.path = path
+
+       # Get candidate files
+        self.candidate_files = self._get_candidate_files()
+
+        self.root = self._get_root()
+
+    def _callgito(self, args):
+        """Call git for this repository
+        """
+        # These two flags are the equivalent of -C in newer versions of Git
+        # but we use these to support versions back to ~1.8
+        return callo(['git', '--git-dir', os.path.join(self.path, ".git"),
+                        '--work-tree', self.path] + args)
+
+    def _get_local_dir(self, path):
+        """Get a directory path relative to the git root directory
+        """
+        if os.path.isabs(path):
+            return os.path.relpath(path, self.root)
+        return path
+
+    def get_candidates(self, candidates):
+        """Get the set of candidate files to check by doing an intersection
+        between the input list, and the list of candidates in the repository
+
+        Returns the full path to the file for clang-format to consume.
+        """
+        # NOTE: Files may have an absolute root (i.e. leading /)
+
+        if candidates is not None and len(candidates) > 0:
+            candidates = [self._get_local_dir(f) for f in candidates]
+            valid_files = list(set(candidates).intersection(self.get_candidate_files()))
+        else:
+            valid_files = list(self.get_candidate_files())
+
+        # Get the full file name here
+        valid_files = [os.path.normpath(os.path.join(self.root, f)) for f in valid_files]
+        return valid_files
+
+    def get_root(self):
+        """Get the root directory for this repository
+        """
+        return self.root
+
+    def _get_root(self):
+        """Gets the root directory for this repository from git
+        """
+        gito = self._callgito(['rev-parse', '--show-toplevel'])
+
+        return gito.rstrip()
+
+    def get_candidate_files(self):
+        """Get a list of candidate files
+        """
+        return self._get_candidate_files()
+
+    def _get_candidate_files(self):
+        """Query git to get a list of all files in the repo to consider for analysis
+        """
+        gito = self._callgito(["ls-files"])
+
+        # This allows us to pick all the interesting files
+        # in the mongo and mongo-enterprise repos
+        file_list = [line.rstrip()
+                for line in gito.splitlines() if "src" in line and not "src/third_party" in line]
+
+        files_match = re.compile('\\.(h|cpp)$')
+
+        file_list = [a for a in file_list if files_match.search(a)]
+
+        return file_list
+
+
+def expand_file_string(glob_pattern):
+    """Expand a string that represents a set of files
+    """
+    return [os.path.abspath(f) for f in globstar.iglob(glob_pattern)]
+
+def get_files_to_check(files):
+    """Filter the specified list of files to check down to the actual
+        list of files that need to be checked."""
+    candidates = []
+
+    # Get a list of candidate_files
+    candidates = [expand_file_string(f) for f in files]
+    candidates = list(itertools.chain.from_iterable(candidates))
+
+    repos = get_repos()
+
+    valid_files = list(itertools.chain.from_iterable([r.get_candidates(candidates) for r in repos]))
+
+    return valid_files
+
+def get_files_to_check_from_patch(patches):
+    """Take a patch file generated by git diff, and scan the patch for a list of files to check.
+    """
+    candidates = []
+
+    # Get a list of candidate_files
+    check = re.compile(r"^diff --git a\/([a-z\/\.\-_0-9]+) b\/[a-z\/\.\-_0-9]+")
+
+    lines = []
+    for patch in patches:
+        with open(patch, "rb") as infile:
+            lines += infile.readlines()
+
+    candidates = [check.match(line).group(1) for line in lines if check.match(line)]
+
+    repos = get_repos()
+
+    valid_files = list(itertools.chain.from_iterable([r.get_candidates(candidates) for r in repos]))
+
+    return valid_files
+
+def _get_build_dir():
+    """Get the location of the scons' build directory in case we need to download clang-format
+    """
+    return os.path.join(get_base_dir(), "build")
+
+def _lint_files(clang_format, files):
+    """Lint a list of files with clang-format
+    """
+    clang_format = ClangFormat(clang_format, _get_build_dir())
+
+    lint_clean = parallel_process([os.path.abspath(f) for f in files], clang_format.lint)
+
+    if not lint_clean:
+        print("ERROR: Code Style does not match coding style")
+        sys.exit(1)
+
+def lint_patch(clang_format, infile):
+    """Lint patch command entry point
+    """
+    files = get_files_to_check_from_patch(infile)
+
+    # Patch may have files that we do not want to check which is fine
+    if files:
+        _lint_files(clang_format, files)
+
+def lint(clang_format, glob):
+    """Lint files command entry point
+    """
+    files = get_files_to_check(glob)
+
+    _lint_files(clang_format, files)
+
+    return True
+
+def _format_files(clang_format, files):
+    """Format a list of files with clang-format
+    """
+    clang_format = ClangFormat(clang_format, _get_build_dir())
+
+    format_clean = parallel_process([os.path.abspath(f) for f in files], clang_format.format)
+
+    if not format_clean:
+        print("ERROR: failed to format files")
+        sys.exit(1)
+
+def format_func(clang_format, glob):
+    """Format files command entry point
+    """
+    files = get_files_to_check(glob)
+
+    _format_files(clang_format, files)
+
+def usage():
+    """Print usage
+    """
+    print("clang-format.py supports 3 commands [ lint, lint-patch, format ]. Run "
+            " <command> -? for more information")
+
+def main():
+    """Main entry point
+    """
+    if len(sys.argv) > 1:
+        command = sys.argv[1]
+
+        parser = OptionParser()
+        parser.add_option("-c", "--clang-format", type="string", dest="clang_format")
+
+        if command == "lint":
+            (options, args) = parser.parse_args(args=sys.argv[2:])
+            lint(options.clang_format, args)
+        elif command == "lint-patch":
+            (options, args) = parser.parse_args(args=sys.argv[2:])
+            lint_patch(options.clang_format, args)
+        elif command == "format":
+            (options, args) = parser.parse_args(args=sys.argv[2:])
+            format_func(options.clang_format, args)
+        else:
+            usage()
+    else:
+        usage()
+
+if __name__ == "__main__":
+    main()
diff --git a/buildscripts/moduleconfig.py b/buildscripts/moduleconfig.py
index 111b49545e5..7f98667b6b5 100644
--- a/buildscripts/moduleconfig.py
+++ b/buildscripts/moduleconfig.py
@@ -24,7 +24,8 @@ MongoDB module SConscript files can describe libraries, programs and unit tests,
 MongoDB SConscript files do.
 """
 
-__all__ = ('discover_modules', 'configure_modules', 'register_module_test')
+__all__ = ('discover_modules', 'discover_module_directories', 'configure_modules',
+            'register_module_test')
 
 import imp
 import inspect
@@ -69,6 +70,36 @@ def discover_modules(module_root, allowed_modules):
 
     return found_modules
 
+def discover_module_directories(module_root, allowed_modules):
+    """Scans module_root for subdirectories that look like MongoDB modules.
+
+    Returns a list of directory names.
+    """
+    if not os.path.isdir(module_root):
+        return []
+
+    found_modules = []
+
+    if allowed_modules is not None:
+        allowed_modules = allowed_modules.split(',')
+
+    for name in os.listdir(module_root):
+        root = os.path.join(module_root, name)
+        if name.startswith('.') or not os.path.isdir(root):
+            continue
+
+        build_py = os.path.join(root, 'build.py')
+
+        if allowed_modules is not None and name not in allowed_modules:
+            print "skipping module: %s" % name
+            continue
+
+        if os.path.isfile(build_py):
+            print "adding module: %s" % name
+            found_modules.append(name)
+
+    return found_modules
+
 def configure_modules(modules, conf):
     """ Run the configure() function in the build.py python modules for each module in "modules"
     (as created by discover_modules).
diff --git a/buildscripts/resmokelib/utils/globstar.py b/buildscripts/resmokelib/utils/globstar.py
index 7a744014c7b..644ebfe3e38 100644
--- a/buildscripts/resmokelib/utils/globstar.py
+++ b/buildscripts/resmokelib/utils/globstar.py
@@ -89,6 +89,9 @@ def _split_path(pathname):
     while True:
         (dirname, basename) = os.path.split(pathname)
         parts.append(basename)
+        if pathname == dirname:
+            parts.append(dirname)
+            break
         if not dirname:
             break
         pathname = dirname