DXR is a code search and navigation tool aimed at making sense of large projects. It supports full-text and regex searches as well as structural queries.

Line Code
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075
#!/bin/env python
# This Source Code Form is subject to the terms of the Mozilla Public
# License, v. 2.0. If a copy of the MPL was not distributed with this
# file, You can obtain one at http://mozilla.org/MPL/2.0/.
#
# Usage: symbolstore.py <params> <dump_syms path> <symbol store path>
#                                <debug info files or dirs>
#   Runs dump_syms on each debug info file specified on the command line,
#   then places the resulting symbol file in the proper directory
#   structure in the symbol store path.  Accepts multiple files
#   on the command line, so can be called as part of a pipe using
#   find <dir> | xargs symbolstore.pl <dump_syms> <storepath>
#   But really, you might just want to pass it <dir>.
#
#   Parameters accepted:
#     -c           : Copy debug info files to the same directory structure
#                    as sym files. On Windows, this will also copy
#                    binaries into the symbol store.
#     -a "<archs>" : Run dump_syms -a <arch> for each space separated
#                    cpu architecture in <archs> (only on OS X)
#     -s <srcdir>  : Use <srcdir> as the top source directory to
#                    generate relative filenames.

import buildconfig
import errno
import sys
import platform
import os
import re
import shutil
import textwrap
import fnmatch
import subprocess
import time
import ctypes
import urlparse
import concurrent.futures
import multiprocessing

from optparse import OptionParser
from xml.dom.minidom import parse

from mozpack.copier import FileRegistry
from mozpack.manifests import (
    InstallManifest,
    UnreadableInstallManifest,
)

# Utility classes

class VCSFileInfo:
    """ A base class for version-controlled file information. Ensures that the
        following attributes are generated only once (successfully):

            self.root
            self.clean_root
            self.revision
            self.filename

        The attributes are generated by a single call to the GetRoot,
        GetRevision, and GetFilename methods. Those methods are explicitly not
        implemented here and must be implemented in derived classes. """

    def __init__(self, file):
        if not file:
            raise ValueError
        self.file = file

    def __getattr__(self, name):
        """ __getattr__ is only called for attributes that are not set on self,
            so setting self.[attr] will prevent future calls to the GetRoot,
            GetRevision, and GetFilename methods. We don't set the values on
            failure on the off chance that a future call might succeed. """

        if name == "root":
            root = self.GetRoot()
            if root:
                self.root = root
            return root

        elif name == "clean_root":
            clean_root = self.GetCleanRoot()
            if clean_root:
                self.clean_root = clean_root
            return clean_root

        elif name == "revision":
            revision = self.GetRevision()
            if revision:
                self.revision = revision
            return revision

        elif name == "filename":
            filename = self.GetFilename()
            if filename:
                self.filename = filename
            return filename

        raise AttributeError

    def GetRoot(self):
        """ This method should return the unmodified root for the file or 'None'
            on failure. """
        raise NotImplementedError

    def GetCleanRoot(self):
        """ This method should return the repository root for the file or 'None'
            on failure. """
        raise NotImplementedError

    def GetRevision(self):
        """ This method should return the revision number for the file or 'None'
            on failure. """
        raise NotImplementedError

    def GetFilename(self):
        """ This method should return the repository-specific filename for the
            file or 'None' on failure. """
        raise NotImplementedError


# This regex separates protocol and optional username/password from a url.
# For instance, all the following urls will be transformed into
# 'foo.com/bar':
#
#   http://foo.com/bar
#   svn+ssh://user@foo.com/bar
#   svn+ssh://user:pass@foo.com/bar
#
rootRegex = re.compile(r'^\S+?:/+(?:[^\s/]*@)?(\S+)$')

def read_output(*args):
    (stdout, _) = subprocess.Popen(args=args, stdout=subprocess.PIPE).communicate()
    return stdout.rstrip()

class HGRepoInfo:
    def __init__(self, path):
        self.path = path
        rev = read_output('hg', '-R', path,
                          'parent', '--template={node|short}')
        # Look for the default hg path.  If SRVSRV_ROOT is set, we
        # don't bother asking hg.
        hg_root = os.environ.get("SRCSRV_ROOT")
        if hg_root:
            root = hg_root
        else:
            root = read_output('hg', '-R', path,
                               'showconfig', 'paths.default')
            if not root:
                print >> sys.stderr, "Failed to get HG Repo for %s" % path
        cleanroot = None
        if root:
            match = rootRegex.match(root)
            if match:
                cleanroot = match.group(1)
                if cleanroot.endswith('/'):
                    cleanroot = cleanroot[:-1]
        if cleanroot is None:
            print >> sys.stderr, textwrap.dedent("""\
                Could not determine repo info for %s.  This is either not a clone of the web-based
                repository, or you have not specified SRCSRV_ROOT, or the clone is corrupt.""") % path
            sys.exit(1)
        self.rev = rev
        self.root = root
        self.cleanroot = cleanroot

    def GetFileInfo(self, file):
        return HGFileInfo(file, self)

class HGFileInfo(VCSFileInfo):
    def __init__(self, file, repo):
        VCSFileInfo.__init__(self, file)
        self.repo = repo
        self.file = os.path.relpath(file, repo.path)

    def GetRoot(self):
        return self.repo.root

    def GetCleanRoot(self):
        return self.repo.cleanroot

    def GetRevision(self):
        return self.repo.rev

    def GetFilename(self):
        if self.revision and self.clean_root:
            return "hg:%s:%s:%s" % (self.clean_root, self.file, self.revision)
        return self.file

class GitRepoInfo:
    """
    Info about a local git repository. Does not currently
    support discovering info about a git clone, the info must be
    provided out-of-band.
    """
    def __init__(self, path, rev, root):
        self.path = path
        cleanroot = None
        if root:
            match = rootRegex.match(root)
            if match:
                cleanroot = match.group(1)
                if cleanroot.endswith('/'):
                    cleanroot = cleanroot[:-1]
        if cleanroot is None:
            print >> sys.stderr, textwrap.dedent("""\
                Could not determine repo info for %s (%s).  This is either not a clone of a web-based
                repository, or you have not specified SRCSRV_ROOT, or the clone is corrupt.""") % (path, root)
            sys.exit(1)
        self.rev = rev
        self.cleanroot = cleanroot

    def GetFileInfo(self, file):
        return GitFileInfo(file, self)

class GitFileInfo(VCSFileInfo):
    def __init__(self, file, repo):
        VCSFileInfo.__init__(self, file)
        self.repo = repo
        self.file = os.path.relpath(file, repo.path)

    def GetRoot(self):
        return self.repo.path

    def GetCleanRoot(self):
        return self.repo.cleanroot

    def GetRevision(self):
        return self.repo.rev

    def GetFilename(self):
        if self.revision and self.clean_root:
            return "git:%s:%s:%s" % (self.clean_root, self.file, self.revision)
        return self.file

# Utility functions

# A cache of files for which VCS info has already been determined. Used to
# prevent extra filesystem activity or process launching.
vcsFileInfoCache = {}

def IsInDir(file, dir):
    # the lower() is to handle win32+vc8, where
    # the source filenames come out all lowercase,
    # but the srcdir can be mixed case
    return os.path.abspath(file).lower().startswith(os.path.abspath(dir).lower())

def GetVCSFilenameFromSrcdir(file, srcdir):
    if srcdir not in Dumper.srcdirRepoInfo:
        # Not in cache, so find it adnd cache it
        if os.path.isdir(os.path.join(srcdir, '.hg')):
            Dumper.srcdirRepoInfo[srcdir] = HGRepoInfo(srcdir)
        else:
            # Unknown VCS or file is not in a repo.
            return None
    return Dumper.srcdirRepoInfo[srcdir].GetFileInfo(file)

def GetVCSFilename(file, srcdirs):
    """Given a full path to a file, and the top source directory,
    look for version control information about this file, and return
    a tuple containing
    1) a specially formatted filename that contains the VCS type,
    VCS location, relative filename, and revision number, formatted like:
    vcs:vcs location:filename:revision
    For example:
    cvs:cvs.mozilla.org/cvsroot:mozilla/browser/app/nsBrowserApp.cpp:1.36
    2) the unmodified root information if it exists"""
    (path, filename) = os.path.split(file)
    if path == '' or filename == '':
        return (file, None)

    fileInfo = None
    root = ''
    if file in vcsFileInfoCache:
        # Already cached this info, use it.
        fileInfo = vcsFileInfoCache[file]
    else:
        for srcdir in srcdirs:
            if not IsInDir(file, srcdir):
                continue
            fileInfo = GetVCSFilenameFromSrcdir(file, srcdir)
            if fileInfo:
                vcsFileInfoCache[file] = fileInfo
                break

    if fileInfo:
        file = fileInfo.filename
        root = fileInfo.root

    # we want forward slashes on win32 paths
    return (file.replace("\\", "/"), root)

def validate_install_manifests(install_manifest_args):
    args = []
    for arg in install_manifest_args:
        bits = arg.split(',')
        if len(bits) != 2:
            raise ValueError('Invalid format for --install-manifest: '
                             'specify manifest,target_dir')
        manifest_file, destination = map(os.path.abspath, bits)
        if not os.path.isfile(manifest_file):
            raise IOError(errno.ENOENT, 'Manifest file not found',
                          manifest_file)
        if not os.path.isdir(destination):
            raise IOError(errno.ENOENT, 'Install directory not found',
                          destination)
        try:
            manifest = InstallManifest(manifest_file)
        except UnreadableInstallManifest:
            raise IOError(errno.EINVAL, 'Error parsing manifest file',
                          manifest_file)
        args.append((manifest, destination))
    return args

def make_file_mapping(install_manifests):
    file_mapping = {}
    for manifest, destination in install_manifests:
        destination = os.path.abspath(destination)
        reg = FileRegistry()
        manifest.populate_registry(reg)
        for dst, src in reg:
            if hasattr(src, 'path'):
                abs_dest = os.path.normpath(os.path.join(destination, dst))
                file_mapping[abs_dest] = src.path
    return file_mapping

def GetPlatformSpecificDumper(**kwargs):
    """This function simply returns a instance of a subclass of Dumper
    that is appropriate for the current platform."""
    return {'WINNT': Dumper_Win32,
            'Linux': Dumper_Linux,
            'Darwin': Dumper_Mac}[buildconfig.substs['OS_ARCH']](**kwargs)

def SourceIndex(fileStream, outputPath, vcs_root):
    """Takes a list of files, writes info to a data block in a .stream file"""
    # Creates a .pdb.stream file in the mozilla\objdir to be used for source indexing
    # Create the srcsrv data block that indexes the pdb file
    result = True
    pdbStreamFile = open(outputPath, "w")
    pdbStreamFile.write('''SRCSRV: ini ------------------------------------------------\r\nVERSION=2\r\nINDEXVERSION=2\r\nVERCTRL=http\r\nSRCSRV: variables ------------------------------------------\r\nHGSERVER=''')
    pdbStreamFile.write(vcs_root)
    pdbStreamFile.write('''\r\nSRCSRVVERCTRL=http\r\nHTTP_EXTRACT_TARGET=%hgserver%/raw-file/%var3%/%var2%\r\nSRCSRVTRG=%http_extract_target%\r\nSRCSRV: source files ---------------------------------------\r\n''')
    pdbStreamFile.write(fileStream) # can't do string interpolation because the source server also uses this and so there are % in the above
    pdbStreamFile.write("SRCSRV: end ------------------------------------------------\r\n\n")
    pdbStreamFile.close()
    return result

def StartJob(dumper, lock, srcdirRepoInfo, func_name, args):
    # Windows worker processes won't have run GlobalInit,
    # and due to a lack of fork(), won't inherit the class
    # variables from the parent, so set them here.
    Dumper.lock = lock
    Dumper.srcdirRepoInfo = srcdirRepoInfo
    return getattr(dumper, func_name)(*args)

class JobPool(object):
    jobs = {}
    executor = None

    @classmethod
    def init(cls, executor):
        cls.executor = executor

    @classmethod
    def shutdown(cls):
        cls.executor.shutdown()

    @classmethod
    def submit(cls, args, callback):
        cls.jobs[cls.executor.submit(StartJob, *args)] = callback

    @classmethod
    def as_completed(cls):
        '''Like concurrent.futures.as_completed, but allows adding new futures
        between generator steps. Iteration will end when the generator has
        yielded all completed futures and JobQueue.jobs is empty.
        Yields (future, callback) pairs.
        '''
        while cls.jobs:
            completed, _ = concurrent.futures.wait(cls.jobs.keys(), return_when=concurrent.futures.FIRST_COMPLETED)
            for f in completed:
                callback = cls.jobs[f]
                del cls.jobs[f]
                yield f, callback

class Dumper:
    """This class can dump symbols from a file with debug info, and
    store the output in a directory structure that is valid for use as
    a Breakpad symbol server.  Requires a path to a dump_syms binary--
    |dump_syms| and a directory to store symbols in--|symbol_path|.
    Optionally takes a list of processor architectures to process from
    each debug file--|archs|, the full path to the top source
    directory--|srcdir|, for generating relative source file names,
    and an option to copy debug info files alongside the dumped
    symbol files--|copy_debug|, mostly useful for creating a
    Microsoft Symbol Server from the resulting output.

    You don't want to use this directly if you intend to process files.
    Instead, call GetPlatformSpecificDumper to get an instance of a
    subclass.

    Processing is performed asynchronously via worker processes; in
    order to wait for processing to finish and cleanup correctly, you
    must call Finish after all ProcessFiles calls have been made.
    You must also call Dumper.GlobalInit before creating or using any
    instances."""
    def __init__(self, dump_syms, symbol_path,
                 archs=None,
                 srcdirs=[],
                 copy_debug=False,
                 vcsinfo=False,
                 srcsrv=False,
                 exclude=[],
                 repo_manifest=None,
                 file_mapping=None):
        # popen likes absolute paths, at least on windows
        self.dump_syms = os.path.abspath(dump_syms)
        self.symbol_path = symbol_path
        if archs is None:
            # makes the loop logic simpler
            self.archs = ['']
        else:
            self.archs = ['-a %s' % a for a in archs.split()]
        self.srcdirs = [os.path.normpath(a) for a in srcdirs]
        self.copy_debug = copy_debug
        self.vcsinfo = vcsinfo
        self.srcsrv = srcsrv
        self.exclude = exclude[:]
        if repo_manifest:
            self.parse_repo_manifest(repo_manifest)
        self.file_mapping = file_mapping or {}

        # book-keeping to keep track of the cleanup work per file tuple
        self.files_record = {}

    @classmethod
    def GlobalInit(cls, executor=concurrent.futures.ProcessPoolExecutor):
        """Initialize the class globals for the multiprocessing setup; must
        be called before any Dumper instances are created and used. Test cases
        may pass in a different executor to use, usually
        concurrent.futures.ThreadPoolExecutor."""
        num_cpus = multiprocessing.cpu_count()
        if num_cpus is None:
            # assume a dual core machine if we can't find out for some reason
            # probably better on single core anyway due to I/O constraints
            num_cpus = 2

        # have to create any locks etc before the pool
        manager = multiprocessing.Manager()
        cls.lock = manager.RLock()
        cls.srcdirRepoInfo = manager.dict()
        JobPool.init(executor(max_workers=num_cpus))

    def output(self, dest, output_str):
        """Writes |output_str| to |dest|, holding |lock|;
        terminates with a newline."""
        with Dumper.lock:
            dest.write(output_str + "\n")
            dest.flush()

    def output_pid(self, dest, output_str):
        """Debugging output; prepends the pid to the string."""
        self.output(dest, "%d: %s" % (os.getpid(), output_str))

    def parse_repo_manifest(self, repo_manifest):
        """
        Parse an XML manifest of repository info as produced
        by the `repo manifest -r` command.
        """
        doc = parse(repo_manifest)
        if doc.firstChild.tagName != "manifest":
            return
        # First, get remotes.
        def ensure_slash(u):
            if not u.endswith("/"):
                return u + "/"
            return u
        remotes = dict([(r.getAttribute("name"), ensure_slash(r.getAttribute("fetch"))) for r in doc.getElementsByTagName("remote")])
        # And default remote.
        default_remote = None
        if doc.getElementsByTagName("default"):
            default_remote = doc.getElementsByTagName("default")[0].getAttribute("remote")
        # Now get projects. Assume they're relative to repo_manifest.
        base_dir = os.path.abspath(os.path.dirname(repo_manifest))
        for proj in doc.getElementsByTagName("project"):
            # name is the repository URL relative to the remote path.
            name = proj.getAttribute("name")
            # path is the path on-disk, relative to the manifest file.
            path = proj.getAttribute("path")
            # revision is the changeset ID.
            rev = proj.getAttribute("revision")
            # remote is the base URL to use.
            remote = proj.getAttribute("remote")
            # remote defaults to the <default remote>.
            if not remote:
                remote = default_remote
            # path defaults to name.
            if not path:
                path = name
            if not (name and path and rev and remote):
                print "Skipping project %s" % proj.toxml()
                continue
            remote = remotes[remote]
            # Turn git URLs into http URLs so that urljoin works.
            if remote.startswith("git:"):
                remote = "http" + remote[3:]
            # Add this project to srcdirs.
            srcdir = os.path.join(base_dir, path)
            self.srcdirs.append(srcdir)
            # And cache its VCS file info. Currently all repos mentioned
            # in a repo manifest are assumed to be git.
            root = urlparse.urljoin(remote, name)
            Dumper.srcdirRepoInfo[srcdir] = GitRepoInfo(srcdir, rev, root)

    # subclasses override this
    def ShouldProcess(self, file):
        return not any(fnmatch.fnmatch(os.path.basename(file), exclude) for exclude in self.exclude)

    # and can override this
    def ShouldSkipDir(self, dir):
        return False

    def RunFileCommand(self, file):
        """Utility function, returns the output of file(1)"""
        try:
            # we use -L to read the targets of symlinks,
            # and -b to print just the content, not the filename
            return os.popen("file -Lb " + file).read()
        except:
            return ""

    # This is a no-op except on Win32
    def FixFilenameCase(self, file):
        return file

    # This is a no-op except on Win32
    def SourceServerIndexing(self, debug_file, guid, sourceFileStream, vcs_root):
        return ""

    # subclasses override this if they want to support this
    def CopyDebug(self, file, debug_file, guid, code_file, code_id):
        pass

    def Finish(self, stop_pool=True):
        '''Process all pending jobs and any jobs their callbacks submit.
        By default, will shutdown the executor, but for testcases that
        need multiple runs, pass stop_pool = False.'''
        for job, callback in JobPool.as_completed():
            try:
                res = job.result()
            except Exception as e:
                self.output(sys.stderr, 'Job raised exception: %s' % e)
                continue
            callback(res)
        if stop_pool:
            JobPool.shutdown()

    def Process(self, *args):
        """Process files recursively in args."""
        # We collect all files to process first then sort by size to schedule
        # larger files first because larger files tend to take longer and we
        # don't like long pole stragglers.
        files = set()
        for arg in args:
            for f in self.get_files_to_process(arg):
                files.add(f)

        for f in sorted(files, key=os.path.getsize, reverse=True):
            self.ProcessFiles((f,))

    def get_files_to_process(self, file_or_dir):
        """Generate the files to process from an input."""
        if os.path.isdir(file_or_dir) and not self.ShouldSkipDir(file_or_dir):
            for f in self.get_files_to_process_in_dir(file_or_dir):
                yield f
        elif os.path.isfile(file_or_dir):
            yield file_or_dir

    def get_files_to_process_in_dir(self, path):
        """Generate the files to process in a directory.

        Valid files are are determined by calling ShouldProcess.
        """
        for root, dirs, files in os.walk(path):
            for d in dirs[:]:
                if self.ShouldSkipDir(d):
                    dirs.remove(d)
            for f in files:
                fullpath = os.path.join(root, f)
                if self.ShouldProcess(fullpath):
                    yield fullpath

    def SubmitJob(self, file_key, func_name, args, callback):
        """Submits a job to the pool of workers"""
        JobPool.submit((self, Dumper.lock, Dumper.srcdirRepoInfo, func_name, args), callback)

    def ProcessFilesFinished(self, res):
        """Callback from multiprocesing when ProcessFilesWork finishes;
        run the cleanup work, if any"""
        # only run the cleanup function once per tuple of files
        self.files_record[res['files']] += 1
        if self.files_record[res['files']] == len(self.archs):
            del self.files_record[res['files']]
            if res['after']:
                res['after'](res['status'], res['after_arg'])

    def ProcessFiles(self, files, after=None, after_arg=None):
        """Dump symbols from these files into a symbol file, stored
        in the proper directory structure in  |symbol_path|; processing is performed
        asynchronously, and Finish must be called to wait for it complete and cleanup.
        All files after the first are fallbacks in case the first file does not process
        successfully; if it does, no other files will be touched."""
        self.output_pid(sys.stderr, "Submitting jobs for files: %s" % str(files))

        # tries to get the vcs root from the .mozconfig first - if it's not set
        # the tinderbox vcs path will be assigned further down
        vcs_root = os.environ.get("SRCSRV_ROOT")
        for arch_num, arch in enumerate(self.archs):
            self.files_record[files] = 0 # record that we submitted jobs for this tuple of files
            self.SubmitJob(files[-1], 'ProcessFilesWork', args=(files, arch_num, arch, vcs_root, after, after_arg), callback=self.ProcessFilesFinished)

    def dump_syms_cmdline(self, file, arch, files):
        '''
        Get the commandline used to invoke dump_syms.
        '''
        # The Mac dumper overrides this.
        return [self.dump_syms, file]

    def ProcessFilesWork(self, files, arch_num, arch, vcs_root, after, after_arg):
        t_start = time.time()
        self.output_pid(sys.stderr, "Worker processing files: %s" % (files,))

        # our result is a status, a cleanup function, an argument to that function, and the tuple of files we were called on
        result = { 'status' : False, 'after' : after, 'after_arg' : after_arg, 'files' : files }

        sourceFileStream = ''
        code_id, code_file = None, None
        for file in files:
            # files is a tuple of files, containing fallbacks in case the first file doesn't process successfully
            try:
                cmd = self.dump_syms_cmdline(file, arch, files)
                self.output_pid(sys.stderr, ' '.join(cmd))
                proc = subprocess.Popen(cmd, stdout=subprocess.PIPE,
                                        stderr=open(os.devnull, 'wb'))
                module_line = proc.stdout.next()
                if module_line.startswith("MODULE"):
                    # MODULE os cpu guid debug_file
                    (guid, debug_file) = (module_line.split())[3:5]
                    # strip off .pdb extensions, and append .sym
                    sym_file = re.sub("\.pdb$", "", debug_file) + ".sym"
                    # we do want forward slashes here
                    rel_path = os.path.join(debug_file,
                                            guid,
                                            sym_file).replace("\\", "/")
                    full_path = os.path.normpath(os.path.join(self.symbol_path,
                                                              rel_path))
                    try:
                        os.makedirs(os.path.dirname(full_path))
                    except OSError: # already exists
                        pass
                    f = open(full_path, "w")
                    f.write(module_line)
                    # now process the rest of the output
                    for line in proc.stdout:
                        if line.startswith("FILE"):
                            # FILE index filename
                            (x, index, filename) = line.rstrip().split(None, 2)
                            filename = os.path.normpath(self.FixFilenameCase(filename))
                            # We want original file paths for the source server.
                            sourcepath = filename
                            if filename in self.file_mapping:
                                filename = self.file_mapping[filename]
                            if self.vcsinfo:
                                (filename, rootname) = GetVCSFilename(filename, self.srcdirs)
                                # sets vcs_root in case the loop through files were to end on an empty rootname
                                if vcs_root is None:
                                  if rootname:
                                     vcs_root = rootname
                            # gather up files with hg for indexing
                            if filename.startswith("hg"):
                                (ver, checkout, source_file, revision) = filename.split(":", 3)
                                sourceFileStream += sourcepath + "*" + source_file + '*' + revision + "\r\n"
                            f.write("FILE %s %s\n" % (index, filename))
                        elif line.startswith("INFO CODE_ID "):
                            # INFO CODE_ID code_id code_file
                            # This gives some info we can use to
                            # store binaries in the symbol store.
                            bits = line.rstrip().split(None, 3)
                            if len(bits) == 4:
                                code_id, code_file = bits[2:]
                            f.write(line)
                        else:
                            # pass through all other lines unchanged
                            f.write(line)
                            # we want to return true only if at least one line is not a MODULE or FILE line
                            result['status'] = True
                    f.close()
                    proc.wait()
                    # we output relative paths so callers can get a list of what
                    # was generated
                    self.output(sys.stdout, rel_path)
                    if self.srcsrv and vcs_root:
                        # add source server indexing to the pdb file
                        self.SourceServerIndexing(file, guid, sourceFileStream, vcs_root)
                    # only copy debug the first time if we have multiple architectures
                    if self.copy_debug and arch_num == 0:
                        self.CopyDebug(file, debug_file, guid,
                                       code_file, code_id)
            except StopIteration:
                pass
            except Exception as e:
                self.output(sys.stderr, "Unexpected error: %s" % (str(e),))
                raise
            if result['status']:
                # we only need 1 file to work
                break

        elapsed = time.time() - t_start
        self.output_pid(sys.stderr, 'Worker finished processing %s in %.2fs' %
                        (files, elapsed))
        return result

# Platform-specific subclasses.  For the most part, these just have
# logic to determine what files to extract symbols from.

class Dumper_Win32(Dumper):
    fixedFilenameCaseCache = {}

    def ShouldProcess(self, file):
        """This function will allow processing of pdb files that have dll
        or exe files with the same base name next to them."""
        if not Dumper.ShouldProcess(self, file):
            return False
        if file.endswith(".pdb"):
            (path,ext) = os.path.splitext(file)
            if os.path.isfile(path + ".exe") or os.path.isfile(path + ".dll"):
                return True
        return False

    def FixFilenameCase(self, file):
        """Recent versions of Visual C++ put filenames into
        PDB files as all lowercase.  If the file exists
        on the local filesystem, fix it."""

        # Use a cached version if we have one.
        if file in self.fixedFilenameCaseCache:
            return self.fixedFilenameCaseCache[file]

        result = file

        ctypes.windll.kernel32.SetErrorMode(ctypes.c_uint(1))
        if not isinstance(file, unicode):
            file = unicode(file, sys.getfilesystemencoding())
        handle = ctypes.windll.kernel32.CreateFileW(file,
                                                    # GENERIC_READ
                                                    0x80000000,
                                                    # FILE_SHARE_READ
                                                    1,
                                                    None,
                                                    # OPEN_EXISTING
                                                    3,
                                                    0,
                                                    None)
        if handle != -1:
            size = ctypes.windll.kernel32.GetFinalPathNameByHandleW(handle,
                                                                    None,
                                                                    0,
                                                                    0)
            buf = ctypes.create_unicode_buffer(size)
            if ctypes.windll.kernel32.GetFinalPathNameByHandleW(handle,
                                                                buf,
                                                                size,
                                                                0) > 0:
                # The return value of GetFinalPathNameByHandleW uses the
                # '\\?\' prefix.
                result = buf.value.encode(sys.getfilesystemencoding())[4:]
            ctypes.windll.kernel32.CloseHandle(handle)

        # Cache the corrected version to avoid future filesystem hits.
        self.fixedFilenameCaseCache[file] = result
        return result

    def CopyDebug(self, file, debug_file, guid, code_file, code_id):
        def compress(path):
            compressed_file = path[:-1] + '_'
            # ignore makecab's output
            success = subprocess.call(["makecab.exe", "/D",
                                       "CompressionType=LZX", "/D",
                                       "CompressionMemory=21",
                                       path, compressed_file],
                                      stdout=open(os.devnull, 'w'),
                                      stderr=subprocess.STDOUT)
            if success == 0 and os.path.exists(compressed_file):
                os.unlink(path)
                return True
            return False

        rel_path = os.path.join(debug_file,
                                guid,
                                debug_file).replace("\\", "/")
        full_path = os.path.normpath(os.path.join(self.symbol_path,
                                                  rel_path))
        shutil.copyfile(file, full_path)
        if compress(full_path):
            self.output(sys.stdout, rel_path[:-1] + '_')
        else:
            self.output(sys.stdout, rel_path)

        # Copy the binary file as well
        if code_file and code_id:
            full_code_path = os.path.join(os.path.dirname(file),
                                          code_file)
            if os.path.exists(full_code_path):
                rel_path = os.path.join(code_file,
                                        code_id,
                                        code_file).replace("\\", "/")
                full_path = os.path.normpath(os.path.join(self.symbol_path,
                                                          rel_path))
                try:
                    os.makedirs(os.path.dirname(full_path))
                except OSError as e:
                    if e.errno != errno.EEXIST:
                        raise
                shutil.copyfile(full_code_path, full_path)
                if compress(full_path):
                    self.output(sys.stdout, rel_path[:-1] + '_')
                else:
                    self.output(sys.stdout, rel_path)

    def SourceServerIndexing(self, debug_file, guid, sourceFileStream, vcs_root):
        # Creates a .pdb.stream file in the mozilla\objdir to be used for source indexing
        debug_file = os.path.abspath(debug_file)
        streamFilename = debug_file + ".stream"
        stream_output_path = os.path.abspath(streamFilename)
        # Call SourceIndex to create the .stream file
        result = SourceIndex(sourceFileStream, stream_output_path, vcs_root)
        if self.copy_debug:
            pdbstr_path = os.environ.get("PDBSTR_PATH")
            pdbstr = os.path.normpath(pdbstr_path)
            subprocess.call([pdbstr, "-w", "-p:" + os.path.basename(debug_file),
                             "-i:" + os.path.basename(streamFilename), "-s:srcsrv"],
                            cwd=os.path.dirname(stream_output_path))
            # clean up all the .stream files when done
            os.remove(stream_output_path)
        return result

class Dumper_Linux(Dumper):
    objcopy = os.environ['OBJCOPY'] if 'OBJCOPY' in os.environ else 'objcopy'
    def ShouldProcess(self, file):
        """This function will allow processing of files that are
        executable, or end with the .so extension, and additionally
        file(1) reports as being ELF files.  It expects to find the file
        command in PATH."""
        if not Dumper.ShouldProcess(self, file):
            return False
        if file.endswith(".so") or os.access(file, os.X_OK):
            return self.RunFileCommand(file).startswith("ELF")
        return False

    def CopyDebug(self, file, debug_file, guid, code_file, code_id):
        # We want to strip out the debug info, and add a
        # .gnu_debuglink section to the object, so the debugger can
        # actually load our debug info later.
        file_dbg = file + ".dbg"
        if subprocess.call([self.objcopy, '--only-keep-debug', file, file_dbg]) == 0 and \
           subprocess.call([self.objcopy, '--add-gnu-debuglink=%s' % file_dbg, file]) == 0:
            rel_path = os.path.join(debug_file,
                                    guid,
                                    debug_file + ".dbg")
            full_path = os.path.normpath(os.path.join(self.symbol_path,
                                                      rel_path))
            shutil.move(file_dbg, full_path)
            # gzip the shipped debug files
            os.system("gzip -f %s" % full_path)
            self.output(sys.stdout, rel_path + ".gz")
        else:
            if os.path.isfile(file_dbg):
                os.unlink(file_dbg)

class Dumper_Solaris(Dumper):
    def RunFileCommand(self, file):
        """Utility function, returns the output of file(1)"""
        try:
            output = os.popen("file " + file).read()
            return output.split('\t')[1];
        except:
            return ""

    def ShouldProcess(self, file):
        """This function will allow processing of files that are
        executable, or end with the .so extension, and additionally
        file(1) reports as being ELF files.  It expects to find the file
        command in PATH."""
        if not Dumper.ShouldProcess(self, file):
            return False
        if file.endswith(".so") or os.access(file, os.X_OK):
            return self.RunFileCommand(file).startswith("ELF")
        return False

def AfterMac(status, dsymbundle):
    """Cleanup function to run on Macs after we process the file(s)."""
    # CopyDebug will already have been run from Dumper.ProcessFiles
    shutil.rmtree(dsymbundle)

class Dumper_Mac(Dumper):
    def ShouldProcess(self, file):
        """This function will allow processing of files that are
        executable, or end with the .dylib extension, and additionally
        file(1) reports as being Mach-O files.  It expects to find the file
        command in PATH."""
        if not Dumper.ShouldProcess(self, file):
            return False
        if file.endswith(".dylib") or os.access(file, os.X_OK):
            return self.RunFileCommand(file).startswith("Mach-O")
        return False

    def ShouldSkipDir(self, dir):
        """We create .dSYM bundles on the fly, but if someone runs
        buildsymbols twice, we should skip any bundles we created
        previously, otherwise we'll recurse into them and try to 
        dump the inner bits again."""
        if dir.endswith(".dSYM"):
            return True
        return False

    def ProcessFiles(self, files, after=None, after_arg=None):
        # also note, files must be len 1 here, since we're the only ones
        # that ever add more than one file to the list
        self.output_pid(sys.stderr, "Submitting job for Mac pre-processing on file: %s" % (files[0]))
        self.SubmitJob(files[0], 'ProcessFilesWorkMac', args=(files[0],), callback=self.ProcessFilesMacFinished)

    def ProcessFilesMacFinished(self, result):
        if result['status']:
            # kick off new jobs per-arch with our new list of files
            Dumper.ProcessFiles(self, result['files'], after=AfterMac, after_arg=result['files'][0])

    def dump_syms_cmdline(self, file, arch, files):
        '''
        Get the commandline used to invoke dump_syms.
        '''
        # dump_syms wants the path to the original binary and the .dSYM
        # in order to dump all the symbols.
        if len(files) == 2 and file == files[0] and file.endswith('.dSYM'):
            # This is the .dSYM bundle.
            return [self.dump_syms] + arch.split() + ['-g', file, files[1]]
        return Dumper.dump_syms_cmdline(self, file, arch, files)

    def ProcessFilesWorkMac(self, file):
        """dump_syms on Mac needs to be run on a dSYM bundle produced
        by dsymutil(1), so run dsymutil here and pass the bundle name
        down to the superclass method instead."""
        t_start = time.time()
        self.output_pid(sys.stderr, "Worker running Mac pre-processing on file: %s" % (file,))

        # our return is a status and a tuple of files to dump symbols for
        # the extra files are fallbacks; as soon as one is dumped successfully, we stop
        result = { 'status' : False, 'files' : None, 'file_key' : file }
        dsymbundle = file + ".dSYM"
        if os.path.exists(dsymbundle):
            shutil.rmtree(dsymbundle)
        dsymutil = buildconfig.substs['DSYMUTIL']
        # dsymutil takes --arch=foo instead of -a foo like everything else
        try:
            cmd = ([dsymutil] +
                   [a.replace('-a ', '--arch=') for a in self.archs if a] +
                   [file])
            self.output_pid(sys.stderr, ' '.join(cmd))
            subprocess.check_call(cmd, stdout=open(os.devnull, 'w'))
        except subprocess.CalledProcessError as e:
            self.output_pid(sys.stderr, 'Error running dsymutil: %s' % str(e))

        if not os.path.exists(dsymbundle):
            # dsymutil won't produce a .dSYM for files without symbols
            self.output_pid(sys.stderr, "No symbols found in file: %s" % (file,))
            result['status'] = False
            result['files'] = (file, )
            return result

        result['status'] = True
        result['files'] = (dsymbundle, file)
        elapsed = time.time() - t_start
        self.output_pid(sys.stderr, 'Worker finished processing %s in %.2fs' %
                        (file, elapsed))
        return result

    def CopyDebug(self, file, debug_file, guid, code_file, code_id):
        """ProcessFiles has already produced a dSYM bundle, so we should just
        copy that to the destination directory. However, we'll package it
        into a .tar.bz2 because the debug symbols are pretty huge, and
        also because it's a bundle, so it's a directory. |file| here is the
        dSYM bundle, and |debug_file| is the original filename."""
        rel_path = os.path.join(debug_file,
                                guid,
                                os.path.basename(file) + ".tar.bz2")
        full_path = os.path.abspath(os.path.join(self.symbol_path,
                                                  rel_path))
        success = subprocess.call(["tar", "cjf", full_path, os.path.basename(file)],
                                  cwd=os.path.dirname(file),
                                  stdout=open(os.devnull, 'w'), stderr=subprocess.STDOUT)
        if success == 0 and os.path.exists(full_path):
            self.output(sys.stdout, rel_path)

# Entry point if called as a standalone program
def main():
    parser = OptionParser(usage="usage: %prog [options] <dump_syms binary> <symbol store path> <debug info files>")
    parser.add_option("-c", "--copy",
                      action="store_true", dest="copy_debug", default=False,
                      help="Copy debug info files into the same directory structure as symbol files")
    parser.add_option("-a", "--archs",
                      action="store", dest="archs",
                      help="Run dump_syms -a <arch> for each space separated cpu architecture in ARCHS (only on OS X)")
    parser.add_option("-s", "--srcdir",
                      action="append", dest="srcdir", default=[],
                      help="Use SRCDIR to determine relative paths to source files")
    parser.add_option("-v", "--vcs-info",
                      action="store_true", dest="vcsinfo",
                      help="Try to retrieve VCS info for each FILE listed in the output")
    parser.add_option("-i", "--source-index",
                      action="store_true", dest="srcsrv", default=False,
                      help="Add source index information to debug files, making them suitable for use in a source server.")
    parser.add_option("-x", "--exclude",
                      action="append", dest="exclude", default=[], metavar="PATTERN",
                      help="Skip processing files matching PATTERN.")
    parser.add_option("--repo-manifest",
                      action="store", dest="repo_manifest",
                      help="""Get source information from this XML manifest
produced by the `repo manifest -r` command.
""")
    parser.add_option("--install-manifest",
                      action="append", dest="install_manifests",
                      default=[],
                      help="""Use this install manifest to map filenames back
to canonical locations in the source repository. Specify
<install manifest filename>,<install destination> as a comma-separated pair.
""")
    (options, args) = parser.parse_args()

    #check to see if the pdbstr.exe exists
    if options.srcsrv:
        pdbstr = os.environ.get("PDBSTR_PATH")
        if not os.path.exists(pdbstr):
            print >> sys.stderr, "Invalid path to pdbstr.exe - please set/check PDBSTR_PATH.\n"
            sys.exit(1)

    if len(args) < 3:
        parser.error("not enough arguments")
        exit(1)

    try:
        manifests = validate_install_manifests(options.install_manifests)
    except (IOError, ValueError) as e:
        parser.error(str(e))
        exit(1)
    file_mapping = make_file_mapping(manifests)
    dumper = GetPlatformSpecificDumper(dump_syms=args[0],
                                       symbol_path=args[1],
                                       copy_debug=options.copy_debug,
                                       archs=options.archs,
                                       srcdirs=options.srcdir,
                                       vcsinfo=options.vcsinfo,
                                       srcsrv=options.srcsrv,
                                       exclude=options.exclude,
                                       repo_manifest=options.repo_manifest,
                                       file_mapping=file_mapping)

    dumper.Process(*args[2:])
    dumper.Finish()

# run main if run directly
if __name__ == "__main__":
    # set up the multiprocessing infrastructure before we start;
    # note that this needs to be in the __main__ guard, or else Windows will choke
    Dumper.GlobalInit()

    main()