toolkit/crashreporter/tools/symbolstore.py@6474c204b198
toolkit/crashreporter/tools/symbolstore.py
Wed, 31 Dec 2014 06:09:35 +0100
- author
- Michael Schloh von Bennewitz <michael@schloh.com>
- date
- Wed, 31 Dec 2014 06:09:35 +0100
- changeset 0
- 6474c204b198
- permissions
- -rwxr-xr-x
Cloned upstream origin tor-browser at tor-browser-31.3.0esr-4.5-1-build1
revision ID fc1c9ff7c1b2defdbc039f12214767608f46423f for hacking purpose.
1 #!/bin/env python
2 # This Source Code Form is subject to the terms of the Mozilla Public
3 # License, v. 2.0. If a copy of the MPL was not distributed with this
4 # file, You can obtain one at http://mozilla.org/MPL/2.0/.
5 #
6 # Usage: symbolstore.py <params> <dump_syms path> <symbol store path>
7 # <debug info files or dirs>
8 # Runs dump_syms on each debug info file specified on the command line,
9 # then places the resulting symbol file in the proper directory
10 # structure in the symbol store path. Accepts multiple files
11 # on the command line, so can be called as part of a pipe using
12 # find <dir> | xargs symbolstore.pl <dump_syms> <storepath>
13 # But really, you might just want to pass it <dir>.
14 #
15 # Parameters accepted:
16 # -c : Copy debug info files to the same directory structure
17 # as sym files
18 # -a "<archs>" : Run dump_syms -a <arch> for each space separated
19 # cpu architecture in <archs> (only on OS X)
20 # -s <srcdir> : Use <srcdir> as the top source directory to
21 # generate relative filenames.
23 import sys
24 import platform
25 import os
26 import re
27 import shutil
28 import textwrap
29 import fnmatch
30 import subprocess
31 import urlparse
32 import multiprocessing
33 import collections
34 from optparse import OptionParser
35 from xml.dom.minidom import parse
37 # Utility classes
39 class VCSFileInfo:
40 """ A base class for version-controlled file information. Ensures that the
41 following attributes are generated only once (successfully):
43 self.root
44 self.clean_root
45 self.revision
46 self.filename
48 The attributes are generated by a single call to the GetRoot,
49 GetRevision, and GetFilename methods. Those methods are explicitly not
50 implemented here and must be implemented in derived classes. """
52 def __init__(self, file):
53 if not file:
54 raise ValueError
55 self.file = file
57 def __getattr__(self, name):
58 """ __getattr__ is only called for attributes that are not set on self,
59 so setting self.[attr] will prevent future calls to the GetRoot,
60 GetRevision, and GetFilename methods. We don't set the values on
61 failure on the off chance that a future call might succeed. """
63 if name == "root":
64 root = self.GetRoot()
65 if root:
66 self.root = root
67 return root
69 elif name == "clean_root":
70 clean_root = self.GetCleanRoot()
71 if clean_root:
72 self.clean_root = clean_root
73 return clean_root
75 elif name == "revision":
76 revision = self.GetRevision()
77 if revision:
78 self.revision = revision
79 return revision
81 elif name == "filename":
82 filename = self.GetFilename()
83 if filename:
84 self.filename = filename
85 return filename
87 raise AttributeError
89 def GetRoot(self):
90 """ This method should return the unmodified root for the file or 'None'
91 on failure. """
92 raise NotImplementedError
94 def GetCleanRoot(self):
95 """ This method should return the repository root for the file or 'None'
96 on failure. """
97 raise NotImplementedErrors
99 def GetRevision(self):
100 """ This method should return the revision number for the file or 'None'
101 on failure. """
102 raise NotImplementedError
104 def GetFilename(self):
105 """ This method should return the repository-specific filename for the
106 file or 'None' on failure. """
107 raise NotImplementedError
110 # This regex separates protocol and optional username/password from a url.
111 # For instance, all the following urls will be transformed into
112 # 'foo.com/bar':
113 #
114 # http://foo.com/bar
115 # svn+ssh://user@foo.com/bar
116 # svn+ssh://user:pass@foo.com/bar
117 #
118 rootRegex = re.compile(r'^\S+?:/+(?:[^\s/]*@)?(\S+)$')
120 def read_output(*args):
121 (stdout, _) = subprocess.Popen(args=args, stdout=subprocess.PIPE).communicate()
122 return stdout.rstrip()
124 class HGRepoInfo:
125 def __init__(self, path):
126 self.path = path
127 rev = read_output('hg', '-R', path,
128 'parent', '--template={node|short}')
129 # Look for the default hg path. If SRVSRV_ROOT is set, we
130 # don't bother asking hg.
131 hg_root = os.environ.get("SRCSRV_ROOT")
132 if hg_root:
133 root = hg_root
134 else:
135 root = read_output('hg', '-R', path,
136 'showconfig', 'paths.default')
137 if not root:
138 print >> sys.stderr, "Failed to get HG Repo for %s" % path
139 cleanroot = None
140 if root:
141 match = rootRegex.match(root)
142 if match:
143 cleanroot = match.group(1)
144 if cleanroot.endswith('/'):
145 cleanroot = cleanroot[:-1]
146 if cleanroot is None:
147 print >> sys.stderr, textwrap.dedent("""\
148 Could not determine repo info for %s. This is either not a clone of the web-based
149 repository, or you have not specified SRCSRV_ROOT, or the clone is corrupt.""") % path
150 sys.exit(1)
151 self.rev = rev
152 self.root = root
153 self.cleanroot = cleanroot
155 def GetFileInfo(self, file):
156 return HGFileInfo(file, self)
158 class HGFileInfo(VCSFileInfo):
159 def __init__(self, file, repo):
160 VCSFileInfo.__init__(self, file)
161 self.repo = repo
162 self.file = os.path.relpath(file, repo.path)
164 def GetRoot(self):
165 return self.repo.root
167 def GetCleanRoot(self):
168 return self.repo.cleanroot
170 def GetRevision(self):
171 return self.repo.rev
173 def GetFilename(self):
174 if self.revision and self.clean_root:
175 return "hg:%s:%s:%s" % (self.clean_root, self.file, self.revision)
176 return self.file
178 class GitRepoInfo:
179 """
180 Info about a local git repository. Does not currently
181 support discovering info about a git clone, the info must be
182 provided out-of-band.
183 """
184 def __init__(self, path, rev, root):
185 self.path = path
186 cleanroot = None
187 if root:
188 match = rootRegex.match(root)
189 if match:
190 cleanroot = match.group(1)
191 if cleanroot.endswith('/'):
192 cleanroot = cleanroot[:-1]
193 if cleanroot is None:
194 print >> sys.stderr, textwrap.dedent("""\
195 Could not determine repo info for %s (%s). This is either not a clone of a web-based
196 repository, or you have not specified SRCSRV_ROOT, or the clone is corrupt.""") % (path, root)
197 sys.exit(1)
198 self.rev = rev
199 self.cleanroot = cleanroot
201 def GetFileInfo(self, file):
202 return GitFileInfo(file, self)
204 class GitFileInfo(VCSFileInfo):
205 def __init__(self, file, repo):
206 VCSFileInfo.__init__(self, file)
207 self.repo = repo
208 self.file = os.path.relpath(file, repo.path)
210 def GetRoot(self):
211 return self.repo.path
213 def GetCleanRoot(self):
214 return self.repo.cleanroot
216 def GetRevision(self):
217 return self.repo.rev
219 def GetFilename(self):
220 if self.revision and self.clean_root:
221 return "git:%s:%s:%s" % (self.clean_root, self.file, self.revision)
222 return self.file
224 # Utility functions
226 # A cache of files for which VCS info has already been determined. Used to
227 # prevent extra filesystem activity or process launching.
228 vcsFileInfoCache = {}
230 def IsInDir(file, dir):
231 # the lower() is to handle win32+vc8, where
232 # the source filenames come out all lowercase,
233 # but the srcdir can be mixed case
234 return os.path.abspath(file).lower().startswith(os.path.abspath(dir).lower())
236 def GetVCSFilenameFromSrcdir(file, srcdir):
237 if srcdir not in Dumper.srcdirRepoInfo:
238 # Not in cache, so find it adnd cache it
239 if os.path.isdir(os.path.join(srcdir, '.hg')):
240 Dumper.srcdirRepoInfo[srcdir] = HGRepoInfo(srcdir)
241 else:
242 # Unknown VCS or file is not in a repo.
243 return None
244 return Dumper.srcdirRepoInfo[srcdir].GetFileInfo(file)
246 def GetVCSFilename(file, srcdirs):
247 """Given a full path to a file, and the top source directory,
248 look for version control information about this file, and return
249 a tuple containing
250 1) a specially formatted filename that contains the VCS type,
251 VCS location, relative filename, and revision number, formatted like:
252 vcs:vcs location:filename:revision
253 For example:
254 cvs:cvs.mozilla.org/cvsroot:mozilla/browser/app/nsBrowserApp.cpp:1.36
255 2) the unmodified root information if it exists"""
256 (path, filename) = os.path.split(file)
257 if path == '' or filename == '':
258 return (file, None)
260 fileInfo = None
261 root = ''
262 if file in vcsFileInfoCache:
263 # Already cached this info, use it.
264 fileInfo = vcsFileInfoCache[file]
265 else:
266 for srcdir in srcdirs:
267 if not IsInDir(file, srcdir):
268 continue
269 fileInfo = GetVCSFilenameFromSrcdir(file, srcdir)
270 if fileInfo:
271 vcsFileInfoCache[file] = fileInfo
272 break
274 if fileInfo:
275 file = fileInfo.filename
276 root = fileInfo.root
278 # we want forward slashes on win32 paths
279 return (file.replace("\\", "/"), root)
281 def GetPlatformSpecificDumper(**kwargs):
282 """This function simply returns a instance of a subclass of Dumper
283 that is appropriate for the current platform."""
284 # Python 2.5 has a bug where platform.system() returns 'Microsoft'.
285 # Remove this when we no longer support Python 2.5.
286 return {'Windows': Dumper_Win32,
287 'Microsoft': Dumper_Win32,
288 'Linux': Dumper_Linux,
289 'Sunos5': Dumper_Solaris,
290 'Darwin': Dumper_Mac}[platform.system()](**kwargs)
292 def SourceIndex(fileStream, outputPath, vcs_root):
293 """Takes a list of files, writes info to a data block in a .stream file"""
294 # Creates a .pdb.stream file in the mozilla\objdir to be used for source indexing
295 # Create the srcsrv data block that indexes the pdb file
296 result = True
297 pdbStreamFile = open(outputPath, "w")
298 pdbStreamFile.write('''SRCSRV: ini ------------------------------------------------\r\nVERSION=2\r\nINDEXVERSION=2\r\nVERCTRL=http\r\nSRCSRV: variables ------------------------------------------\r\nHGSERVER=''')
299 pdbStreamFile.write(vcs_root)
300 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''')
301 pdbStreamFile.write(fileStream) # can't do string interpolation because the source server also uses this and so there are % in the above
302 pdbStreamFile.write("SRCSRV: end ------------------------------------------------\r\n\n")
303 pdbStreamFile.close()
304 return result
306 def WorkerInitializer(cls, lock, srcdirRepoInfo):
307 """Windows worker processes won't have run GlobalInit, and due to a lack of fork(),
308 won't inherit the class variables from the parent. They only need a few variables,
309 so we run an initializer to set them. Redundant but harmless on other platforms."""
310 cls.lock = lock
311 cls.srcdirRepoInfo = srcdirRepoInfo
313 def StartProcessFilesWork(dumper, files, arch_num, arch, vcs_root, after, after_arg):
314 """multiprocessing can't handle methods as Process targets, so we define
315 a simple wrapper function around the work method."""
316 return dumper.ProcessFilesWork(files, arch_num, arch, vcs_root, after, after_arg)
318 class Dumper:
319 """This class can dump symbols from a file with debug info, and
320 store the output in a directory structure that is valid for use as
321 a Breakpad symbol server. Requires a path to a dump_syms binary--
322 |dump_syms| and a directory to store symbols in--|symbol_path|.
323 Optionally takes a list of processor architectures to process from
324 each debug file--|archs|, the full path to the top source
325 directory--|srcdir|, for generating relative source file names,
326 and an option to copy debug info files alongside the dumped
327 symbol files--|copy_debug|, mostly useful for creating a
328 Microsoft Symbol Server from the resulting output.
330 You don't want to use this directly if you intend to call
331 ProcessDir. Instead, call GetPlatformSpecificDumper to
332 get an instance of a subclass.
334 Processing is performed asynchronously via worker processes; in
335 order to wait for processing to finish and cleanup correctly, you
336 must call Finish after all Process/ProcessDir calls have been made.
337 You must also call Dumper.GlobalInit before creating or using any
338 instances."""
339 def __init__(self, dump_syms, symbol_path,
340 archs=None,
341 srcdirs=[],
342 copy_debug=False,
343 vcsinfo=False,
344 srcsrv=False,
345 exclude=[],
346 repo_manifest=None):
347 # popen likes absolute paths, at least on windows
348 self.dump_syms = os.path.abspath(dump_syms)
349 self.symbol_path = symbol_path
350 if archs is None:
351 # makes the loop logic simpler
352 self.archs = ['']
353 else:
354 self.archs = ['-a %s' % a for a in archs.split()]
355 self.srcdirs = [os.path.normpath(a) for a in srcdirs]
356 self.copy_debug = copy_debug
357 self.vcsinfo = vcsinfo
358 self.srcsrv = srcsrv
359 self.exclude = exclude[:]
360 if repo_manifest:
361 self.parse_repo_manifest(repo_manifest)
363 # book-keeping to keep track of our jobs and the cleanup work per file tuple
364 self.files_record = {}
365 self.jobs_record = collections.defaultdict(int)
367 @classmethod
368 def GlobalInit(cls, module=multiprocessing):
369 """Initialize the class globals for the multiprocessing setup; must
370 be called before any Dumper instances are created and used. Test cases
371 may pass in a different module to supply Manager and Pool objects,
372 usually multiprocessing.dummy."""
373 num_cpus = module.cpu_count()
374 if num_cpus is None:
375 # assume a dual core machine if we can't find out for some reason
376 # probably better on single core anyway due to I/O constraints
377 num_cpus = 2
379 # have to create any locks etc before the pool
380 cls.manager = module.Manager()
381 cls.jobs_condition = Dumper.manager.Condition()
382 cls.lock = Dumper.manager.RLock()
383 cls.srcdirRepoInfo = Dumper.manager.dict()
384 cls.pool = module.Pool(num_cpus, WorkerInitializer,
385 (cls, cls.lock, cls.srcdirRepoInfo))
387 def JobStarted(self, file_key):
388 """Increments the number of submitted jobs for the specified key file,
389 defined as the original file we processed; note that a single key file
390 can generate up to 1 + len(self.archs) jobs in the Mac case."""
391 with Dumper.jobs_condition:
392 self.jobs_record[file_key] += 1
393 Dumper.jobs_condition.notify_all()
395 def JobFinished(self, file_key):
396 """Decrements the number of submitted jobs for the specified key file,
397 defined as the original file we processed; once the count is back to 0,
398 remove the entry from our record."""
399 with Dumper.jobs_condition:
400 self.jobs_record[file_key] -= 1
402 if self.jobs_record[file_key] == 0:
403 del self.jobs_record[file_key]
405 Dumper.jobs_condition.notify_all()
407 def output(self, dest, output_str):
408 """Writes |output_str| to |dest|, holding |lock|;
409 terminates with a newline."""
410 with Dumper.lock:
411 dest.write(output_str + "\n")
412 dest.flush()
414 def output_pid(self, dest, output_str):
415 """Debugging output; prepends the pid to the string."""
416 self.output(dest, "%d: %s" % (os.getpid(), output_str))
418 def parse_repo_manifest(self, repo_manifest):
419 """
420 Parse an XML manifest of repository info as produced
421 by the `repo manifest -r` command.
422 """
423 doc = parse(repo_manifest)
424 if doc.firstChild.tagName != "manifest":
425 return
426 # First, get remotes.
427 def ensure_slash(u):
428 if not u.endswith("/"):
429 return u + "/"
430 return u
431 remotes = dict([(r.getAttribute("name"), ensure_slash(r.getAttribute("fetch"))) for r in doc.getElementsByTagName("remote")])
432 # And default remote.
433 default_remote = None
434 if doc.getElementsByTagName("default"):
435 default_remote = doc.getElementsByTagName("default")[0].getAttribute("remote")
436 # Now get projects. Assume they're relative to repo_manifest.
437 base_dir = os.path.abspath(os.path.dirname(repo_manifest))
438 for proj in doc.getElementsByTagName("project"):
439 # name is the repository URL relative to the remote path.
440 name = proj.getAttribute("name")
441 # path is the path on-disk, relative to the manifest file.
442 path = proj.getAttribute("path")
443 # revision is the changeset ID.
444 rev = proj.getAttribute("revision")
445 # remote is the base URL to use.
446 remote = proj.getAttribute("remote")
447 # remote defaults to the <default remote>.
448 if not remote:
449 remote = default_remote
450 # path defaults to name.
451 if not path:
452 path = name
453 if not (name and path and rev and remote):
454 print "Skipping project %s" % proj.toxml()
455 continue
456 remote = remotes[remote]
457 # Turn git URLs into http URLs so that urljoin works.
458 if remote.startswith("git:"):
459 remote = "http" + remote[3:]
460 # Add this project to srcdirs.
461 srcdir = os.path.join(base_dir, path)
462 self.srcdirs.append(srcdir)
463 # And cache its VCS file info. Currently all repos mentioned
464 # in a repo manifest are assumed to be git.
465 root = urlparse.urljoin(remote, name)
466 Dumper.srcdirRepoInfo[srcdir] = GitRepoInfo(srcdir, rev, root)
468 # subclasses override this
469 def ShouldProcess(self, file):
470 return not any(fnmatch.fnmatch(os.path.basename(file), exclude) for exclude in self.exclude)
472 # and can override this
473 def ShouldSkipDir(self, dir):
474 return False
476 def RunFileCommand(self, file):
477 """Utility function, returns the output of file(1)"""
478 try:
479 # we use -L to read the targets of symlinks,
480 # and -b to print just the content, not the filename
481 return os.popen("file -Lb " + file).read()
482 except:
483 return ""
485 # This is a no-op except on Win32
486 def FixFilenameCase(self, file):
487 return file
489 # This is a no-op except on Win32
490 def SourceServerIndexing(self, debug_file, guid, sourceFileStream, vcs_root):
491 return ""
493 # subclasses override this if they want to support this
494 def CopyDebug(self, file, debug_file, guid):
495 pass
497 def Finish(self, stop_pool=True):
498 """Wait for the expected number of jobs to be submitted, and then
499 wait for the pool to finish processing them. By default, will close
500 and clear the pool, but for testcases that need multiple runs, pass
501 stop_pool = False."""
502 with Dumper.jobs_condition:
503 while len(self.jobs_record) != 0:
504 Dumper.jobs_condition.wait()
505 if stop_pool:
506 Dumper.pool.close()
507 Dumper.pool.join()
509 def Process(self, file_or_dir):
510 """Process a file or all the (valid) files in a directory; processing is performed
511 asynchronously, and Finish must be called to wait for it complete and cleanup."""
512 if os.path.isdir(file_or_dir) and not self.ShouldSkipDir(file_or_dir):
513 self.ProcessDir(file_or_dir)
514 elif os.path.isfile(file_or_dir):
515 self.ProcessFiles((file_or_dir,))
517 def ProcessDir(self, dir):
518 """Process all the valid files in this directory. Valid files
519 are determined by calling ShouldProcess; processing is performed
520 asynchronously, and Finish must be called to wait for it complete and cleanup."""
521 for root, dirs, files in os.walk(dir):
522 for d in dirs[:]:
523 if self.ShouldSkipDir(d):
524 dirs.remove(d)
525 for f in files:
526 fullpath = os.path.join(root, f)
527 if self.ShouldProcess(fullpath):
528 self.ProcessFiles((fullpath,))
530 def SubmitJob(self, file_key, func, args, callback):
531 """Submits a job to the pool of workers; increments the number of submitted jobs."""
532 self.JobStarted(file_key)
533 res = Dumper.pool.apply_async(func, args=args, callback=callback)
535 def ProcessFilesFinished(self, res):
536 """Callback from multiprocesing when ProcessFilesWork finishes;
537 run the cleanup work, if any"""
538 self.JobFinished(res['files'][-1])
539 # only run the cleanup function once per tuple of files
540 self.files_record[res['files']] += 1
541 if self.files_record[res['files']] == len(self.archs):
542 del self.files_record[res['files']]
543 if res['after']:
544 res['after'](res['status'], res['after_arg'])
546 def ProcessFiles(self, files, after=None, after_arg=None):
547 """Dump symbols from these files into a symbol file, stored
548 in the proper directory structure in |symbol_path|; processing is performed
549 asynchronously, and Finish must be called to wait for it complete and cleanup.
550 All files after the first are fallbacks in case the first file does not process
551 successfully; if it does, no other files will be touched."""
552 self.output_pid(sys.stderr, "Submitting jobs for files: %s" % str(files))
554 # tries to get the vcs root from the .mozconfig first - if it's not set
555 # the tinderbox vcs path will be assigned further down
556 vcs_root = os.environ.get("SRCSRV_ROOT")
557 for arch_num, arch in enumerate(self.archs):
558 self.files_record[files] = 0 # record that we submitted jobs for this tuple of files
559 self.SubmitJob(files[-1], StartProcessFilesWork, args=(self, files, arch_num, arch, vcs_root, after, after_arg), callback=self.ProcessFilesFinished)
561 def ProcessFilesWork(self, files, arch_num, arch, vcs_root, after, after_arg):
562 self.output_pid(sys.stderr, "Worker processing files: %s" % (files,))
564 # our result is a status, a cleanup function, an argument to that function, and the tuple of files we were called on
565 result = { 'status' : False, 'after' : after, 'after_arg' : after_arg, 'files' : files }
567 sourceFileStream = ''
568 for file in files:
569 # files is a tuple of files, containing fallbacks in case the first file doesn't process successfully
570 try:
571 proc = subprocess.Popen([self.dump_syms] + arch.split() + [file],
572 stdout=subprocess.PIPE)
573 module_line = proc.stdout.next()
574 if module_line.startswith("MODULE"):
575 # MODULE os cpu guid debug_file
576 (guid, debug_file) = (module_line.split())[3:5]
577 # strip off .pdb extensions, and append .sym
578 sym_file = re.sub("\.pdb$", "", debug_file) + ".sym"
579 # we do want forward slashes here
580 rel_path = os.path.join(debug_file,
581 guid,
582 sym_file).replace("\\", "/")
583 full_path = os.path.normpath(os.path.join(self.symbol_path,
584 rel_path))
585 try:
586 os.makedirs(os.path.dirname(full_path))
587 except OSError: # already exists
588 pass
589 f = open(full_path, "w")
590 f.write(module_line)
591 # now process the rest of the output
592 for line in proc.stdout:
593 if line.startswith("FILE"):
594 # FILE index filename
595 (x, index, filename) = line.rstrip().split(None, 2)
596 if sys.platform == "sunos5":
597 for srcdir in self.srcdirs:
598 start = filename.find(self.srcdir)
599 if start != -1:
600 filename = filename[start:]
601 break
602 filename = self.FixFilenameCase(filename)
603 sourcepath = filename
604 if self.vcsinfo:
605 (filename, rootname) = GetVCSFilename(filename, self.srcdirs)
606 # sets vcs_root in case the loop through files were to end on an empty rootname
607 if vcs_root is None:
608 if rootname:
609 vcs_root = rootname
610 # gather up files with hg for indexing
611 if filename.startswith("hg"):
612 (ver, checkout, source_file, revision) = filename.split(":", 3)
613 sourceFileStream += sourcepath + "*" + source_file + '*' + revision + "\r\n"
614 f.write("FILE %s %s\n" % (index, filename))
615 else:
616 # pass through all other lines unchanged
617 f.write(line)
618 # we want to return true only if at least one line is not a MODULE or FILE line
619 result['status'] = True
620 f.close()
621 proc.wait()
622 # we output relative paths so callers can get a list of what
623 # was generated
624 self.output(sys.stdout, rel_path)
625 if self.srcsrv and vcs_root:
626 # add source server indexing to the pdb file
627 self.SourceServerIndexing(file, guid, sourceFileStream, vcs_root)
628 # only copy debug the first time if we have multiple architectures
629 if self.copy_debug and arch_num == 0:
630 self.CopyDebug(file, debug_file, guid)
631 except StopIteration:
632 pass
633 except e:
634 self.output(sys.stderr, "Unexpected error: %s" % (str(e),))
635 raise
636 if result['status']:
637 # we only need 1 file to work
638 break
639 return result
641 # Platform-specific subclasses. For the most part, these just have
642 # logic to determine what files to extract symbols from.
644 class Dumper_Win32(Dumper):
645 fixedFilenameCaseCache = {}
647 def ShouldProcess(self, file):
648 """This function will allow processing of pdb files that have dll
649 or exe files with the same base name next to them."""
650 if not Dumper.ShouldProcess(self, file):
651 return False
652 if file.endswith(".pdb"):
653 (path,ext) = os.path.splitext(file)
654 if os.path.isfile(path + ".exe") or os.path.isfile(path + ".dll"):
655 return True
656 return False
658 def FixFilenameCase(self, file):
659 """Recent versions of Visual C++ put filenames into
660 PDB files as all lowercase. If the file exists
661 on the local filesystem, fix it."""
663 # Use a cached version if we have one.
664 if file in self.fixedFilenameCaseCache:
665 return self.fixedFilenameCaseCache[file]
667 result = file
669 (path, filename) = os.path.split(file)
670 if os.path.isdir(path):
671 lc_filename = filename.lower()
672 for f in os.listdir(path):
673 if f.lower() == lc_filename:
674 result = os.path.join(path, f)
675 break
677 # Cache the corrected version to avoid future filesystem hits.
678 self.fixedFilenameCaseCache[file] = result
679 return result
681 def CopyDebug(self, file, debug_file, guid):
682 rel_path = os.path.join(debug_file,
683 guid,
684 debug_file).replace("\\", "/")
685 full_path = os.path.normpath(os.path.join(self.symbol_path,
686 rel_path))
687 shutil.copyfile(file, full_path)
688 # try compressing it
689 compressed_file = os.path.splitext(full_path)[0] + ".pd_"
690 # ignore makecab's output
691 success = subprocess.call(["makecab.exe", "/D", "CompressionType=LZX", "/D",
692 "CompressionMemory=21",
693 full_path, compressed_file],
694 stdout=open("NUL:","w"), stderr=subprocess.STDOUT)
695 if success == 0 and os.path.exists(compressed_file):
696 os.unlink(full_path)
697 self.output(sys.stdout, os.path.splitext(rel_path)[0] + ".pd_")
698 else:
699 self.output(sys.stdout, rel_path)
701 def SourceServerIndexing(self, debug_file, guid, sourceFileStream, vcs_root):
702 # Creates a .pdb.stream file in the mozilla\objdir to be used for source indexing
703 debug_file = os.path.abspath(debug_file)
704 streamFilename = debug_file + ".stream"
705 stream_output_path = os.path.abspath(streamFilename)
706 # Call SourceIndex to create the .stream file
707 result = SourceIndex(sourceFileStream, stream_output_path, vcs_root)
708 if self.copy_debug:
709 pdbstr_path = os.environ.get("PDBSTR_PATH")
710 pdbstr = os.path.normpath(pdbstr_path)
711 subprocess.call([pdbstr, "-w", "-p:" + os.path.basename(debug_file),
712 "-i:" + os.path.basename(streamFilename), "-s:srcsrv"],
713 cwd=os.path.dirname(stream_output_path))
714 # clean up all the .stream files when done
715 os.remove(stream_output_path)
716 return result
718 class Dumper_Linux(Dumper):
719 objcopy = os.environ['OBJCOPY'] if 'OBJCOPY' in os.environ else 'objcopy'
720 def ShouldProcess(self, file):
721 """This function will allow processing of files that are
722 executable, or end with the .so extension, and additionally
723 file(1) reports as being ELF files. It expects to find the file
724 command in PATH."""
725 if not Dumper.ShouldProcess(self, file):
726 return False
727 if file.endswith(".so") or os.access(file, os.X_OK):
728 return self.RunFileCommand(file).startswith("ELF")
729 return False
731 def CopyDebug(self, file, debug_file, guid):
732 # We want to strip out the debug info, and add a
733 # .gnu_debuglink section to the object, so the debugger can
734 # actually load our debug info later.
735 file_dbg = file + ".dbg"
736 if subprocess.call([self.objcopy, '--only-keep-debug', file, file_dbg]) == 0 and \
737 subprocess.call([self.objcopy, '--add-gnu-debuglink=%s' % file_dbg, file]) == 0:
738 rel_path = os.path.join(debug_file,
739 guid,
740 debug_file + ".dbg")
741 full_path = os.path.normpath(os.path.join(self.symbol_path,
742 rel_path))
743 shutil.move(file_dbg, full_path)
744 # gzip the shipped debug files
745 os.system("gzip %s" % full_path)
746 self.output(sys.stdout, rel_path + ".gz")
747 else:
748 if os.path.isfile(file_dbg):
749 os.unlink(file_dbg)
751 class Dumper_Solaris(Dumper):
752 def RunFileCommand(self, file):
753 """Utility function, returns the output of file(1)"""
754 try:
755 output = os.popen("file " + file).read()
756 return output.split('\t')[1];
757 except:
758 return ""
760 def ShouldProcess(self, file):
761 """This function will allow processing of files that are
762 executable, or end with the .so extension, and additionally
763 file(1) reports as being ELF files. It expects to find the file
764 command in PATH."""
765 if not Dumper.ShouldProcess(self, file):
766 return False
767 if file.endswith(".so") or os.access(file, os.X_OK):
768 return self.RunFileCommand(file).startswith("ELF")
769 return False
771 def StartProcessFilesWorkMac(dumper, file):
772 """multiprocessing can't handle methods as Process targets, so we define
773 a simple wrapper function around the work method."""
774 return dumper.ProcessFilesWorkMac(file)
776 def AfterMac(status, dsymbundle):
777 """Cleanup function to run on Macs after we process the file(s)."""
778 # CopyDebug will already have been run from Dumper.ProcessFiles
779 shutil.rmtree(dsymbundle)
781 class Dumper_Mac(Dumper):
782 def ShouldProcess(self, file):
783 """This function will allow processing of files that are
784 executable, or end with the .dylib extension, and additionally
785 file(1) reports as being Mach-O files. It expects to find the file
786 command in PATH."""
787 if not Dumper.ShouldProcess(self, file):
788 return False
789 if file.endswith(".dylib") or os.access(file, os.X_OK):
790 return self.RunFileCommand(file).startswith("Mach-O")
791 return False
793 def ShouldSkipDir(self, dir):
794 """We create .dSYM bundles on the fly, but if someone runs
795 buildsymbols twice, we should skip any bundles we created
796 previously, otherwise we'll recurse into them and try to
797 dump the inner bits again."""
798 if dir.endswith(".dSYM"):
799 return True
800 return False
802 def ProcessFiles(self, files, after=None, after_arg=None):
803 # also note, files must be len 1 here, since we're the only ones
804 # that ever add more than one file to the list
805 self.output_pid(sys.stderr, "Submitting job for Mac pre-processing on file: %s" % (files[0]))
806 self.SubmitJob(files[0], StartProcessFilesWorkMac, args=(self, files[0]), callback=self.ProcessFilesMacFinished)
808 def ProcessFilesMacFinished(self, result):
809 if result['status']:
810 # kick off new jobs per-arch with our new list of files
811 Dumper.ProcessFiles(self, result['files'], after=AfterMac, after_arg=result['files'][0])
812 # only decrement jobs *after* that, since otherwise we'll remove the record for this file
813 self.JobFinished(result['files'][-1])
815 def ProcessFilesWorkMac(self, file):
816 """dump_syms on Mac needs to be run on a dSYM bundle produced
817 by dsymutil(1), so run dsymutil here and pass the bundle name
818 down to the superclass method instead."""
819 self.output_pid(sys.stderr, "Worker running Mac pre-processing on file: %s" % (file,))
821 # our return is a status and a tuple of files to dump symbols for
822 # the extra files are fallbacks; as soon as one is dumped successfully, we stop
823 result = { 'status' : False, 'files' : None, 'file_key' : file }
824 dsymbundle = file + ".dSYM"
825 if os.path.exists(dsymbundle):
826 shutil.rmtree(dsymbundle)
827 # dsymutil takes --arch=foo instead of -a foo like everything else
828 subprocess.call(["dsymutil"] + [a.replace('-a ', '--arch=') for a in self.archs if a]
829 + [file],
830 stdout=open("/dev/null","w"))
831 if not os.path.exists(dsymbundle):
832 # dsymutil won't produce a .dSYM for files without symbols
833 self.output_pid(sys.stderr, "No symbols found in file: %s" % (file,))
834 result['status'] = False
835 result['files'] = (file, )
836 return result
838 result['status'] = True
839 result['files'] = (dsymbundle, file)
840 return result
842 def CopyDebug(self, file, debug_file, guid):
843 """ProcessFiles has already produced a dSYM bundle, so we should just
844 copy that to the destination directory. However, we'll package it
845 into a .tar.bz2 because the debug symbols are pretty huge, and
846 also because it's a bundle, so it's a directory. |file| here is the
847 dSYM bundle, and |debug_file| is the original filename."""
848 rel_path = os.path.join(debug_file,
849 guid,
850 os.path.basename(file) + ".tar.bz2")
851 full_path = os.path.abspath(os.path.join(self.symbol_path,
852 rel_path))
853 success = subprocess.call(["tar", "cjf", full_path, os.path.basename(file)],
854 cwd=os.path.dirname(file),
855 stdout=open("/dev/null","w"), stderr=subprocess.STDOUT)
856 if success == 0 and os.path.exists(full_path):
857 self.output(sys.stdout, rel_path)
859 # Entry point if called as a standalone program
860 def main():
861 parser = OptionParser(usage="usage: %prog [options] <dump_syms binary> <symbol store path> <debug info files>")
862 parser.add_option("-c", "--copy",
863 action="store_true", dest="copy_debug", default=False,
864 help="Copy debug info files into the same directory structure as symbol files")
865 parser.add_option("-a", "--archs",
866 action="store", dest="archs",
867 help="Run dump_syms -a <arch> for each space separated cpu architecture in ARCHS (only on OS X)")
868 parser.add_option("-s", "--srcdir",
869 action="append", dest="srcdir", default=[],
870 help="Use SRCDIR to determine relative paths to source files")
871 parser.add_option("-v", "--vcs-info",
872 action="store_true", dest="vcsinfo",
873 help="Try to retrieve VCS info for each FILE listed in the output")
874 parser.add_option("-i", "--source-index",
875 action="store_true", dest="srcsrv", default=False,
876 help="Add source index information to debug files, making them suitable for use in a source server.")
877 parser.add_option("-x", "--exclude",
878 action="append", dest="exclude", default=[], metavar="PATTERN",
879 help="Skip processing files matching PATTERN.")
880 parser.add_option("--repo-manifest",
881 action="store", dest="repo_manifest",
882 help="""Get source information from this XML manifest
883 produced by the `repo manifest -r` command.
884 """)
885 (options, args) = parser.parse_args()
887 #check to see if the pdbstr.exe exists
888 if options.srcsrv:
889 pdbstr = os.environ.get("PDBSTR_PATH")
890 if not os.path.exists(pdbstr):
891 print >> sys.stderr, "Invalid path to pdbstr.exe - please set/check PDBSTR_PATH.\n"
892 sys.exit(1)
894 if len(args) < 3:
895 parser.error("not enough arguments")
896 exit(1)
898 dumper = GetPlatformSpecificDumper(dump_syms=args[0],
899 symbol_path=args[1],
900 copy_debug=options.copy_debug,
901 archs=options.archs,
902 srcdirs=options.srcdir,
903 vcsinfo=options.vcsinfo,
904 srcsrv=options.srcsrv,
905 exclude=options.exclude,
906 repo_manifest=options.repo_manifest)
907 for arg in args[2:]:
908 dumper.Process(arg)
909 dumper.Finish()
911 # run main if run directly
912 if __name__ == "__main__":
913 # set up the multiprocessing infrastructure before we start;
914 # note that this needs to be in the __main__ guard, or else Windows will choke
915 Dumper.GlobalInit()
917 main()
