# -*- Mode: Python; test-case-name: whipper.test.test_program_cdparanoia -*-
# vi:si:et:sw=4:sts=4:ts=4
# Copyright (C) 2009 Thomas Vander Stichele
# This file is part of whipper.
#
# whipper is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# whipper is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with whipper. If not, see .
import errno
import os
import re
import shutil
import stat
import subprocess
import tempfile
import time
from whipper.common import common
from whipper.common import task as ctask
from whipper.extern import asyncsub
from whipper.extern.task import task
import logging
logger = logging.getLogger(__name__)
class FileSizeError(Exception):
"""The given path does not have the expected size."""
message = None
def __init__(self, path, message):
self.args = (path, message)
self.path = path
self.message = message
class ReturnCodeError(Exception):
"""The program had a non-zero return code."""
def __init__(self, returncode):
self.args = (returncode, )
self.returncode = returncode
class ChecksumException(Exception):
pass
# example:
# ##: 0 [read] @ 24696
_PROGRESS_RE = re.compile(r"""
^\#\#: (?P.+)\s # function code
\[(?P.*)\]\s@\s # [function name] @
(?P\d+) # offset in words (2-byte one channel value)
""", re.VERBOSE)
_ERROR_RE = re.compile("^scsi_read error:")
# from reading cdparanoia source code, it looks like offset is reported in
# number of single-channel samples, ie. 2 bytes (word) per unit, and absolute
class ProgressParser:
"""Parse cdparanoia's output information.
:cvar read:
:vartype read:
:cvar wrote:
:vartype wrote:
:cvar errors:
:vartype errors:
:ivar reads:
:vartype reads:
:ivar start: first frame to rip.
:vartype start: int
:ivar stop: last frame to rip (inclusive).
:vartype stop: int
"""
read = 0 # last [read] frame
wrote = 0 # last [wrote] frame
errors = 0 # count of number of scsi errors
_nframes = None # number of frames read on each [read]
_firstFrames = None # number of frames read on first [read]
reads = 0 # total number of reads
def __init__(self, start, stop):
self.start = start
self.stop = stop
# FIXME: privatize
self.read = start
self._reads = {} # read count for each sector
def parse(self, line):
"""Parse a line.
:param line:
:type line:
"""
m = _PROGRESS_RE.search(line)
if m:
# code = int(m.group('code'))
function = m.group('function')
wordOffset = int(m.group('offset'))
if function == 'read':
self._parse_read(wordOffset)
elif function == 'wrote':
self._parse_wrote(wordOffset)
m = _ERROR_RE.search(line)
if m:
self.errors += 1
def _parse_read(self, wordOffset):
if wordOffset % common.WORDS_PER_FRAME != 0:
logger.debug('THOMAS: not a multiple of %d: %d' % (
common.WORDS_PER_FRAME, wordOffset))
return
frameOffset = wordOffset / common.WORDS_PER_FRAME
# set nframes if not yet set
if self._nframes is None and self.read != 0:
self._nframes = frameOffset - self.read
logger.debug('set nframes to %r', self._nframes)
# set firstFrames if not yet set
if self._firstFrames is None:
self._firstFrames = frameOffset - self.start
logger.debug('set firstFrames to %r', self._firstFrames)
markStart = None
markEnd = None # the next unread frame (half-inclusive)
# verify it either read nframes more or went back for verify
if frameOffset > self.read:
delta = frameOffset - self.read
if self._nframes and delta != self._nframes:
# my drive either reads 7 or 13 frames
pass
# update our read sectors hash
markStart = self.read
markEnd = frameOffset
else:
# went back to verify
# we could use firstFrames as an estimate on how many frames this
# read, but this lowers our track quality needlessly where
# EAC still reports 100% track quality
markStart = frameOffset # - self._firstFrames
markEnd = frameOffset
# FIXME: doing this is way too slow even for a testcase, so disable
if False:
for frame in range(markStart, markEnd):
if frame not in self._reads.keys():
self._reads[frame] = 0
self._reads[frame] += 1
# cdparanoia reads quite a bit beyond the current track before it
# goes back to verify; don't count those
# markStart, markEnd of 0, 21 with stop 0 should give 1 read
if markEnd > self.stop + 1:
markEnd = self.stop + 1
if markStart > self.stop + 1:
markStart = self.stop + 1
self.reads += markEnd - markStart
# update our read pointer
self.read = frameOffset
def _parse_wrote(self, wordOffset):
# cdparanoia outputs most [wrote] calls with one word less than a frame
frameOffset = (wordOffset + 1) / common.WORDS_PER_FRAME
self.wrote = frameOffset
def getTrackQuality(self):
"""Each frame gets read twice.
More than two reads for a frame reduce track quality.
:returns:
:rtype: float or int
"""
frames = self.stop - self.start + 1 # + 1 since stop is inclusive
reads = self.reads
logger.debug('getTrackQuality: frames %d, reads %d' % (frames, reads))
# don't go over a 100%; we know cdparanoia reads each frame at least
# twice
try:
return min(frames * 2.0 / reads, 1.0)
except ZeroDivisionError:
return 0
# FIXME: handle errors
class ReadTrackTask(task.Task):
"""I am a task that reads a track using cdparanoia.
:cvar description:
:vartype description:
:cvar quality:
:cvar speed:
:cvar duration:
:ivar path: where to store the ripped track.
:vartype path: unicode
:ivar table: table of contents of CD.
:vartype table: L{table.Table}
:ivar start: first frame to rip.
:vartype start: int
:ivar stop: last frame to rip (inclusive); >= start.
:vartype stop: int
:ivar offset: read offset, in samples.
:vartype offset: int
:ivar parser:
:vartype parser:
:ivar device: the device to rip from.
:vartype device: str
:ivar start_time:
:vartype start_time:
:ivar overread:
:vartype overread:
:ivar buffer:
:vartype buffer:
:ivar errors:
:vartype errors:
"""
description = "Reading track"
quality = None # set at end of reading
speed = None
duration = None # in seconds
_MAXERROR = 100 # number of errors detected by parser
def __init__(self, path, table, start, stop, overread, offset=0,
device=None, action="Reading", what="track"):
assert type(path) is unicode, "%r is not unicode" % path
self.path = path
self._table = table
self._start = start
self._stop = stop
self._offset = offset
self._parser = ProgressParser(start, stop)
self._device = device
self._start_time = None
self._overread = overread
self._buffer = "" # accumulate characters
self._errors = []
self.description = "%s %s" % (action, what)
def start(self, runner):
task.Task.start(self, runner)
# find on which track the range starts and stops
startTrack = 0
startOffset = 0
stopTrack = 0
stopOffset = self._stop
for i, t in enumerate(self._table.tracks):
if self._table.getTrackStart(i + 1) <= self._start:
startTrack = i + 1
startOffset = self._start - self._table.getTrackStart(i + 1)
if self._table.getTrackEnd(i + 1) <= self._stop:
stopTrack = i + 1
stopOffset = self._stop - self._table.getTrackStart(i + 1)
logger.debug('Ripping from %d to %d (inclusive)',
self._start, self._stop)
logger.debug('Starting at track %d, offset %d',
startTrack, startOffset)
logger.debug('Stopping at track %d, offset %d',
stopTrack, stopOffset)
bufsize = 1024
if self._overread:
argv = ["cd-paranoia", "--stderr-progress",
"--sample-offset=%d" % self._offset, "--force-overread", ]
else:
argv = ["cd-paranoia", "--stderr-progress",
"--sample-offset=%d" % self._offset, ]
if self._device:
argv.extend(["--force-cdrom-device", self._device, ])
argv.extend(["%d[%s]-%d[%s]" % (
startTrack, common.framesToHMSF(startOffset),
stopTrack, common.framesToHMSF(stopOffset)),
self.path])
logger.debug('Running %s' % (" ".join(argv), ))
try:
self._popen = asyncsub.Popen(argv,
bufsize=bufsize,
stdin=subprocess.PIPE,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE,
close_fds=True)
except OSError as e:
import errno
if e.errno == errno.ENOENT:
raise common.MissingDependencyException('cd-paranoia')
raise
self._start_time = time.time()
self.schedule(1.0, self._read, runner)
def _read(self, runner):
ret = self._popen.recv_err()
if not ret:
if self._popen.poll() is not None:
self._done()
return
self.schedule(0.01, self._read, runner)
return
self._buffer += ret
# parse buffer into lines if possible, and parse them
if "\n" in self._buffer:
lines = self._buffer.split('\n')
if lines[-1] != "\n":
# last line didn't end yet
self._buffer = lines[-1]
del lines[-1]
else:
self._buffer = ""
for line in lines:
self._parser.parse(line)
# fail if too many errors
if self._parser.errors > self._MAXERROR:
logger.debug('%d errors, terminating', self._parser.errors)
self._popen.terminate()
num = self._parser.wrote - self._start + 1
den = self._stop - self._start + 1
assert den != 0, "stop %d should be >= start %d" % (
self._stop, self._start)
progress = float(num) / float(den)
if progress < 1.0:
self.setProgress(progress)
# 0 does not give us output before we complete, 1.0 gives us output
# too late
self.schedule(0.01, self._read, runner)
def _poll(self, runner):
if self._popen.poll() is None:
self.schedule(1.0, self._poll, runner)
return
self._done()
def _done(self):
end_time = time.time()
self.setProgress(1.0)
# check if the length matches
size = os.stat(self.path)[stat.ST_SIZE]
# wav header is 44 bytes
offsetLength = self._stop - self._start + 1
expected = offsetLength * common.BYTES_PER_FRAME + 44
if size != expected:
# FIXME: handle errors better
logger.warning('file size %d did not match expected size %d',
size, expected)
if (size - expected) % common.BYTES_PER_FRAME == 0:
logger.warning('%d frames difference' % (
(size - expected) / common.BYTES_PER_FRAME))
else:
logger.warning('non-integral amount of frames difference')
self.setAndRaiseException(FileSizeError(self.path,
"File size %d did not "
"match expected "
"size %d" % (
size, expected)))
if not self.exception and self._popen.returncode != 0:
if self._errors:
print "\n".join(self._errors)
else:
logger.warning('exit code %r', self._popen.returncode)
self.exception = ReturnCodeError(self._popen.returncode)
self.quality = self._parser.getTrackQuality()
self.duration = end_time - self._start_time
self.speed = (offsetLength / 75.0) / self.duration
self.stop()
return
class ReadVerifyTrackTask(task.MultiSeparateTask):
"""I am a task that reads and verifies a track using cdparanoia.
I also encode the track.
The path where the file is stored can be changed if necessary, for
example if the file name is too long.
:cvar checksum: the checksum of the track; set if they match.
:vartype checksum:
:cvar testchecksum: the test checksum of the track.
:vartype testchecksum:
:cvar copychecksum: the copy checksum of the track.
:vartype copychecksum:
:cvar peak: the peak level of the track
:vartype peak:
:cvar quality:
:vartype quality:
:cvar testspeed: the test speed of the track, as a multiple of
track duration.
:vartype testspeed:
:cvar copyspeed: the copy speed of the track, as a multiple of
track duration.
:vartype copyspeed:
:cvar testduration: the test duration of the track, in seconds.
:vartype testduration:
:cvar copyduration: the copy duration of the track, in seconds.
:vartype copyduration:
:ivar path: the path where the file is to be stored.
:vartype path: str
:ivar table: table of contents of CD.
:vartype table: L{table.Table}
:ivar stop: last frame to rip (inclusive).
:vartype stop: int
:ivar offset: read offset, in samples.
:vartype offset: int
:ivar device: the device to rip from.
:vartype device: str
:ivar taglist: a dict of tags.
:vartype taglist: dict
"""
checksum = None
testchecksum = None
copychecksum = None
peak = None
quality = None
testspeed = None
copyspeed = None
testduration = None
copyduration = None
_tmpwavpath = None
_tmppath = None
def __init__(self, path, table, start, stop, overread, offset=0,
device=None, taglist=None, what="track"):
task.MultiSeparateTask.__init__(self)
logger.debug('Creating read and verify task on %r', path)
if taglist:
logger.debug('read and verify with taglist %r', taglist)
# FIXME: choose a dir on the same disk/dir as the final path
fd, tmppath = tempfile.mkstemp(suffix='.whipper.wav')
tmppath = unicode(tmppath)
os.close(fd)
self._tmpwavpath = tmppath
from whipper.common import checksum
self.tasks = []
self.tasks.append(
ReadTrackTask(tmppath, table, start, stop, overread,
offset=offset, device=device, what=what))
self.tasks.append(checksum.CRC32Task(tmppath))
t = ReadTrackTask(tmppath, table, start, stop, overread,
offset=offset, device=device, action="Verifying",
what=what)
self.tasks.append(t)
self.tasks.append(checksum.CRC32Task(tmppath))
# encode to the final path + '.part'
try:
tmpoutpath = path + u'.part'
open(tmpoutpath, 'wb').close()
except IOError as e:
if errno.ENAMETOOLONG != e.errno:
raise
path = common.shrinkPath(path)
tmpoutpath = path + u'.part'
open(tmpoutpath, 'wb').close()
self._tmppath = tmpoutpath
self.path = path
from whipper.common import encode
self.tasks.append(encode.FlacEncodeTask(tmppath, tmpoutpath))
# MerlijnWajer: XXX: We run the CRC32Task on the wav file, because it's
# in general stupid to run the CRC32 on the flac file since it already
# has --verify. We should just get rid of this CRC32 step.
# make sure our encoding is accurate
self.tasks.append(checksum.CRC32Task(tmppath))
self.tasks.append(encode.SoxPeakTask(tmppath))
# TODO: Move tagging outside of cdparanoia
self.tasks.append(encode.TaggingTask(tmpoutpath, taglist))
self.checksum = None
def stop(self):
# FIXME: maybe this kind of try-wrapping to make sure
# we chain up should be handled by a parent class function ?
try:
if not self.exception:
self.quality = max(self.tasks[0].quality,
self.tasks[2].quality)
self.peak = self.tasks[6].peak
logger.debug('peak: %r', self.peak)
self.testspeed = self.tasks[0].speed
self.copyspeed = self.tasks[2].speed
self.testduration = self.tasks[0].duration
self.copyduration = self.tasks[2].duration
self.testchecksum = c1 = self.tasks[1].checksum
self.copychecksum = c2 = self.tasks[3].checksum
if c1 == c2:
logger.info('Checksums match, %08x' % c1)
self.checksum = self.testchecksum
else:
# FIXME: detect this before encoding
logger.info('Checksums do not match, %08x %08x' % (
c1, c2))
self.exception = ChecksumException(
'read and verify failed: test checksum')
if self.tasks[5].checksum != self.checksum:
self.exception = ChecksumException(
'Encoding failed, checksum does not match')
# delete the unencoded file
os.unlink(self._tmpwavpath)
if not self.exception:
try:
logger.debug('Moving to final path %r', self.path)
os.rename(self._tmppath, self.path)
except Exception as e:
logger.debug('Exception while moving to final '
'path %r: %r', self.path, str(e))
self.exception = e
else:
os.unlink(self._tmppath)
else:
logger.debug('stop: exception %r', self.exception)
except Exception as e:
print 'WARNING: unhandled exception %r' % (e, )
task.MultiSeparateTask.stop(self)
_VERSION_RE = re.compile(
"^cdparanoia (?P.+) release (?P.+)")
def getCdParanoiaVersion():
getter = common.VersionGetter('cd-paranoia',
["cd-paranoia", "-V"],
_VERSION_RE,
"%(version)s %(release)s")
return getter.get()
_OK_RE = re.compile(r'Drive tests OK with Paranoia.')
_WARNING_RE = re.compile(r'WARNING! PARANOIA MAY NOT BE')
class AnalyzeTask(ctask.PopenTask):
logCategory = 'AnalyzeTask'
description = 'Analyzing drive caching behaviour'
defeatsCache = None
cwd = None
_output = []
def __init__(self, device=None):
# cdparanoia -A *always* writes cdparanoia.log
self.cwd = tempfile.mkdtemp(suffix='.whipper.cache')
self.command = ['cd-paranoia', '-A']
if device:
self.command += ['-d', device]
def commandMissing(self):
raise common.MissingDependencyException('cd-paranoia')
def readbyteserr(self, bytes):
self._output.append(bytes)
def done(self):
if self.cwd:
shutil.rmtree(self.cwd)
output = "".join(self._output)
m = _OK_RE.search(output)
if m:
self.defeatsCache = True
else:
self.defeatsCache = False
def failed(self):
# cdparanoia exits with return code 1 if it can't determine
# whether it can defeat the audio cache
output = "".join(self._output)
m = _WARNING_RE.search(output)
if m:
self.defeatsCache = False
if self.cwd:
shutil.rmtree(self.cwd)