aboutsummaryrefslogtreecommitdiffstats
path: root/fusearchive.py
blob: ccfb38405b0dc32db7e78470669c5678b3f55847 (plain)
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
#!/usr/bin/env python

#    Copyright (C) 2001  Jeff Epler  <jepler@unpythonic.dhs.org>
#    Copyright (C) 2006  Csaba Henk  <csaba.henk@creo.hu>
#    Copyright (C) 2009  Steve Slaven  <bpk@hoopajoo.net>
#
#    This program can be distributed under the terms of the GNU LGPL.
#    See the file COPYING.
#

import os, sys, fcntl, fuse, sha, cPickle, gzip, errno
import zipfile, stat, logging
from fuse import Fuse

import pdb

if not hasattr(fuse, '__version__'):
    raise RuntimeError, \
        "your fuse-py doesn't know of fuse.__version__, probably it's too old."

fuse.fuse_python_api = (0, 2)

fuse.feature_assert('stateful_files', 'has_init')


logging.basicConfig( level = logging.DEBUG,
                     format = '%(asctime)s %(levelname)s %(filename)s:%(lineno)d %(module)s:%(funcName)s() %(message)s',
                     stream = sys.stderr,
                     filemode = 'w' )

magic_profiling = False
enable_stats = False
enable_psyco = False

# These control some of the file output
magic_blocksize = 1024 * 128
# Use a tiny block size to debug writes, so you can use a smaller test file
#magic_blocksize = 1024
magic_depth = 5
gzip_compress_level = 6
chunkstyle = 'fs'

# Memory for dirty blocks, per file (1M)
dirty_size = 1024 * 1024 * 1;
# This is the number of actualy blocks in that size
dirty_flush = int( dirty_size / magic_blocksize )

def flag2mode(flags):
    md = {os.O_RDONLY: 'r', os.O_WRONLY: 'w', os.O_RDWR: 'w+'}
    m = md[flags & (os.O_RDONLY | os.O_WRONLY | os.O_RDWR)]

    if flags | os.O_APPEND:
        m = m.replace('w', 'a', 1)

    return m

def save_chunk( chunk ):
    if chunkstyle == 'fs':
        return _save_chunk_fs( chunk )
    elif chunkstyle == 'zip':
        return _save_chunk_zip( chunk )
    else:
        raise ValueError( 'Unknown chunk style' )

def load_chunk( key ):
    if chunkstyle == 'fs':
        return _load_chunk_fs( key )
    elif chunkstyle == 'zip':
        return _load_chunk_zip( key )
    else:
        raise ValueError( 'Unknown chunk style' )

# This will write out a data block, it will return a key that can get this
# data back later
def _save_chunk_fs( chunk ):
    if magic_profiling:
        return( [ 0, 0 ] )

    logging.debug( "Begin save_chunk, length: " + str( len( chunk ) ) )
    if debug_level > 4:
        logging.debug( "Chunk: " + str( chunk ) )

    # Save this hash string, similar to the backuppc algo
    digest = sha.new( chunk ).digest()

    # Write out our chunk
    chars = list( digest )
    logging.debug( chars )

    # We make the hexdigest here, yeah we could just call hexdigest()
    # but we need to essentially do this same thing to reassemble the
    # file anyway
    hexdigest = ''.join( [ "%02x" % ord( x ) for x in chars ] )

    # Subparts just needs the first N chars
    subparts = [ "%02x" % ord( x ) for x in chars[ :magic_depth ] ]

    logging.debug( subparts )
    subpath = '/'.join( subparts )
    logging.debug( "Subpath: " + subpath )

    # Make sure this sub path exists
    nextpart = "./storage"
    for part in subparts:
        nextpart += "/" + part
        if not os.path.exists( nextpart ):
            logging.debug( "Creating subdir: " + nextpart )
            os.mkdir( nextpart )

    # Find a chunk slot
    sub = 0
    while True:
        checkpath = "./storage/" + subpath + "/" + hexdigest + "_" + str( sub )
        logging.debug( "Checking: " + checkpath )
        if os.path.exists( checkpath ):
            # Check if this is our data
            verify = FuseArchiveStream.open( checkpath, "rb" )
            verify_contents = verify.read()
            verify.close()

            if verify_contents == chunk:
                logging.debug( "Found existing block" )
                break
            else:
                logging.debug( "Block exists but is not the same" )
                sub += 1
        else:
            # We found a spot, dump our data here
            logging.debug( "No block here, creating new block" )
            savechunk = FuseArchiveStream.open( checkpath, "wb" )
            savechunk.write( chunk )
            savechunk.close()
            break

    logging.debug( "Got chunk slot: " + str( sub ) )
    return( [ digest, sub ] )

def _save_chunk_zip( chunk ):
    if magic_profiling:
        return( [ 0, 0 ] )

    logging.debug( "Begin save_chunk, length: " + str( len( chunk ) ) )
    if debug_level > 4:
        logging.debug( "Chunk: " + str( chunk ) )

    # Save this hash string, similar to the backuppc algo
    digest = sha.new( chunk ).digest()

    # Write out our chunk
    chars = list( digest )
    logging.debug( chars )

    # We make the hexdigest here, yeah we could just call hexdigest()
    # but we need to essentially do this same thing to reassemble the
    # file anyway
    hexdigest = ''.join( [ "%02x" % ord( x ) for x in chars ] )

    # Should be about max of 32k zip files
    zipname = hexdigest[ 0:4 ] + ".zip"
    logging.debug( "Zip name: " + zipname )
    if not os.path.exists( "./storage/" + zipname ):
        logging.debug( "Creating intial empty zip" )
        z = zipfile.ZipFile( "./storage/" + zipname, 'w', zipfile.ZIP_DEFLATED, True )
        # append mode throws an exception if it's not zip, or maybe it's
        # just zero-length files
        z.writestr( 'junk', 'junk' )
        z.close()

    z = zipfile.ZipFile( "./storage/" + zipname, 'a', zipfile.ZIP_DEFLATED, True )

    # Find a chunk slot
    sub = 0
    while True:
        checkpath = hexdigest + "_" + str( sub )
        logging.debug( "Checking: " + checkpath )
        try:
            data = z.read( checkpath )
        except:
            data = ''

        if len(data):
            if data == chunk:
                logging.debug( "Found existing block" )
                break
            else:
                logging.debug( "Block exists but is not the same" )
                sub += 1
        else:
            # We found a spot, dump our data here
            logging.debug( "No block here, creating new block" )
            z.writestr( checkpath, chunk )
            break

    z.close()
    logging.debug( "Got chunk slot: " + str( sub ) )
    return( [ digest, sub ] )

# This will return a data block by key that was saved previously
def _load_chunk_fs( key ):
    if magic_profiling:
        return ''

    ( thash, seq ) = key
    logging.debug( "Begin load_chunk" )

    chars = list( thash )
    logging.debug( chars )

    # Todo: make a digest -> path function to share with deflate
    hexdigest = ''.join( [ "%02x" % ord( x ) for x in chars ] )
    logging.debug( "Hash is: " + str(hexdigest) + " sub " + str(seq) )
    subparts = [ "%02x" % ord( x ) for x in chars[ :magic_depth ] ]
    subpath = '/'.join( subparts )
    logging.debug( "Subpath: " + subpath )

    subpath += "/" + hexdigest + "_" + str( seq )

    logging.debug( "Chunk path: " + subpath )

    if os.path.exists( "./storage/" + subpath ):
        logging.debug( "Exporting chunk" )
        readchunk = FuseArchiveStream.open( "./storage/" + subpath )
        chunk = readchunk.read()
        readchunk.close()
    else:
        raise IOError

    if debug_level > 4:
        logging.debug( "Load-Chunk: " + str( chunk ) )

    return chunk

def _load_chunk_zip( key ):
    if magic_profiling:
        return ''

    ( thash, seq ) = key
    logging.debug( "Begin load_chunk" )

    chars = list( thash )
    logging.debug( chars )

    # Todo: make a digest -> path function to share with deflate
    hexdigest = ''.join( [ "%02x" % ord( x ) for x in chars ] )

    zipname = hexdigest[ 0:4 ] + ".zip"
    logging.debug( "Zip name: " + zipname )
    z = zipfile.ZipFile( "./storage/" + zipname, 'r', zipfile.ZIP_DEFLATED, True )

    subpath = hexdigest + "_" + str( seq )
    logging.debug( "Chunk path: " + subpath )
    data = z.read( subpath )
    if len( data ):
        logging.debug( "Exporting chunk" )
        chunk = data
    else:
        z.close()
        raise IOError

    if debug_level > 4:
        logging.debug( "Load-Chunk: " + str( chunk ) )

    z.close()
    return chunk


class FuseArchiveStream:
    """This just allows switching out writer classes easily"""
    @staticmethod
    def open( path, mode = 'r' ):
        fh = gzip.open( path, mode, gzip_compress_level )
        #fh = open( path, mode )
        return fh

class FuseArchiveSerializer:
    """This lets us experiment with different main file serializers"""
    @staticmethod
    def dump( file, obj ):
        out = FuseArchiveStream.open( file, "wb" )
        FuseArchiveSerializer.dumpfh( obj, out, -1 ) # new file format
        out.close()

    @staticmethod
    def dumpfh( fh, obj ):
        fh.truncate( 0 )
        file = gzip.GzipFile( None, "wb", gzip_compress_level, fh )
        #file = fh
        cPickle.dump( obj, file, -1 ) # new file format
        file.flush()

    @staticmethod
    def load( file ):
        if magic_profiling:
            return { 'size': 0, 'chunks': 0, 'chunk_size': 0 }

        inp = FuseArchiveStream.open( file, "rb" )
        magic = FuseArchiveSerialize.loadfh( inp )
        inp.close()
        return magic

    @staticmethod
    def loadfh( fh ):
        fh.seek( 0 )
        file = gzip.GzipFile( None, "rb", gzip_compress_level, fh )
        #file = fh
        #pdb.set_trace()
        magic = cPickle.load( file )
        return( magic )

class FuseArchiveStat(fuse.Stat):
    def __init__(self, stat):
        self.st_mode = stat.st_mode
        self.st_ino = stat.st_ino
        self.st_dev = stat.st_dev
        self.st_rdev = stat.st_rdev
        self.st_nlink = stat.st_nlink
        self.st_uid = stat.st_uid
        self.st_gid = stat.st_gid
        self.st_size = stat.st_size
        self.st_atime = stat.st_atime
        self.st_mtime = stat.st_mtime
        self.st_ctime = stat.st_mtime
        self.st_blocks = stat.st_blocks
        self.st_blksize = stat.st_blksize

    def overstat( self, size ):
        self.st_size = size
        # Yeah we shouldn't always just add 1
        self.st_blocks = int( self.st_size / 512 ) + 1

class FuseArchive(Fuse):

    def __init__(self, *args, **kw):

        Fuse.__init__(self, *args, **kw)
        self.root = None

    # Fix getattr and fgetattr to?
    def getattr(self, path):
        treefile = "./tree" + path

        if os.path.isfile( treefile ):
            logging.debug( "Delegating getattr to FuserArchiveFile" )

            f = self.FuseArchiveFile( path, os.O_RDONLY, 0 )
            stats = f.fgetattr()
            f.release( 0 )
        else:
            logging.debug( "Using os.lstat to get stats" )
            stats = os.lstat( treefile )

        return stats

    def readlink(self, path):
        return os.readlink("./tree" + path)

    def readdir(self, path, offset):
        for e in os.listdir("./tree" + path):
            yield fuse.Direntry(e)

    def unlink(self, path):
        os.unlink("./tree" + path)

    def rmdir(self, path):
        os.rmdir("./tree" + path)

    def symlink(self, path, path1):
        os.symlink(path, "./tree" + path1)

    def rename(self, path, path1):
        os.rename("./tree" + path, "./tree" + path1)

    def link(self, path, path1):
        os.link("./tree" + path, "./tree" + path1)

    def chmod(self, path, mode):
        os.chmod("./tree" + path, mode)

    def chown(self, path, user, group):
        os.chown("./tree" + path, user, group)

    def truncate(self, path, len):
        # Truncate using the ftruncate on the file
        logging.debug( "Using FuseArchiveFile to truncate " + path + " to " + str(len) )
        f = self.FuseArchiveFile( path, os.O_WRONLY | os.O_APPEND, 0 )
        f.ftruncate(len)
        f.release( 0 )

    def mknod(self, path, mode, dev):
        os.mknod("./tree" + path, mode, dev)

    def mkdir(self, path, mode):
        os.mkdir("./tree" + path, mode)

    def utime(self, path, times):
        os.utime("./tree" + path, times)

#    The following utimens method would do the same as the above utime method.
#    We can't make it better though as the Python stdlib doesn't know of
#    subsecond preciseness in acces/modify times.
#  
#    def utimens(self, path, ts_acc, ts_mod):
#      os.utime("." + path, (ts_acc.tv_sec, ts_mod.tv_sec))

    def access(self, path, mode):
        if not os.access("./tree" + path, mode):
            return -errno.EACCES

#    This is how we could add stub extended attribute handlers...
#    (We can't have ones which aptly delegate requests to the underlying fs
#    because Python lacks a standard xattr interface.)
#
#    def getxattr(self, path, name, size):
#        val = name.swapcase() + '@' + path
#        if size == 0:
#            # We are asked for size of the value.
#            return len(val)
#        return val
#
#    def listxattr(self, path, size):
#        # We use the "user" namespace to please XFS utils
#        aa = ["user." + a for a in ("foo", "bar")]
#        if size == 0:
#            # We are asked for size of the attr list, ie. joint size of attrs
#            # plus null separators.
#            return len("".join(aa)) + len(aa)
#        return aa

    def statfs(self):
        """
        Should return an object with statvfs attributes (f_bsize, f_frsize...).
        Eg., the return value of os.statvfs() is such a thing (since py 2.2).
        If you are not reusing an existing statvfs object, start with
        fuse.StatVFS(), and define the attributes.

        To provide usable information (ie., you want sensible df(1)
        output, you are suggested to specify the following attributes:

            - f_bsize - preferred size of file blocks, in bytes
            - f_frsize - fundamental size of file blcoks, in bytes
                [if you have no idea, use the same as blocksize]
            - f_blocks - total number of blocks in the filesystem
            - f_bfree - number of free blocks
            - f_files - total number of file inodes
            - f_ffree - nunber of free file inodes
        """

        return os.statvfs(".")

    def fsinit(self):
        os.chdir(self.root)

    class FuseArchiveFile(object):

        def __init__(self, path, flags, *mode):
            # Inflate the file
            logging.debug( "Init file: " + path )
            self.orig_path = path

            # init rw and offset
            self.offset = 0
            self.rd = False
            self.wr = False
            self.size = 0
            self.modified = False

            # This is the current in-memory chunk and offset in to data[]
            self.chunk_cache = {};
            self.chunk = ''
            self.chunk_index = -1
            self.chunk_modified = False
            self.chunk_size = magic_blocksize
            self.dirty_chunks = 0

            # The chunk table
            self.chunks = []

            # TODO: Better flag handling here?
            if flags | os.O_RDONLY:
                self.rd = True

            if flags | os.O_RDWR:
                self.rd = True
                self.wr = True

            if flags | os.O_WRONLY:
                self.wr = True

            if flags | os.O_APPEND:
                self.wr = True
                # TODO: handle offset -1
                self.offset = -1

            if os.path.exists( "./tree" + self.orig_path ):
                preexist = True
            else:
                preexist = False

            # Open the file now and keep the fh around so that cp -a on r/o
            # files works (in the create a read-only file for writing case)
            src = "./tree" + path
            logging.debug( "Saving fh for " + src )
            self.file = os.fdopen( os.open( src, flags, *mode ),
                flag2mode( flags ) )

            if preexist:
                # Read in file info table
                logging.debug( "Unpickling: " + str( self.file ) )
                # TODO: return an IO error if inflating fails
                magic = FuseArchiveSerializer.loadfh( self.file )
                logging.debug( "Got data: " + str( magic ) )
                self.size = magic[ 'size' ]
                self.chunks = magic[ 'chunks' ]
                self.chunk_size = magic[ 'chunk_size' ]
            else:
                if self.wr:
                    logging.debug( "File doesn't exist and we're going to write, creating temp empty file" )
                    self.modified = True
                    self.flush()

            self.direct_io = False
            self.keep_cache = False

            #pdb.set_trace()
            logging.debug( str(self) + " init complete" )

        def _load_chunk( self, index ):
            # If the current chunk is the same as the chunk we're loading
            # just return
            if index == self.chunk_index:
                logging.debug( "Load chunk is same as current chunk, all done" )
                return

            # Save this chunk if modified
            self._save_chunk()

            logging.debug( "Loading chunk " + str(index) )
            key = None

            size = len( self.chunks )
            if index >= size:
                logging.debug( "Index doesn't exist" )
            else:
                key = self.chunks[ index ]

            if key:
                if isinstance( key, str ):
                    logging.debug( "Found cached dirty page" )
                    self.chunk = key
                else:
                    logging.debug( "Index: " + str( key ) )
                    self.chunk = load_chunk( key )
            else:
                logging.debug( "No chunk at this index, loading nothing" )
                self.chunk = ''

            logging.debug( "Loaded chunk of length: " + str( len( self.chunk ) ) )

            self.chunk_index = index
            self.chunk_modified = False

        # This simply puts the chunk data inside our current chunks at chunk_index
        def _save_chunk(self):
            if self.chunk_modified:
                # Make sure we have room for this chunk
                size = len( self.chunks )
                if self.chunk_index >= size:
                    self.chunks.extend( [ '' ] * ( self.chunk_index  -size + 1 ) )

                # Increment dirty chunks if we had a key here already
                if isinstance( self.chunks[ self.chunk_index ], list ) or \
                        len( self.chunks[ self.chunk_index ] ) == 0:
                    self.dirty_chunks += 1
                    logging.debug( "Dirty chunks is now: " + str( self.dirty_chunks ) )
                    logging.debug( "Dirty flush at: " + str( dirty_flush ) )

                # Save the dirty chunk temporarily in memory
                self.chunks[ self.chunk_index ] = self.chunk

                # Flush if we have too many dirty chunks
                if self.dirty_chunks > dirty_flush:
                    self._flush_chunks()

        # This flushes any cached chunks
        def _flush_chunks(self):
            for index in range( len( self.chunks ) ):
                if isinstance( self.chunks[ index ], str ):
                    logging.debug( "Flushing chunk at " + str( index ) )
                    key = save_chunk( self.chunks[ index ] )
                    self.chunks[ index ] = key
                    logging.debug( "Key was " + str( key ) )
                    self.dirty_chunks = 0

        def read(self, length, offset):
            logging.debug( "Reading from " + self.orig_path + " offset: " + str( offset )
                + " length: " + str( length ) )

            data_read = 0
            data = ''
            index = int( offset / self.chunk_size )
            rest = offset % self.chunk_size
            is_eof = False

            # Keep reading chunks until we have at least this much data
            while data_read < length and not is_eof:
                logging.debug( "Pulling chunk data: " + str( index ) )
                self._load_chunk( index )
                if len(self.chunk):
                    chunk_remaining = len(self.chunk) - rest
                    to_read = chunk_remaining
                    data_left = length - data_read
                    if data_left < chunk_remaining:
                        to_read = data_left

                    logging.debug( "chunk_remaining: " + str( chunk_remaining ) )
                    logging.debug( "data_left: " + str( data_left ) )
                    logging.debug( "data_read: " + str( data_read ) )
                    logging.debug( "rest: " + str( rest ) )
                    logging.debug( "Copying " + str(to_read) + " bytes" )

                    data += self.chunk[ rest:(rest+to_read) ]
                    data_read += to_read
                    index += 1
                    rest = 0
                else:
                    logging.debug( "No more chunk data, bye" )
                    is_eof = True

            logging.debug( "Returning " + str( len( data ) ) + " bytes of data" )
            logging.debug( "Internal count was: " + str( data_read ) )
            return data

        def write(self, buf, offset):
            if magic_profiling:
                return len( buf )

            logging.debug( "Writing to " + self.orig_path + " offset: " + str( offset ) )

            index = int( offset / self.chunk_size )
            rest = offset % self.chunk_size

            buf_offset = 0
            buf_len = len(buf)

            logging.debug( "Length: " + str( buf_len ) )
            while( buf_offset < buf_len ):
                logging.debug( "Pulling in chunk for writing: " + str(index) )
                self._load_chunk( index )
                buf_remain = buf_len - buf_offset
                chunk_remain = self.chunk_size - rest

                logging.debug( "buf_remain: " + str(buf_remain) )
                logging.debug( "chunk_remain: " + str(chunk_remain) )

                if chunk_remain < buf_remain:
                    logging.debug( "Writing " + str( chunk_remain ) + " bytes, buffer boundry" )
                    this_len = chunk_remain
                else:
                    logging.debug( "Writing final " + str( buf_remain ) + " bytes" )
                    this_len = buf_remain

                logging.debug( "Bytes to copy: " + str(this_len) )
                logging.debug( " buf offset: " + str(buf_offset) )
                logging.debug( " chunk offset: " + str(rest) )


                if debug_level > 4:
                    logging.debug( "Pre-Buf: " + str(buf) )
                    logging.debug( "Pre-Chunk: " + str(self.chunk) )

                # Since python doesn't do in-place reassignment like you
                # can with splice() we will reconstruct the data by joining
                # stuff by offsets (first chars to skip, then our joining
                # buf chunk, the everything that would have been after it)
                #pdb.set_trace()
                self.chunk = self.chunk[ :rest ] + \
                    buf[ buf_offset:(buf_offset+this_len) ] + \
                    self.chunk[ (rest + this_len): ]

                if debug_level > 4:
                    logging.debug( "Post-Buf: " + str(buf) )
                    logging.debug( "Post-Chunk: " + str(self.chunk) )

                buf_offset += this_len

                # Advance to next block
                rest = 0
                index += 1
                self.chunk_modified = True

            self.modified = True
            logging.debug( "Chunk size is now: " + str(len(self.chunk)) )
            return len(buf)

        # BUG: If you cp -a a file then quickly ls -l sometimes it doesn't show
        # up right?  like wrong size and stuff?
        # Maybe because release doesn't return a fuse message and is async?
        def release(self, flags):
            # Deflate the file
            logging.debug( "Release: " + self.orig_path )
            self.flush()
            self.file.close()

        def _fflush(self):
            if self.wr and self.modified:
                logging.debug( "_fflush!" )
                # Save our main data
                self._save_chunk()

                # And flush any cached chunks
                self._flush_chunks()

                # Figure out our size based on the number of chunks + the
                # len of the final chunk
                numchunks = len( self.chunks )
                if numchunks > 0:
                    # Load the last chunk
                    logging.debug( "We have " + str(numchunks) + " chunks, calculating size" )
                    self._load_chunk( numchunks - 1 )
                    self.size = ( numchunks - 1 ) * self.chunk_size + \
                        len( self.chunk )
                else:
                    logging.debug( "No chunks, setting size to zero" )
                    self.size = 0

                logging.debug( "Size calculated is: " + str( self.size ) )

                FuseArchiveSerializer.dumpfh( self.file, {
                    'size': self.size,
                    'chunks': self.chunks,
                    'chunk_size': self.chunk_size
                } )



            return 1


        # Currently we treat fsync as flush since we don't keep any data
        # hanging around anyway in fh stuff
        def fsync(self, isfsyncfile):
            logging.debug( "fsync " + self.orig_path )
            self._fflush()
            #if isfsyncfile and hasattr(os, 'fdatasync'):
            #    os.fdatasync(self.fd)
            #else:
            #    os.fsync(self.fd)

        def flush(self):
            logging.debug( "flush " + self.orig_path )
            self._fflush()

        def fgetattr(self):
            logging.debug( "Overridding fgetattr" )
            stats = FuseArchiveStat( os.lstat( "./tree" + self.orig_path ) )

            if self.modified:
                # We would need to fsync here to recalc size, but don't do
                # it unless modified? otherwise simple getattr will be
                # rewriting a ton of files
                print "WARNING: self.modified causes fgetattr to be incorrect!"

            stats.overstat( self.size )
            return stats

        def ftruncate(self, length):
            if not self.wr:
                return errno.IOError

            curr_chunks = len( self.chunks )
            need_chunks = ( length / self.chunk_size )
            extra_bytes = length % self.chunk_size
            logging.debug( "Ftruncate - " + str(length) )
            logging.debug( " - curr_chunks: " + str( curr_chunks ) )
            logging.debug( " - need_chunks: " + str( need_chunks ) )
            logging.debug( " - extra_bytes: " + str( extra_bytes ) )

            if extra_bytes:
                logging.debug( "Need an extra chunk" )
                need_chunks += 1

            self._load_chunk( 0 )

            pdb.set_trace()
            if need_chunks == 0:
                logging.debug( "Creating 0 chunk file" )
                self.chunks = []
            elif need_chunks > curr_chunks:
                logging.debug( "Need to pad out file" )

                # Just write out null bytes to the length requested
                self.flush()
                tmp_size = self.size
                while tmp_size < length:
                    to_write = 4096
                    if tmp_size + to_write > length:
                        to_write = length - tmp_size

                    logging.debug( "Writing " + str( to_write ) + " bytes to extend file to " + str( length ) )
                    self.write( "\0" * to_write, tmp_size + 1 )
                    tmp_size += to_write

            else:
                logging.debug( "Truncating chunks" )
                while True:
                    logging.debug( "Need chunks: " + str( need_chunks ) + " curr: " + str( curr_chunks ) )
                    if need_chunks == curr_chunks:
                        break

                    logging.debug( "Deleting chunk " + str( self.chunk_index ) )
                    self.chunks.pop()
                    curr_chunks = len( self.chunks )

            # Now make sure this chunk is the right size, first load the
            # last chunk
            if len( self.chunks ):
                self._load_chunk( len( self.chunks ) - 1 )
                logging.debug( "Loaded final chunk, len: " + str( len( self.chunk ) ) )

            # Now truncate or extend this item
            if len( self.chunk ) < extra_bytes:
                remaining = extra_bytes - len(self.chunk)
                logging.debug( "Extending final chunk " + str( remaining ) + " bytes" )
                self.chunk += "\0" * remaining
            else:
                logging.debug( "Truncating final chunk to " + str( extra_bytes ) )
                self.chunk = self.chunk[ :extra_bytes ]

            self.modified = True
            self._load_chunk( 0 )
            self._fflush()

        def lock(self, cmd, owner, **kw):
            logging.debug( "WARNING: locking unsupported" )
            return 1

            # The code here is much rather just a demonstration of the locking
            # API than something which actually was seen to be useful.

            # Advisory file locking is pretty messy in Unix, and the Python
            # interface to this doesn't make it better.
            # We can't do fcntl(2)/F_GETLK from Python in a platfrom independent
            # way. The following implementation *might* work under Linux. 
            #
            # if cmd == fcntl.F_GETLK:
            #     import struct
            # 
            #     lockdata = struct.pack('hhQQi', kw['l_type'], os.SEEK_SET,
            #                            kw['l_start'], kw['l_len'], kw['l_pid'])
            #     ld2 = fcntl.fcntl(self.fd, fcntl.F_GETLK, lockdata)
            #     flockfields = ('l_type', 'l_whence', 'l_start', 'l_len', 'l_pid')
            #     uld2 = struct.unpack('hhQQi', ld2)
            #     res = {}
            #     for i in xrange(len(uld2)):
            #          res[flockfields[i]] = uld2[i]
            #  
            #     return fuse.Flock(**res)

            # Convert fcntl-ish lock parameters to Python's weird
            # lockf(3)/flock(2) medley locking API...
            op = { fcntl.F_UNLCK : fcntl.LOCK_UN,
                   fcntl.F_RDLCK : fcntl.LOCK_SH,
                   fcntl.F_WRLCK : fcntl.LOCK_EX }[kw['l_type']]
            if cmd == fcntl.F_GETLK:
                return -errno.EOPNOTSUPP
            elif cmd == fcntl.F_SETLK:
                if op != fcntl.LOCK_UN:
                    op |= fcntl.LOCK_NB
            elif cmd == fcntl.F_SETLKW:
                pass
            else:
                return -errno.EINVAL

            fcntl.lockf(self.fd, op, kw['l_start'], kw['l_len'])


    def main(self, *a, **kw):

        self.file_class = self.FuseArchiveFile

        # This is where fragments go
        if not os.path.exists( 'storage' ):
            os.mkdir( 'storage' )

        # This is where the real files exist
        if not os.path.exists( 'tree' ):
            os.mkdir( 'tree' )

        return Fuse.main(self, *a, **kw)


def main():

    usage = """
Userspace nullfs-alike: mirror the filesystem tree from some point on.

""" + Fuse.fusage

    server = FuseArchive(version="%prog " + fuse.__version__,
                 usage=usage,
                 dash_s_do='setsingle')

    server.multithreaded = False

    server.parse(values=server, errex=1)

    if len(server.parser.largs) != 2:
        print "Usage: " + sys.argv[0] + " storageDirectory mountDirectory"
        sys.exit(1)

    server.root = server.parser.largs[0]

    try:
        if server.fuse_args.mount_expected():
            os.chdir(server.root)
    except OSError:
        print >> sys.stderr, "can't enter root of underlying filesystem"
        sys.exit(1)

    server.main()


if __name__ == '__main__':
    if enable_psyco:
        # Import Psyco if available
        try:
            import psyco
            psyco.full()
        except ImportError:
            pass

    if enable_stats:
        import hotshot
        prof = hotshot.Profile( "fusearchive_stats" )
        prof.runcall(main)
        prof.close()
    else:
        main()