720
767
reader._parse_dirblocks()
721
768
state._dirblock_state = DirState.IN_MEMORY_UNMODIFIED
771
cdef int minikind_from_mode(int mode):
772
# in order of frequency:
782
_encode = binascii.b2a_base64
785
from struct import pack
786
cdef _pack_stat(stat_value):
787
"""return a string representing the stat value's key fields.
789
:param stat_value: A stat oject with st_size, st_mtime, st_ctime, st_dev,
790
st_ino and st_mode fields.
792
cdef char result[6*4] # 6 long ints
794
aliased = <int *>result
795
aliased[0] = htonl(stat_value.st_size)
796
aliased[1] = htonl(int(stat_value.st_mtime))
797
aliased[2] = htonl(int(stat_value.st_ctime))
798
aliased[3] = htonl(stat_value.st_dev)
799
aliased[4] = htonl(stat_value.st_ino & 0xFFFFFFFF)
800
aliased[5] = htonl(stat_value.st_mode)
801
packed = PyString_FromStringAndSize(result, 6*4)
802
return _encode(packed)[:-1]
805
def update_entry(self, entry, abspath, stat_value):
806
"""Update the entry based on what is actually on disk.
808
This function only calculates the sha if it needs to - if the entry is
809
uncachable, or clearly different to the first parent's entry, no sha
810
is calculated, and None is returned.
812
:param entry: This is the dirblock entry for the file in question.
813
:param abspath: The path on disk for this file.
814
:param stat_value: (optional) if we already have done a stat on the
816
:return: None, or The sha1 hexdigest of the file (40 bytes) or link
819
return _update_entry(self, entry, abspath, stat_value)
822
cdef _update_entry(self, entry, abspath, stat_value):
823
"""Update the entry based on what is actually on disk.
825
This function only calculates the sha if it needs to - if the entry is
826
uncachable, or clearly different to the first parent's entry, no sha
827
is calculated, and None is returned.
829
:param self: The dirstate object this is operating on.
830
:param entry: This is the dirblock entry for the file in question.
831
:param abspath: The path on disk for this file.
832
:param stat_value: The stat value done on the path.
833
:return: None, or The sha1 hexdigest of the file (40 bytes) or link
836
# TODO - require pyrex 0.9.8, then use a pyd file to define access to the
837
# _st mode of the compiled stat objects.
838
cdef int minikind, saved_minikind
840
minikind = minikind_from_mode(stat_value.st_mode)
843
packed_stat = _pack_stat(stat_value)
844
details = PyList_GetItem_void_void(PyTuple_GetItem_void_void(<void *>entry, 1), 0)
845
saved_minikind = PyString_AsString_obj(<PyObject *>PyTuple_GetItem_void_void(details, 0))[0]
846
if minikind == c'd' and saved_minikind == c't':
848
saved_link_or_sha1 = PyTuple_GetItem_void_object(details, 1)
849
saved_file_size = PyTuple_GetItem_void_object(details, 2)
850
saved_executable = PyTuple_GetItem_void_object(details, 3)
851
saved_packed_stat = PyTuple_GetItem_void_object(details, 4)
852
# Deal with pyrex decrefing the objects
853
Py_INCREF(saved_link_or_sha1)
854
Py_INCREF(saved_file_size)
855
Py_INCREF(saved_executable)
856
Py_INCREF(saved_packed_stat)
857
#(saved_minikind, saved_link_or_sha1, saved_file_size,
858
# saved_executable, saved_packed_stat) = entry[1][0]
860
if (minikind == saved_minikind
861
and packed_stat == saved_packed_stat):
862
# The stat hasn't changed since we saved, so we can re-use the
867
# size should also be in packed_stat
868
if saved_file_size == stat_value.st_size:
869
return saved_link_or_sha1
871
# If we have gotten this far, that means that we need to actually
872
# process this entry.
875
executable = self._is_executable(stat_value.st_mode,
877
if self._cutoff_time is None:
878
self._sha_cutoff_time()
879
if (stat_value.st_mtime < self._cutoff_time
880
and stat_value.st_ctime < self._cutoff_time
881
and len(entry[1]) > 1
882
and entry[1][1][0] != 'a'):
883
# Could check for size changes for further optimised
884
# avoidance of sha1's. However the most prominent case of
885
# over-shaing is during initial add, which this catches.
886
link_or_sha1 = self._sha1_file(abspath)
887
entry[1][0] = ('f', link_or_sha1, stat_value.st_size,
888
executable, packed_stat)
890
entry[1][0] = ('f', '', stat_value.st_size,
891
executable, DirState.NULLSTAT)
892
elif minikind == c'd':
894
entry[1][0] = ('d', '', 0, False, packed_stat)
895
if saved_minikind != c'd':
896
# This changed from something into a directory. Make sure we
897
# have a directory block for it. This doesn't happen very
898
# often, so this doesn't have to be super fast.
899
block_index, entry_index, dir_present, file_present = \
900
self._get_block_entry_index(entry[0][0], entry[0][1], 0)
901
self._ensure_block(block_index, entry_index,
902
pathjoin(entry[0][0], entry[0][1]))
903
elif minikind == c'l':
904
link_or_sha1 = self._read_link(abspath, saved_link_or_sha1)
905
if self._cutoff_time is None:
906
self._sha_cutoff_time()
907
if (stat_value.st_mtime < self._cutoff_time
908
and stat_value.st_ctime < self._cutoff_time):
909
entry[1][0] = ('l', link_or_sha1, stat_value.st_size,
912
entry[1][0] = ('l', '', stat_value.st_size,
913
False, DirState.NULLSTAT)
914
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
918
cdef char _minikind_from_string(object string):
919
"""Convert a python string to a char."""
920
return PyString_AsString(string)[0]
923
cdef object _kind_absent
924
cdef object _kind_file
925
cdef object _kind_directory
926
cdef object _kind_symlink
927
cdef object _kind_relocated
928
cdef object _kind_tree_reference
929
_kind_absent = "absent"
931
_kind_directory = "directory"
932
_kind_symlink = "symlink"
933
_kind_relocated = "relocated"
934
_kind_tree_reference = "tree-reference"
937
cdef object _minikind_to_kind(char minikind):
938
"""Create a string kind for minikind."""
939
cdef char _minikind[1]
942
elif minikind == c'd':
943
return _kind_directory
944
elif minikind == c'a':
946
elif minikind == c'r':
947
return _kind_relocated
948
elif minikind == c'l':
950
elif minikind == c't':
951
return _kind_tree_reference
952
_minikind[0] = minikind
953
raise KeyError(PyString_FromStringAndSize(_minikind, 1))
956
cdef int _versioned_minikind(char minikind):
957
"""Return non-zero if minikind is in fltd"""
958
return (minikind == c'f' or
964
cdef class ProcessEntryC:
966
cdef int doing_consistency_expansion
967
cdef object old_dirname_to_file_id # dict
968
cdef object new_dirname_to_file_id # dict
969
cdef object last_source_parent
970
cdef object last_target_parent
971
cdef int include_unchanged
973
cdef object use_filesystem_for_exec
974
cdef object utf8_decode
975
cdef readonly object searched_specific_files
976
cdef readonly object searched_exact_paths
977
cdef object search_specific_files
978
# The parents up to the root of the paths we are searching.
979
# After all normal paths are returned, these specific items are returned.
980
cdef object search_specific_file_parents
982
# Current iteration variables:
983
cdef object current_root
984
cdef object current_root_unicode
985
cdef object root_entries
986
cdef int root_entries_pos, root_entries_len
987
cdef object root_abspath
988
cdef int source_index, target_index
989
cdef int want_unversioned
991
cdef object dir_iterator
993
cdef object current_block
994
cdef int current_block_pos
995
cdef object current_block_list
996
cdef object current_dir_info
997
cdef object current_dir_list
998
cdef object _pending_consistent_entries # list
1000
cdef object root_dir_info
1001
cdef object bisect_left
1002
cdef object pathjoin
1004
# A set of the ids we've output when doing partial output.
1005
cdef object seen_ids
1006
cdef object sha_file
1008
def __init__(self, include_unchanged, use_filesystem_for_exec,
1009
search_specific_files, state, source_index, target_index,
1010
want_unversioned, tree):
1011
self.doing_consistency_expansion = 0
1012
self.old_dirname_to_file_id = {}
1013
self.new_dirname_to_file_id = {}
1014
# Are we doing a partial iter_changes?
1015
self.partial = set(['']).__ne__(search_specific_files)
1016
# Using a list so that we can access the values and change them in
1017
# nested scope. Each one is [path, file_id, entry]
1018
self.last_source_parent = [None, None]
1019
self.last_target_parent = [None, None]
1020
if include_unchanged is None:
1021
self.include_unchanged = False
1023
self.include_unchanged = int(include_unchanged)
1024
self.use_filesystem_for_exec = use_filesystem_for_exec
1025
self.utf8_decode = cache_utf8._utf8_decode
1026
# for all search_indexs in each path at or under each element of
1027
# search_specific_files, if the detail is relocated: add the id, and
1028
# add the relocated path as one to search if its not searched already.
1029
# If the detail is not relocated, add the id.
1030
self.searched_specific_files = set()
1031
# When we search exact paths without expanding downwards, we record
1033
self.searched_exact_paths = set()
1034
self.search_specific_files = search_specific_files
1035
# The parents up to the root of the paths we are searching.
1036
# After all normal paths are returned, these specific items are returned.
1037
self.search_specific_file_parents = set()
1038
# The ids we've sent out in the delta.
1039
self.seen_ids = set()
1041
self.current_root = None
1042
self.current_root_unicode = None
1043
self.root_entries = None
1044
self.root_entries_pos = 0
1045
self.root_entries_len = 0
1046
self.root_abspath = None
1047
if source_index is None:
1048
self.source_index = -1
1050
self.source_index = source_index
1051
self.target_index = target_index
1052
self.want_unversioned = want_unversioned
1054
self.dir_iterator = None
1055
self.block_index = -1
1056
self.current_block = None
1057
self.current_block_list = None
1058
self.current_block_pos = -1
1059
self.current_dir_info = None
1060
self.current_dir_list = None
1061
self._pending_consistent_entries = []
1063
self.root_dir_info = None
1064
self.bisect_left = bisect.bisect_left
1065
self.pathjoin = osutils.pathjoin
1066
self.fstat = os.fstat
1067
self.sha_file = osutils.sha_file
1068
if target_index != 0:
1069
# A lot of code in here depends on target_index == 0
1070
raise errors.BzrError('unsupported target index')
1072
cdef _process_entry(self, entry, path_info):
1073
"""Compare an entry and real disk to generate delta information.
1075
:param path_info: top_relpath, basename, kind, lstat, abspath for
1076
the path of entry. If None, then the path is considered absent in
1077
the target (Perhaps we should pass in a concrete entry for this ?)
1078
Basename is returned as a utf8 string because we expect this
1079
tuple will be ignored, and don't want to take the time to
1081
:return: (iter_changes_result, changed). If the entry has not been
1082
handled then changed is None. Otherwise it is False if no content
1083
or metadata changes have occured, and True if any content or
1084
metadata change has occurred. If self.include_unchanged is True then
1085
if changed is not None, iter_changes_result will always be a result
1086
tuple. Otherwise, iter_changes_result is None unless changed is
1089
cdef char target_minikind
1090
cdef char source_minikind
1092
cdef int content_change
1093
cdef object details_list
1095
details_list = entry[1]
1096
if -1 == self.source_index:
1097
source_details = DirState.NULL_PARENT_DETAILS
1099
source_details = details_list[self.source_index]
1100
target_details = details_list[self.target_index]
1101
target_minikind = _minikind_from_string(target_details[0])
1102
if path_info is not None and _versioned_minikind(target_minikind):
1103
if self.target_index != 0:
1104
raise AssertionError("Unsupported target index %d" %
1106
link_or_sha1 = _update_entry(self.state, entry, path_info[4], path_info[3])
1107
# The entry may have been modified by update_entry
1108
target_details = details_list[self.target_index]
1109
target_minikind = _minikind_from_string(target_details[0])
1112
# the rest of this function is 0.3 seconds on 50K paths, or
1113
# 0.000006 seconds per call.
1114
source_minikind = _minikind_from_string(source_details[0])
1115
if ((_versioned_minikind(source_minikind) or source_minikind == c'r')
1116
and _versioned_minikind(target_minikind)):
1117
# claimed content in both: diff
1118
# r | fdlt | | add source to search, add id path move and perform
1119
# | | | diff check on source-target
1120
# r | fdlt | a | dangling file that was present in the basis.
1122
if source_minikind != c'r':
1123
old_dirname = entry[0][0]
1124
old_basename = entry[0][1]
1125
old_path = path = None
1127
# add the source to the search path to find any children it
1128
# has. TODO ? : only add if it is a container ?
1129
if (not self.doing_consistency_expansion and
1130
not osutils.is_inside_any(self.searched_specific_files,
1131
source_details[1])):
1132
self.search_specific_files.add(source_details[1])
1133
# expanding from a user requested path, parent expansion
1134
# for delta consistency happens later.
1135
# generate the old path; this is needed for stating later
1137
old_path = source_details[1]
1138
old_dirname, old_basename = os.path.split(old_path)
1139
path = self.pathjoin(entry[0][0], entry[0][1])
1140
old_entry = self.state._get_entry(self.source_index,
1142
# update the source details variable to be the real
1144
if old_entry == (None, None):
1145
raise errors.CorruptDirstate(self.state._filename,
1146
"entry '%s/%s' is considered renamed from %r"
1147
" but source does not exist\n"
1148
"entry: %s" % (entry[0][0], entry[0][1], old_path, entry))
1149
source_details = old_entry[1][self.source_index]
1150
source_minikind = _minikind_from_string(source_details[0])
1151
if path_info is None:
1152
# the file is missing on disk, show as removed.
1157
# source and target are both versioned and disk file is present.
1158
target_kind = path_info[2]
1159
if target_kind == 'directory':
1161
old_path = path = self.pathjoin(old_dirname, old_basename)
1162
file_id = entry[0][2]
1163
self.new_dirname_to_file_id[path] = file_id
1164
if source_minikind != c'd':
1167
# directories have no fingerprint
1170
elif target_kind == 'file':
1171
if source_minikind != c'f':
1174
# Check the sha. We can't just rely on the size as
1175
# content filtering may mean differ sizes actually
1176
# map to the same content
1177
if link_or_sha1 is None:
1179
statvalue, link_or_sha1 = \
1180
self.state._sha1_provider.stat_and_sha1(
1182
self.state._observed_sha1(entry, link_or_sha1,
1184
content_change = (link_or_sha1 != source_details[1])
1185
# Target details is updated at update_entry time
1186
if self.use_filesystem_for_exec:
1187
# We don't need S_ISREG here, because we are sure
1188
# we are dealing with a file.
1189
target_exec = bool(S_IXUSR & path_info[3].st_mode)
1191
target_exec = target_details[3]
1192
elif target_kind == 'symlink':
1193
if source_minikind != c'l':
1196
content_change = (link_or_sha1 != source_details[1])
1198
elif target_kind == 'tree-reference':
1199
if source_minikind != c't':
1205
raise Exception, "unknown kind %s" % path_info[2]
1206
if source_minikind == c'd':
1208
old_path = path = self.pathjoin(old_dirname, old_basename)
1210
file_id = entry[0][2]
1211
self.old_dirname_to_file_id[old_path] = file_id
1212
# parent id is the entry for the path in the target tree
1213
if old_basename and old_dirname == self.last_source_parent[0]:
1214
# use a cached hit for non-root source entries.
1215
source_parent_id = self.last_source_parent[1]
1218
source_parent_id = self.old_dirname_to_file_id[old_dirname]
1220
source_parent_entry = self.state._get_entry(self.source_index,
1221
path_utf8=old_dirname)
1222
source_parent_id = source_parent_entry[0][2]
1223
if source_parent_id == entry[0][2]:
1224
# This is the root, so the parent is None
1225
source_parent_id = None
1227
self.last_source_parent[0] = old_dirname
1228
self.last_source_parent[1] = source_parent_id
1229
new_dirname = entry[0][0]
1230
if entry[0][1] and new_dirname == self.last_target_parent[0]:
1231
# use a cached hit for non-root target entries.
1232
target_parent_id = self.last_target_parent[1]
1235
target_parent_id = self.new_dirname_to_file_id[new_dirname]
1237
# TODO: We don't always need to do the lookup, because the
1238
# parent entry will be the same as the source entry.
1239
target_parent_entry = self.state._get_entry(self.target_index,
1240
path_utf8=new_dirname)
1241
if target_parent_entry == (None, None):
1242
raise AssertionError(
1243
"Could not find target parent in wt: %s\nparent of: %s"
1244
% (new_dirname, entry))
1245
target_parent_id = target_parent_entry[0][2]
1246
if target_parent_id == entry[0][2]:
1247
# This is the root, so the parent is None
1248
target_parent_id = None
1250
self.last_target_parent[0] = new_dirname
1251
self.last_target_parent[1] = target_parent_id
1253
source_exec = source_details[3]
1254
changed = (content_change
1255
or source_parent_id != target_parent_id
1256
or old_basename != entry[0][1]
1257
or source_exec != target_exec
1259
if not changed and not self.include_unchanged:
1262
if old_path is None:
1263
path = self.pathjoin(old_dirname, old_basename)
1265
old_path_u = self.utf8_decode(old_path)[0]
1268
old_path_u = self.utf8_decode(old_path)[0]
1269
if old_path == path:
1272
path_u = self.utf8_decode(path)[0]
1273
source_kind = _minikind_to_kind(source_minikind)
1274
return (entry[0][2],
1275
(old_path_u, path_u),
1278
(source_parent_id, target_parent_id),
1279
(self.utf8_decode(old_basename)[0], self.utf8_decode(entry[0][1])[0]),
1280
(source_kind, target_kind),
1281
(source_exec, target_exec)), changed
1282
elif source_minikind == c'a' and _versioned_minikind(target_minikind):
1283
# looks like a new file
1284
path = self.pathjoin(entry[0][0], entry[0][1])
1285
# parent id is the entry for the path in the target tree
1286
# TODO: these are the same for an entire directory: cache em.
1287
parent_entry = self.state._get_entry(self.target_index,
1288
path_utf8=entry[0][0])
1289
if parent_entry is None:
1290
raise errors.DirstateCorrupt(self.state,
1291
"We could not find the parent entry in index %d"
1292
" for the entry: %s"
1293
% (self.target_index, entry[0]))
1294
parent_id = parent_entry[0][2]
1295
if parent_id == entry[0][2]:
1297
if path_info is not None:
1299
if self.use_filesystem_for_exec:
1300
# We need S_ISREG here, because we aren't sure if this
1303
S_ISREG(path_info[3].st_mode)
1304
and S_IXUSR & path_info[3].st_mode)
1306
target_exec = target_details[3]
1307
return (entry[0][2],
1308
(None, self.utf8_decode(path)[0]),
1312
(None, self.utf8_decode(entry[0][1])[0]),
1313
(None, path_info[2]),
1314
(None, target_exec)), True
1316
# Its a missing file, report it as such.
1317
return (entry[0][2],
1318
(None, self.utf8_decode(path)[0]),
1322
(None, self.utf8_decode(entry[0][1])[0]),
1324
(None, False)), True
1325
elif _versioned_minikind(source_minikind) and target_minikind == c'a':
1326
# unversioned, possibly, or possibly not deleted: we dont care.
1327
# if its still on disk, *and* theres no other entry at this
1328
# path [we dont know this in this routine at the moment -
1329
# perhaps we should change this - then it would be an unknown.
1330
old_path = self.pathjoin(entry[0][0], entry[0][1])
1331
# parent id is the entry for the path in the target tree
1332
parent_id = self.state._get_entry(self.source_index, path_utf8=entry[0][0])[0][2]
1333
if parent_id == entry[0][2]:
1335
return (entry[0][2],
1336
(self.utf8_decode(old_path)[0], None),
1340
(self.utf8_decode(entry[0][1])[0], None),
1341
(_minikind_to_kind(source_minikind), None),
1342
(source_details[3], None)), True
1343
elif _versioned_minikind(source_minikind) and target_minikind == c'r':
1344
# a rename; could be a true rename, or a rename inherited from
1345
# a renamed parent. TODO: handle this efficiently. Its not
1346
# common case to rename dirs though, so a correct but slow
1347
# implementation will do.
1348
if (not self.doing_consistency_expansion and
1349
not osutils.is_inside_any(self.searched_specific_files,
1350
target_details[1])):
1351
self.search_specific_files.add(target_details[1])
1352
# We don't expand the specific files parents list here as
1353
# the path is absent in target and won't create a delta with
1355
elif ((source_minikind == c'r' or source_minikind == c'a') and
1356
(target_minikind == c'r' or target_minikind == c'a')):
1357
# neither of the selected trees contain this path,
1358
# so skip over it. This is not currently directly tested, but
1359
# is indirectly via test_too_much.TestCommands.test_conflicts.
1362
raise AssertionError("don't know how to compare "
1363
"source_minikind=%r, target_minikind=%r"
1364
% (source_minikind, target_minikind))
1365
## import pdb;pdb.set_trace()
1371
def iter_changes(self):
1374
cdef void _gather_result_for_consistency(self, result):
1375
"""Check a result we will yield to make sure we are consistent later.
1377
This gathers result's parents into a set to output later.
1379
:param result: A result tuple.
1381
if not self.partial or not result[0]:
1383
self.seen_ids.add(result[0])
1384
new_path = result[1][1]
1386
# Not the root and not a delete: queue up the parents of the path.
1387
self.search_specific_file_parents.update(
1388
osutils.parent_directories(new_path.encode('utf8')))
1389
# Add the root directory which parent_directories does not
1391
self.search_specific_file_parents.add('')
1393
cdef void _update_current_block(self):
1394
if (self.block_index < len(self.state._dirblocks) and
1395
osutils.is_inside(self.current_root, self.state._dirblocks[self.block_index][0])):
1396
self.current_block = self.state._dirblocks[self.block_index]
1397
self.current_block_list = self.current_block[1]
1398
self.current_block_pos = 0
1400
self.current_block = None
1401
self.current_block_list = None
1404
# Simple thunk to allow tail recursion without pyrex confusion
1405
return self._iter_next()
1407
cdef _iter_next(self):
1408
"""Iterate over the changes."""
1409
# This function single steps through an iterator. As such while loops
1410
# are often exited by 'return' - the code is structured so that the
1411
# next call into the function will return to the same while loop. Note
1412
# that all flow control needed to re-reach that step is reexecuted,
1413
# which can be a performance problem. It has not yet been tuned to
1414
# minimise this; a state machine is probably the simplest restructuring
1415
# to both minimise this overhead and make the code considerably more
1419
# compare source_index and target_index at or under each element of search_specific_files.
1420
# follow the following comparison table. Note that we only want to do diff operations when
1421
# the target is fdl because thats when the walkdirs logic will have exposed the pathinfo
1425
# Source | Target | disk | action
1426
# r | fdlt | | add source to search, add id path move and perform
1427
# | | | diff check on source-target
1428
# r | fdlt | a | dangling file that was present in the basis.
1430
# r | a | | add source to search
1432
# r | r | | this path is present in a non-examined tree, skip.
1433
# r | r | a | this path is present in a non-examined tree, skip.
1434
# a | fdlt | | add new id
1435
# a | fdlt | a | dangling locally added file, skip
1436
# a | a | | not present in either tree, skip
1437
# a | a | a | not present in any tree, skip
1438
# a | r | | not present in either tree at this path, skip as it
1439
# | | | may not be selected by the users list of paths.
1440
# a | r | a | not present in either tree at this path, skip as it
1441
# | | | may not be selected by the users list of paths.
1442
# fdlt | fdlt | | content in both: diff them
1443
# fdlt | fdlt | a | deleted locally, but not unversioned - show as deleted ?
1444
# fdlt | a | | unversioned: output deleted id for now
1445
# fdlt | a | a | unversioned and deleted: output deleted id
1446
# fdlt | r | | relocated in this tree, so add target to search.
1447
# | | | Dont diff, we will see an r,fd; pair when we reach
1448
# | | | this id at the other path.
1449
# fdlt | r | a | relocated in this tree, so add target to search.
1450
# | | | Dont diff, we will see an r,fd; pair when we reach
1451
# | | | this id at the other path.
1453
# TODO: jam 20070516 - Avoid the _get_entry lookup overhead by
1454
# keeping a cache of directories that we have seen.
1455
cdef object current_dirname, current_blockname
1456
cdef char * current_dirname_c, * current_blockname_c
1457
cdef int advance_entry, advance_path
1458
cdef int path_handled
1459
searched_specific_files = self.searched_specific_files
1460
# Are we walking a root?
1461
while self.root_entries_pos < self.root_entries_len:
1462
entry = self.root_entries[self.root_entries_pos]
1463
self.root_entries_pos = self.root_entries_pos + 1
1464
result, changed = self._process_entry(entry, self.root_dir_info)
1465
if changed is not None:
1467
self._gather_result_for_consistency(result)
1468
if changed or self.include_unchanged:
1470
# Have we finished the prior root, or never started one ?
1471
if self.current_root is None:
1472
# TODO: the pending list should be lexically sorted? the
1473
# interface doesn't require it.
1475
self.current_root = self.search_specific_files.pop()
1477
raise StopIteration()
1478
self.searched_specific_files.add(self.current_root)
1479
# process the entries for this containing directory: the rest will be
1480
# found by their parents recursively.
1481
self.root_entries = self.state._entries_for_path(self.current_root)
1482
self.root_entries_len = len(self.root_entries)
1483
self.current_root_unicode = self.current_root.decode('utf8')
1484
self.root_abspath = self.tree.abspath(self.current_root_unicode)
1486
root_stat = os.lstat(self.root_abspath)
1488
if e.errno == errno.ENOENT:
1489
# the path does not exist: let _process_entry know that.
1490
self.root_dir_info = None
1492
# some other random error: hand it up.
1495
self.root_dir_info = ('', self.current_root,
1496
osutils.file_kind_from_stat_mode(root_stat.st_mode), root_stat,
1498
if self.root_dir_info[2] == 'directory':
1499
if self.tree._directory_is_tree_reference(
1500
self.current_root_unicode):
1501
self.root_dir_info = self.root_dir_info[:2] + \
1502
('tree-reference',) + self.root_dir_info[3:]
1503
if not self.root_entries and not self.root_dir_info:
1504
# this specified path is not present at all, skip it.
1505
# (tail recursion, can do a loop once the full structure is
1507
return self._iter_next()
1509
self.root_entries_pos = 0
1510
# XXX Clarity: This loop is duplicated a out the self.current_root
1511
# is None guard above: if we return from it, it completes there
1512
# (and the following if block cannot trigger because
1513
# path_handled must be true, so the if block is not # duplicated.
1514
while self.root_entries_pos < self.root_entries_len:
1515
entry = self.root_entries[self.root_entries_pos]
1516
self.root_entries_pos = self.root_entries_pos + 1
1517
result, changed = self._process_entry(entry, self.root_dir_info)
1518
if changed is not None:
1521
self._gather_result_for_consistency(result)
1522
if changed or self.include_unchanged:
1524
# handle unversioned specified paths:
1525
if self.want_unversioned and not path_handled and self.root_dir_info:
1526
new_executable = bool(
1527
stat.S_ISREG(self.root_dir_info[3].st_mode)
1528
and stat.S_IEXEC & self.root_dir_info[3].st_mode)
1530
(None, self.current_root_unicode),
1534
(None, splitpath(self.current_root_unicode)[-1]),
1535
(None, self.root_dir_info[2]),
1536
(None, new_executable)
1538
# If we reach here, the outer flow continues, which enters into the
1539
# per-root setup logic.
1540
if (self.current_dir_info is None and self.current_block is None and not
1541
self.doing_consistency_expansion):
1542
# setup iteration of this root:
1543
self.current_dir_list = None
1544
if self.root_dir_info and self.root_dir_info[2] == 'tree-reference':
1545
self.current_dir_info = None
1547
self.dir_iterator = osutils._walkdirs_utf8(self.root_abspath,
1548
prefix=self.current_root)
1551
self.current_dir_info = self.dir_iterator.next()
1552
self.current_dir_list = self.current_dir_info[1]
1554
# there may be directories in the inventory even though
1555
# this path is not a file on disk: so mark it as end of
1557
if e.errno in (errno.ENOENT, errno.ENOTDIR, errno.EINVAL):
1558
self.current_dir_info = None
1559
elif sys.platform == 'win32':
1560
# on win32, python2.4 has e.errno == ERROR_DIRECTORY, but
1561
# python 2.5 has e.errno == EINVAL,
1562
# and e.winerror == ERROR_DIRECTORY
1564
e_winerror = e.winerror
1565
except AttributeError:
1567
win_errors = (ERROR_DIRECTORY, ERROR_PATH_NOT_FOUND)
1568
if (e.errno in win_errors or e_winerror in win_errors):
1569
self.current_dir_info = None
1571
# Will this really raise the right exception ?
1576
if self.current_dir_info[0][0] == '':
1577
# remove .bzr from iteration
1578
bzr_index = self.bisect_left(self.current_dir_list, ('.bzr',))
1579
if self.current_dir_list[bzr_index][0] != '.bzr':
1580
raise AssertionError()
1581
del self.current_dir_list[bzr_index]
1582
initial_key = (self.current_root, '', '')
1583
self.block_index, _ = self.state._find_block_index_from_key(initial_key)
1584
if self.block_index == 0:
1585
# we have processed the total root already, but because the
1586
# initial key matched it we should skip it here.
1587
self.block_index = self.block_index + 1
1588
self._update_current_block()
1589
# walk until both the directory listing and the versioned metadata
1591
while (self.current_dir_info is not None
1592
or self.current_block is not None):
1593
# Uncommon case - a missing directory or an unversioned directory:
1594
if (self.current_dir_info and self.current_block
1595
and self.current_dir_info[0][0] != self.current_block[0]):
1596
# Work around pyrex broken heuristic - current_dirname has
1597
# the same scope as current_dirname_c
1598
current_dirname = self.current_dir_info[0][0]
1599
current_dirname_c = PyString_AS_STRING_void(
1600
<void *>current_dirname)
1601
current_blockname = self.current_block[0]
1602
current_blockname_c = PyString_AS_STRING_void(
1603
<void *>current_blockname)
1604
# In the python generator we evaluate this if block once per
1605
# dir+block; because we reenter in the pyrex version its being
1606
# evaluated once per path: we could cache the result before
1607
# doing the while loop and probably save time.
1608
if _cmp_by_dirs(current_dirname_c,
1609
PyString_Size(current_dirname),
1610
current_blockname_c,
1611
PyString_Size(current_blockname)) < 0:
1612
# filesystem data refers to paths not covered by the
1613
# dirblock. this has two possibilities:
1614
# A) it is versioned but empty, so there is no block for it
1615
# B) it is not versioned.
1617
# if (A) then we need to recurse into it to check for
1618
# new unknown files or directories.
1619
# if (B) then we should ignore it, because we don't
1620
# recurse into unknown directories.
1621
# We are doing a loop
1622
while self.path_index < len(self.current_dir_list):
1623
current_path_info = self.current_dir_list[self.path_index]
1624
# dont descend into this unversioned path if it is
1626
if current_path_info[2] in ('directory',
1628
del self.current_dir_list[self.path_index]
1629
self.path_index = self.path_index - 1
1630
self.path_index = self.path_index + 1
1631
if self.want_unversioned:
1632
if current_path_info[2] == 'directory':
1633
if self.tree._directory_is_tree_reference(
1634
self.utf8_decode(current_path_info[0])[0]):
1635
current_path_info = current_path_info[:2] + \
1636
('tree-reference',) + current_path_info[3:]
1637
new_executable = bool(
1638
stat.S_ISREG(current_path_info[3].st_mode)
1639
and stat.S_IEXEC & current_path_info[3].st_mode)
1641
(None, self.utf8_decode(current_path_info[0])[0]),
1645
(None, self.utf8_decode(current_path_info[1])[0]),
1646
(None, current_path_info[2]),
1647
(None, new_executable))
1648
# This dir info has been handled, go to the next
1650
self.current_dir_list = None
1652
self.current_dir_info = self.dir_iterator.next()
1653
self.current_dir_list = self.current_dir_info[1]
1654
except StopIteration:
1655
self.current_dir_info = None
1657
# We have a dirblock entry for this location, but there
1658
# is no filesystem path for this. This is most likely
1659
# because a directory was removed from the disk.
1660
# We don't have to report the missing directory,
1661
# because that should have already been handled, but we
1662
# need to handle all of the files that are contained
1664
while self.current_block_pos < len(self.current_block_list):
1665
current_entry = self.current_block_list[self.current_block_pos]
1666
self.current_block_pos = self.current_block_pos + 1
1667
# entry referring to file not present on disk.
1668
# advance the entry only, after processing.
1669
result, changed = self._process_entry(current_entry, None)
1670
if changed is not None:
1672
self._gather_result_for_consistency(result)
1673
if changed or self.include_unchanged:
1675
self.block_index = self.block_index + 1
1676
self._update_current_block()
1677
continue # next loop-on-block/dir
1678
result = self._loop_one_block()
1679
if result is not None:
1681
if len(self.search_specific_files):
1682
# More supplied paths to process
1683
self.current_root = None
1684
return self._iter_next()
1685
# Start expanding more conservatively, adding paths the user may not
1686
# have intended but required for consistent deltas.
1687
self.doing_consistency_expansion = 1
1688
if not self._pending_consistent_entries:
1689
self._pending_consistent_entries = self._next_consistent_entries()
1690
while self._pending_consistent_entries:
1691
result, changed = self._pending_consistent_entries.pop()
1692
if changed is not None:
1694
raise StopIteration()
1696
cdef object _maybe_tree_ref(self, current_path_info):
1697
if self.tree._directory_is_tree_reference(
1698
self.utf8_decode(current_path_info[0])[0]):
1699
return current_path_info[:2] + \
1700
('tree-reference',) + current_path_info[3:]
1702
return current_path_info
1704
cdef object _loop_one_block(self):
1705
# current_dir_info and current_block refer to the same directory -
1706
# this is the common case code.
1707
# Assign local variables for current path and entry:
1708
cdef object current_entry
1709
cdef object current_path_info
1710
cdef int path_handled
1713
# cdef char * temp_str
1714
# cdef Py_ssize_t temp_str_length
1715
# PyString_AsStringAndSize(disk_kind, &temp_str, &temp_str_length)
1716
# if not strncmp(temp_str, "directory", temp_str_length):
1717
if (self.current_block is not None and
1718
self.current_block_pos < PyList_GET_SIZE(self.current_block_list)):
1719
current_entry = PyList_GET_ITEM(self.current_block_list,
1720
self.current_block_pos)
1722
Py_INCREF(current_entry)
1724
current_entry = None
1725
if (self.current_dir_info is not None and
1726
self.path_index < PyList_GET_SIZE(self.current_dir_list)):
1727
current_path_info = PyList_GET_ITEM(self.current_dir_list,
1730
Py_INCREF(current_path_info)
1731
disk_kind = PyTuple_GET_ITEM(current_path_info, 2)
1733
Py_INCREF(disk_kind)
1734
if disk_kind == "directory":
1735
current_path_info = self._maybe_tree_ref(current_path_info)
1737
current_path_info = None
1738
while (current_entry is not None or current_path_info is not None):
1743
if current_entry is None:
1744
# unversioned - the check for path_handled when the path
1745
# is advanced will yield this path if needed.
1747
elif current_path_info is None:
1748
# no path is fine: the per entry code will handle it.
1749
result, changed = self._process_entry(current_entry,
1752
minikind = _minikind_from_string(
1753
current_entry[1][self.target_index][0])
1754
cmp_result = cmp(current_path_info[1], current_entry[0][1])
1755
if (cmp_result or minikind == c'a' or minikind == c'r'):
1756
# The current path on disk doesn't match the dirblock
1757
# record. Either the dirblock record is marked as
1758
# absent/renamed, or the file on disk is not present at all
1759
# in the dirblock. Either way, report about the dirblock
1760
# entry, and let other code handle the filesystem one.
1762
# Compare the basename for these files to determine
1765
# extra file on disk: pass for now, but only
1766
# increment the path, not the entry
1769
# entry referring to file not present on disk.
1770
# advance the entry only, after processing.
1771
result, changed = self._process_entry(current_entry,
1775
# paths are the same,and the dirstate entry is not
1776
# absent or renamed.
1777
result, changed = self._process_entry(current_entry,
1779
if changed is not None:
1781
if not changed and not self.include_unchanged:
1783
# >- loop control starts here:
1785
if advance_entry and current_entry is not None:
1786
self.current_block_pos = self.current_block_pos + 1
1787
if self.current_block_pos < PyList_GET_SIZE(self.current_block_list):
1788
current_entry = self.current_block_list[self.current_block_pos]
1790
current_entry = None
1792
if advance_path and current_path_info is not None:
1793
if not path_handled:
1794
# unversioned in all regards
1795
if self.want_unversioned:
1796
new_executable = bool(
1797
stat.S_ISREG(current_path_info[3].st_mode)
1798
and stat.S_IEXEC & current_path_info[3].st_mode)
1800
relpath_unicode = self.utf8_decode(current_path_info[0])[0]
1801
except UnicodeDecodeError:
1802
raise errors.BadFilenameEncoding(
1803
current_path_info[0], osutils._fs_enc)
1804
if changed is not None:
1805
raise AssertionError(
1806
"result is not None: %r" % result)
1808
(None, relpath_unicode),
1812
(None, self.utf8_decode(current_path_info[1])[0]),
1813
(None, current_path_info[2]),
1814
(None, new_executable))
1816
# dont descend into this unversioned path if it is
1818
if current_path_info[2] in ('directory'):
1819
del self.current_dir_list[self.path_index]
1820
self.path_index = self.path_index - 1
1821
# dont descend the disk iterator into any tree
1823
if current_path_info[2] == 'tree-reference':
1824
del self.current_dir_list[self.path_index]
1825
self.path_index = self.path_index - 1
1826
self.path_index = self.path_index + 1
1827
if self.path_index < len(self.current_dir_list):
1828
current_path_info = self.current_dir_list[self.path_index]
1829
if current_path_info[2] == 'directory':
1830
current_path_info = self._maybe_tree_ref(
1833
current_path_info = None
1834
if changed is not None:
1835
# Found a result on this pass, yield it
1837
self._gather_result_for_consistency(result)
1838
if changed or self.include_unchanged:
1840
if self.current_block is not None:
1841
self.block_index = self.block_index + 1
1842
self._update_current_block()
1843
if self.current_dir_info is not None:
1845
self.current_dir_list = None
1847
self.current_dir_info = self.dir_iterator.next()
1848
self.current_dir_list = self.current_dir_info[1]
1849
except StopIteration:
1850
self.current_dir_info = None
1852
cdef object _next_consistent_entries(self):
1853
"""Grabs the next specific file parent case to consider.
1855
:return: A list of the results, each of which is as for _process_entry.
1858
while self.search_specific_file_parents:
1859
# Process the parent directories for the paths we were iterating.
1860
# Even in extremely large trees this should be modest, so currently
1861
# no attempt is made to optimise.
1862
path_utf8 = self.search_specific_file_parents.pop()
1863
if path_utf8 in self.searched_exact_paths:
1864
# We've examined this path.
1866
if osutils.is_inside_any(self.searched_specific_files, path_utf8):
1867
# We've examined this path.
1869
path_entries = self.state._entries_for_path(path_utf8)
1870
# We need either one or two entries. If the path in
1871
# self.target_index has moved (so the entry in source_index is in
1872
# 'ar') then we need to also look for the entry for this path in
1873
# self.source_index, to output the appropriate delete-or-rename.
1874
selected_entries = []
1876
for candidate_entry in path_entries:
1877
# Find entries present in target at this path:
1878
if candidate_entry[1][self.target_index][0] not in 'ar':
1880
selected_entries.append(candidate_entry)
1881
# Find entries present in source at this path:
1882
elif (self.source_index is not None and
1883
candidate_entry[1][self.source_index][0] not in 'ar'):
1885
if candidate_entry[1][self.target_index][0] == 'a':
1886
# Deleted, emit it here.
1887
selected_entries.append(candidate_entry)
1889
# renamed, emit it when we process the directory it
1891
self.search_specific_file_parents.add(
1892
candidate_entry[1][self.target_index][1])
1894
raise AssertionError(
1895
"Missing entry for specific path parent %r, %r" % (
1896
path_utf8, path_entries))
1897
path_info = self._path_info(path_utf8, path_utf8.decode('utf8'))
1898
for entry in selected_entries:
1899
if entry[0][2] in self.seen_ids:
1901
result, changed = self._process_entry(entry, path_info)
1903
raise AssertionError(
1904
"Got entry<->path mismatch for specific path "
1905
"%r entry %r path_info %r " % (
1906
path_utf8, entry, path_info))
1907
# Only include changes - we're outside the users requested
1910
self._gather_result_for_consistency(result)
1911
if (result[6][0] == 'directory' and
1912
result[6][1] != 'directory'):
1913
# This stopped being a directory, the old children have
1915
if entry[1][self.source_index][0] == 'r':
1916
# renamed, take the source path
1917
entry_path_utf8 = entry[1][self.source_index][1]
1919
entry_path_utf8 = path_utf8
1920
initial_key = (entry_path_utf8, '', '')
1921
block_index, _ = self.state._find_block_index_from_key(
1923
if block_index == 0:
1924
# The children of the root are in block index 1.
1925
block_index = block_index + 1
1926
current_block = None
1927
if block_index < len(self.state._dirblocks):
1928
current_block = self.state._dirblocks[block_index]
1929
if not osutils.is_inside(
1930
entry_path_utf8, current_block[0]):
1931
# No entries for this directory at all.
1932
current_block = None
1933
if current_block is not None:
1934
for entry in current_block[1]:
1935
if entry[1][self.source_index][0] in 'ar':
1936
# Not in the source tree, so doesn't have to be
1939
# Path of the entry itself.
1940
self.search_specific_file_parents.add(
1941
self.pathjoin(*entry[0][:2]))
1942
if changed or self.include_unchanged:
1943
results.append((result, changed))
1944
self.searched_exact_paths.add(path_utf8)
1947
cdef object _path_info(self, utf8_path, unicode_path):
1948
"""Generate path_info for unicode_path.
1950
:return: None if unicode_path does not exist, or a path_info tuple.
1952
abspath = self.tree.abspath(unicode_path)
1954
stat = os.lstat(abspath)
1956
if e.errno == errno.ENOENT:
1957
# the path does not exist.
1961
utf8_basename = utf8_path.rsplit('/', 1)[-1]
1962
dir_info = (utf8_path, utf8_basename,
1963
osutils.file_kind_from_stat_mode(stat.st_mode), stat,
1965
if dir_info[2] == 'directory':
1966
if self.tree._directory_is_tree_reference(
1968
self.root_dir_info = self.root_dir_info[:2] + \
1969
('tree-reference',) + self.root_dir_info[3:]
added
removed
bzrlib/_dirstate_helpers_pyx.pyx
