1776
2059
if not all_versioned:
1777
2060
raise errors.PathsNotVersionedError(specific_files)
1778
2061
# -- remove redundancy in supplied specific_files to prevent over-scanning --
1779
search_specific_files = set()
1780
for path in specific_files:
1781
other_specific_files = specific_files.difference(set([path]))
1782
if not osutils.is_inside_any(other_specific_files, path):
1783
# this is a top level path, we must check it.
1784
search_specific_files.add(path)
1786
# compare source_index and target_index at or under each element of search_specific_files.
1787
# follow the following comparison table. Note that we only want to do diff operations when
1788
# the target is fdl because thats when the walkdirs logic will have exposed the pathinfo
1792
# Source | Target | disk | action
1793
# r | fdlt | | add source to search, add id path move and perform
1794
# | | | diff check on source-target
1795
# r | fdlt | a | dangling file that was present in the basis.
1797
# r | a | | add source to search
1799
# r | r | | this path is present in a non-examined tree, skip.
1800
# r | r | a | this path is present in a non-examined tree, skip.
1801
# a | fdlt | | add new id
1802
# a | fdlt | a | dangling locally added file, skip
1803
# a | a | | not present in either tree, skip
1804
# a | a | a | not present in any tree, skip
1805
# a | r | | not present in either tree at this path, skip as it
1806
# | | | may not be selected by the users list of paths.
1807
# a | r | a | not present in either tree at this path, skip as it
1808
# | | | may not be selected by the users list of paths.
1809
# fdlt | fdlt | | content in both: diff them
1810
# fdlt | fdlt | a | deleted locally, but not unversioned - show as deleted ?
1811
# fdlt | a | | unversioned: output deleted id for now
1812
# fdlt | a | a | unversioned and deleted: output deleted id
1813
# fdlt | r | | relocated in this tree, so add target to search.
1814
# | | | Dont diff, we will see an r,fd; pair when we reach
1815
# | | | this id at the other path.
1816
# fdlt | r | a | relocated in this tree, so add target to search.
1817
# | | | Dont diff, we will see an r,fd; pair when we reach
1818
# | | | this id at the other path.
1820
# for all search_indexs in each path at or under each element of
1821
# search_specific_files, if the detail is relocated: add the id, and add the
1822
# relocated path as one to search if its not searched already. If the
1823
# detail is not relocated, add the id.
1824
searched_specific_files = set()
1825
NULL_PARENT_DETAILS = dirstate.DirState.NULL_PARENT_DETAILS
1826
# Using a list so that we can access the values and change them in
1827
# nested scope. Each one is [path, file_id, entry]
1828
last_source_parent = [None, None]
1829
last_target_parent = [None, None]
2062
search_specific_files = osutils.minimum_path_selection(specific_files)
1831
2064
use_filesystem_for_exec = (sys.platform != 'win32')
1833
# Just a sentry, so that _process_entry can say that this
1834
# record is handled, but isn't interesting to process (unchanged)
1835
uninteresting = object()
1838
old_dirname_to_file_id = {}
1839
new_dirname_to_file_id = {}
1840
# TODO: jam 20070516 - Avoid the _get_entry lookup overhead by
1841
# keeping a cache of directories that we have seen.
1843
def _process_entry(entry, path_info):
1844
"""Compare an entry and real disk to generate delta information.
1846
:param path_info: top_relpath, basename, kind, lstat, abspath for
1847
the path of entry. If None, then the path is considered absent.
1848
(Perhaps we should pass in a concrete entry for this ?)
1849
Basename is returned as a utf8 string because we expect this
1850
tuple will be ignored, and don't want to take the time to
1852
:return: None if these don't match
1853
A tuple of information about the change, or
1854
the object 'uninteresting' if these match, but are
1855
basically identical.
1857
if source_index is None:
1858
source_details = NULL_PARENT_DETAILS
1860
source_details = entry[1][source_index]
1861
target_details = entry[1][target_index]
1862
target_minikind = target_details[0]
1863
if path_info is not None and target_minikind in 'fdlt':
1864
assert target_index == 0
1865
link_or_sha1 = state.update_entry(entry, abspath=path_info[4],
1866
stat_value=path_info[3])
1867
# The entry may have been modified by update_entry
1868
target_details = entry[1][target_index]
1869
target_minikind = target_details[0]
1872
file_id = entry[0][2]
1873
source_minikind = source_details[0]
1874
if source_minikind in 'fdltr' and target_minikind in 'fdlt':
1875
# claimed content in both: diff
1876
# r | fdlt | | add source to search, add id path move and perform
1877
# | | | diff check on source-target
1878
# r | fdlt | a | dangling file that was present in the basis.
1880
if source_minikind in 'r':
1881
# add the source to the search path to find any children it
1882
# has. TODO ? : only add if it is a container ?
1883
if not osutils.is_inside_any(searched_specific_files,
1885
search_specific_files.add(source_details[1])
1886
# generate the old path; this is needed for stating later
1888
old_path = source_details[1]
1889
old_dirname, old_basename = os.path.split(old_path)
1890
path = pathjoin(entry[0][0], entry[0][1])
1891
old_entry = state._get_entry(source_index,
1893
# update the source details variable to be the real
1895
source_details = old_entry[1][source_index]
1896
source_minikind = source_details[0]
1898
old_dirname = entry[0][0]
1899
old_basename = entry[0][1]
1900
old_path = path = None
1901
if path_info is None:
1902
# the file is missing on disk, show as removed.
1903
content_change = True
1907
# source and target are both versioned and disk file is present.
1908
target_kind = path_info[2]
1909
if target_kind == 'directory':
1911
old_path = path = pathjoin(old_dirname, old_basename)
1912
new_dirname_to_file_id[path] = file_id
1913
if source_minikind != 'd':
1914
content_change = True
1916
# directories have no fingerprint
1917
content_change = False
1919
elif target_kind == 'file':
1920
if source_minikind != 'f':
1921
content_change = True
1923
# We could check the size, but we already have the
1925
content_change = (link_or_sha1 != source_details[1])
1926
# Target details is updated at update_entry time
1927
if use_filesystem_for_exec:
1928
# We don't need S_ISREG here, because we are sure
1929
# we are dealing with a file.
1930
target_exec = bool(stat.S_IEXEC & path_info[3].st_mode)
1932
target_exec = target_details[3]
1933
elif target_kind == 'symlink':
1934
if source_minikind != 'l':
1935
content_change = True
1937
content_change = (link_or_sha1 != source_details[1])
1939
elif target_kind == 'tree-reference':
1940
if source_minikind != 't':
1941
content_change = True
1943
content_change = False
1946
raise Exception, "unknown kind %s" % path_info[2]
1947
if source_minikind == 'd':
1949
old_path = path = pathjoin(old_dirname, old_basename)
1950
old_dirname_to_file_id[old_path] = file_id
1951
# parent id is the entry for the path in the target tree
1952
if old_dirname == last_source_parent[0]:
1953
source_parent_id = last_source_parent[1]
1956
source_parent_id = old_dirname_to_file_id[old_dirname]
1958
source_parent_entry = state._get_entry(source_index,
1959
path_utf8=old_dirname)
1960
source_parent_id = source_parent_entry[0][2]
1961
if source_parent_id == entry[0][2]:
1962
# This is the root, so the parent is None
1963
source_parent_id = None
1965
last_source_parent[0] = old_dirname
1966
last_source_parent[1] = source_parent_id
1967
new_dirname = entry[0][0]
1968
if new_dirname == last_target_parent[0]:
1969
target_parent_id = last_target_parent[1]
1972
target_parent_id = new_dirname_to_file_id[new_dirname]
1974
# TODO: We don't always need to do the lookup, because the
1975
# parent entry will be the same as the source entry.
1976
target_parent_entry = state._get_entry(target_index,
1977
path_utf8=new_dirname)
1978
assert target_parent_entry != (None, None), (
1979
"Could not find target parent in wt: %s\nparent of: %s"
1980
% (new_dirname, entry))
1981
target_parent_id = target_parent_entry[0][2]
1982
if target_parent_id == entry[0][2]:
1983
# This is the root, so the parent is None
1984
target_parent_id = None
1986
last_target_parent[0] = new_dirname
1987
last_target_parent[1] = target_parent_id
1989
source_exec = source_details[3]
1990
if (include_unchanged
1992
or source_parent_id != target_parent_id
1993
or old_basename != entry[0][1]
1994
or source_exec != target_exec
1996
if old_path is None:
1997
old_path = path = pathjoin(old_dirname, old_basename)
1998
old_path_u = utf8_decode(old_path)[0]
2001
old_path_u = utf8_decode(old_path)[0]
2002
if old_path == path:
2005
path_u = utf8_decode(path)[0]
2006
source_kind = _minikind_to_kind[source_minikind]
2007
return (entry[0][2],
2008
(old_path_u, path_u),
2011
(source_parent_id, target_parent_id),
2012
(utf8_decode(old_basename)[0], utf8_decode(entry[0][1])[0]),
2013
(source_kind, target_kind),
2014
(source_exec, target_exec))
2016
return uninteresting
2017
elif source_minikind in 'a' and target_minikind in 'fdlt':
2018
# looks like a new file
2019
if path_info is not None:
2020
path = pathjoin(entry[0][0], entry[0][1])
2021
# parent id is the entry for the path in the target tree
2022
# TODO: these are the same for an entire directory: cache em.
2023
parent_id = state._get_entry(target_index,
2024
path_utf8=entry[0][0])[0][2]
2025
if parent_id == entry[0][2]:
2027
if use_filesystem_for_exec:
2028
# We need S_ISREG here, because we aren't sure if this
2031
stat.S_ISREG(path_info[3].st_mode)
2032
and stat.S_IEXEC & path_info[3].st_mode)
2034
target_exec = target_details[3]
2035
return (entry[0][2],
2036
(None, utf8_decode(path)[0]),
2040
(None, utf8_decode(entry[0][1])[0]),
2041
(None, path_info[2]),
2042
(None, target_exec))
2044
# but its not on disk: we deliberately treat this as just
2045
# never-present. (Why ?! - RBC 20070224)
2047
elif source_minikind in 'fdlt' and target_minikind in 'a':
2048
# unversioned, possibly, or possibly not deleted: we dont care.
2049
# if its still on disk, *and* theres no other entry at this
2050
# path [we dont know this in this routine at the moment -
2051
# perhaps we should change this - then it would be an unknown.
2052
old_path = pathjoin(entry[0][0], entry[0][1])
2053
# parent id is the entry for the path in the target tree
2054
parent_id = state._get_entry(source_index, path_utf8=entry[0][0])[0][2]
2055
if parent_id == entry[0][2]:
2057
return (entry[0][2],
2058
(utf8_decode(old_path)[0], None),
2062
(utf8_decode(entry[0][1])[0], None),
2063
(_minikind_to_kind[source_minikind], None),
2064
(source_details[3], None))
2065
elif source_minikind in 'fdlt' and target_minikind in 'r':
2066
# a rename; could be a true rename, or a rename inherited from
2067
# a renamed parent. TODO: handle this efficiently. Its not
2068
# common case to rename dirs though, so a correct but slow
2069
# implementation will do.
2070
if not osutils.is_inside_any(searched_specific_files, target_details[1]):
2071
search_specific_files.add(target_details[1])
2072
elif source_minikind in 'ra' and target_minikind in 'ra':
2073
# neither of the selected trees contain this file,
2074
# so skip over it. This is not currently directly tested, but
2075
# is indirectly via test_too_much.TestCommands.test_conflicts.
2078
raise AssertionError("don't know how to compare "
2079
"source_minikind=%r, target_minikind=%r"
2080
% (source_minikind, target_minikind))
2081
## import pdb;pdb.set_trace()
2084
while search_specific_files:
2085
# TODO: the pending list should be lexically sorted? the
2086
# interface doesn't require it.
2087
current_root = search_specific_files.pop()
2088
current_root_unicode = current_root.decode('utf8')
2089
searched_specific_files.add(current_root)
2090
# process the entries for this containing directory: the rest will be
2091
# found by their parents recursively.
2092
root_entries = _entries_for_path(current_root)
2093
root_abspath = self.target.abspath(current_root_unicode)
2095
root_stat = os.lstat(root_abspath)
2097
if e.errno == errno.ENOENT:
2098
# the path does not exist: let _process_entry know that.
2099
root_dir_info = None
2101
# some other random error: hand it up.
2104
root_dir_info = ('', current_root,
2105
osutils.file_kind_from_stat_mode(root_stat.st_mode), root_stat,
2107
if root_dir_info[2] == 'directory':
2108
if self.target._directory_is_tree_reference(
2109
current_root.decode('utf8')):
2110
root_dir_info = root_dir_info[:2] + \
2111
('tree-reference',) + root_dir_info[3:]
2113
if not root_entries and not root_dir_info:
2114
# this specified path is not present at all, skip it.
2116
path_handled = False
2117
for entry in root_entries:
2118
result = _process_entry(entry, root_dir_info)
2119
if result is not None:
2121
if result is not uninteresting:
2123
if want_unversioned and not path_handled and root_dir_info:
2124
new_executable = bool(
2125
stat.S_ISREG(root_dir_info[3].st_mode)
2126
and stat.S_IEXEC & root_dir_info[3].st_mode)
2128
(None, current_root_unicode),
2132
(None, splitpath(current_root_unicode)[-1]),
2133
(None, root_dir_info[2]),
2134
(None, new_executable)
2136
initial_key = (current_root, '', '')
2137
block_index, _ = state._find_block_index_from_key(initial_key)
2138
if block_index == 0:
2139
# we have processed the total root already, but because the
2140
# initial key matched it we should skip it here.
2142
if root_dir_info and root_dir_info[2] == 'tree-reference':
2143
current_dir_info = None
2145
dir_iterator = osutils._walkdirs_utf8(root_abspath, prefix=current_root)
2147
current_dir_info = dir_iterator.next()
2149
# on win32, python2.4 has e.errno == ERROR_DIRECTORY, but
2150
# python 2.5 has e.errno == EINVAL,
2151
# and e.winerror == ERROR_DIRECTORY
2152
e_winerror = getattr(e, 'winerror', None)
2153
win_errors = (ERROR_DIRECTORY, ERROR_PATH_NOT_FOUND)
2154
# there may be directories in the inventory even though
2155
# this path is not a file on disk: so mark it as end of
2157
if e.errno in (errno.ENOENT, errno.ENOTDIR, errno.EINVAL):
2158
current_dir_info = None
2159
elif (sys.platform == 'win32'
2160
and (e.errno in win_errors
2161
or e_winerror in win_errors)):
2162
current_dir_info = None
2166
if current_dir_info[0][0] == '':
2167
# remove .bzr from iteration
2168
bzr_index = bisect_left(current_dir_info[1], ('.bzr',))
2169
assert current_dir_info[1][bzr_index][0] == '.bzr'
2170
del current_dir_info[1][bzr_index]
2171
# walk until both the directory listing and the versioned metadata
2173
if (block_index < len(state._dirblocks) and
2174
osutils.is_inside(current_root, state._dirblocks[block_index][0])):
2175
current_block = state._dirblocks[block_index]
2177
current_block = None
2178
while (current_dir_info is not None or
2179
current_block is not None):
2180
if (current_dir_info and current_block
2181
and current_dir_info[0][0] != current_block[0]):
2182
if cmp_by_dirs(current_dir_info[0][0], current_block[0]) < 0:
2183
# filesystem data refers to paths not covered by the dirblock.
2184
# this has two possibilities:
2185
# A) it is versioned but empty, so there is no block for it
2186
# B) it is not versioned.
2188
# if (A) then we need to recurse into it to check for
2189
# new unknown files or directories.
2190
# if (B) then we should ignore it, because we don't
2191
# recurse into unknown directories.
2193
while path_index < len(current_dir_info[1]):
2194
current_path_info = current_dir_info[1][path_index]
2195
if want_unversioned:
2196
if current_path_info[2] == 'directory':
2197
if self.target._directory_is_tree_reference(
2198
current_path_info[0].decode('utf8')):
2199
current_path_info = current_path_info[:2] + \
2200
('tree-reference',) + current_path_info[3:]
2201
new_executable = bool(
2202
stat.S_ISREG(current_path_info[3].st_mode)
2203
and stat.S_IEXEC & current_path_info[3].st_mode)
2205
(None, utf8_decode(current_path_info[0])[0]),
2209
(None, utf8_decode(current_path_info[1])[0]),
2210
(None, current_path_info[2]),
2211
(None, new_executable))
2212
# dont descend into this unversioned path if it is
2214
if current_path_info[2] in ('directory',
2216
del current_dir_info[1][path_index]
2220
# This dir info has been handled, go to the next
2222
current_dir_info = dir_iterator.next()
2223
except StopIteration:
2224
current_dir_info = None
2226
# We have a dirblock entry for this location, but there
2227
# is no filesystem path for this. This is most likely
2228
# because a directory was removed from the disk.
2229
# We don't have to report the missing directory,
2230
# because that should have already been handled, but we
2231
# need to handle all of the files that are contained
2233
for current_entry in current_block[1]:
2234
# entry referring to file not present on disk.
2235
# advance the entry only, after processing.
2236
result = _process_entry(current_entry, None)
2237
if result is not None:
2238
if result is not uninteresting:
2241
if (block_index < len(state._dirblocks) and
2242
osutils.is_inside(current_root,
2243
state._dirblocks[block_index][0])):
2244
current_block = state._dirblocks[block_index]
2246
current_block = None
2249
if current_block and entry_index < len(current_block[1]):
2250
current_entry = current_block[1][entry_index]
2252
current_entry = None
2253
advance_entry = True
2255
if current_dir_info and path_index < len(current_dir_info[1]):
2256
current_path_info = current_dir_info[1][path_index]
2257
if current_path_info[2] == 'directory':
2258
if self.target._directory_is_tree_reference(
2259
current_path_info[0].decode('utf8')):
2260
current_path_info = current_path_info[:2] + \
2261
('tree-reference',) + current_path_info[3:]
2263
current_path_info = None
2265
path_handled = False
2266
while (current_entry is not None or
2267
current_path_info is not None):
2268
if current_entry is None:
2269
# the check for path_handled when the path is adnvaced
2270
# will yield this path if needed.
2272
elif current_path_info is None:
2273
# no path is fine: the per entry code will handle it.
2274
result = _process_entry(current_entry, current_path_info)
2275
if result is not None:
2276
if result is not uninteresting:
2278
elif (current_entry[0][1] != current_path_info[1]
2279
or current_entry[1][target_index][0] in 'ar'):
2280
# The current path on disk doesn't match the dirblock
2281
# record. Either the dirblock is marked as absent, or
2282
# the file on disk is not present at all in the
2283
# dirblock. Either way, report about the dirblock
2284
# entry, and let other code handle the filesystem one.
2286
# Compare the basename for these files to determine
2288
if current_path_info[1] < current_entry[0][1]:
2289
# extra file on disk: pass for now, but only
2290
# increment the path, not the entry
2291
advance_entry = False
2293
# entry referring to file not present on disk.
2294
# advance the entry only, after processing.
2295
result = _process_entry(current_entry, None)
2296
if result is not None:
2297
if result is not uninteresting:
2299
advance_path = False
2301
result = _process_entry(current_entry, current_path_info)
2302
if result is not None:
2304
if result is not uninteresting:
2306
if advance_entry and current_entry is not None:
2308
if entry_index < len(current_block[1]):
2309
current_entry = current_block[1][entry_index]
2311
current_entry = None
2313
advance_entry = True # reset the advance flaga
2314
if advance_path and current_path_info is not None:
2315
if not path_handled:
2316
# unversioned in all regards
2317
if want_unversioned:
2318
new_executable = bool(
2319
stat.S_ISREG(current_path_info[3].st_mode)
2320
and stat.S_IEXEC & current_path_info[3].st_mode)
2322
(None, utf8_decode(current_path_info[0])[0]),
2326
(None, utf8_decode(current_path_info[1])[0]),
2327
(None, current_path_info[2]),
2328
(None, new_executable))
2329
# dont descend into this unversioned path if it is
2331
if current_path_info[2] in ('directory'):
2332
del current_dir_info[1][path_index]
2334
# dont descend the disk iterator into any tree
2336
if current_path_info[2] == 'tree-reference':
2337
del current_dir_info[1][path_index]
2340
if path_index < len(current_dir_info[1]):
2341
current_path_info = current_dir_info[1][path_index]
2342
if current_path_info[2] == 'directory':
2343
if self.target._directory_is_tree_reference(
2344
current_path_info[0].decode('utf8')):
2345
current_path_info = current_path_info[:2] + \
2346
('tree-reference',) + current_path_info[3:]
2348
current_path_info = None
2349
path_handled = False
2351
advance_path = True # reset the advance flagg.
2352
if current_block is not None:
2354
if (block_index < len(state._dirblocks) and
2355
osutils.is_inside(current_root, state._dirblocks[block_index][0])):
2356
current_block = state._dirblocks[block_index]
2358
current_block = None
2359
if current_dir_info is not None:
2361
current_dir_info = dir_iterator.next()
2362
except StopIteration:
2363
current_dir_info = None
2065
iter_changes = self.target._iter_changes(include_unchanged,
2066
use_filesystem_for_exec, search_specific_files, state,
2067
source_index, target_index, want_unversioned, self.target)
2068
return iter_changes.iter_changes()
2367
2071
def is_compatible(source, target):
2368
2072
# the target must be a dirstate working tree
2369
if not isinstance(target, WorkingTree4):
2073
if not isinstance(target, DirStateWorkingTree):
2371
# the source must be a revtreee or dirstate rev tree.
2075
# the source must be a revtree or dirstate rev tree.
2372
2076
if not isinstance(source,
2373
2077
(revisiontree.RevisionTree, DirStateRevisionTree)):
2375
2079
# the source revid must be in the target dirstate
2376
2080
if not (source._revision_id == NULL_REVISION or
2377
2081
source._revision_id in target.get_parent_ids()):
2378
# TODO: what about ghosts? it may well need to
2082
# TODO: what about ghosts? it may well need to
2379
2083
# check for them explicitly.
added
removed
bzrlib/workingtree_4.py
