167
184
delta = TreeDelta()
168
185
mutter('start compare_trees')
170
# TODO: match for specific files can be rather smarter by finding
171
# the IDs of those files up front and then considering only that.
173
for file_id in old_tree:
174
if file_id in new_tree:
175
old_ie = old_inv[file_id]
176
new_ie = new_inv[file_id]
179
assert kind == new_ie.kind
181
assert kind in InventoryEntry.known_kinds, \
182
'invalid file kind %r' % kind
184
if kind == 'root_directory':
188
if (not is_inside_any(specific_files, old_inv.id2path(file_id))
189
and not is_inside_any(specific_files, new_inv.id2path(file_id))):
192
# temporary hack until all entries are populated before clients
194
old_path = old_inv.id2path(file_id)
195
new_path = new_inv.id2path(file_id)
196
old_ie._read_tree_state(old_path, old_tree)
197
new_ie._read_tree_state(new_path, new_tree)
198
text_modified, meta_modified = new_ie.detect_changes(old_ie)
200
# TODO: Can possibly avoid calculating path strings if the
201
# two files are unchanged and their names and parents are
202
# the same and the parents are unchanged all the way up.
203
# May not be worthwhile.
205
if (old_ie.name != new_ie.name
206
or old_ie.parent_id != new_ie.parent_id):
207
delta.renamed.append((old_path,
210
text_modified, meta_modified))
211
elif text_modified or meta_modified:
212
delta.modified.append((new_path, file_id, kind,
213
text_modified, meta_modified))
215
delta.unchanged.append((new_path, file_id, kind))
217
kind = old_inv.get_file_kind(file_id)
218
if kind == 'root_directory':
220
old_path = old_inv.id2path(file_id)
222
if not is_inside_any(specific_files, old_path):
224
delta.removed.append((old_path, file_id, kind))
226
mutter('start looking for new files')
227
for file_id in new_inv:
228
if file_id in old_inv or file_id not in new_tree:
230
kind = new_inv.get_file_kind(file_id)
231
if kind == 'root_directory':
233
new_path = new_inv.id2path(file_id)
187
# TODO: Rather than iterating over the whole tree and then filtering, we
188
# could diff just the specified files (if any) and their subtrees.
189
# Perhaps should take a list of file-ids instead? Need to indicate any
190
# ids or names which were not found in the trees.
192
old_files = old_tree.list_files()
193
new_files = new_tree.list_files()
204
except StopIteration:
205
return None, None, None, None, None
206
old_path, old_class, old_kind, old_file_id, old_entry = get_next(old_files)
207
new_path, new_class, new_kind, new_file_id, new_entry = get_next(new_files)
210
def check_matching(old_path, old_entry, new_path, new_entry):
211
"""We have matched up 2 file_ids, check for changes."""
212
assert old_entry.kind == new_entry.kind
214
if old_entry.kind == 'root_directory':
218
if (not is_inside_any(specific_files, old_path)
219
and not is_inside_any(specific_files, new_path)):
222
# temporary hack until all entries are populated before clients
224
old_entry._read_tree_state(old_path, old_tree)
225
new_entry._read_tree_state(new_path, new_tree)
226
text_modified, meta_modified = new_entry.detect_changes(old_entry)
228
# If the name changes, or the parent_id changes, we have a rename
229
# (if we move a parent, that doesn't count as a rename for the file)
230
if (old_entry.name != new_entry.name
231
or old_entry.parent_id != new_entry.parent_id):
232
delta.renamed.append((old_path,
234
old_entry.file_id, old_entry.kind,
235
text_modified, meta_modified))
236
elif text_modified or meta_modified:
237
delta.modified.append((new_path, new_entry.file_id, new_entry.kind,
238
text_modified, meta_modified))
240
delta.unchanged.append((new_path, new_entry.file_id, new_entry.kind))
243
def handle_old(path, entry):
244
"""old entry without a new entry match
246
Check to see if a matching new entry was already seen as an
247
added file, and switch the pair into being a rename.
248
Otherwise just mark the old entry being removed.
250
if entry.file_id in added:
251
# Actually this is a rename, we found a new file_id earlier
252
# at a different location, so it is no-longer added
253
x_new_path, x_new_entry = added.pop(entry.file_id)
254
check_matching(path, entry, x_new_path, x_new_entry)
256
# We have an old_file_id which doesn't line up with a new_file_id
257
# So this file looks to be removed
258
assert entry.file_id not in removed
259
removed[entry.file_id] = path, entry
261
def handle_new(path, entry):
262
"""new entry without an old entry match
264
Check to see if a matching old entry was already seen as a
265
removal, and change the pair into a rename.
266
Otherwise just mark the new entry as an added file.
268
if entry.file_id in removed:
269
# We saw this file_id earlier at an old different location
270
# it is no longer removed, just renamed
271
x_old_path, x_old_entry = removed.pop(entry.file_id)
272
check_matching(x_old_path, x_old_entry, path, entry)
274
# We have a new file which does not match an old file
276
assert entry.file_id not in added
277
added[entry.file_id] = path, entry
279
while old_path or new_path:
280
# list_files() returns files in alphabetical path sorted order
281
if old_path == new_path:
282
if old_file_id == new_file_id:
283
# This is the common case, the files are in the same place
284
# check if there were any content changes
286
if old_file_id is None:
287
# We have 2 unversioned files, no deltas possible???
290
check_matching(old_path, old_entry, new_path, new_entry)
292
# The ids don't match, so we have to handle them both
294
if old_file_id is not None:
295
handle_old(old_path, old_entry)
297
if new_file_id is not None:
298
handle_new(new_path, new_entry)
300
# The two entries were at the same path, so increment both sides
301
old_path, old_class, old_kind, old_file_id, old_entry = get_next(old_files)
302
new_path, new_class, new_kind, new_file_id, new_entry = get_next(new_files)
303
elif new_path is None or (old_path is not None and old_path < new_path):
304
# Assume we don't match, only process old_path
305
if old_file_id is not None:
306
handle_old(old_path, old_entry)
307
# old_path came first, so increment it, trying to match up
308
old_path, old_class, old_kind, old_file_id, old_entry = get_next(old_files)
309
elif new_path is not None:
310
# new_path came first, so increment it, trying to match up
311
if new_file_id is not None:
312
handle_new(new_path, new_entry)
313
new_path, new_class, new_kind, new_file_id, new_entry = get_next(new_files)
315
# Now we have a set of added and removed files, mark them all
316
for old_path, old_entry in removed.itervalues():
318
if not is_inside_any(specific_files, old_path):
320
delta.removed.append((old_path, old_entry.file_id, old_entry.kind))
321
for new_path, new_entry in added.itervalues():
234
322
if specific_files:
235
323
if not is_inside_any(specific_files, new_path):
237
delta.added.append((new_path, file_id, kind))
325
delta.added.append((new_path, new_entry.file_id, new_entry.kind))
239
327
delta.removed.sort()
240
328
delta.added.sort()
241
329
delta.renamed.sort()
330
# TODO: jam 20060529 These lists shouldn't need to be sorted
331
# since we added them in alphabetical order.
242
332
delta.modified.sort()
243
333
delta.unchanged.sort()
added
removed
bzrlib/delta.py
