201
181
delta = TreeDelta()
202
182
mutter('start compare_trees')
204
# TODO: Rather than iterating over the whole tree and then filtering, we
205
# could diff just the specified files (if any) and their subtrees.
207
old_files = old_tree.list_files()
208
new_files = new_tree.list_files()
219
except StopIteration:
220
return None, None, None, None, None
221
old_path, old_class, old_kind, old_file_id, old_entry = get_next(old_files)
222
new_path, new_class, new_kind, new_file_id, new_entry = get_next(new_files)
225
def check_matching(old_path, old_entry, new_path, new_entry):
226
"""We have matched up 2 file_ids, check for changes."""
227
assert old_entry.kind == new_entry.kind
229
if old_entry.kind == 'root_directory':
232
if specific_file_ids:
233
if (old_file_id not in specific_file_ids and
234
new_file_id not in specific_file_ids):
237
# temporary hack until all entries are populated before clients
239
old_entry._read_tree_state(old_path, old_tree)
240
new_entry._read_tree_state(new_path, new_tree)
241
text_modified, meta_modified = new_entry.detect_changes(old_entry)
243
# If the name changes, or the parent_id changes, we have a rename
244
# (if we move a parent, that doesn't count as a rename for the file)
245
if (old_entry.name != new_entry.name
246
or old_entry.parent_id != new_entry.parent_id):
247
delta.renamed.append((old_path,
249
old_entry.file_id, old_entry.kind,
250
text_modified, meta_modified))
251
elif text_modified or meta_modified:
252
delta.modified.append((new_path, new_entry.file_id, new_entry.kind,
253
text_modified, meta_modified))
255
delta.unchanged.append((new_path, new_entry.file_id, new_entry.kind))
258
def handle_old(path, entry):
259
"""old entry without a new entry match
261
Check to see if a matching new entry was already seen as an
262
added file, and switch the pair into being a rename.
263
Otherwise just mark the old entry being removed.
265
if entry.file_id in added:
266
# Actually this is a rename, we found a new file_id earlier
267
# at a different location, so it is no-longer added
268
x_new_path, x_new_entry = added.pop(entry.file_id)
269
check_matching(path, entry, x_new_path, x_new_entry)
271
# We have an old_file_id which doesn't line up with a new_file_id
272
# So this file looks to be removed
273
assert entry.file_id not in removed
274
removed[entry.file_id] = path, entry
276
def handle_new(path, entry):
277
"""new entry without an old entry match
279
Check to see if a matching old entry was already seen as a
280
removal, and change the pair into a rename.
281
Otherwise just mark the new entry as an added file.
283
if entry.file_id in removed:
284
# We saw this file_id earlier at an old different location
285
# it is no longer removed, just renamed
286
x_old_path, x_old_entry = removed.pop(entry.file_id)
287
check_matching(x_old_path, x_old_entry, path, entry)
289
# We have a new file which does not match an old file
291
assert entry.file_id not in added
292
added[entry.file_id] = path, entry
294
while old_path or new_path:
295
# list_files() returns files in alphabetical path sorted order
296
if old_path == new_path:
297
if old_file_id == new_file_id:
298
# This is the common case, the files are in the same place
299
# check if there were any content changes
301
if old_file_id is None:
302
# We have 2 unversioned files, no deltas possible???
305
check_matching(old_path, old_entry, new_path, new_entry)
307
# The ids don't match, so we have to handle them both
309
if old_file_id is not None:
310
handle_old(old_path, old_entry)
312
if new_file_id is not None:
313
handle_new(new_path, new_entry)
315
# The two entries were at the same path, so increment both sides
316
old_path, old_class, old_kind, old_file_id, old_entry = get_next(old_files)
317
new_path, new_class, new_kind, new_file_id, new_entry = get_next(new_files)
318
elif new_path is None or (old_path is not None and old_path < new_path):
319
# Assume we don't match, only process old_path
320
if old_file_id is not None:
321
handle_old(old_path, old_entry)
322
# old_path came first, so increment it, trying to match up
323
old_path, old_class, old_kind, old_file_id, old_entry = get_next(old_files)
324
elif new_path is not None:
325
# new_path came first, so increment it, trying to match up
326
if new_file_id is not None:
327
handle_new(new_path, new_entry)
328
new_path, new_class, new_kind, new_file_id, new_entry = get_next(new_files)
330
# Now we have a set of added and removed files, mark them all
331
for old_path, old_entry in removed.itervalues():
332
if specific_file_ids:
333
if not old_entry.file_id in specific_file_ids:
335
delta.removed.append((old_path, old_entry.file_id, old_entry.kind))
336
for new_path, new_entry in added.itervalues():
337
if specific_file_ids:
338
if not new_entry.file_id in specific_file_ids:
340
delta.added.append((new_path, new_entry.file_id, new_entry.kind))
184
# TODO: match for specific files can be rather smarter by finding
185
# the IDs of those files up front and then considering only that.
187
for file_id in old_tree:
188
if file_id in new_tree:
189
old_ie = old_inv[file_id]
190
new_ie = new_inv[file_id]
193
assert kind == new_ie.kind
195
assert kind in InventoryEntry.known_kinds, \
196
'invalid file kind %r' % kind
198
if kind == 'root_directory':
202
if (not is_inside_any(specific_files, old_inv.id2path(file_id))
203
and not is_inside_any(specific_files, new_inv.id2path(file_id))):
206
# temporary hack until all entries are populated before clients
208
old_path = old_inv.id2path(file_id)
209
new_path = new_inv.id2path(file_id)
210
old_ie._read_tree_state(old_path, old_tree)
211
new_ie._read_tree_state(new_path, new_tree)
212
text_modified, meta_modified = new_ie.detect_changes(old_ie)
214
# TODO: Can possibly avoid calculating path strings if the
215
# two files are unchanged and their names and parents are
216
# the same and the parents are unchanged all the way up.
217
# May not be worthwhile.
219
if (old_ie.name != new_ie.name
220
or old_ie.parent_id != new_ie.parent_id):
221
delta.renamed.append((old_path,
224
text_modified, meta_modified))
225
elif text_modified or meta_modified:
226
delta.modified.append((new_path, file_id, kind,
227
text_modified, meta_modified))
229
delta.unchanged.append((new_path, file_id, kind))
231
kind = old_inv.get_file_kind(file_id)
232
if kind == 'root_directory':
234
old_path = old_inv.id2path(file_id)
236
if not is_inside_any(specific_files, old_path):
238
delta.removed.append((old_path, file_id, kind))
240
mutter('start looking for new files')
241
for file_id in new_inv:
242
if file_id in old_inv or file_id not in new_tree:
244
kind = new_inv.get_file_kind(file_id)
245
if kind == 'root_directory':
247
new_path = new_inv.id2path(file_id)
249
if not is_inside_any(specific_files, new_path):
251
delta.added.append((new_path, file_id, kind))
342
253
delta.removed.sort()
343
254
delta.added.sort()
344
255
delta.renamed.sort()
345
# TODO: jam 20060529 These lists shouldn't need to be sorted
346
# since we added them in alphabetical order.
347
256
delta.modified.sort()
348
257
delta.unchanged.sort()
added
removed
bzrlib/delta.py
