2
# -*- coding: UTF-8 -*-
4
# This program is free software; you can redistribute it and/or modify
5
# it under the terms of the GNU General Public License as published by
6
# the Free Software Foundation; either version 2 of the License, or
7
# (at your option) any later version.
9
# This program is distributed in the hope that it will be useful,
10
# but WITHOUT ANY WARRANTY; without even the implied warranty of
11
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12
# GNU General Public License for more details.
14
# You should have received a copy of the GNU General Public License
15
# along with this program; if not, write to the Free Software
16
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20
from trace import mutter
21
from errors import BzrError
24
def diff_trees(old_tree, new_tree):
25
"""Compute diff between two trees.
27
They may be in different branches and may be working or historical
30
This only compares the versioned files, paying no attention to
31
files which are ignored or unknown. Those can only be present in
32
working trees and can be reported on separately.
34
Yields a sequence of (state, id, old_name, new_name, kind).
35
Each filename and each id is listed only once.
37
## TODO: Allow specifying a list of files to compare, rather than
38
## doing the whole tree? (Not urgent.)
40
## TODO: Allow diffing any two inventories, not just the
41
## current one against one. We mgiht need to specify two
42
## stores to look for the files if diffing two branches. That
43
## might imply this shouldn't be primarily a Branch method.
45
sha_match_cnt = modified_cnt = 0
47
old_it = old_tree.list_files()
48
new_it = new_tree.list_files()
56
old_item = next(old_it)
57
new_item = next(new_it)
59
# We step through the two sorted iterators in parallel, trying to
62
while (old_item != None) or (new_item != None):
63
# OK, we still have some remaining on both, but they may be
66
old_name, old_class, old_kind, old_id = old_item
71
new_name, new_class, new_kind, new_id = new_item
76
# can't handle the old tree being a WorkingTree
77
assert old_class == 'V'
79
if new_item and (new_class != 'V'):
80
yield new_class, None, None, new_name, new_kind
81
new_item = next(new_it)
82
elif (not new_item) or (old_item and (old_name < new_name)):
83
if new_tree.has_id(old_id):
84
# will be mentioned as renamed under new name
87
yield 'D', old_id, old_name, None, old_kind
88
old_item = next(old_it)
89
elif (not old_item) or (new_item and (new_name < old_name)):
90
if old_tree.has_id(new_id):
91
yield 'R', new_id, old_tree.id2path(new_id), new_name, new_kind
93
yield 'A', new_id, None, new_name, new_kind
94
new_item = next(new_it)
95
elif old_id != new_id:
96
assert old_name == new_name
97
# both trees have a file of this name, but it is not the
98
# same file. in other words, the old filename has been
99
# overwritten by either a newly-added or a renamed file.
100
# (should we return something about the overwritten file?)
101
if old_tree.has_id(new_id):
102
# renaming, overlying a deleted file
103
yield 'R', new_id, old_tree.id2path(new_id), new_name, new_kind
105
yield 'A', new_id, None, new_name, new_kind
107
new_item = next(new_it)
108
old_item = next(old_it)
110
assert old_id == new_id
111
assert old_id != None
112
assert old_name == new_name
113
assert old_kind == new_kind
115
if old_kind == 'directory':
116
yield '.', new_id, old_name, new_name, new_kind
117
elif old_tree.get_file_sha1(old_id) == new_tree.get_file_sha1(old_id):
119
yield '.', new_id, old_name, new_name, new_kind
122
yield 'M', new_id, old_name, new_name, new_kind
124
new_item = next(new_it)
125
old_item = next(old_it)
128
mutter("diff finished: %d SHA matches, %d modified"
129
% (sha_match_cnt, modified_cnt))
133
def show_diff(b, revision, file_list):
134
import difflib, sys, types
137
old_tree = b.basis_tree()
139
old_tree = b.revision_tree(b.lookup_revision(revision))
141
new_tree = b.working_tree()
143
# TODO: Options to control putting on a prefix or suffix, perhaps as a format string
147
DEVNULL = '/dev/null'
148
# Windows users, don't panic about this filename -- it is a
149
# special signal to GNU patch that the file should be created or
150
# deleted respectively.
152
# TODO: Generation of pseudo-diffs for added/deleted files could
153
# be usefully made into a much faster special case.
155
# TODO: Better to return them in sorted order I think.
158
file_list = [b.relpath(f) for f in file_list]
160
# FIXME: If given a file list, compare only those files rather
161
# than comparing everything and then throwing stuff away.
163
for file_state, fid, old_name, new_name, kind in diff_trees(old_tree, new_tree):
165
if file_list and (new_name not in file_list):
168
# Don't show this by default; maybe do it if an option is passed
169
# idlabel = ' {%s}' % fid
172
def diffit(oldlines, newlines, **kw):
174
# FIXME: difflib is wrong if there is no trailing newline.
175
# The syntax used by patch seems to be "\ No newline at
176
# end of file" following the last diff line from that
177
# file. This is not trivial to insert into the
178
# unified_diff output and it might be better to just fix
179
# or replace that function.
181
# In the meantime we at least make sure the patch isn't
185
# Special workaround for Python2.3, where difflib fails if
186
# both sequences are empty.
187
if not oldlines and not newlines:
192
if oldlines and (oldlines[-1][-1] != '\n'):
195
if newlines and (newlines[-1][-1] != '\n'):
199
ud = difflib.unified_diff(oldlines, newlines, **kw)
201
# work-around for difflib being too smart for its own good
202
# if /dev/null is "1,0", patch won't recognize it as /dev/null
205
ud[2] = ud[2].replace('-1,0', '-0,0')
208
ud[2] = ud[2].replace('+1,0', '+0,0')
210
sys.stdout.writelines(ud)
212
print "\\ No newline at end of file"
213
sys.stdout.write('\n')
215
if file_state in ['.', '?', 'I']:
217
elif file_state == 'A':
218
print '*** added %s %r' % (kind, new_name)
221
new_tree.get_file(fid).readlines(),
223
tofile=new_label + new_name + idlabel)
224
elif file_state == 'D':
225
assert isinstance(old_name, types.StringTypes)
226
print '*** deleted %s %r' % (kind, old_name)
228
diffit(old_tree.get_file(fid).readlines(), [],
229
fromfile=old_label + old_name + idlabel,
231
elif file_state in ['M', 'R']:
232
if file_state == 'M':
233
assert kind == 'file'
234
assert old_name == new_name
235
print '*** modified %s %r' % (kind, new_name)
236
elif file_state == 'R':
237
print '*** renamed %s %r => %r' % (kind, old_name, new_name)
240
diffit(old_tree.get_file(fid).readlines(),
241
new_tree.get_file(fid).readlines(),
242
fromfile=old_label + old_name + idlabel,
243
tofile=new_label + new_name)
245
raise BzrError("can't represent state %s {%s}" % (file_state, fid))
250
"""Describes changes from one tree to another.
259
(oldpath, newpath, id, text_modified)
265
Each id is listed only once.
267
Files that are both modified and renamed are listed only in
268
renamed, with the text_modified flag true.
270
The lists are normally sorted when the delta is created.
279
def show(self, to_file, show_ids=False, show_unchanged=False):
280
def show_list(files):
281
for path, fid in files:
283
print >>to_file, ' %-30s %s' % (path, fid)
285
print >>to_file, ' ', path
288
print >>to_file, 'removed files:'
289
show_list(self.removed)
292
print >>to_file, 'added files:'
293
show_list(self.added)
296
print >>to_file, 'renamed files:'
297
for oldpath, newpath, fid, text_modified in self.renamed:
299
print >>to_file, ' %s => %s %s' % (oldpath, newpath, fid)
301
print >>to_file, ' %s => %s' % (oldpath, newpath)
304
print >>to_file, 'modified files:'
305
show_list(self.modified)
307
if show_unchanged and self.unchanged:
308
print >>to_file, 'unchanged files:'
309
show_list(self.unchanged)
313
def compare_trees(old_tree, new_tree, want_unchanged):
314
old_inv = old_tree.inventory
315
new_inv = new_tree.inventory
317
for file_id in old_tree:
318
if file_id in new_tree:
319
old_path = old_inv.id2path(file_id)
320
new_path = new_inv.id2path(file_id)
322
kind = old_inv.get_file_kind(file_id)
323
assert kind in ('file', 'directory', 'symlink', 'root_directory'), \
324
'invalid file kind %r' % kind
326
old_sha1 = old_tree.get_file_sha1(file_id)
327
new_sha1 = new_tree.get_file_sha1(file_id)
328
text_modified = (old_sha1 != new_sha1)
330
## mutter("no text to check for %r %r" % (file_id, kind))
331
text_modified = False
333
# TODO: Can possibly avoid calculating path strings if the
334
# two files are unchanged and their names and parents are
335
# the same and the parents are unchanged all the way up.
336
# May not be worthwhile.
338
if old_path != new_path:
339
delta.renamed.append((old_path, new_path, file_id, text_modified))
341
delta.modified.append((new_path, file_id))
343
delta.unchanged.append((new_path, file_id))
345
delta.removed.append((old_inv.id2path(file_id), file_id))
346
for file_id in new_inv:
347
if file_id in old_inv:
349
delta.added.append((new_inv.id2path(file_id), file_id))
354
delta.modified.sort()
added
removed
bzrlib/diff.py
