1
# Copyright (C) 2005, 2006 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
18
from cStringIO import StringIO
21
from tempfile import TemporaryFile
23
from bzrlib.diff import internal_diff, external_diff, show_diff_trees
24
from bzrlib.errors import BinaryFile, NoDiff
25
import bzrlib.osutils as osutils
26
import bzrlib.patiencediff
27
import bzrlib._patiencediff_py
28
from bzrlib.tests import (Feature, TestCase, TestCaseWithTransport,
29
TestCaseInTempDir, TestSkipped)
32
class _CompiledPatienceDiffFeature(Feature):
36
import bzrlib._patiencediff_c
41
def feature_name(self):
42
return 'bzrlib._patiencediff_c'
44
CompiledPatienceDiffFeature = _CompiledPatienceDiffFeature()
47
class _UnicodeFilename(Feature):
48
"""Does the filesystem support Unicode filenames?"""
53
except UnicodeEncodeError:
55
except (IOError, OSError):
56
# The filesystem allows the Unicode filename but the file doesn't
60
# The filesystem allows the Unicode filename and the file exists,
64
UnicodeFilename = _UnicodeFilename()
67
class TestUnicodeFilename(TestCase):
69
def test_probe_passes(self):
70
"""UnicodeFilename._probe passes."""
71
# We can't test much more than that because the behaviour depends
73
UnicodeFilename._probe()
76
def udiff_lines(old, new, allow_binary=False):
78
internal_diff('old', old, 'new', new, output, allow_binary)
80
return output.readlines()
83
def external_udiff_lines(old, new, use_stringio=False):
85
# StringIO has no fileno, so it tests a different codepath
88
output = TemporaryFile()
90
external_diff('old', old, 'new', new, output, diff_opts=['-u'])
92
raise TestSkipped('external "diff" not present to test')
94
lines = output.readlines()
99
class TestDiff(TestCase):
101
def test_add_nl(self):
102
"""diff generates a valid diff for patches that add a newline"""
103
lines = udiff_lines(['boo'], ['boo\n'])
104
self.check_patch(lines)
105
self.assertEquals(lines[4], '\\ No newline at end of file\n')
106
## "expected no-nl, got %r" % lines[4]
108
def test_add_nl_2(self):
109
"""diff generates a valid diff for patches that change last line and
112
lines = udiff_lines(['boo'], ['goo\n'])
113
self.check_patch(lines)
114
self.assertEquals(lines[4], '\\ No newline at end of file\n')
115
## "expected no-nl, got %r" % lines[4]
117
def test_remove_nl(self):
118
"""diff generates a valid diff for patches that change last line and
121
lines = udiff_lines(['boo\n'], ['boo'])
122
self.check_patch(lines)
123
self.assertEquals(lines[5], '\\ No newline at end of file\n')
124
## "expected no-nl, got %r" % lines[5]
126
def check_patch(self, lines):
127
self.assert_(len(lines) > 1)
128
## "Not enough lines for a file header for patch:\n%s" % "".join(lines)
129
self.assert_(lines[0].startswith ('---'))
130
## 'No orig line for patch:\n%s' % "".join(lines)
131
self.assert_(lines[1].startswith ('+++'))
132
## 'No mod line for patch:\n%s' % "".join(lines)
133
self.assert_(len(lines) > 2)
134
## "No hunks for patch:\n%s" % "".join(lines)
135
self.assert_(lines[2].startswith('@@'))
136
## "No hunk header for patch:\n%s" % "".join(lines)
137
self.assert_('@@' in lines[2][2:])
138
## "Unterminated hunk header for patch:\n%s" % "".join(lines)
140
def test_binary_lines(self):
141
self.assertRaises(BinaryFile, udiff_lines, [1023 * 'a' + '\x00'], [])
142
self.assertRaises(BinaryFile, udiff_lines, [], [1023 * 'a' + '\x00'])
143
udiff_lines([1023 * 'a' + '\x00'], [], allow_binary=True)
144
udiff_lines([], [1023 * 'a' + '\x00'], allow_binary=True)
146
def test_external_diff(self):
147
lines = external_udiff_lines(['boo\n'], ['goo\n'])
148
self.check_patch(lines)
149
self.assertEqual('\n', lines[-1])
151
def test_external_diff_no_fileno(self):
152
# Make sure that we can handle not having a fileno, even
153
# if the diff is large
154
lines = external_udiff_lines(['boo\n']*10000,
157
self.check_patch(lines)
159
def test_external_diff_binary_lang_c(self):
161
for lang in ('LANG', 'LC_ALL', 'LANGUAGE'):
162
old_env[lang] = osutils.set_or_unset_env(lang, 'C')
164
lines = external_udiff_lines(['\x00foobar\n'], ['foo\x00bar\n'])
165
# Older versions of diffutils say "Binary files", newer
166
# versions just say "Files".
167
self.assertContainsRe(lines[0],
168
'(Binary f|F)iles old and new differ\n')
169
self.assertEquals(lines[1:], ['\n'])
171
for lang, old_val in old_env.iteritems():
172
osutils.set_or_unset_env(lang, old_val)
174
def test_no_external_diff(self):
175
"""Check that NoDiff is raised when diff is not available"""
176
# Use os.environ['PATH'] to make sure no 'diff' command is available
177
orig_path = os.environ['PATH']
179
os.environ['PATH'] = ''
180
self.assertRaises(NoDiff, external_diff,
181
'old', ['boo\n'], 'new', ['goo\n'],
182
StringIO(), diff_opts=['-u'])
184
os.environ['PATH'] = orig_path
186
def test_internal_diff_default(self):
187
# Default internal diff encoding is utf8
189
internal_diff(u'old_\xb5', ['old_text\n'],
190
u'new_\xe5', ['new_text\n'], output)
191
lines = output.getvalue().splitlines(True)
192
self.check_patch(lines)
193
self.assertEquals(['--- old_\xc2\xb5\n',
194
'+++ new_\xc3\xa5\n',
202
def test_internal_diff_utf8(self):
204
internal_diff(u'old_\xb5', ['old_text\n'],
205
u'new_\xe5', ['new_text\n'], output,
206
path_encoding='utf8')
207
lines = output.getvalue().splitlines(True)
208
self.check_patch(lines)
209
self.assertEquals(['--- old_\xc2\xb5\n',
210
'+++ new_\xc3\xa5\n',
218
def test_internal_diff_iso_8859_1(self):
220
internal_diff(u'old_\xb5', ['old_text\n'],
221
u'new_\xe5', ['new_text\n'], output,
222
path_encoding='iso-8859-1')
223
lines = output.getvalue().splitlines(True)
224
self.check_patch(lines)
225
self.assertEquals(['--- old_\xb5\n',
234
def test_internal_diff_returns_bytes(self):
236
output = StringIO.StringIO()
237
internal_diff(u'old_\xb5', ['old_text\n'],
238
u'new_\xe5', ['new_text\n'], output)
239
self.failUnless(isinstance(output.getvalue(), str),
240
'internal_diff should return bytestrings')
243
class TestDiffFiles(TestCaseInTempDir):
245
def test_external_diff_binary(self):
246
"""The output when using external diff should use diff's i18n error"""
247
# Make sure external_diff doesn't fail in the current LANG
248
lines = external_udiff_lines(['\x00foobar\n'], ['foo\x00bar\n'])
250
cmd = ['diff', '-u', '--binary', 'old', 'new']
251
open('old', 'wb').write('\x00foobar\n')
252
open('new', 'wb').write('foo\x00bar\n')
253
pipe = subprocess.Popen(cmd, stdout=subprocess.PIPE,
254
stdin=subprocess.PIPE)
255
out, err = pipe.communicate()
256
# Diff returns '2' on Binary files.
257
self.assertEqual(2, pipe.returncode)
258
# We should output whatever diff tells us, plus a trailing newline
259
self.assertEqual(out.splitlines(True) + ['\n'], lines)
262
class TestShowDiffTreesHelper(TestCaseWithTransport):
263
"""Has a helper for running show_diff_trees"""
265
def get_diff(self, tree1, tree2, specific_files=None, working_tree=None):
267
if working_tree is not None:
268
extra_trees = (working_tree,)
271
show_diff_trees(tree1, tree2, output, specific_files=specific_files,
272
extra_trees=extra_trees, old_label='old/',
274
return output.getvalue()
277
class TestDiffDates(TestShowDiffTreesHelper):
280
super(TestDiffDates, self).setUp()
281
self.wt = self.make_branch_and_tree('.')
282
self.b = self.wt.branch
283
self.build_tree_contents([
284
('file1', 'file1 contents at rev 1\n'),
285
('file2', 'file2 contents at rev 1\n')
287
self.wt.add(['file1', 'file2'])
289
message='Revision 1',
290
timestamp=1143849600, # 2006-04-01 00:00:00 UTC
293
self.build_tree_contents([('file1', 'file1 contents at rev 2\n')])
295
message='Revision 2',
296
timestamp=1143936000, # 2006-04-02 00:00:00 UTC
299
self.build_tree_contents([('file2', 'file2 contents at rev 3\n')])
301
message='Revision 3',
302
timestamp=1144022400, # 2006-04-03 00:00:00 UTC
305
self.wt.remove(['file2'])
307
message='Revision 4',
308
timestamp=1144108800, # 2006-04-04 00:00:00 UTC
311
self.build_tree_contents([
312
('file1', 'file1 contents in working tree\n')
314
# set the date stamps for files in the working tree to known values
315
os.utime('file1', (1144195200, 1144195200)) # 2006-04-05 00:00:00 UTC
317
def test_diff_rev_tree_working_tree(self):
318
output = self.get_diff(self.wt.basis_tree(), self.wt)
319
# note that the date for old/file1 is from rev 2 rather than from
320
# the basis revision (rev 4)
321
self.assertEqualDiff(output, '''\
322
=== modified file 'file1'
323
--- old/file1\t2006-04-02 00:00:00 +0000
324
+++ new/file1\t2006-04-05 00:00:00 +0000
326
-file1 contents at rev 2
327
+file1 contents in working tree
331
def test_diff_rev_tree_rev_tree(self):
332
tree1 = self.b.repository.revision_tree('rev-2')
333
tree2 = self.b.repository.revision_tree('rev-3')
334
output = self.get_diff(tree1, tree2)
335
self.assertEqualDiff(output, '''\
336
=== modified file 'file2'
337
--- old/file2\t2006-04-01 00:00:00 +0000
338
+++ new/file2\t2006-04-03 00:00:00 +0000
340
-file2 contents at rev 1
341
+file2 contents at rev 3
345
def test_diff_add_files(self):
346
tree1 = self.b.repository.revision_tree(None)
347
tree2 = self.b.repository.revision_tree('rev-1')
348
output = self.get_diff(tree1, tree2)
349
# the files have the epoch time stamp for the tree in which
351
self.assertEqualDiff(output, '''\
352
=== added file 'file1'
353
--- old/file1\t1970-01-01 00:00:00 +0000
354
+++ new/file1\t2006-04-01 00:00:00 +0000
356
+file1 contents at rev 1
358
=== added file 'file2'
359
--- old/file2\t1970-01-01 00:00:00 +0000
360
+++ new/file2\t2006-04-01 00:00:00 +0000
362
+file2 contents at rev 1
366
def test_diff_remove_files(self):
367
tree1 = self.b.repository.revision_tree('rev-3')
368
tree2 = self.b.repository.revision_tree('rev-4')
369
output = self.get_diff(tree1, tree2)
370
# the file has the epoch time stamp for the tree in which
372
self.assertEqualDiff(output, '''\
373
=== removed file 'file2'
374
--- old/file2\t2006-04-03 00:00:00 +0000
375
+++ new/file2\t1970-01-01 00:00:00 +0000
377
-file2 contents at rev 3
381
def test_show_diff_specified(self):
382
"""A working tree filename can be used to identify a file"""
383
self.wt.rename_one('file1', 'file1b')
384
old_tree = self.b.repository.revision_tree('rev-1')
385
new_tree = self.b.repository.revision_tree('rev-4')
386
out = self.get_diff(old_tree, new_tree, specific_files=['file1b'],
387
working_tree=self.wt)
388
self.assertContainsRe(out, 'file1\t')
390
def test_recursive_diff(self):
391
"""Children of directories are matched"""
394
self.wt.add(['dir1', 'dir2'])
395
self.wt.rename_one('file1', 'dir1/file1')
396
old_tree = self.b.repository.revision_tree('rev-1')
397
new_tree = self.b.repository.revision_tree('rev-4')
398
out = self.get_diff(old_tree, new_tree, specific_files=['dir1'],
399
working_tree=self.wt)
400
self.assertContainsRe(out, 'file1\t')
401
out = self.get_diff(old_tree, new_tree, specific_files=['dir2'],
402
working_tree=self.wt)
403
self.assertNotContainsRe(out, 'file1\t')
407
class TestShowDiffTrees(TestShowDiffTreesHelper):
408
"""Direct tests for show_diff_trees"""
410
def test_modified_file(self):
411
"""Test when a file is modified."""
412
tree = self.make_branch_and_tree('tree')
413
self.build_tree_contents([('tree/file', 'contents\n')])
414
tree.add(['file'], ['file-id'])
415
tree.commit('one', rev_id='rev-1')
417
self.build_tree_contents([('tree/file', 'new contents\n')])
418
diff = self.get_diff(tree.basis_tree(), tree)
419
self.assertContainsRe(diff, "=== modified file 'file'\n")
420
self.assertContainsRe(diff, '--- old/file\t')
421
self.assertContainsRe(diff, '\\+\\+\\+ new/file\t')
422
self.assertContainsRe(diff, '-contents\n'
425
def test_modified_file_in_renamed_dir(self):
426
"""Test when a file is modified in a renamed directory."""
427
tree = self.make_branch_and_tree('tree')
428
self.build_tree(['tree/dir/'])
429
self.build_tree_contents([('tree/dir/file', 'contents\n')])
430
tree.add(['dir', 'dir/file'], ['dir-id', 'file-id'])
431
tree.commit('one', rev_id='rev-1')
433
tree.rename_one('dir', 'other')
434
self.build_tree_contents([('tree/other/file', 'new contents\n')])
435
diff = self.get_diff(tree.basis_tree(), tree)
436
self.assertContainsRe(diff, "=== renamed directory 'dir' => 'other'\n")
437
self.assertContainsRe(diff, "=== modified file 'other/file'\n")
438
# XXX: This is technically incorrect, because it used to be at another
439
# location. What to do?
440
self.assertContainsRe(diff, '--- old/dir/file\t')
441
self.assertContainsRe(diff, '\\+\\+\\+ new/other/file\t')
442
self.assertContainsRe(diff, '-contents\n'
445
def test_renamed_directory(self):
446
"""Test when only a directory is only renamed."""
447
tree = self.make_branch_and_tree('tree')
448
self.build_tree(['tree/dir/'])
449
self.build_tree_contents([('tree/dir/file', 'contents\n')])
450
tree.add(['dir', 'dir/file'], ['dir-id', 'file-id'])
451
tree.commit('one', rev_id='rev-1')
453
tree.rename_one('dir', 'newdir')
454
diff = self.get_diff(tree.basis_tree(), tree)
455
# Renaming a directory should be a single "you renamed this dir" even
456
# when there are files inside.
457
self.assertEqual("=== renamed directory 'dir' => 'newdir'\n", diff)
459
def test_renamed_file(self):
460
"""Test when a file is only renamed."""
461
tree = self.make_branch_and_tree('tree')
462
self.build_tree_contents([('tree/file', 'contents\n')])
463
tree.add(['file'], ['file-id'])
464
tree.commit('one', rev_id='rev-1')
466
tree.rename_one('file', 'newname')
467
diff = self.get_diff(tree.basis_tree(), tree)
468
self.assertContainsRe(diff, "=== renamed file 'file' => 'newname'\n")
469
# We shouldn't have a --- or +++ line, because there is no content
471
self.assertNotContainsRe(diff, '---')
473
def test_renamed_and_modified_file(self):
474
"""Test when a file is only renamed."""
475
tree = self.make_branch_and_tree('tree')
476
self.build_tree_contents([('tree/file', 'contents\n')])
477
tree.add(['file'], ['file-id'])
478
tree.commit('one', rev_id='rev-1')
480
tree.rename_one('file', 'newname')
481
self.build_tree_contents([('tree/newname', 'new contents\n')])
482
diff = self.get_diff(tree.basis_tree(), tree)
483
self.assertContainsRe(diff, "=== renamed file 'file' => 'newname'\n")
484
self.assertContainsRe(diff, '--- old/file\t')
485
self.assertContainsRe(diff, '\\+\\+\\+ new/newname\t')
486
self.assertContainsRe(diff, '-contents\n'
489
def test_binary_unicode_filenames(self):
490
"""Test that contents of files are *not* encoded in UTF-8 when there
491
is a binary file in the diff.
493
# See https://bugs.launchpad.net/bugs/110092.
494
self.requireFeature(UnicodeFilename)
496
# This bug isn't triggered with cStringIO.
497
from StringIO import StringIO
498
tree = self.make_branch_and_tree('tree')
499
alpha, omega = u'\u03b1', u'\u03c9'
500
alpha_utf8, omega_utf8 = alpha.encode('utf8'), omega.encode('utf8')
501
self.build_tree_contents(
502
[('tree/' + alpha, chr(0)),
504
('The %s and the %s\n' % (alpha_utf8, omega_utf8)))])
505
tree.add([alpha], ['file-id'])
506
tree.add([omega], ['file-id-2'])
507
diff_content = StringIO()
508
show_diff_trees(tree.basis_tree(), tree, diff_content)
509
diff = diff_content.getvalue()
510
self.assertContainsRe(diff, r"=== added file '%s'" % alpha_utf8)
511
self.assertContainsRe(
512
diff, "Binary files a/%s.*and b/%s.* differ\n" % (alpha_utf8, alpha_utf8))
513
self.assertContainsRe(diff, r"=== added file '%s'" % omega_utf8)
514
self.assertContainsRe(diff, r"--- a/%s" % (omega_utf8,))
515
self.assertContainsRe(diff, r"\+\+\+ b/%s" % (omega_utf8,))
517
def test_unicode_filename(self):
518
"""Test when the filename are unicode."""
519
self.requireFeature(UnicodeFilename)
521
alpha, omega = u'\u03b1', u'\u03c9'
522
autf8, outf8 = alpha.encode('utf8'), omega.encode('utf8')
524
tree = self.make_branch_and_tree('tree')
525
self.build_tree_contents([('tree/ren_'+alpha, 'contents\n')])
526
tree.add(['ren_'+alpha], ['file-id-2'])
527
self.build_tree_contents([('tree/del_'+alpha, 'contents\n')])
528
tree.add(['del_'+alpha], ['file-id-3'])
529
self.build_tree_contents([('tree/mod_'+alpha, 'contents\n')])
530
tree.add(['mod_'+alpha], ['file-id-4'])
532
tree.commit('one', rev_id='rev-1')
534
tree.rename_one('ren_'+alpha, 'ren_'+omega)
535
tree.remove('del_'+alpha)
536
self.build_tree_contents([('tree/add_'+alpha, 'contents\n')])
537
tree.add(['add_'+alpha], ['file-id'])
538
self.build_tree_contents([('tree/mod_'+alpha, 'contents_mod\n')])
540
diff = self.get_diff(tree.basis_tree(), tree)
541
self.assertContainsRe(diff,
542
"=== renamed file 'ren_%s' => 'ren_%s'\n"%(autf8, outf8))
543
self.assertContainsRe(diff, "=== added file 'add_%s'"%autf8)
544
self.assertContainsRe(diff, "=== modified file 'mod_%s'"%autf8)
545
self.assertContainsRe(diff, "=== removed file 'del_%s'"%autf8)
547
class TestPatienceDiffLib(TestCase):
550
super(TestPatienceDiffLib, self).setUp()
551
self._unique_lcs = bzrlib._patiencediff_py.unique_lcs_py
552
self._recurse_matches = bzrlib._patiencediff_py.recurse_matches_py
553
self._PatienceSequenceMatcher = \
554
bzrlib._patiencediff_py.PatienceSequenceMatcher_py
556
def test_unique_lcs(self):
557
unique_lcs = self._unique_lcs
558
self.assertEquals(unique_lcs('', ''), [])
559
self.assertEquals(unique_lcs('', 'a'), [])
560
self.assertEquals(unique_lcs('a', ''), [])
561
self.assertEquals(unique_lcs('a', 'a'), [(0,0)])
562
self.assertEquals(unique_lcs('a', 'b'), [])
563
self.assertEquals(unique_lcs('ab', 'ab'), [(0,0), (1,1)])
564
self.assertEquals(unique_lcs('abcde', 'cdeab'), [(2,0), (3,1), (4,2)])
565
self.assertEquals(unique_lcs('cdeab', 'abcde'), [(0,2), (1,3), (2,4)])
566
self.assertEquals(unique_lcs('abXde', 'abYde'), [(0,0), (1,1),
568
self.assertEquals(unique_lcs('acbac', 'abc'), [(2,1)])
570
def test_recurse_matches(self):
571
def test_one(a, b, matches):
573
self._recurse_matches(
574
a, b, 0, 0, len(a), len(b), test_matches, 10)
575
self.assertEquals(test_matches, matches)
577
test_one(['a', '', 'b', '', 'c'], ['a', 'a', 'b', 'c', 'c'],
578
[(0, 0), (2, 2), (4, 4)])
579
test_one(['a', 'c', 'b', 'a', 'c'], ['a', 'b', 'c'],
580
[(0, 0), (2, 1), (4, 2)])
581
# Even though 'bc' is not unique globally, and is surrounded by
582
# non-matching lines, we should still match, because they are locally
584
test_one('abcdbce', 'afbcgdbce', [(0,0), (1, 2), (2, 3), (3, 5),
585
(4, 6), (5, 7), (6, 8)])
587
# recurse_matches doesn't match non-unique
588
# lines surrounded by bogus text.
589
# The update has been done in patiencediff.SequenceMatcher instead
591
# This is what it could be
592
#test_one('aBccDe', 'abccde', [(0,0), (2,2), (3,3), (5,5)])
594
# This is what it currently gives:
595
test_one('aBccDe', 'abccde', [(0,0), (5,5)])
597
def test_matching_blocks(self):
598
def chk_blocks(a, b, expected_blocks):
599
# difflib always adds a signature of the total
600
# length, with no matching entries at the end
601
s = self._PatienceSequenceMatcher(None, a, b)
602
blocks = s.get_matching_blocks()
603
self.assertEquals((len(a), len(b), 0), blocks[-1])
604
self.assertEquals(expected_blocks, blocks[:-1])
606
# Some basic matching tests
607
chk_blocks('', '', [])
608
chk_blocks([], [], [])
609
chk_blocks('abc', '', [])
610
chk_blocks('', 'abc', [])
611
chk_blocks('abcd', 'abcd', [(0, 0, 4)])
612
chk_blocks('abcd', 'abce', [(0, 0, 3)])
613
chk_blocks('eabc', 'abce', [(1, 0, 3)])
614
chk_blocks('eabce', 'abce', [(1, 0, 4)])
615
chk_blocks('abcde', 'abXde', [(0, 0, 2), (3, 3, 2)])
616
chk_blocks('abcde', 'abXYZde', [(0, 0, 2), (3, 5, 2)])
617
chk_blocks('abde', 'abXYZde', [(0, 0, 2), (2, 5, 2)])
618
# This may check too much, but it checks to see that
619
# a copied block stays attached to the previous section,
621
# difflib would tend to grab the trailing longest match
622
# which would make the diff not look right
623
chk_blocks('abcdefghijklmnop', 'abcdefxydefghijklmnop',
624
[(0, 0, 6), (6, 11, 10)])
626
# make sure it supports passing in lists
630
'how are you today?\n'],
632
'how are you today?\n'],
633
[(0, 0, 1), (2, 1, 1)])
635
# non unique lines surrounded by non-matching lines
637
chk_blocks('aBccDe', 'abccde', [(0,0,1), (5,5,1)])
639
# But they only need to be locally unique
640
chk_blocks('aBcDec', 'abcdec', [(0,0,1), (2,2,1), (4,4,2)])
642
# non unique blocks won't be matched
643
chk_blocks('aBcdEcdFg', 'abcdecdfg', [(0,0,1), (8,8,1)])
645
# but locally unique ones will
646
chk_blocks('aBcdEeXcdFg', 'abcdecdfg', [(0,0,1), (2,2,2),
647
(5,4,1), (7,5,2), (10,8,1)])
649
chk_blocks('abbabbXd', 'cabbabxd', [(7,7,1)])
650
chk_blocks('abbabbbb', 'cabbabbc', [])
651
chk_blocks('bbbbbbbb', 'cbbbbbbc', [])
653
def test_opcodes(self):
654
def chk_ops(a, b, expected_codes):
655
s = self._PatienceSequenceMatcher(None, a, b)
656
self.assertEquals(expected_codes, s.get_opcodes())
660
chk_ops('abc', '', [('delete', 0,3, 0,0)])
661
chk_ops('', 'abc', [('insert', 0,0, 0,3)])
662
chk_ops('abcd', 'abcd', [('equal', 0,4, 0,4)])
663
chk_ops('abcd', 'abce', [('equal', 0,3, 0,3),
664
('replace', 3,4, 3,4)
666
chk_ops('eabc', 'abce', [('delete', 0,1, 0,0),
670
chk_ops('eabce', 'abce', [('delete', 0,1, 0,0),
673
chk_ops('abcde', 'abXde', [('equal', 0,2, 0,2),
674
('replace', 2,3, 2,3),
677
chk_ops('abcde', 'abXYZde', [('equal', 0,2, 0,2),
678
('replace', 2,3, 2,5),
681
chk_ops('abde', 'abXYZde', [('equal', 0,2, 0,2),
682
('insert', 2,2, 2,5),
685
chk_ops('abcdefghijklmnop', 'abcdefxydefghijklmnop',
686
[('equal', 0,6, 0,6),
687
('insert', 6,6, 6,11),
688
('equal', 6,16, 11,21)
693
, 'how are you today?\n'],
695
, 'how are you today?\n'],
696
[('equal', 0,1, 0,1),
697
('delete', 1,2, 1,1),
700
chk_ops('aBccDe', 'abccde',
701
[('equal', 0,1, 0,1),
702
('replace', 1,5, 1,5),
705
chk_ops('aBcDec', 'abcdec',
706
[('equal', 0,1, 0,1),
707
('replace', 1,2, 1,2),
709
('replace', 3,4, 3,4),
712
chk_ops('aBcdEcdFg', 'abcdecdfg',
713
[('equal', 0,1, 0,1),
714
('replace', 1,8, 1,8),
717
chk_ops('aBcdEeXcdFg', 'abcdecdfg',
718
[('equal', 0,1, 0,1),
719
('replace', 1,2, 1,2),
721
('delete', 4,5, 4,4),
723
('delete', 6,7, 5,5),
725
('replace', 9,10, 7,8),
726
('equal', 10,11, 8,9)
729
def test_grouped_opcodes(self):
730
def chk_ops(a, b, expected_codes, n=3):
731
s = self._PatienceSequenceMatcher(None, a, b)
732
self.assertEquals(expected_codes, list(s.get_grouped_opcodes(n)))
736
chk_ops('abc', '', [[('delete', 0,3, 0,0)]])
737
chk_ops('', 'abc', [[('insert', 0,0, 0,3)]])
738
chk_ops('abcd', 'abcd', [])
739
chk_ops('abcd', 'abce', [[('equal', 0,3, 0,3),
740
('replace', 3,4, 3,4)
742
chk_ops('eabc', 'abce', [[('delete', 0,1, 0,0),
746
chk_ops('abcdefghijklmnop', 'abcdefxydefghijklmnop',
747
[[('equal', 3,6, 3,6),
748
('insert', 6,6, 6,11),
749
('equal', 6,9, 11,14)
751
chk_ops('abcdefghijklmnop', 'abcdefxydefghijklmnop',
752
[[('equal', 2,6, 2,6),
753
('insert', 6,6, 6,11),
754
('equal', 6,10, 11,15)
756
chk_ops('Xabcdef', 'abcdef',
757
[[('delete', 0,1, 0,0),
760
chk_ops('abcdef', 'abcdefX',
761
[[('equal', 3,6, 3,6),
766
def test_multiple_ranges(self):
767
# There was an earlier bug where we used a bad set of ranges,
768
# this triggers that specific bug, to make sure it doesn't regress
769
def chk_blocks(a, b, expected_blocks):
770
# difflib always adds a signature of the total
771
# length, with no matching entries at the end
772
s = self._PatienceSequenceMatcher(None, a, b)
773
blocks = s.get_matching_blocks()
775
self.assertEquals(x, (len(a), len(b), 0))
776
self.assertEquals(expected_blocks, blocks)
778
chk_blocks('abcdefghijklmnop'
779
, 'abcXghiYZQRSTUVWXYZijklmnop'
780
, [(0, 0, 3), (6, 4, 3), (9, 20, 7)])
782
chk_blocks('ABCd efghIjk L'
783
, 'AxyzBCn mo pqrstuvwI1 2 L'
784
, [(0,0,1), (1, 4, 2), (9, 19, 1), (12, 23, 3)])
786
# These are rot13 code snippets.
788
trg nqqrq jura lbh nqq n svyr va gur qverpgbel.
790
gnxrf_netf = ['svyr*']
791
gnxrf_bcgvbaf = ['ab-erphefr']
793
qrs eha(frys, svyr_yvfg, ab_erphefr=Snyfr):
794
sebz omeyvo.nqq vzcbeg fzneg_nqq, nqq_ercbegre_cevag, nqq_ercbegre_ahyy
796
ercbegre = nqq_ercbegre_ahyy
798
ercbegre = nqq_ercbegre_cevag
799
fzneg_nqq(svyr_yvfg, abg ab_erphefr, ercbegre)
802
pynff pzq_zxqve(Pbzznaq):
803
'''.splitlines(True), '''\
804
trg nqqrq jura lbh nqq n svyr va gur qverpgbel.
806
--qel-eha jvyy fubj juvpu svyrf jbhyq or nqqrq, ohg abg npghnyyl
809
gnxrf_netf = ['svyr*']
810
gnxrf_bcgvbaf = ['ab-erphefr', 'qel-eha']
812
qrs eha(frys, svyr_yvfg, ab_erphefr=Snyfr, qel_eha=Snyfr):
817
# Guvf vf cbvagyrff, ohg V'q engure abg envfr na reebe
818
npgvba = omeyvo.nqq.nqq_npgvba_ahyy
820
npgvba = omeyvo.nqq.nqq_npgvba_cevag
822
npgvba = omeyvo.nqq.nqq_npgvba_nqq
824
npgvba = omeyvo.nqq.nqq_npgvba_nqq_naq_cevag
826
omeyvo.nqq.fzneg_nqq(svyr_yvfg, abg ab_erphefr, npgvba)
829
pynff pzq_zxqve(Pbzznaq):
831
, [(0,0,1), (1, 4, 2), (9, 19, 1), (12, 23, 3)])
833
def test_patience_unified_diff(self):
834
txt_a = ['hello there\n',
836
'how are you today?\n']
837
txt_b = ['hello there\n',
838
'how are you today?\n']
839
unified_diff = bzrlib.patiencediff.unified_diff
840
psm = self._PatienceSequenceMatcher
841
self.assertEquals([ '--- \n',
846
' how are you today?\n'
848
, list(unified_diff(txt_a, txt_b,
849
sequencematcher=psm)))
850
txt_a = map(lambda x: x+'\n', 'abcdefghijklmnop')
851
txt_b = map(lambda x: x+'\n', 'abcdefxydefghijklmnop')
852
# This is the result with LongestCommonSubstring matching
853
self.assertEquals(['--- \n',
855
'@@ -1,6 +1,11 @@\n',
867
, list(unified_diff(txt_a, txt_b)))
868
# And the patience diff
869
self.assertEquals(['--- \n',
871
'@@ -4,6 +4,11 @@\n',
884
, list(unified_diff(txt_a, txt_b,
885
sequencematcher=psm)))
888
class TestPatienceDiffLib_c(TestPatienceDiffLib):
890
_test_needs_features = [CompiledPatienceDiffFeature]
893
super(TestPatienceDiffLib_c, self).setUp()
894
import bzrlib._patiencediff_c
895
self._unique_lcs = bzrlib._patiencediff_c.unique_lcs_c
896
self._recurse_matches = bzrlib._patiencediff_c.recurse_matches_c
897
self._PatienceSequenceMatcher = \
898
bzrlib._patiencediff_c.PatienceSequenceMatcher_c
901
class TestPatienceDiffLibFiles(TestCaseInTempDir):
904
super(TestPatienceDiffLibFiles, self).setUp()
905
self._PatienceSequenceMatcher = \
906
bzrlib._patiencediff_py.PatienceSequenceMatcher_py
908
def test_patience_unified_diff_files(self):
909
txt_a = ['hello there\n',
911
'how are you today?\n']
912
txt_b = ['hello there\n',
913
'how are you today?\n']
914
open('a1', 'wb').writelines(txt_a)
915
open('b1', 'wb').writelines(txt_b)
917
unified_diff_files = bzrlib.patiencediff.unified_diff_files
918
psm = self._PatienceSequenceMatcher
919
self.assertEquals(['--- a1 \n',
924
' how are you today?\n',
926
, list(unified_diff_files('a1', 'b1',
927
sequencematcher=psm)))
929
txt_a = map(lambda x: x+'\n', 'abcdefghijklmnop')
930
txt_b = map(lambda x: x+'\n', 'abcdefxydefghijklmnop')
931
open('a2', 'wb').writelines(txt_a)
932
open('b2', 'wb').writelines(txt_b)
934
# This is the result with LongestCommonSubstring matching
935
self.assertEquals(['--- a2 \n',
937
'@@ -1,6 +1,11 @@\n',
949
, list(unified_diff_files('a2', 'b2')))
951
# And the patience diff
952
self.assertEquals(['--- a2 \n',
954
'@@ -4,6 +4,11 @@\n',
967
, list(unified_diff_files('a2', 'b2',
968
sequencematcher=psm)))
971
class TestPatienceDiffLibFiles_c(TestPatienceDiffLibFiles):
973
_test_needs_features = [CompiledPatienceDiffFeature]
976
super(TestPatienceDiffLibFiles_c, self).setUp()
977
import bzrlib._patiencediff_c
978
self._PatienceSequenceMatcher = \
979
bzrlib._patiencediff_c.PatienceSequenceMatcher_c
982
class TestUsingCompiledIfAvailable(TestCase):
984
def test_PatienceSequenceMatcher(self):
985
if CompiledPatienceDiffFeature.available():
986
from bzrlib._patiencediff_c import PatienceSequenceMatcher_c
987
self.assertIs(PatienceSequenceMatcher_c,
988
bzrlib.patiencediff.PatienceSequenceMatcher)
990
from bzrlib._patiencediff_py import PatienceSequenceMatcher_py
991
self.assertIs(PatienceSequenceMatcher_py,
992
bzrlib.patiencediff.PatienceSequenceMatcher)
994
def test_unique_lcs(self):
995
if CompiledPatienceDiffFeature.available():
996
from bzrlib._patiencediff_c import unique_lcs_c
997
self.assertIs(unique_lcs_c,
998
bzrlib.patiencediff.unique_lcs)
1000
from bzrlib._patiencediff_py import unique_lcs_py
1001
self.assertIs(unique_lcs_py,
1002
bzrlib.patiencediff.unique_lcs)
1004
def test_recurse_matches(self):
1005
if CompiledPatienceDiffFeature.available():
1006
from bzrlib._patiencediff_c import recurse_matches_c
1007
self.assertIs(recurse_matches_c,
1008
bzrlib.patiencediff.recurse_matches)
1010
from bzrlib._patiencediff_py import recurse_matches_py
1011
self.assertIs(recurse_matches_py,
1012
bzrlib.patiencediff.recurse_matches)
added
removed
bzrlib/tests/test_diff.py
