1
# Copyright (C) 2004, 2005 by Canonical Ltd
1
# Copyright (C) 2005-2010 Canonical Ltd
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
5
5
# the Free Software Foundation; either version 2 of the License, or
6
6
# (at your option) any later version.
8
8
# This program is distributed in the hope that it will be useful,
9
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
11
# GNU General Public License for more details.
13
13
# You should have received a copy of the GNU General Public License
14
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
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
18
18
# mbp: "you know that thing where cvs gives you conflict markers?"
19
19
# s: "i hate that."
23
28
def intersect(ra, rb):
63
67
Given BASE, OTHER, THIS, tries to produce a combined text
64
68
incorporating the changes from both BASE->OTHER and BASE->THIS.
65
69
All three will typically be sequences of lines."""
66
def __init__(self, base, a, b):
71
def __init__(self, base, a, b, is_cherrypick=False, allow_objects=False):
74
:param base: lines in BASE
77
:param is_cherrypick: flag indicating if this merge is a cherrypick.
78
When cherrypicking b => a, matches with b and base do not conflict.
79
:param allow_objects: if True, do not require that base, a and b are
80
plain Python strs. Also prevents BinaryFile from being raised.
81
Lines can be any sequence of comparable and hashable Python
85
textfile.check_text_lines(base)
86
textfile.check_text_lines(a)
87
textfile.check_text_lines(b)
70
from difflib import SequenceMatcher
71
self.a_ops = SequenceMatcher(None, base, a).get_opcodes()
72
self.b_ops = SequenceMatcher(None, base, b).get_opcodes()
91
self.is_cherrypick = is_cherrypick
76
93
def merge_lines(self,
79
start_marker='<<<<<<<<',
80
mid_marker='========',
81
end_marker='>>>>>>>>',
97
start_marker='<<<<<<<',
83
102
"""Return merge in cvs-like form.
106
if self.a[0].endswith('\r\n'):
108
elif self.a[0].endswith('\r'):
110
if base_marker and reprocess:
111
raise errors.CantReprocessAndShowBase()
86
113
start_marker = start_marker + ' ' + name_a
88
115
end_marker = end_marker + ' ' + name_b
90
for t in self.merge_regions():
116
if name_base and base_marker:
117
base_marker = base_marker + ' ' + name_base
118
merge_regions = self.merge_regions()
119
if reprocess is True:
120
merge_regions = self.reprocess_merge_regions(merge_regions)
121
for t in merge_regions:
92
123
if what == 'unchanged':
93
124
for i in range(t[1], t[2]):
99
130
for i in range(t[1], t[2]):
101
132
elif what == 'conflict':
102
yield start_marker + '\n'
133
yield start_marker + newline
103
134
for i in range(t[3], t[4]):
105
yield mid_marker + '\n'
136
if base_marker is not None:
137
yield base_marker + newline
138
for i in range(t[1], t[2]):
140
yield mid_marker + newline
106
141
for i in range(t[5], t[6]):
108
yield end_marker + '\n'
143
yield end_marker + newline
110
145
raise ValueError(what)
116
147
def merge_annotated(self):
117
148
"""Return merge with conflicts, showing origin of lines.
119
Most useful for debugging merge.
150
Most useful for debugging merge.
121
152
for t in self.merge_regions():
209
235
# section a[0:ia] has been disposed of, etc
212
238
for zmatch, zend, amatch, aend, bmatch, bend in self.find_sync_regions():
213
#print 'match base [%d:%d]' % (zmatch, zend)
215
239
matchlen = zend - zmatch
217
assert matchlen == (aend - amatch)
218
assert matchlen == (bend - bmatch)
242
# matchlen == (aend - amatch)
243
# matchlen == (bend - bmatch)
220
244
len_a = amatch - ia
221
245
len_b = bmatch - ib
222
246
len_base = zmatch - iz
250
# assert len_base >= 0
227
252
#print 'unmatched a=%d, b=%d' % (len_a, len_b)
229
254
if len_a or len_b:
230
255
# try to avoid actually slicing the lists
231
equal_a = compare_range(self.a, ia, amatch,
232
self.base, iz, zmatch)
233
equal_b = compare_range(self.b, ib, bmatch,
234
self.base, iz, zmatch)
235
256
same = compare_range(self.a, ia, amatch,
236
257
self.b, ib, bmatch)
239
260
yield 'same', ia, amatch
240
elif equal_a and not equal_b:
241
yield 'b', ib, bmatch
242
elif equal_b and not equal_a:
243
yield 'a', ia, amatch
244
elif not equal_a and not equal_b:
245
yield 'conflict', iz, zmatch, ia, amatch, ib, bmatch
247
raise AssertionError("can't handle a=b=base but unmatched")
262
equal_a = compare_range(self.a, ia, amatch,
263
self.base, iz, zmatch)
264
equal_b = compare_range(self.b, ib, bmatch,
265
self.base, iz, zmatch)
266
if equal_a and not equal_b:
267
yield 'b', ib, bmatch
268
elif equal_b and not equal_a:
269
yield 'a', ia, amatch
270
elif not equal_a and not equal_b:
271
if self.is_cherrypick:
272
for node in self._refine_cherrypick_conflict(
273
iz, zmatch, ia, amatch,
277
yield 'conflict', iz, zmatch, ia, amatch, ib, bmatch
279
raise AssertionError("can't handle a=b=base but unmatched")
253
285
# if the same part of the base was deleted on both sides
254
286
# that's OK, we can just skip it.
290
# assert ia == amatch
291
# assert ib == bmatch
292
# assert iz == zmatch
262
294
yield 'unchanged', zmatch, zend
299
def _refine_cherrypick_conflict(self, zstart, zend, astart, aend, bstart, bend):
300
"""When cherrypicking b => a, ignore matches with b and base."""
301
# Do not emit regions which match, only regions which do not match
302
matches = patiencediff.PatienceSequenceMatcher(None,
303
self.base[zstart:zend], self.b[bstart:bend]).get_matching_blocks()
308
for base_idx, b_idx, match_len in matches:
309
conflict_z_len = base_idx - last_base_idx
310
conflict_b_len = b_idx - last_b_idx
311
if conflict_b_len == 0: # There are no lines in b which conflict,
317
zstart + last_base_idx, zstart + base_idx,
318
aend, aend, bstart + last_b_idx, bstart + b_idx)
320
# The first conflict gets the a-range
322
yield ('conflict', zstart + last_base_idx, zstart +
324
astart, aend, bstart + last_b_idx, bstart + b_idx)
325
last_base_idx = base_idx + match_len
326
last_b_idx = b_idx + match_len
327
if last_base_idx != zend - zstart or last_b_idx != bend - bstart:
329
yield ('conflict', zstart + last_base_idx, zstart + base_idx,
330
aend, aend, bstart + last_b_idx, bstart + b_idx)
332
# The first conflict gets the a-range
334
yield ('conflict', zstart + last_base_idx, zstart + base_idx,
335
astart, aend, bstart + last_b_idx, bstart + b_idx)
337
yield ('conflict', zstart, zend, astart, aend, bstart, bend)
339
def reprocess_merge_regions(self, merge_regions):
340
"""Where there are conflict regions, remove the agreed lines.
342
Lines where both A and B have made the same changes are
345
for region in merge_regions:
346
if region[0] != "conflict":
349
type, iz, zmatch, ia, amatch, ib, bmatch = region
350
a_region = self.a[ia:amatch]
351
b_region = self.b[ib:bmatch]
352
matches = patiencediff.PatienceSequenceMatcher(
353
None, a_region, b_region).get_matching_blocks()
356
for region_ia, region_ib, region_len in matches[:-1]:
359
reg = self.mismatch_region(next_a, region_ia, next_b,
363
yield 'same', region_ia, region_len+region_ia
364
next_a = region_ia + region_len
365
next_b = region_ib + region_len
366
reg = self.mismatch_region(next_a, amatch, next_b, bmatch)
371
def mismatch_region(next_a, region_ia, next_b, region_ib):
372
if next_a < region_ia or next_b < region_ib:
373
return 'conflict', None, None, next_a, region_ia, next_b, region_ib
269
375
def find_sync_regions(self):
270
376
"""Return a list of sync regions, where both descendents match the base.
272
378
Generates a list of (base1, base2, a1, a2, b1, b2). There is
273
379
always a zero-length sync region at the end of all the files.
275
from difflib import SequenceMatcher
278
amatches = SequenceMatcher(None, self.base, self.a).get_matching_blocks()
279
bmatches = SequenceMatcher(None, self.base, self.b).get_matching_blocks()
383
amatches = patiencediff.PatienceSequenceMatcher(
384
None, self.base, self.a).get_matching_blocks()
385
bmatches = patiencediff.PatienceSequenceMatcher(
386
None, self.base, self.b).get_matching_blocks()
280
387
len_a = len(amatches)
281
388
len_b = len(bmatches)
297
404
# found a match of base[i[0], i[1]]; this may be less than
298
405
# the region that matches in either one
299
assert intlen <= alen
300
assert intlen <= blen
301
assert abase <= intbase
302
assert bbase <= intbase
406
# assert intlen <= alen
407
# assert intlen <= blen
408
# assert abase <= intbase
409
# assert bbase <= intbase
304
411
asub = amatch + (intbase - abase)
305
412
bsub = bmatch + (intbase - bbase)
306
413
aend = asub + intlen
307
414
bend = bsub + intlen
309
assert self.base[intbase:intend] == self.a[asub:aend], \
310
(self.base[intbase:intend], self.a[asub:aend])
312
assert self.base[intbase:intend] == self.b[bsub:bend]
416
# assert self.base[intbase:intend] == self.a[asub:aend], \
417
# (self.base[intbase:intend], self.a[asub:aend])
418
# assert self.base[intbase:intend] == self.b[bsub:bend]
314
420
sl.append((intbase, intend,
318
423
# advance whichever one ends first in the base text
319
424
if (abase + alen) < (bbase + blen):
324
429
intbase = len(self.base)
325
430
abase = len(self.a)
326
431
bbase = len(self.b)