1
# Copyright (C) 2005-2010 Canonical Ltd
1
# Copyright (C) 2004, 2005 by 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from __future__ import absolute_import
19
18
# mbp: "you know that thing where cvs gives you conflict markers?"
20
19
# s: "i hate that."
31
class CantReprocessAndShowBase(errors.BzrError):
33
_fmt = ("Can't reprocess and show base, because reprocessing obscures "
34
"the relationship of conflicting lines to the base")
37
23
def intersect(ra, rb):
74
63
Given BASE, OTHER, THIS, tries to produce a combined text
75
64
incorporating the changes from both BASE->OTHER and BASE->THIS.
76
65
All three will typically be sequences of lines."""
78
def __init__(self, base, a, b, is_cherrypick=False, allow_objects=False):
81
:param base: lines in BASE
84
:param is_cherrypick: flag indicating if this merge is a cherrypick.
85
When cherrypicking b => a, matches with b and base do not conflict.
86
:param allow_objects: if True, do not require that base, a and b are
87
plain Python strs. Also prevents BinaryFile from being raised.
88
Lines can be any sequence of comparable and hashable Python
92
textfile.check_text_lines(base)
93
textfile.check_text_lines(a)
94
textfile.check_text_lines(b)
66
def __init__(self, base, a, b):
98
self.is_cherrypick = is_cherrypick
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()
100
76
def merge_lines(self,
104
start_marker=b'<<<<<<<',
105
mid_marker=b'=======',
106
end_marker=b'>>>>>>>',
79
start_marker='<<<<<<<',
109
83
"""Return merge in cvs-like form.
113
if self.a[0].endswith(b'\r\n'):
115
elif self.a[0].endswith(b'\r'):
117
if base_marker and reprocess:
118
raise CantReprocessAndShowBase()
120
start_marker = start_marker + b' ' + name_a
86
start_marker = start_marker + ' ' + name_a
122
end_marker = end_marker + b' ' + name_b
123
if name_base and base_marker:
124
base_marker = base_marker + b' ' + name_base
125
merge_regions = self.merge_regions()
126
if reprocess is True:
127
merge_regions = self.reprocess_merge_regions(merge_regions)
128
for t in merge_regions:
88
end_marker = end_marker + ' ' + name_b
90
for t in self.merge_regions():
130
92
if what == 'unchanged':
131
93
for i in range(t[1], t[2]):
137
99
for i in range(t[1], t[2]):
139
101
elif what == 'conflict':
140
yield start_marker + newline
102
yield start_marker + '\n'
141
103
for i in range(t[3], t[4]):
143
if base_marker is not None:
144
yield base_marker + newline
145
for i in range(t[1], t[2]):
147
yield mid_marker + newline
105
yield mid_marker + '\n'
148
106
for i in range(t[5], t[6]):
150
yield end_marker + newline
108
yield end_marker + '\n'
152
110
raise ValueError(what)
154
116
def merge_annotated(self):
155
117
"""Return merge with conflicts, showing origin of lines.
157
Most useful for debugging merge.
119
Most useful for debugging merge.
159
121
for t in self.merge_regions():
242
209
# section a[0:ia] has been disposed of, etc
245
212
for zmatch, zend, amatch, aend, bmatch, bend in self.find_sync_regions():
213
#print 'match base [%d:%d]' % (zmatch, zend)
246
215
matchlen = zend - zmatch
249
# matchlen == (aend - amatch)
250
# matchlen == (bend - bmatch)
217
assert matchlen == (aend - amatch)
218
assert matchlen == (bend - bmatch)
251
220
len_a = amatch - ia
252
221
len_b = bmatch - ib
222
len_base = zmatch - iz
257
# print 'unmatched a=%d, b=%d' % (len_a, len_b)
227
#print 'unmatched a=%d, b=%d' % (len_a, len_b)
259
229
if len_a or len_b:
260
230
# 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)
261
235
same = compare_range(self.a, ia, amatch,
262
236
self.b, ib, bmatch)
265
239
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
267
equal_a = compare_range(self.a, ia, amatch,
268
self.base, iz, zmatch)
269
equal_b = compare_range(self.b, ib, bmatch,
270
self.base, iz, zmatch)
271
if equal_a and not equal_b:
272
yield 'b', ib, bmatch
273
elif equal_b and not equal_a:
274
yield 'a', ia, amatch
275
elif not equal_a and not equal_b:
276
if self.is_cherrypick:
277
for node in self._refine_cherrypick_conflict(
278
iz, zmatch, ia, amatch,
282
yield ('conflict', iz, zmatch, ia, amatch, ib,
285
raise AssertionError(
286
"can't handle a=b=base but unmatched")
247
raise AssertionError("can't handle a=b=base but unmatched")
292
253
# if the same part of the base was deleted on both sides
293
254
# that's OK, we can just skip it.
297
# assert ia == amatch
298
# assert ib == bmatch
299
# assert iz == zmatch
301
262
yield 'unchanged', zmatch, zend
306
def _refine_cherrypick_conflict(self, zstart, zend, astart, aend, bstart,
308
"""When cherrypicking b => a, ignore matches with b and base."""
309
# Do not emit regions which match, only regions which do not match
310
matches = patiencediff.PatienceSequenceMatcher(
311
None, self.base[zstart:zend], self.b[bstart:bend]
312
).get_matching_blocks()
317
for base_idx, b_idx, match_len in matches:
318
conflict_b_len = b_idx - last_b_idx
319
if conflict_b_len == 0:
320
# There are no lines in b which conflict, so skip it
325
zstart + last_base_idx, zstart + base_idx,
326
aend, aend, bstart + last_b_idx, bstart + b_idx)
328
# The first conflict gets the a-range
330
yield ('conflict', zstart + last_base_idx, zstart +
332
astart, aend, bstart + last_b_idx, bstart + b_idx)
333
last_base_idx = base_idx + match_len
334
last_b_idx = b_idx + match_len
335
if last_base_idx != zend - zstart or last_b_idx != bend - bstart:
337
yield ('conflict', zstart + last_base_idx, zstart + base_idx,
338
aend, aend, bstart + last_b_idx, bstart + b_idx)
340
# The first conflict gets the a-range
342
yield ('conflict', zstart + last_base_idx, zstart + base_idx,
343
astart, aend, bstart + last_b_idx, bstart + b_idx)
345
yield ('conflict', zstart, zend, astart, aend, bstart, bend)
347
def reprocess_merge_regions(self, merge_regions):
348
"""Where there are conflict regions, remove the agreed lines.
350
Lines where both A and B have made the same changes are
353
for region in merge_regions:
354
if region[0] != "conflict":
357
type, iz, zmatch, ia, amatch, ib, bmatch = region
358
a_region = self.a[ia:amatch]
359
b_region = self.b[ib:bmatch]
360
matches = patiencediff.PatienceSequenceMatcher(
361
None, a_region, b_region).get_matching_blocks()
364
for region_ia, region_ib, region_len in matches[:-1]:
367
reg = self.mismatch_region(next_a, region_ia, next_b,
371
yield 'same', region_ia, region_len + region_ia
372
next_a = region_ia + region_len
373
next_b = region_ib + region_len
374
reg = self.mismatch_region(next_a, amatch, next_b, bmatch)
379
def mismatch_region(next_a, region_ia, next_b, region_ib):
380
if next_a < region_ia or next_b < region_ib:
381
return 'conflict', None, None, next_a, region_ia, next_b, region_ib
383
269
def find_sync_regions(self):
384
270
"""Return a list of sync regions, where both descendents match the base.
386
272
Generates a list of (base1, base2, a1, a2, b1, b2). There is
387
273
always a zero-length sync region at the end of all the files.
275
from difflib import SequenceMatcher
391
amatches = patiencediff.PatienceSequenceMatcher(
392
None, self.base, self.a).get_matching_blocks()
393
bmatches = patiencediff.PatienceSequenceMatcher(
394
None, self.base, self.b).get_matching_blocks()
278
amatches = SequenceMatcher(None, self.base, self.a).get_matching_blocks()
279
bmatches = SequenceMatcher(None, self.base, self.b).get_matching_blocks()
395
280
len_a = len(amatches)
396
281
len_b = len(bmatches)
412
297
# found a match of base[i[0], i[1]]; this may be less than
413
298
# the region that matches in either one
414
# assert intlen <= alen
415
# assert intlen <= blen
416
# assert abase <= intbase
417
# assert bbase <= intbase
299
assert intlen <= alen
300
assert intlen <= blen
301
assert abase <= intbase
302
assert bbase <= intbase
419
304
asub = amatch + (intbase - abase)
420
305
bsub = bmatch + (intbase - bbase)
421
306
aend = asub + intlen
422
307
bend = bsub + intlen
424
# assert self.base[intbase:intend] == self.a[asub:aend], \
425
# (self.base[intbase:intend], self.a[asub:aend])
426
# assert self.base[intbase:intend] == self.b[bsub:bend]
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]
428
314
sl.append((intbase, intend,
431
318
# advance whichever one ends first in the base text
432
319
if (abase + alen) < (bbase + blen):
437
324
intbase = len(self.base)
438
325
abase = len(self.a)
439
326
bbase = len(self.b)