1
# Copyright (C) 2004, 2005 by Canonical Ltd
1
# Copyright (C) 2004, 2005 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."
22
from bzrlib.errors import CantReprocessAndShowBase
23
import bzrlib.patiencediff
24
from bzrlib.textfile import check_text_lines
23
27
def intersect(ra, rb):
24
28
"""Given two ranges return the range where they intersect or None.
63
66
Given BASE, OTHER, THIS, tries to produce a combined text
64
67
incorporating the changes from both BASE->OTHER and BASE->THIS.
65
68
All three will typically be sequences of lines."""
66
def __init__(self, base, a, b):
70
def __init__(self, base, a, b, is_cherrypick=False, allow_objects=False):
73
:param base: lines in BASE
76
:param is_cherrypick: flag indicating if this merge is a cherrypick.
77
When cherrypicking b => a, matches with b and base do not conflict.
78
:param allow_objects: if True, do not require that base, a and b are
79
plain Python strs. Also prevents BinaryFile from being raised.
80
Lines can be any sequence of comparable and hashable Python
84
check_text_lines(base)
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()
90
self.is_cherrypick = is_cherrypick
76
92
def merge_lines(self,
79
start_marker='<<<<<<<<',
80
mid_marker='========',
81
end_marker='>>>>>>>>',
96
start_marker='<<<<<<<',
83
101
"""Return merge in cvs-like form.
105
if self.a[0].endswith('\r\n'):
107
elif self.a[0].endswith('\r'):
109
if base_marker and reprocess:
110
raise CantReprocessAndShowBase()
86
112
start_marker = start_marker + ' ' + name_a
88
114
end_marker = end_marker + ' ' + name_b
90
for t in self.merge_regions():
115
if name_base and base_marker:
116
base_marker = base_marker + ' ' + name_base
117
merge_regions = self.merge_regions()
118
if reprocess is True:
119
merge_regions = self.reprocess_merge_regions(merge_regions)
120
for t in merge_regions:
92
122
if what == 'unchanged':
93
123
for i in range(t[1], t[2]):
99
129
for i in range(t[1], t[2]):
101
131
elif what == 'conflict':
102
yield start_marker + '\n'
132
yield start_marker + newline
103
133
for i in range(t[3], t[4]):
105
yield mid_marker + '\n'
135
if base_marker is not None:
136
yield base_marker + newline
137
for i in range(t[1], t[2]):
139
yield mid_marker + newline
106
140
for i in range(t[5], t[6]):
108
yield end_marker + '\n'
142
yield end_marker + newline
110
144
raise ValueError(what)
116
146
def merge_annotated(self):
117
147
"""Return merge with conflicts, showing origin of lines.
119
Most useful for debugging merge.
149
Most useful for debugging merge.
121
151
for t in self.merge_regions():
209
234
# section a[0:ia] has been disposed of, etc
212
237
for zmatch, zend, amatch, aend, bmatch, bend in self.find_sync_regions():
213
#print 'match base [%d:%d]' % (zmatch, zend)
215
238
matchlen = zend - zmatch
217
assert matchlen == (aend - amatch)
218
assert matchlen == (bend - bmatch)
241
# matchlen == (aend - amatch)
242
# matchlen == (bend - bmatch)
220
243
len_a = amatch - ia
221
244
len_b = bmatch - ib
222
245
len_base = zmatch - iz
249
# assert len_base >= 0
227
251
#print 'unmatched a=%d, b=%d' % (len_a, len_b)
229
253
if len_a or len_b:
230
254
# 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
255
same = compare_range(self.a, ia, amatch,
236
256
self.b, ib, bmatch)
239
259
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")
261
equal_a = compare_range(self.a, ia, amatch,
262
self.base, iz, zmatch)
263
equal_b = compare_range(self.b, ib, bmatch,
264
self.base, iz, zmatch)
265
if equal_a and not equal_b:
266
yield 'b', ib, bmatch
267
elif equal_b and not equal_a:
268
yield 'a', ia, amatch
269
elif not equal_a and not equal_b:
270
if self.is_cherrypick:
271
for node in self._refine_cherrypick_conflict(
272
iz, zmatch, ia, amatch,
276
yield 'conflict', iz, zmatch, ia, amatch, ib, bmatch
278
raise AssertionError("can't handle a=b=base but unmatched")
253
284
# if the same part of the base was deleted on both sides
254
285
# that's OK, we can just skip it.
289
# assert ia == amatch
290
# assert ib == bmatch
291
# assert iz == zmatch
262
293
yield 'unchanged', zmatch, zend
298
def _refine_cherrypick_conflict(self, zstart, zend, astart, aend, bstart, bend):
299
"""When cherrypicking b => a, ignore matches with b and base."""
300
# Do not emit regions which match, only regions which do not match
301
matches = bzrlib.patiencediff.PatienceSequenceMatcher(None,
302
self.base[zstart:zend], self.b[bstart:bend]).get_matching_blocks()
307
for base_idx, b_idx, match_len in matches:
308
conflict_z_len = base_idx - last_base_idx
309
conflict_b_len = b_idx - last_b_idx
310
if conflict_b_len == 0: # There are no lines in b which conflict,
316
zstart + last_base_idx, zstart + base_idx,
317
aend, aend, bstart + last_b_idx, bstart + b_idx)
319
# The first conflict gets the a-range
321
yield ('conflict', zstart + last_base_idx, zstart +
323
astart, aend, bstart + last_b_idx, bstart + b_idx)
324
last_base_idx = base_idx + match_len
325
last_b_idx = b_idx + match_len
326
if last_base_idx != zend - zstart or last_b_idx != bend - bstart:
328
yield ('conflict', zstart + last_base_idx, zstart + base_idx,
329
aend, aend, bstart + last_b_idx, bstart + b_idx)
331
# The first conflict gets the a-range
333
yield ('conflict', zstart + last_base_idx, zstart + base_idx,
334
astart, aend, bstart + last_b_idx, bstart + b_idx)
336
yield ('conflict', zstart, zend, astart, aend, bstart, bend)
338
def reprocess_merge_regions(self, merge_regions):
339
"""Where there are conflict regions, remove the agreed lines.
341
Lines where both A and B have made the same changes are
344
for region in merge_regions:
345
if region[0] != "conflict":
348
type, iz, zmatch, ia, amatch, ib, bmatch = region
349
a_region = self.a[ia:amatch]
350
b_region = self.b[ib:bmatch]
351
matches = bzrlib.patiencediff.PatienceSequenceMatcher(
352
None, a_region, b_region).get_matching_blocks()
355
for region_ia, region_ib, region_len in matches[:-1]:
358
reg = self.mismatch_region(next_a, region_ia, next_b,
362
yield 'same', region_ia, region_len+region_ia
363
next_a = region_ia + region_len
364
next_b = region_ib + region_len
365
reg = self.mismatch_region(next_a, amatch, next_b, bmatch)
370
def mismatch_region(next_a, region_ia, next_b, region_ib):
371
if next_a < region_ia or next_b < region_ib:
372
return 'conflict', None, None, next_a, region_ia, next_b, region_ib
269
374
def find_sync_regions(self):
270
375
"""Return a list of sync regions, where both descendents match the base.
272
377
Generates a list of (base1, base2, a1, a2, b1, b2). There is
273
378
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()
382
amatches = bzrlib.patiencediff.PatienceSequenceMatcher(
383
None, self.base, self.a).get_matching_blocks()
384
bmatches = bzrlib.patiencediff.PatienceSequenceMatcher(
385
None, self.base, self.b).get_matching_blocks()
280
386
len_a = len(amatches)
281
387
len_b = len(bmatches)
297
403
# found a match of base[i[0], i[1]]; this may be less than
298
404
# the region that matches in either one
299
assert intlen <= alen
300
assert intlen <= blen
301
assert abase <= intbase
302
assert bbase <= intbase
405
# assert intlen <= alen
406
# assert intlen <= blen
407
# assert abase <= intbase
408
# assert bbase <= intbase
304
410
asub = amatch + (intbase - abase)
305
411
bsub = bmatch + (intbase - bbase)
306
412
aend = asub + intlen
307
413
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]
415
# assert self.base[intbase:intend] == self.a[asub:aend], \
416
# (self.base[intbase:intend], self.a[asub:aend])
417
# assert self.base[intbase:intend] == self.b[bsub:bend]
314
419
sl.append((intbase, intend,
318
422
# advance whichever one ends first in the base text
319
423
if (abase + alen) < (bbase + blen):
324
428
intbase = len(self.base)
325
429
abase = len(self.a)
326
430
bbase = len(self.b)