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
17
"""File annotate based on weave storage"""
19
19
# TODO: Choice of more or less verbose formats:
21
21
# interposed: show more details between blocks of modified lines
23
23
# TODO: Show which revision caused a line to merge into the parent
25
25
# TODO: perhaps abbreviate timescales depending on how recent they are
26
# e.g. "3:12 Tue", "13 Oct", "Oct 2005", etc.
26
# e.g. "3:12 Tue", "13 Oct", "Oct 2005", etc.
31
from bzrlib.lazy_import import lazy_import
32
lazy_import(globals(), """
42
33
from bzrlib.config import extract_email_address
43
from bzrlib.repository import _strip_NULL_ghosts
44
from bzrlib.revision import (
48
from bzrlib.symbol_versioning import (
54
@deprecated_function(deprecated_in((2, 4, 0)))
34
from bzrlib.errors import BzrError
55
37
def annotate_file(branch, rev_id, file_id, verbose=False, full=False,
56
to_file=None, show_ids=False):
57
"""Annotate file_id at revision rev_id in branch.
59
The branch should already be read_locked() when annotate_file is called.
61
:param branch: The branch to look for revision numbers and history from.
62
:param rev_id: The revision id to annotate.
63
:param file_id: The file_id to annotate.
64
:param verbose: Show all details rather than truncating to ensure
65
reasonable text width.
66
:param full: XXXX Not sure what this does.
67
:param to_file: The file to output the annotation to; if None stdout is
69
:param show_ids: Show revision ids in the annotation output.
71
tree = branch.repository.revision_tree(rev_id)
72
annotate_file_tree(tree, file_id, to_file, verbose=verbose,
73
full=full, show_ids=show_ids, branch=branch)
76
def annotate_file_tree(tree, file_id, to_file, verbose=False, full=False,
77
show_ids=False, branch=None):
78
"""Annotate file_id in a tree.
80
The tree should already be read_locked() when annotate_file_tree is called.
82
:param tree: The tree to look for revision numbers and history from.
83
:param file_id: The file_id to annotate.
84
:param to_file: The file to output the annotation to.
85
:param verbose: Show all details rather than truncating to ensure
86
reasonable text width.
87
:param full: XXXX Not sure what this does.
88
:param show_ids: Show revision ids in the annotation output.
89
:param branch: Branch to use for revision revno lookups
93
39
if to_file is None:
94
40
to_file = sys.stdout
96
# Handle the show_ids case
97
annotations = list(tree.annotate_iter(file_id))
99
return _show_id_annotations(annotations, to_file, full)
101
if not getattr(tree, "get_revision_id", False):
102
# Create a virtual revision to represent the current tree state.
103
# Should get some more pending commit attributes, like pending tags,
105
current_rev = Revision(CURRENT_REVISION)
106
current_rev.parent_ids = tree.get_parent_ids()
108
current_rev.committer = branch.get_config().username()
109
except errors.NoWhoami:
110
current_rev.committer = 'local user'
111
current_rev.message = "?"
112
current_rev.timestamp = round(time.time(), 3)
113
current_rev.timezone = osutils.local_time_offset()
116
annotation = list(_expand_annotations(annotations, branch,
118
_print_annotations(annotation, verbose, to_file, full)
121
def _print_annotations(annotation, verbose, to_file, full):
122
"""Print annotations to to_file.
124
:param to_file: The file to output the annotation to.
125
:param verbose: Show all details rather than truncating to ensure
126
reasonable text width.
127
:param full: XXXX Not sure what this does.
129
if len(annotation) == 0:
130
max_origin_len = max_revno_len = max_revid_len = 0
132
max_origin_len = max(len(x[1]) for x in annotation)
133
max_revno_len = max(len(x[0]) for x in annotation)
134
max_revid_len = max(len(x[3]) for x in annotation)
136
max_revno_len = min(max_revno_len, 12)
137
max_revno_len = max(max_revno_len, 3)
139
# Output the annotations
141
encoding = getattr(to_file, 'encoding', None) or \
142
osutils.get_terminal_encoding()
143
for (revno_str, author, date_str, line_rev_id, text) in annotation:
43
annotation = list(_annotate_file(branch, rev_id, file_id))
44
max_origin_len = max(len(origin) for origin in set(x[1] for x in annotation))
45
for (revno_str, author, date_str, line_rev_id, text ) in annotation:
145
anno = '%-*s %-*s %8s ' % (max_revno_len, revno_str,
146
max_origin_len, author, date_str)
148
if len(revno_str) > max_revno_len:
149
revno_str = revno_str[:max_revno_len-1] + '>'
150
anno = "%-*s %-7s " % (max_revno_len, revno_str, author[:7])
151
if anno.lstrip() == "" and full:
155
except UnicodeEncodeError:
156
# cmd_annotate should be passing in an 'exact' object, which means
157
# we have a direct handle to sys.stdout or equivalent. It may not
158
# be able to handle the exact Unicode characters, but 'annotate' is
159
# a user function (non-scripting), so shouldn't die because of
160
# unrepresentable annotation characters. So encode using 'replace',
161
# and write them again.
162
to_file.write(anno.encode(encoding, 'replace'))
163
to_file.write('| %s\n' % (text,))
167
def _show_id_annotations(annotations, to_file, full):
171
max_origin_len = max(len(origin) for origin, text in annotations)
172
for origin, text in annotations:
173
if full or last_rev_id != origin:
177
to_file.write('%*s | %s' % (max_origin_len, this, text))
182
def _expand_annotations(annotations, branch, current_rev=None):
183
"""Expand a file's annotations into command line UI ready tuples.
185
Each tuple includes detailed information, such as the author name, and date
186
string for the commit, rather than just the revision id.
188
:param annotations: The annotations to expand.
189
:param revision_id_to_revno: A map from id to revision numbers.
190
:param branch: A locked branch to query for revision details.
192
repository = branch.repository
193
if current_rev is not None:
194
# This can probably become a function on MutableTree, get_revno_map
195
# there, or something.
196
last_revision = current_rev.revision_id
197
# XXX: Partially Cloned from branch, uses the old_get_graph, eep.
198
# XXX: The main difficulty is that we need to inject a single new node
199
# (current_rev) into the graph before it gets numbered, etc.
200
# Once KnownGraph gets an 'add_node()' function, we can use
201
# VF.get_known_graph_ancestry().
202
graph = repository.get_graph()
203
revision_graph = dict(((key, value) for key, value in
204
graph.iter_ancestry(current_rev.parent_ids) if value is not None))
205
revision_graph = _strip_NULL_ghosts(revision_graph)
206
revision_graph[last_revision] = current_rev.parent_ids
207
merge_sorted_revisions = tsort.merge_sort(
212
revision_id_to_revno = dict((rev_id, revno)
213
for seq_num, rev_id, depth, revno, end_of_merge in
214
merge_sorted_revisions)
216
revision_id_to_revno = branch.get_revision_id_to_revno_map()
47
anno = '%5s %-*s %8s ' % (revno_str, max_origin_len, author, date_str)
49
anno = "%5s %-7s " % ( revno_str, author[:7] )
51
if anno.lstrip() == "" and full: anno = prevanno
52
print >>to_file, '%s| %s' % (anno, text)
55
def _annotate_file(branch, rev_id, file_id ):
57
rh = branch.revision_history()
58
w = branch.repository.weave_store.get_weave(file_id,
59
branch.repository.get_transaction())
217
60
last_origin = None
218
revision_ids = set(o for o, t in annotations)
220
if CURRENT_REVISION in revision_ids:
221
revision_id_to_revno[CURRENT_REVISION] = (
222
"%d?" % (branch.revno() + 1),)
223
revisions[CURRENT_REVISION] = current_rev
224
revision_ids = [o for o in revision_ids if
225
repository.has_revision(o)]
226
revisions.update((r.revision_id, r) for r in
227
repository.get_revisions(revision_ids))
228
for origin, text in annotations:
61
for origin, text in w.annotate_iter(rev_id):
229
62
text = text.rstrip('\r\n')
230
63
if origin == last_origin:
231
64
(revno_str, author, date_str) = ('','','')
233
66
last_origin = origin
234
if origin not in revisions:
67
line_rev_id = w.idx_to_name(origin)
68
if not branch.repository.has_revision(line_rev_id):
235
69
(revno_str, author, date_str) = ('?','?','?')
237
revno_str = '.'.join(str(i) for i in
238
revision_id_to_revno[origin])
239
rev = revisions[origin]
72
revno_str = str(rh.index(line_rev_id) + 1)
75
rev = branch.repository.get_revision(line_rev_id)
240
76
tz = rev.timezone or 0
241
date_str = time.strftime('%Y%m%d',
77
date_str = time.strftime('%Y%m%d',
242
78
time.gmtime(rev.timestamp + tz))
243
79
# a lazy way to get something like the email address
244
80
# TODO: Get real email address
245
author = rev.get_apparent_authors()[0]
81
author = rev.committer
247
83
author = extract_email_address(author)
248
except errors.NoEmailInUsername:
249
85
pass # use the whole name
250
yield (revno_str, author, date_str, origin, text)
253
def reannotate(parents_lines, new_lines, new_revision_id,
254
_left_matching_blocks=None,
255
heads_provider=None):
256
"""Create a new annotated version from new lines and parent annotations.
258
:param parents_lines: List of annotated lines for all parents
259
:param new_lines: The un-annotated new lines
260
:param new_revision_id: The revision-id to associate with new lines
261
(will often be CURRENT_REVISION)
262
:param left_matching_blocks: a hint about which areas are common
263
between the text and its left-hand-parent. The format is
264
the SequenceMatcher.get_matching_blocks format
265
(start_left, start_right, length_of_match).
266
:param heads_provider: An object which provides a .heads() call to resolve
267
if any revision ids are children of others.
268
If None, then any ancestry disputes will be resolved with
271
if len(parents_lines) == 0:
272
lines = [(new_revision_id, line) for line in new_lines]
273
elif len(parents_lines) == 1:
274
lines = _reannotate(parents_lines[0], new_lines, new_revision_id,
275
_left_matching_blocks)
276
elif len(parents_lines) == 2:
277
left = _reannotate(parents_lines[0], new_lines, new_revision_id,
278
_left_matching_blocks)
279
lines = _reannotate_annotated(parents_lines[1], new_lines,
280
new_revision_id, left,
283
reannotations = [_reannotate(parents_lines[0], new_lines,
284
new_revision_id, _left_matching_blocks)]
285
reannotations.extend(_reannotate(p, new_lines, new_revision_id)
286
for p in parents_lines[1:])
288
for annos in zip(*reannotations):
289
origins = set(a for a, l in annos)
290
if len(origins) == 1:
291
# All the parents agree, so just return the first one
292
lines.append(annos[0])
295
if len(origins) == 2 and new_revision_id in origins:
296
origins.remove(new_revision_id)
297
if len(origins) == 1:
298
lines.append((origins.pop(), line))
300
lines.append((new_revision_id, line))
304
def _reannotate(parent_lines, new_lines, new_revision_id,
305
matching_blocks=None):
307
if matching_blocks is None:
308
plain_parent_lines = [l for r, l in parent_lines]
309
matcher = patiencediff.PatienceSequenceMatcher(None,
310
plain_parent_lines, new_lines)
311
matching_blocks = matcher.get_matching_blocks()
313
for i, j, n in matching_blocks:
314
for line in new_lines[new_cur:j]:
315
lines.append((new_revision_id, line))
316
lines.extend(parent_lines[i:i+n])
321
def _get_matching_blocks(old, new):
322
matcher = patiencediff.PatienceSequenceMatcher(None, old, new)
323
return matcher.get_matching_blocks()
326
_break_annotation_tie = None
328
def _old_break_annotation_tie(annotated_lines):
329
"""Chose an attribution between several possible ones.
331
:param annotated_lines: A list of tuples ((file_id, rev_id), line) where
332
the lines are identical but the revids different while no parent
333
relation exist between them
335
:return : The "winning" line. This must be one with a revid that
336
guarantees that further criss-cross merges will converge. Failing to
337
do so have performance implications.
339
# sort lexicographically so that we always get a stable result.
341
# TODO: while 'sort' is the easiest (and nearly the only possible solution)
342
# with the current implementation, chosing the oldest revision is known to
343
# provide better results (as in matching user expectations). The most
344
# common use case being manual cherry-pick from an already existing
346
return sorted(annotated_lines)[0]
349
def _find_matching_unannotated_lines(output_lines, plain_child_lines,
350
child_lines, start_child, end_child,
351
right_lines, start_right, end_right,
352
heads_provider, revision_id):
353
"""Find lines in plain_right_lines that match the existing lines.
355
:param output_lines: Append final annotated lines to this list
356
:param plain_child_lines: The unannotated new lines for the child text
357
:param child_lines: Lines for the child text which have been annotated
360
:param start_child: Position in plain_child_lines and child_lines to start
362
:param end_child: Last position in plain_child_lines and child_lines to
364
:param right_lines: The annotated lines for the whole text for the right
366
:param start_right: Position in right_lines to start the match
367
:param end_right: Last position in right_lines to search for a match
368
:param heads_provider: When parents disagree on the lineage of a line, we
369
need to check if one side supersedes the other
370
:param revision_id: The label to give if a line should be labeled 'tip'
372
output_extend = output_lines.extend
373
output_append = output_lines.append
374
# We need to see if any of the unannotated lines match
375
plain_right_subset = [l for a,l in right_lines[start_right:end_right]]
376
plain_child_subset = plain_child_lines[start_child:end_child]
377
match_blocks = _get_matching_blocks(plain_right_subset, plain_child_subset)
381
for right_idx, child_idx, match_len in match_blocks:
382
# All the lines that don't match are just passed along
383
if child_idx > last_child_idx:
384
output_extend(child_lines[start_child + last_child_idx
385
:start_child + child_idx])
386
for offset in xrange(match_len):
387
left = child_lines[start_child+child_idx+offset]
388
right = right_lines[start_right+right_idx+offset]
389
if left[0] == right[0]:
390
# The annotations match, just return the left one
392
elif left[0] == revision_id:
393
# The left parent marked this as unmatched, so let the
394
# right parent claim it
397
# Left and Right both claim this line
398
if heads_provider is None:
399
output_append((revision_id, left[1]))
401
heads = heads_provider.heads((left[0], right[0]))
403
output_append((iter(heads).next(), left[1]))
405
# Both claim different origins, get a stable result.
406
# If the result is not stable, there is a risk a
407
# performance degradation as criss-cross merges will
408
# flip-flop the attribution.
409
if _break_annotation_tie is None:
411
_old_break_annotation_tie([left, right]))
413
output_append(_break_annotation_tie([left, right]))
414
last_child_idx = child_idx + match_len
417
def _reannotate_annotated(right_parent_lines, new_lines, new_revision_id,
418
annotated_lines, heads_provider):
419
"""Update the annotations for a node based on another parent.
421
:param right_parent_lines: A list of annotated lines for the right-hand
423
:param new_lines: The unannotated new lines.
424
:param new_revision_id: The revision_id to attribute to lines which are not
425
present in either parent.
426
:param annotated_lines: A list of annotated lines. This should be the
427
annotation of new_lines based on parents seen so far.
428
:param heads_provider: When parents disagree on the lineage of a line, we
429
need to check if one side supersedes the other.
431
if len(new_lines) != len(annotated_lines):
432
raise AssertionError("mismatched new_lines and annotated_lines")
433
# First compare the newly annotated lines with the right annotated lines.
434
# Lines which were not changed in left or right should match. This tends to
435
# be the bulk of the lines, and they will need no further processing.
437
lines_extend = lines.extend
438
last_right_idx = 0 # The line just after the last match from the right side
440
matching_left_and_right = _get_matching_blocks(right_parent_lines,
442
for right_idx, left_idx, match_len in matching_left_and_right:
443
# annotated lines from last_left_idx to left_idx did not match the
444
# lines from last_right_idx to right_idx, the raw lines should be
445
# compared to determine what annotations need to be updated
446
if last_right_idx == right_idx or last_left_idx == left_idx:
447
# One of the sides is empty, so this is a pure insertion
448
lines_extend(annotated_lines[last_left_idx:left_idx])
450
# We need to see if any of the unannotated lines match
451
_find_matching_unannotated_lines(lines,
452
new_lines, annotated_lines,
453
last_left_idx, left_idx,
455
last_right_idx, right_idx,
458
last_right_idx = right_idx + match_len
459
last_left_idx = left_idx + match_len
460
# If left and right agree on a range, just push that into the output
461
lines_extend(annotated_lines[left_idx:left_idx + match_len])
466
from bzrlib._annotator_pyx import Annotator
467
except ImportError, e:
468
osutils.failed_to_load_extension(e)
469
from bzrlib._annotator_py import Annotator
86
yield (revno_str, author, date_str, line_rev_id, text)
added
removed
bzrlib/annotate.py
