1
# Copyright (C) 2004, 2005, 2006, 2007, 2008, 2009 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.
37
33
from bzrlib.config import extract_email_address
38
from bzrlib.repository import _strip_NULL_ghosts
39
from bzrlib.revision import CURRENT_REVISION, Revision
34
from bzrlib.errors import BzrError
42
37
def annotate_file(branch, rev_id, file_id, verbose=False, full=False,
43
to_file=None, show_ids=False):
44
"""Annotate file_id at revision rev_id in branch.
46
The branch should already be read_locked() when annotate_file is called.
48
:param branch: The branch to look for revision numbers and history from.
49
:param rev_id: The revision id to annotate.
50
:param file_id: The file_id to annotate.
51
:param verbose: Show all details rather than truncating to ensure
52
reasonable text width.
53
:param full: XXXX Not sure what this does.
54
:param to_file: The file to output the annotation to; if None stdout is
56
:param show_ids: Show revision ids in the annotation output.
58
39
if to_file is None:
59
40
to_file = sys.stdout
61
# Handle the show_ids case
62
annotations = _annotations(branch.repository, file_id, rev_id)
64
return _show_id_annotations(annotations, to_file, full)
66
# Calculate the lengths of the various columns
67
annotation = list(_expand_annotations(annotations, branch))
68
_print_annotations(annotation, verbose, to_file, full)
71
def annotate_file_tree(tree, file_id, to_file, verbose=False, full=False,
73
"""Annotate file_id in a tree.
75
The tree should already be read_locked() when annotate_file_tree is called.
77
:param tree: The tree to look for revision numbers and history from.
78
:param file_id: The file_id to annotate.
79
:param to_file: The file to output the annotation to.
80
:param verbose: Show all details rather than truncating to ensure
81
reasonable text width.
82
:param full: XXXX Not sure what this does.
83
:param show_ids: Show revision ids in the annotation output.
85
rev_id = tree.last_revision()
88
# Handle the show_ids case
89
annotations = list(tree.annotate_iter(file_id))
91
return _show_id_annotations(annotations, to_file, full)
93
# Create a virtual revision to represent the current tree state.
94
# Should get some more pending commit attributes, like pending tags,
96
current_rev = Revision(CURRENT_REVISION)
97
current_rev.parent_ids = tree.get_parent_ids()
98
current_rev.committer = tree.branch.get_config().username()
99
current_rev.message = "?"
100
current_rev.timestamp = round(time.time(), 3)
101
current_rev.timezone = osutils.local_time_offset()
102
annotation = list(_expand_annotations(annotations, tree.branch,
104
_print_annotations(annotation, verbose, to_file, full)
107
def _print_annotations(annotation, verbose, to_file, full):
108
"""Print annotations to to_file.
110
:param to_file: The file to output the annotation to.
111
:param verbose: Show all details rather than truncating to ensure
112
reasonable text width.
113
:param full: XXXX Not sure what this does.
115
if len(annotation) == 0:
116
max_origin_len = max_revno_len = max_revid_len = 0
118
max_origin_len = max(len(x[1]) for x in annotation)
119
max_revno_len = max(len(x[0]) for x in annotation)
120
max_revid_len = max(len(x[3]) for x in annotation)
122
max_revno_len = min(max_revno_len, 12)
123
max_revno_len = max(max_revno_len, 3)
125
# Output the annotations
127
encoding = getattr(to_file, 'encoding', None) or \
128
osutils.get_terminal_encoding()
129
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:
131
anno = '%-*s %-*s %8s ' % (max_revno_len, revno_str,
132
max_origin_len, author, date_str)
134
if len(revno_str) > max_revno_len:
135
revno_str = revno_str[:max_revno_len-1] + '>'
136
anno = "%-*s %-7s " % (max_revno_len, revno_str, author[:7])
137
if anno.lstrip() == "" and full:
141
except UnicodeEncodeError:
142
# cmd_annotate should be passing in an 'exact' object, which means
143
# we have a direct handle to sys.stdout or equivalent. It may not
144
# be able to handle the exact Unicode characters, but 'annotate' is
145
# a user function (non-scripting), so shouldn't die because of
146
# unrepresentable annotation characters. So encode using 'replace',
147
# and write them again.
148
to_file.write(anno.encode(encoding, 'replace'))
149
to_file.write('| %s\n' % (text,))
153
def _show_id_annotations(annotations, to_file, full):
157
max_origin_len = max(len(origin) for origin, text in annotations)
158
for origin, text in annotations:
159
if full or last_rev_id != origin:
163
to_file.write('%*s | %s' % (max_origin_len, this, text))
168
def _annotations(repo, file_id, rev_id):
169
"""Return the list of (origin_revision_id, line_text) for a revision of a file in a repository."""
170
annotations = repo.texts.annotate((file_id, rev_id))
172
return [(key[-1], line) for (key, line) in annotations]
175
def _expand_annotations(annotations, branch, current_rev=None):
176
"""Expand a file's annotations into command line UI ready tuples.
178
Each tuple includes detailed information, such as the author name, and date
179
string for the commit, rather than just the revision id.
181
:param annotations: The annotations to expand.
182
:param revision_id_to_revno: A map from id to revision numbers.
183
:param branch: A locked branch to query for revision details.
185
repository = branch.repository
186
if current_rev is not None:
187
# This can probably become a function on MutableTree, get_revno_map there,
189
last_revision = current_rev.revision_id
190
# XXX: Partially Cloned from branch, uses the old_get_graph, eep.
191
# XXX: The main difficulty is that we need to inject a single new node
192
# (current_rev) into the graph before it gets numbered, etc.
193
# Once KnownGraph gets an 'add_node()' function, we can use
194
# VF.get_known_graph_ancestry().
195
graph = repository.get_graph()
196
revision_graph = dict(((key, value) for key, value in
197
graph.iter_ancestry(current_rev.parent_ids) if value is not None))
198
revision_graph = _strip_NULL_ghosts(revision_graph)
199
revision_graph[last_revision] = current_rev.parent_ids
200
merge_sorted_revisions = tsort.merge_sort(
205
revision_id_to_revno = dict((rev_id, revno)
206
for seq_num, rev_id, depth, revno, end_of_merge in
207
merge_sorted_revisions)
209
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())
210
60
last_origin = None
211
revision_ids = set(o for o, t in annotations)
213
if CURRENT_REVISION in revision_ids:
214
revision_id_to_revno[CURRENT_REVISION] = (
215
"%d?" % (branch.revno() + 1),)
216
revisions[CURRENT_REVISION] = current_rev
217
revision_ids = [o for o in revision_ids if
218
repository.has_revision(o)]
219
revisions.update((r.revision_id, r) for r in
220
repository.get_revisions(revision_ids))
221
for origin, text in annotations:
61
for origin, text in w.annotate_iter(rev_id):
222
62
text = text.rstrip('\r\n')
223
63
if origin == last_origin:
224
64
(revno_str, author, date_str) = ('','','')
226
66
last_origin = origin
227
if origin not in revisions:
67
line_rev_id = w.idx_to_name(origin)
68
if not branch.repository.has_revision(line_rev_id):
228
69
(revno_str, author, date_str) = ('?','?','?')
230
revno_str = '.'.join(str(i) for i in
231
revision_id_to_revno[origin])
232
rev = revisions[origin]
72
revno_str = str(rh.index(line_rev_id) + 1)
75
rev = branch.repository.get_revision(line_rev_id)
233
76
tz = rev.timezone or 0
234
date_str = time.strftime('%Y%m%d',
77
date_str = time.strftime('%Y%m%d',
235
78
time.gmtime(rev.timestamp + tz))
236
79
# a lazy way to get something like the email address
237
80
# TODO: Get real email address
238
author = rev.get_apparent_authors()[0]
81
author = rev.committer
240
83
author = extract_email_address(author)
241
except errors.NoEmailInUsername:
242
85
pass # use the whole name
243
yield (revno_str, author, date_str, origin, text)
246
def reannotate(parents_lines, new_lines, new_revision_id,
247
_left_matching_blocks=None,
248
heads_provider=None):
249
"""Create a new annotated version from new lines and parent annotations.
251
:param parents_lines: List of annotated lines for all parents
252
:param new_lines: The un-annotated new lines
253
:param new_revision_id: The revision-id to associate with new lines
254
(will often be CURRENT_REVISION)
255
:param left_matching_blocks: a hint about which areas are common
256
between the text and its left-hand-parent. The format is
257
the SequenceMatcher.get_matching_blocks format
258
(start_left, start_right, length_of_match).
259
:param heads_provider: An object which provides a .heads() call to resolve
260
if any revision ids are children of others.
261
If None, then any ancestry disputes will be resolved with
264
if len(parents_lines) == 0:
265
lines = [(new_revision_id, line) for line in new_lines]
266
elif len(parents_lines) == 1:
267
lines = _reannotate(parents_lines[0], new_lines, new_revision_id,
268
_left_matching_blocks)
269
elif len(parents_lines) == 2:
270
left = _reannotate(parents_lines[0], new_lines, new_revision_id,
271
_left_matching_blocks)
272
lines = _reannotate_annotated(parents_lines[1], new_lines,
273
new_revision_id, left,
276
reannotations = [_reannotate(parents_lines[0], new_lines,
277
new_revision_id, _left_matching_blocks)]
278
reannotations.extend(_reannotate(p, new_lines, new_revision_id)
279
for p in parents_lines[1:])
281
for annos in zip(*reannotations):
282
origins = set(a for a, l in annos)
283
if len(origins) == 1:
284
# All the parents agree, so just return the first one
285
lines.append(annos[0])
288
if len(origins) == 2 and new_revision_id in origins:
289
origins.remove(new_revision_id)
290
if len(origins) == 1:
291
lines.append((origins.pop(), line))
293
lines.append((new_revision_id, line))
297
def _reannotate(parent_lines, new_lines, new_revision_id,
298
matching_blocks=None):
300
if matching_blocks is None:
301
plain_parent_lines = [l for r, l in parent_lines]
302
matcher = patiencediff.PatienceSequenceMatcher(None,
303
plain_parent_lines, new_lines)
304
matching_blocks = matcher.get_matching_blocks()
306
for i, j, n in matching_blocks:
307
for line in new_lines[new_cur:j]:
308
lines.append((new_revision_id, line))
309
lines.extend(parent_lines[i:i+n])
314
def _get_matching_blocks(old, new):
315
matcher = patiencediff.PatienceSequenceMatcher(None,
317
return matcher.get_matching_blocks()
320
_break_annotation_tie = None
322
def _old_break_annotation_tie(annotated_lines):
323
"""Chose an attribution between several possible ones.
325
:param annotated_lines: A list of tuples ((file_id, rev_id), line) where
326
the lines are identical but the revids different while no parent
327
relation exist between them
329
:return : The "winning" line. This must be one with a revid that
330
guarantees that further criss-cross merges will converge. Failing to
331
do so have performance implications.
333
# sort lexicographically so that we always get a stable result.
335
# TODO: while 'sort' is the easiest (and nearly the only possible solution)
336
# with the current implementation, chosing the oldest revision is known to
337
# provide better results (as in matching user expectations). The most
338
# common use case being manual cherry-pick from an already existing
340
return sorted(annotated_lines)[0]
343
def _find_matching_unannotated_lines(output_lines, plain_child_lines,
344
child_lines, start_child, end_child,
345
right_lines, start_right, end_right,
346
heads_provider, revision_id):
347
"""Find lines in plain_right_lines that match the existing lines.
349
:param output_lines: Append final annotated lines to this list
350
:param plain_child_lines: The unannotated new lines for the child text
351
:param child_lines: Lines for the child text which have been annotated
354
:param start_child: Position in plain_child_lines and child_lines to start
356
:param end_child: Last position in plain_child_lines and child_lines to
358
:param right_lines: The annotated lines for the whole text for the right
360
:param start_right: Position in right_lines to start the match
361
:param end_right: Last position in right_lines to search for a match
362
:param heads_provider: When parents disagree on the lineage of a line, we
363
need to check if one side supersedes the other
364
:param revision_id: The label to give if a line should be labeled 'tip'
366
output_extend = output_lines.extend
367
output_append = output_lines.append
368
# We need to see if any of the unannotated lines match
369
plain_right_subset = [l for a,l in right_lines[start_right:end_right]]
370
plain_child_subset = plain_child_lines[start_child:end_child]
371
match_blocks = _get_matching_blocks(plain_right_subset, plain_child_subset)
375
for right_idx, child_idx, match_len in match_blocks:
376
# All the lines that don't match are just passed along
377
if child_idx > last_child_idx:
378
output_extend(child_lines[start_child + last_child_idx
379
:start_child + child_idx])
380
for offset in xrange(match_len):
381
left = child_lines[start_child+child_idx+offset]
382
right = right_lines[start_right+right_idx+offset]
383
if left[0] == right[0]:
384
# The annotations match, just return the left one
386
elif left[0] == revision_id:
387
# The left parent marked this as unmatched, so let the
388
# right parent claim it
391
# Left and Right both claim this line
392
if heads_provider is None:
393
output_append((revision_id, left[1]))
395
heads = heads_provider.heads((left[0], right[0]))
397
output_append((iter(heads).next(), left[1]))
399
# Both claim different origins, get a stable result.
400
# If the result is not stable, there is a risk a
401
# performance degradation as criss-cross merges will
402
# flip-flop the attribution.
403
if _break_annotation_tie is None:
405
_old_break_annotation_tie([left, right]))
407
output_append(_break_annotation_tie([left, right]))
408
last_child_idx = child_idx + match_len
411
def _reannotate_annotated(right_parent_lines, new_lines, new_revision_id,
412
annotated_lines, heads_provider):
413
"""Update the annotations for a node based on another parent.
415
:param right_parent_lines: A list of annotated lines for the right-hand
417
:param new_lines: The unannotated new lines.
418
:param new_revision_id: The revision_id to attribute to lines which are not
419
present in either parent.
420
:param annotated_lines: A list of annotated lines. This should be the
421
annotation of new_lines based on parents seen so far.
422
:param heads_provider: When parents disagree on the lineage of a line, we
423
need to check if one side supersedes the other.
425
if len(new_lines) != len(annotated_lines):
426
raise AssertionError("mismatched new_lines and annotated_lines")
427
# First compare the newly annotated lines with the right annotated lines.
428
# Lines which were not changed in left or right should match. This tends to
429
# be the bulk of the lines, and they will need no further processing.
431
lines_extend = lines.extend
432
last_right_idx = 0 # The line just after the last match from the right side
434
matching_left_and_right = _get_matching_blocks(right_parent_lines,
436
for right_idx, left_idx, match_len in matching_left_and_right:
437
# annotated lines from last_left_idx to left_idx did not match the
438
# lines from last_right_idx to right_idx, the raw lines should be
439
# compared to determine what annotations need to be updated
440
if last_right_idx == right_idx or last_left_idx == left_idx:
441
# One of the sides is empty, so this is a pure insertion
442
lines_extend(annotated_lines[last_left_idx:left_idx])
444
# We need to see if any of the unannotated lines match
445
_find_matching_unannotated_lines(lines,
446
new_lines, annotated_lines,
447
last_left_idx, left_idx,
449
last_right_idx, right_idx,
452
last_right_idx = right_idx + match_len
453
last_left_idx = left_idx + match_len
454
# If left and right agree on a range, just push that into the output
455
lines_extend(annotated_lines[left_idx:left_idx + match_len])
460
from bzrlib._annotator_pyx import Annotator
461
except ImportError, e:
462
osutils.failed_to_load_extension(e)
463
from bzrlib._annotator_py import Annotator
86
yield (revno_str, author, date_str, line_rev_id, text)
added
removed
bzrlib/annotate.py
