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
# short: just show revno
22
# long: revno, author, date
21
23
# interposed: show more details between blocks of modified lines
23
25
# TODO: Show which revision caused a line to merge into the parent
25
# TODO: perhaps abbreviate timescales depending on how recent they are
26
# e.g. "3:12 Tue", "13 Oct", "Oct 2005", etc.
31
from bzrlib.lazy_import import lazy_import
32
lazy_import(globals(), """
42
from bzrlib.config import extract_email_address
43
from bzrlib.repository import _strip_NULL_ghosts
44
from bzrlib.revision import CURRENT_REVISION, Revision
47
def annotate_file(branch, rev_id, file_id, verbose=False, full=False,
48
to_file=None, show_ids=False):
49
"""Annotate file_id at revision rev_id in branch.
51
The branch should already be read_locked() when annotate_file is called.
53
:param branch: The branch to look for revision numbers and history from.
54
:param rev_id: The revision id to annotate.
55
:param file_id: The file_id to annotate.
56
:param verbose: Show all details rather than truncating to ensure
57
reasonable text width.
58
:param full: XXXX Not sure what this does.
59
:param to_file: The file to output the annotation to; if None stdout is
61
:param show_ids: Show revision ids in the annotation output.
33
def annotate_file(branch, rev_id, file_id, to_file=None):
63
34
if to_file is None:
64
35
to_file = sys.stdout
66
# Handle the show_ids case
67
annotations = _annotations(branch.repository, file_id, rev_id)
69
return _show_id_annotations(annotations, to_file, full)
71
# Calculate the lengths of the various columns
72
annotation = list(_expand_annotations(annotations, branch))
73
_print_annotations(annotation, verbose, to_file, full)
76
def annotate_file_tree(tree, file_id, to_file, verbose=False, full=False,
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.
90
rev_id = tree.last_revision()
93
# Handle the show_ids case
94
annotations = list(tree.annotate_iter(file_id))
96
return _show_id_annotations(annotations, to_file, full)
98
# Create a virtual revision to represent the current tree state.
99
# Should get some more pending commit attributes, like pending tags,
101
current_rev = Revision(CURRENT_REVISION)
102
current_rev.parent_ids = tree.get_parent_ids()
103
current_rev.committer = tree.branch.get_config().username()
104
current_rev.message = "?"
105
current_rev.timestamp = round(time.time(), 3)
106
current_rev.timezone = osutils.local_time_offset()
107
annotation = list(_expand_annotations(annotations, tree.branch,
109
_print_annotations(annotation, verbose, to_file, full)
112
def _print_annotations(annotation, verbose, to_file, full):
113
"""Print annotations to to_file.
115
:param to_file: The file to output the annotation to.
116
:param verbose: Show all details rather than truncating to ensure
117
reasonable text width.
118
:param full: XXXX Not sure what this does.
120
if len(annotation) == 0:
121
max_origin_len = max_revno_len = max_revid_len = 0
123
max_origin_len = max(len(x[1]) for x in annotation)
124
max_revno_len = max(len(x[0]) for x in annotation)
125
max_revid_len = max(len(x[3]) for x in annotation)
127
max_revno_len = min(max_revno_len, 12)
128
max_revno_len = max(max_revno_len, 3)
130
# Output the annotations
132
encoding = getattr(to_file, 'encoding', None) or \
133
osutils.get_terminal_encoding()
134
for (revno_str, author, date_str, line_rev_id, text) in annotation:
136
anno = '%-*s %-*s %8s ' % (max_revno_len, revno_str,
137
max_origin_len, author, date_str)
139
if len(revno_str) > max_revno_len:
140
revno_str = revno_str[:max_revno_len-1] + '>'
141
anno = "%-*s %-7s " % (max_revno_len, revno_str, author[:7])
142
if anno.lstrip() == "" and full:
146
except UnicodeEncodeError:
147
# cmd_annotate should be passing in an 'exact' object, which means
148
# we have a direct handle to sys.stdout or equivalent. It may not
149
# be able to handle the exact Unicode characters, but 'annotate' is
150
# a user function (non-scripting), so shouldn't die because of
151
# unrepresentable annotation characters. So encode using 'replace',
152
# and write them again.
153
to_file.write(anno.encode(encoding, 'replace'))
154
to_file.write('| %s\n' % (text,))
158
def _show_id_annotations(annotations, to_file, full):
162
max_origin_len = max(len(origin) for origin, text in annotations)
163
for origin, text in annotations:
164
if full or last_rev_id != origin:
168
to_file.write('%*s | %s' % (max_origin_len, this, text))
173
def _annotations(repo, file_id, rev_id):
174
"""Return the list of (origin_revision_id, line_text) for a revision of a file in a repository."""
175
annotations = repo.texts.annotate((file_id, rev_id))
177
return [(key[-1], line) for (key, line) in annotations]
180
def _expand_annotations(annotations, branch, current_rev=None):
181
"""Expand a file's annotations into command line UI ready tuples.
183
Each tuple includes detailed information, such as the author name, and date
184
string for the commit, rather than just the revision id.
186
:param annotations: The annotations to expand.
187
:param revision_id_to_revno: A map from id to revision numbers.
188
:param branch: A locked branch to query for revision details.
190
repository = branch.repository
191
if current_rev is not None:
192
# This can probably become a function on MutableTree, get_revno_map there,
194
last_revision = current_rev.revision_id
195
# XXX: Partially Cloned from branch, uses the old_get_graph, eep.
196
# XXX: The main difficulty is that we need to inject a single new node
197
# (current_rev) into the graph before it gets numbered, etc.
198
# Once KnownGraph gets an 'add_node()' function, we can use
199
# VF.get_known_graph_ancestry().
200
graph = repository.get_graph()
201
revision_graph = dict(((key, value) for key, value in
202
graph.iter_ancestry(current_rev.parent_ids) if value is not None))
203
revision_graph = _strip_NULL_ghosts(revision_graph)
204
revision_graph[last_revision] = current_rev.parent_ids
205
merge_sorted_revisions = tsort.merge_sort(
210
revision_id_to_revno = dict((rev_id, revno)
211
for seq_num, rev_id, depth, revno, end_of_merge in
212
merge_sorted_revisions)
214
revision_id_to_revno = branch.get_revision_id_to_revno_map()
36
rh = branch.revision_history()
37
w = branch.weave_store.get_weave(file_id, branch.get_transaction())
215
38
last_origin = None
216
revision_ids = set(o for o, t in annotations)
218
if CURRENT_REVISION in revision_ids:
219
revision_id_to_revno[CURRENT_REVISION] = (
220
"%d?" % (branch.revno() + 1),)
221
revisions[CURRENT_REVISION] = current_rev
222
revision_ids = [o for o in revision_ids if
223
repository.has_revision(o)]
224
revisions.update((r.revision_id, r) for r in
225
repository.get_revisions(revision_ids))
226
for origin, text in annotations:
39
for origin, text in w.annotate_iter(rev_id):
227
40
text = text.rstrip('\r\n')
228
41
if origin == last_origin:
229
(revno_str, author, date_str) = ('','','')
231
44
last_origin = origin
232
if origin not in revisions:
233
(revno_str, author, date_str) = ('?','?','?')
45
line_rev_id = w.idx_to_name(origin)
46
if not branch.has_revision(line_rev_id):
235
revno_str = '.'.join(str(i) for i in
236
revision_id_to_revno[origin])
237
rev = revisions[origin]
50
revno_str = str(rh.index(line_rev_id) + 1)
53
rev = branch.get_revision(line_rev_id)
238
54
tz = rev.timezone or 0
239
date_str = time.strftime('%Y%m%d',
55
date_str = time.strftime('%Y%m%d',
240
56
time.gmtime(rev.timestamp + tz))
241
57
# a lazy way to get something like the email address
242
58
# TODO: Get real email address
243
author = rev.get_apparent_authors()[0]
245
author = extract_email_address(author)
246
except errors.NoEmailInUsername:
247
pass # use the whole name
248
yield (revno_str, author, date_str, origin, text)
251
def reannotate(parents_lines, new_lines, new_revision_id,
252
_left_matching_blocks=None,
253
heads_provider=None):
254
"""Create a new annotated version from new lines and parent annotations.
256
:param parents_lines: List of annotated lines for all parents
257
:param new_lines: The un-annotated new lines
258
:param new_revision_id: The revision-id to associate with new lines
259
(will often be CURRENT_REVISION)
260
:param left_matching_blocks: a hint about which areas are common
261
between the text and its left-hand-parent. The format is
262
the SequenceMatcher.get_matching_blocks format
263
(start_left, start_right, length_of_match).
264
:param heads_provider: An object which provides a .heads() call to resolve
265
if any revision ids are children of others.
266
If None, then any ancestry disputes will be resolved with
269
if len(parents_lines) == 0:
270
lines = [(new_revision_id, line) for line in new_lines]
271
elif len(parents_lines) == 1:
272
lines = _reannotate(parents_lines[0], new_lines, new_revision_id,
273
_left_matching_blocks)
274
elif len(parents_lines) == 2:
275
left = _reannotate(parents_lines[0], new_lines, new_revision_id,
276
_left_matching_blocks)
277
lines = _reannotate_annotated(parents_lines[1], new_lines,
278
new_revision_id, left,
281
reannotations = [_reannotate(parents_lines[0], new_lines,
282
new_revision_id, _left_matching_blocks)]
283
reannotations.extend(_reannotate(p, new_lines, new_revision_id)
284
for p in parents_lines[1:])
286
for annos in zip(*reannotations):
287
origins = set(a for a, l in annos)
288
if len(origins) == 1:
289
# All the parents agree, so just return the first one
290
lines.append(annos[0])
293
if len(origins) == 2 and new_revision_id in origins:
294
origins.remove(new_revision_id)
295
if len(origins) == 1:
296
lines.append((origins.pop(), line))
298
lines.append((new_revision_id, line))
302
def _reannotate(parent_lines, new_lines, new_revision_id,
303
matching_blocks=None):
305
if matching_blocks is None:
306
plain_parent_lines = [l for r, l in parent_lines]
307
matcher = patiencediff.PatienceSequenceMatcher(None,
308
plain_parent_lines, new_lines)
309
matching_blocks = matcher.get_matching_blocks()
311
for i, j, n in matching_blocks:
312
for line in new_lines[new_cur:j]:
313
lines.append((new_revision_id, line))
314
lines.extend(parent_lines[i:i+n])
319
def _get_matching_blocks(old, new):
320
matcher = patiencediff.PatienceSequenceMatcher(None, old, new)
321
return matcher.get_matching_blocks()
324
_break_annotation_tie = None
326
def _old_break_annotation_tie(annotated_lines):
327
"""Chose an attribution between several possible ones.
329
:param annotated_lines: A list of tuples ((file_id, rev_id), line) where
330
the lines are identical but the revids different while no parent
331
relation exist between them
333
:return : The "winning" line. This must be one with a revid that
334
guarantees that further criss-cross merges will converge. Failing to
335
do so have performance implications.
337
# sort lexicographically so that we always get a stable result.
339
# TODO: while 'sort' is the easiest (and nearly the only possible solution)
340
# with the current implementation, chosing the oldest revision is known to
341
# provide better results (as in matching user expectations). The most
342
# common use case being manual cherry-pick from an already existing
344
return sorted(annotated_lines)[0]
347
def _find_matching_unannotated_lines(output_lines, plain_child_lines,
348
child_lines, start_child, end_child,
349
right_lines, start_right, end_right,
350
heads_provider, revision_id):
351
"""Find lines in plain_right_lines that match the existing lines.
353
:param output_lines: Append final annotated lines to this list
354
:param plain_child_lines: The unannotated new lines for the child text
355
:param child_lines: Lines for the child text which have been annotated
358
:param start_child: Position in plain_child_lines and child_lines to start
360
:param end_child: Last position in plain_child_lines and child_lines to
362
:param right_lines: The annotated lines for the whole text for the right
364
:param start_right: Position in right_lines to start the match
365
:param end_right: Last position in right_lines to search for a match
366
:param heads_provider: When parents disagree on the lineage of a line, we
367
need to check if one side supersedes the other
368
:param revision_id: The label to give if a line should be labeled 'tip'
370
output_extend = output_lines.extend
371
output_append = output_lines.append
372
# We need to see if any of the unannotated lines match
373
plain_right_subset = [l for a,l in right_lines[start_right:end_right]]
374
plain_child_subset = plain_child_lines[start_child:end_child]
375
match_blocks = _get_matching_blocks(plain_right_subset, plain_child_subset)
379
for right_idx, child_idx, match_len in match_blocks:
380
# All the lines that don't match are just passed along
381
if child_idx > last_child_idx:
382
output_extend(child_lines[start_child + last_child_idx
383
:start_child + child_idx])
384
for offset in xrange(match_len):
385
left = child_lines[start_child+child_idx+offset]
386
right = right_lines[start_right+right_idx+offset]
387
if left[0] == right[0]:
388
# The annotations match, just return the left one
390
elif left[0] == revision_id:
391
# The left parent marked this as unmatched, so let the
392
# right parent claim it
395
# Left and Right both claim this line
396
if heads_provider is None:
397
output_append((revision_id, left[1]))
399
heads = heads_provider.heads((left[0], right[0]))
401
output_append((iter(heads).next(), left[1]))
403
# Both claim different origins, get a stable result.
404
# If the result is not stable, there is a risk a
405
# performance degradation as criss-cross merges will
406
# flip-flop the attribution.
407
if _break_annotation_tie is None:
409
_old_break_annotation_tie([left, right]))
411
output_append(_break_annotation_tie([left, right]))
412
last_child_idx = child_idx + match_len
415
def _reannotate_annotated(right_parent_lines, new_lines, new_revision_id,
416
annotated_lines, heads_provider):
417
"""Update the annotations for a node based on another parent.
419
:param right_parent_lines: A list of annotated lines for the right-hand
421
:param new_lines: The unannotated new lines.
422
:param new_revision_id: The revision_id to attribute to lines which are not
423
present in either parent.
424
:param annotated_lines: A list of annotated lines. This should be the
425
annotation of new_lines based on parents seen so far.
426
:param heads_provider: When parents disagree on the lineage of a line, we
427
need to check if one side supersedes the other.
429
if len(new_lines) != len(annotated_lines):
430
raise AssertionError("mismatched new_lines and annotated_lines")
431
# First compare the newly annotated lines with the right annotated lines.
432
# Lines which were not changed in left or right should match. This tends to
433
# be the bulk of the lines, and they will need no further processing.
435
lines_extend = lines.extend
436
last_right_idx = 0 # The line just after the last match from the right side
438
matching_left_and_right = _get_matching_blocks(right_parent_lines,
440
for right_idx, left_idx, match_len in matching_left_and_right:
441
# annotated lines from last_left_idx to left_idx did not match the
442
# lines from last_right_idx to right_idx, the raw lines should be
443
# compared to determine what annotations need to be updated
444
if last_right_idx == right_idx or last_left_idx == left_idx:
445
# One of the sides is empty, so this is a pure insertion
446
lines_extend(annotated_lines[last_left_idx:left_idx])
448
# We need to see if any of the unannotated lines match
449
_find_matching_unannotated_lines(lines,
450
new_lines, annotated_lines,
451
last_left_idx, left_idx,
453
last_right_idx, right_idx,
456
last_right_idx = right_idx + match_len
457
last_left_idx = left_idx + match_len
458
# If left and right agree on a range, just push that into the output
459
lines_extend(annotated_lines[left_idx:left_idx + match_len])
464
from bzrlib._annotator_pyx import Annotator
465
except ImportError, e:
466
osutils.failed_to_load_extension(e)
467
from bzrlib._annotator_py import Annotator
61
author = author[:author.index('@')]
63
anno = '%5s %-12s %8s' % (revno_str, author, date_str)
64
print '%-27.27s | %s' % (anno, text)
67
if __name__ == '__main__':
68
from bzrlib.branch import Branch
69
from bzrlib.trace import enable_default_logging
71
enable_default_logging()
72
b = Branch.open_containing(sys.argv[1])
73
rp = b.relpath(sys.argv[1])
74
tree = b.revision_tree(b.last_revision())
75
file_id = tree.inventory.path2id(rp)
76
file_version = tree.inventory[file_id].revision
77
annotate_file(b, file_version, file_id, sys.stdout)
added
removed
bzrlib/annotate.py
