1
# Copyright (C) 2004, 2005 by Canonical Ltd
1
# Copyright (C) 2005-2010 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
17
17
"""File annotate based on weave storage"""
19
from __future__ import absolute_import
19
21
# TODO: Choice of more or less verbose formats:
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: With --long, show entire email address, not just the first bit
27
27
# TODO: perhaps abbreviate timescales depending on how recent they are
28
# e.g. "3:12 Tue", "13 Oct", "Oct 2005", etc.
28
# e.g. "3:12 Tue", "13 Oct", "Oct 2005", etc.
33
from bzrlib.lazy_import import lazy_import
34
lazy_import(globals(), """
35
44
from bzrlib.config import extract_email_address
45
from bzrlib.repository import _strip_NULL_ghosts
46
from bzrlib.revision import (
50
from bzrlib.symbol_versioning import (
56
@deprecated_function(deprecated_in((2, 4, 0)))
38
57
def annotate_file(branch, rev_id, file_id, verbose=False, full=False,
58
to_file=None, show_ids=False):
59
"""Annotate file_id at revision rev_id in branch.
61
The branch should already be read_locked() when annotate_file is called.
63
:param branch: The branch to look for revision numbers and history from.
64
:param rev_id: The revision id to annotate.
65
:param file_id: The file_id to annotate.
66
:param verbose: Show all details rather than truncating to ensure
67
reasonable text width.
68
:param full: XXXX Not sure what this does.
69
:param to_file: The file to output the annotation to; if None stdout is
71
:param show_ids: Show revision ids in the annotation output.
73
tree = branch.repository.revision_tree(rev_id)
74
annotate_file_tree(tree, file_id, to_file, verbose=verbose,
75
full=full, show_ids=show_ids, branch=branch)
78
def annotate_file_tree(tree, file_id, to_file, verbose=False, full=False,
79
show_ids=False, branch=None):
80
"""Annotate file_id in a tree.
82
The tree should already be read_locked() when annotate_file_tree is called.
84
:param tree: The tree to look for revision numbers and history from.
85
:param file_id: The file_id to annotate.
86
:param to_file: The file to output the annotation to.
87
:param verbose: Show all details rather than truncating to ensure
88
reasonable text width.
89
:param full: XXXX Not sure what this does.
90
:param show_ids: Show revision ids in the annotation output.
91
:param branch: Branch to use for revision revno lookups
40
95
if to_file is None:
41
96
to_file = sys.stdout
44
for (revno_str, author, date_str, line_rev_id, text ) in \
45
_annotate_file(branch, rev_id, file_id ):
98
# Handle the show_ids case
99
annotations = list(tree.annotate_iter(file_id))
101
return _show_id_annotations(annotations, to_file, full)
103
if not getattr(tree, "get_revision_id", False):
104
# Create a virtual revision to represent the current tree state.
105
# Should get some more pending commit attributes, like pending tags,
107
current_rev = Revision(CURRENT_REVISION)
108
current_rev.parent_ids = tree.get_parent_ids()
110
current_rev.committer = branch.get_config_stack().get('email')
111
except errors.NoWhoami:
112
current_rev.committer = 'local user'
113
current_rev.message = "?"
114
current_rev.timestamp = round(time.time(), 3)
115
current_rev.timezone = osutils.local_time_offset()
118
annotation = list(_expand_annotations(annotations, branch,
120
_print_annotations(annotation, verbose, to_file, full)
123
def _print_annotations(annotation, verbose, to_file, full):
124
"""Print annotations to to_file.
126
:param to_file: The file to output the annotation to.
127
:param verbose: Show all details rather than truncating to ensure
128
reasonable text width.
129
:param full: XXXX Not sure what this does.
131
if len(annotation) == 0:
132
max_origin_len = max_revno_len = max_revid_len = 0
134
max_origin_len = max(len(x[1]) for x in annotation)
135
max_revno_len = max(len(x[0]) for x in annotation)
136
max_revid_len = max(len(x[3]) for x in annotation)
138
max_revno_len = min(max_revno_len, 12)
139
max_revno_len = max(max_revno_len, 3)
141
# Output the annotations
143
encoding = getattr(to_file, 'encoding', None) or \
144
osutils.get_terminal_encoding()
145
for (revno_str, author, date_str, line_rev_id, text) in annotation:
48
anno = '%5s %-12s %8s ' % (revno_str, author[:12], date_str)
50
anno = "%5s %-7s " % ( revno_str, author[:7] )
52
if anno.lstrip() == "" and full: anno = prevanno
53
print >>to_file, '%s| %s' % (anno, text)
56
def _annotate_file(branch, rev_id, file_id ):
58
rh = branch.revision_history()
59
w = branch.weave_store.get_weave(file_id, branch.get_transaction())
147
anno = '%-*s %-*s %8s ' % (max_revno_len, revno_str,
148
max_origin_len, author, date_str)
150
if len(revno_str) > max_revno_len:
151
revno_str = revno_str[:max_revno_len-1] + '>'
152
anno = "%-*s %-7s " % (max_revno_len, revno_str, author[:7])
153
if anno.lstrip() == "" and full:
157
except UnicodeEncodeError:
158
# cmd_annotate should be passing in an 'exact' object, which means
159
# we have a direct handle to sys.stdout or equivalent. It may not
160
# be able to handle the exact Unicode characters, but 'annotate' is
161
# a user function (non-scripting), so shouldn't die because of
162
# unrepresentable annotation characters. So encode using 'replace',
163
# and write them again.
164
to_file.write(anno.encode(encoding, 'replace'))
165
to_file.write('| %s\n' % (text,))
169
def _show_id_annotations(annotations, to_file, full):
173
max_origin_len = max(len(origin) for origin, text in annotations)
174
for origin, text in annotations:
175
if full or last_rev_id != origin:
179
to_file.write('%*s | %s' % (max_origin_len, this, text))
184
def _expand_annotations(annotations, branch, current_rev=None):
185
"""Expand a file's annotations into command line UI ready tuples.
187
Each tuple includes detailed information, such as the author name, and date
188
string for the commit, rather than just the revision id.
190
:param annotations: The annotations to expand.
191
:param revision_id_to_revno: A map from id to revision numbers.
192
:param branch: A locked branch to query for revision details.
194
repository = branch.repository
195
if current_rev is not None:
196
# This can probably become a function on MutableTree, get_revno_map
197
# there, or something.
198
last_revision = current_rev.revision_id
199
# XXX: Partially Cloned from branch, uses the old_get_graph, eep.
200
# XXX: The main difficulty is that we need to inject a single new node
201
# (current_rev) into the graph before it gets numbered, etc.
202
# Once KnownGraph gets an 'add_node()' function, we can use
203
# VF.get_known_graph_ancestry().
204
graph = repository.get_graph()
205
revision_graph = dict(((key, value) for key, value in
206
graph.iter_ancestry(current_rev.parent_ids) if value is not None))
207
revision_graph = _strip_NULL_ghosts(revision_graph)
208
revision_graph[last_revision] = current_rev.parent_ids
209
merge_sorted_revisions = tsort.merge_sort(
214
revision_id_to_revno = dict((rev_id, revno)
215
for seq_num, rev_id, depth, revno, end_of_merge in
216
merge_sorted_revisions)
218
revision_id_to_revno = branch.get_revision_id_to_revno_map()
60
219
last_origin = None
61
for origin, text in w.annotate_iter(rev_id):
220
revision_ids = set(o for o, t in annotations)
222
if CURRENT_REVISION in revision_ids:
223
revision_id_to_revno[CURRENT_REVISION] = (
224
"%d?" % (branch.revno() + 1),)
225
revisions[CURRENT_REVISION] = current_rev
226
revision_ids = [o for o in revision_ids if
227
repository.has_revision(o)]
228
revisions.update((r.revision_id, r) for r in
229
repository.get_revisions(revision_ids))
230
for origin, text in annotations:
62
231
text = text.rstrip('\r\n')
63
232
if origin == last_origin:
64
233
(revno_str, author, date_str) = ('','','')
66
235
last_origin = origin
67
line_rev_id = w.idx_to_name(origin)
68
if not branch.has_revision(line_rev_id):
236
if origin not in revisions:
69
237
(revno_str, author, date_str) = ('?','?','?')
72
revno_str = str(rh.index(line_rev_id) + 1)
75
rev = branch.get_revision(line_rev_id)
239
revno_str = '.'.join(str(i) for i in
240
revision_id_to_revno[origin])
241
rev = revisions[origin]
76
242
tz = rev.timezone or 0
77
date_str = time.strftime('%Y%m%d',
243
date_str = time.strftime('%Y%m%d',
78
244
time.gmtime(rev.timestamp + tz))
79
245
# a lazy way to get something like the email address
80
246
# TODO: Get real email address
81
author = rev.committer
247
author = rev.get_apparent_authors()[0]
83
249
author = extract_email_address(author)
250
except errors.NoEmailInUsername:
85
251
pass # use the whole name
86
yield (revno_str, author, date_str, line_rev_id, text)
89
if __name__ == '__main__':
90
from bzrlib.branch import Branch
91
from bzrlib.trace import enable_default_logging
92
from bzrlib.workingtree import WorkingTree
94
enable_default_logging()
95
b = Branch.open_containing(sys.argv[1])[0]
96
tree = WorkingTree(b.base, b)
97
rp = tree.relpath(sys.argv[1])
98
tree = b.revision_tree(b.last_revision())
99
file_id = tree.inventory.path2id(rp)
100
file_version = tree.inventory[file_id].revision
101
annotate_file(b, file_version, file_id, to_file = sys.stdout)
252
yield (revno_str, author, date_str, origin, text)
255
def reannotate(parents_lines, new_lines, new_revision_id,
256
_left_matching_blocks=None,
257
heads_provider=None):
258
"""Create a new annotated version from new lines and parent annotations.
260
:param parents_lines: List of annotated lines for all parents
261
:param new_lines: The un-annotated new lines
262
:param new_revision_id: The revision-id to associate with new lines
263
(will often be CURRENT_REVISION)
264
:param left_matching_blocks: a hint about which areas are common
265
between the text and its left-hand-parent. The format is
266
the SequenceMatcher.get_matching_blocks format
267
(start_left, start_right, length_of_match).
268
:param heads_provider: An object which provides a .heads() call to resolve
269
if any revision ids are children of others.
270
If None, then any ancestry disputes will be resolved with
273
if len(parents_lines) == 0:
274
lines = [(new_revision_id, line) for line in new_lines]
275
elif len(parents_lines) == 1:
276
lines = _reannotate(parents_lines[0], new_lines, new_revision_id,
277
_left_matching_blocks)
278
elif len(parents_lines) == 2:
279
left = _reannotate(parents_lines[0], new_lines, new_revision_id,
280
_left_matching_blocks)
281
lines = _reannotate_annotated(parents_lines[1], new_lines,
282
new_revision_id, left,
285
reannotations = [_reannotate(parents_lines[0], new_lines,
286
new_revision_id, _left_matching_blocks)]
287
reannotations.extend(_reannotate(p, new_lines, new_revision_id)
288
for p in parents_lines[1:])
290
for annos in zip(*reannotations):
291
origins = set(a for a, l in annos)
292
if len(origins) == 1:
293
# All the parents agree, so just return the first one
294
lines.append(annos[0])
297
if len(origins) == 2 and new_revision_id in origins:
298
origins.remove(new_revision_id)
299
if len(origins) == 1:
300
lines.append((origins.pop(), line))
302
lines.append((new_revision_id, line))
306
def _reannotate(parent_lines, new_lines, new_revision_id,
307
matching_blocks=None):
309
if matching_blocks is None:
310
plain_parent_lines = [l for r, l in parent_lines]
311
matcher = patiencediff.PatienceSequenceMatcher(None,
312
plain_parent_lines, new_lines)
313
matching_blocks = matcher.get_matching_blocks()
315
for i, j, n in matching_blocks:
316
for line in new_lines[new_cur:j]:
317
lines.append((new_revision_id, line))
318
lines.extend(parent_lines[i:i+n])
323
def _get_matching_blocks(old, new):
324
matcher = patiencediff.PatienceSequenceMatcher(None, old, new)
325
return matcher.get_matching_blocks()
328
_break_annotation_tie = None
330
def _old_break_annotation_tie(annotated_lines):
331
"""Chose an attribution between several possible ones.
333
:param annotated_lines: A list of tuples ((file_id, rev_id), line) where
334
the lines are identical but the revids different while no parent
335
relation exist between them
337
:return : The "winning" line. This must be one with a revid that
338
guarantees that further criss-cross merges will converge. Failing to
339
do so have performance implications.
341
# sort lexicographically so that we always get a stable result.
343
# TODO: while 'sort' is the easiest (and nearly the only possible solution)
344
# with the current implementation, chosing the oldest revision is known to
345
# provide better results (as in matching user expectations). The most
346
# common use case being manual cherry-pick from an already existing
348
return sorted(annotated_lines)[0]
351
def _find_matching_unannotated_lines(output_lines, plain_child_lines,
352
child_lines, start_child, end_child,
353
right_lines, start_right, end_right,
354
heads_provider, revision_id):
355
"""Find lines in plain_right_lines that match the existing lines.
357
:param output_lines: Append final annotated lines to this list
358
:param plain_child_lines: The unannotated new lines for the child text
359
:param child_lines: Lines for the child text which have been annotated
362
:param start_child: Position in plain_child_lines and child_lines to start
364
:param end_child: Last position in plain_child_lines and child_lines to
366
:param right_lines: The annotated lines for the whole text for the right
368
:param start_right: Position in right_lines to start the match
369
:param end_right: Last position in right_lines to search for a match
370
:param heads_provider: When parents disagree on the lineage of a line, we
371
need to check if one side supersedes the other
372
:param revision_id: The label to give if a line should be labeled 'tip'
374
output_extend = output_lines.extend
375
output_append = output_lines.append
376
# We need to see if any of the unannotated lines match
377
plain_right_subset = [l for a,l in right_lines[start_right:end_right]]
378
plain_child_subset = plain_child_lines[start_child:end_child]
379
match_blocks = _get_matching_blocks(plain_right_subset, plain_child_subset)
383
for right_idx, child_idx, match_len in match_blocks:
384
# All the lines that don't match are just passed along
385
if child_idx > last_child_idx:
386
output_extend(child_lines[start_child + last_child_idx
387
:start_child + child_idx])
388
for offset in xrange(match_len):
389
left = child_lines[start_child+child_idx+offset]
390
right = right_lines[start_right+right_idx+offset]
391
if left[0] == right[0]:
392
# The annotations match, just return the left one
394
elif left[0] == revision_id:
395
# The left parent marked this as unmatched, so let the
396
# right parent claim it
399
# Left and Right both claim this line
400
if heads_provider is None:
401
output_append((revision_id, left[1]))
403
heads = heads_provider.heads((left[0], right[0]))
405
output_append((iter(heads).next(), left[1]))
407
# Both claim different origins, get a stable result.
408
# If the result is not stable, there is a risk a
409
# performance degradation as criss-cross merges will
410
# flip-flop the attribution.
411
if _break_annotation_tie is None:
413
_old_break_annotation_tie([left, right]))
415
output_append(_break_annotation_tie([left, right]))
416
last_child_idx = child_idx + match_len
419
def _reannotate_annotated(right_parent_lines, new_lines, new_revision_id,
420
annotated_lines, heads_provider):
421
"""Update the annotations for a node based on another parent.
423
:param right_parent_lines: A list of annotated lines for the right-hand
425
:param new_lines: The unannotated new lines.
426
:param new_revision_id: The revision_id to attribute to lines which are not
427
present in either parent.
428
:param annotated_lines: A list of annotated lines. This should be the
429
annotation of new_lines based on parents seen so far.
430
:param heads_provider: When parents disagree on the lineage of a line, we
431
need to check if one side supersedes the other.
433
if len(new_lines) != len(annotated_lines):
434
raise AssertionError("mismatched new_lines and annotated_lines")
435
# First compare the newly annotated lines with the right annotated lines.
436
# Lines which were not changed in left or right should match. This tends to
437
# be the bulk of the lines, and they will need no further processing.
439
lines_extend = lines.extend
440
last_right_idx = 0 # The line just after the last match from the right side
442
matching_left_and_right = _get_matching_blocks(right_parent_lines,
444
for right_idx, left_idx, match_len in matching_left_and_right:
445
# annotated lines from last_left_idx to left_idx did not match the
446
# lines from last_right_idx to right_idx, the raw lines should be
447
# compared to determine what annotations need to be updated
448
if last_right_idx == right_idx or last_left_idx == left_idx:
449
# One of the sides is empty, so this is a pure insertion
450
lines_extend(annotated_lines[last_left_idx:left_idx])
452
# We need to see if any of the unannotated lines match
453
_find_matching_unannotated_lines(lines,
454
new_lines, annotated_lines,
455
last_left_idx, left_idx,
457
last_right_idx, right_idx,
460
last_right_idx = right_idx + match_len
461
last_left_idx = left_idx + match_len
462
# If left and right agree on a range, just push that into the output
463
lines_extend(annotated_lines[left_idx:left_idx + match_len])
468
from bzrlib._annotator_pyx import Annotator
469
except ImportError, e:
470
osutils.failed_to_load_extension(e)
471
from bzrlib._annotator_py import Annotator
added
removed
bzrlib/annotate.py
