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(), """
33
31
from bzrlib import (
42
37
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)))
55
40
def annotate_file(branch, rev_id, file_id, verbose=False, full=False,
56
41
to_file=None, show_ids=False):
57
42
"""Annotate file_id at revision rev_id in branch.
69
54
: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
56
if to_file is None:
94
57
to_file = sys.stdout
96
59
# 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()
107
current_rev.committer = branch.get_config().username()
108
current_rev.message = "?"
109
current_rev.timestamp = round(time.time(), 3)
110
current_rev.timezone = osutils.local_time_offset()
113
annotation = list(_expand_annotations(annotations, branch,
115
_print_annotations(annotation, verbose, to_file, full)
118
def _print_annotations(annotation, verbose, to_file, full):
119
"""Print annotations to to_file.
121
:param to_file: The file to output the annotation to.
122
:param verbose: Show all details rather than truncating to ensure
123
reasonable text width.
124
:param full: XXXX Not sure what this does.
62
annotations = _annotations(branch.repository, file_id, rev_id)
63
max_origin_len = max(len(origin) for origin, text in annotations)
64
for origin, text in annotations:
65
if full or last_rev_id != origin:
69
to_file.write('%*s | %s' % (max_origin_len, this, text))
73
# Calculate the lengths of the various columns
74
annotation = list(_annotate_file(branch, rev_id, file_id))
126
75
if len(annotation) == 0:
127
76
max_origin_len = max_revno_len = max_revid_len = 0
164
def _show_id_annotations(annotations, to_file, full):
168
max_origin_len = max(len(origin) for origin, text in annotations)
169
for origin, text in annotations:
170
if full or last_rev_id != origin:
174
to_file.write('%*s | %s' % (max_origin_len, this, text))
179
def _expand_annotations(annotations, branch, current_rev=None):
180
"""Expand a file's annotations into command line UI ready tuples.
182
Each tuple includes detailed information, such as the author name, and date
183
string for the commit, rather than just the revision id.
185
:param annotations: The annotations to expand.
186
:param revision_id_to_revno: A map from id to revision numbers.
187
:param branch: A locked branch to query for revision details.
113
def _annotations(repo, file_id, rev_id):
114
"""Return the list of (origin,text) for a revision of a file in a repository."""
115
w = repo.weave_store.get_weave(file_id, repo.get_transaction())
116
return list(w.annotate_iter(rev_id))
119
def _annotate_file(branch, rev_id, file_id):
120
"""Yield the origins for each line of a file.
122
This includes detailed information, such as the author name, and
123
date string for the commit, rather than just the revision id.
189
repository = branch.repository
190
if current_rev is not None:
191
# This can probably become a function on MutableTree, get_revno_map
192
# there, or something.
193
last_revision = current_rev.revision_id
194
# XXX: Partially Cloned from branch, uses the old_get_graph, eep.
195
# XXX: The main difficulty is that we need to inject a single new node
196
# (current_rev) into the graph before it gets numbered, etc.
197
# Once KnownGraph gets an 'add_node()' function, we can use
198
# VF.get_known_graph_ancestry().
199
graph = repository.get_graph()
200
revision_graph = dict(((key, value) for key, value in
201
graph.iter_ancestry(current_rev.parent_ids) if value is not None))
202
revision_graph = _strip_NULL_ghosts(revision_graph)
203
revision_graph[last_revision] = current_rev.parent_ids
204
merge_sorted_revisions = tsort.merge_sort(
209
revision_id_to_revno = dict((rev_id, revno)
210
for seq_num, rev_id, depth, revno, end_of_merge in
211
merge_sorted_revisions)
213
revision_id_to_revno = branch.get_revision_id_to_revno_map()
125
revision_id_to_revno = branch.get_revision_id_to_revno_map()
126
annotations = _annotations(branch.repository, file_id, rev_id)
214
127
last_origin = None
215
128
revision_ids = set(o for o, t in annotations)
217
if CURRENT_REVISION in revision_ids:
218
revision_id_to_revno[CURRENT_REVISION] = (
219
"%d?" % (branch.revno() + 1),)
220
revisions[CURRENT_REVISION] = current_rev
221
revision_ids = [o for o in revision_ids if
222
repository.has_revision(o)]
223
revisions.update((r.revision_id, r) for r in
224
repository.get_revisions(revision_ids))
129
revision_ids = [o for o in revision_ids if
130
branch.repository.has_revision(o)]
131
revisions = dict((r.revision_id, r) for r in
132
branch.repository.get_revisions(revision_ids))
225
133
for origin, text in annotations:
226
134
text = text.rstrip('\r\n')
227
135
if origin == last_origin:
250
158
def reannotate(parents_lines, new_lines, new_revision_id,
251
_left_matching_blocks=None,
252
heads_provider=None):
159
_left_matching_blocks=None):
253
160
"""Create a new annotated version from new lines and parent annotations.
255
162
:param parents_lines: List of annotated lines for all parents
256
163
:param new_lines: The un-annotated new lines
257
164
:param new_revision_id: The revision-id to associate with new lines
258
165
(will often be CURRENT_REVISION)
259
166
:param left_matching_blocks: a hint about which areas are common
260
167
between the text and its left-hand-parent. The format is
261
the SequenceMatcher.get_matching_blocks format
262
(start_left, start_right, length_of_match).
263
:param heads_provider: An object which provides a .heads() call to resolve
264
if any revision ids are children of others.
265
If None, then any ancestry disputes will be resolved with
168
the SequenceMatcher.get_matching_blocks format.
268
170
if len(parents_lines) == 0:
269
lines = [(new_revision_id, line) for line in new_lines]
171
for line in new_lines:
172
yield new_revision_id, line
270
173
elif len(parents_lines) == 1:
271
lines = _reannotate(parents_lines[0], new_lines, new_revision_id,
272
_left_matching_blocks)
273
elif len(parents_lines) == 2:
274
left = _reannotate(parents_lines[0], new_lines, new_revision_id,
275
_left_matching_blocks)
276
lines = _reannotate_annotated(parents_lines[1], new_lines,
277
new_revision_id, left,
174
for data in _reannotate(parents_lines[0], new_lines, new_revision_id,
175
_left_matching_blocks):
280
reannotations = [_reannotate(parents_lines[0], new_lines,
281
new_revision_id, _left_matching_blocks)]
282
reannotations.extend(_reannotate(p, new_lines, new_revision_id)
283
for p in parents_lines[1:])
178
block_list = [_left_matching_blocks] + [None] * len(parents_lines)
179
reannotations = [list(_reannotate(p, new_lines, new_revision_id, b))
180
for p, b in zip(parents_lines, block_list)]
285
181
for annos in zip(*reannotations):
286
182
origins = set(a for a, l in annos)
287
184
if len(origins) == 1:
288
# All the parents agree, so just return the first one
289
lines.append(annos[0])
185
yield iter(origins).next(), line
186
elif len(origins) == 2 and new_revision_id in origins:
187
yield (x for x in origins if x != new_revision_id).next(), line
292
if len(origins) == 2 and new_revision_id in origins:
293
origins.remove(new_revision_id)
294
if len(origins) == 1:
295
lines.append((origins.pop(), line))
297
lines.append((new_revision_id, line))
189
yield new_revision_id, line
301
192
def _reannotate(parent_lines, new_lines, new_revision_id,
306
197
matcher = patiencediff.PatienceSequenceMatcher(None,
307
198
plain_parent_lines, new_lines)
308
199
matching_blocks = matcher.get_matching_blocks()
310
200
for i, j, n in matching_blocks:
311
201
for line in new_lines[new_cur:j]:
312
lines.append((new_revision_id, line))
313
lines.extend(parent_lines[i:i+n])
202
yield new_revision_id, line
203
for data in parent_lines[i:i+n]:
318
def _get_matching_blocks(old, new):
319
matcher = patiencediff.PatienceSequenceMatcher(None, old, new)
320
return matcher.get_matching_blocks()
323
_break_annotation_tie = None
325
def _old_break_annotation_tie(annotated_lines):
326
"""Chose an attribution between several possible ones.
328
:param annotated_lines: A list of tuples ((file_id, rev_id), line) where
329
the lines are identical but the revids different while no parent
330
relation exist between them
332
:return : The "winning" line. This must be one with a revid that
333
guarantees that further criss-cross merges will converge. Failing to
334
do so have performance implications.
336
# sort lexicographically so that we always get a stable result.
338
# TODO: while 'sort' is the easiest (and nearly the only possible solution)
339
# with the current implementation, chosing the oldest revision is known to
340
# provide better results (as in matching user expectations). The most
341
# common use case being manual cherry-pick from an already existing
343
return sorted(annotated_lines)[0]
346
def _find_matching_unannotated_lines(output_lines, plain_child_lines,
347
child_lines, start_child, end_child,
348
right_lines, start_right, end_right,
349
heads_provider, revision_id):
350
"""Find lines in plain_right_lines that match the existing lines.
352
:param output_lines: Append final annotated lines to this list
353
:param plain_child_lines: The unannotated new lines for the child text
354
:param child_lines: Lines for the child text which have been annotated
357
:param start_child: Position in plain_child_lines and child_lines to start
359
:param end_child: Last position in plain_child_lines and child_lines to
361
:param right_lines: The annotated lines for the whole text for the right
363
:param start_right: Position in right_lines to start the match
364
:param end_right: Last position in right_lines to search for a match
365
:param heads_provider: When parents disagree on the lineage of a line, we
366
need to check if one side supersedes the other
367
:param revision_id: The label to give if a line should be labeled 'tip'
369
output_extend = output_lines.extend
370
output_append = output_lines.append
371
# We need to see if any of the unannotated lines match
372
plain_right_subset = [l for a,l in right_lines[start_right:end_right]]
373
plain_child_subset = plain_child_lines[start_child:end_child]
374
match_blocks = _get_matching_blocks(plain_right_subset, plain_child_subset)
378
for right_idx, child_idx, match_len in match_blocks:
379
# All the lines that don't match are just passed along
380
if child_idx > last_child_idx:
381
output_extend(child_lines[start_child + last_child_idx
382
:start_child + child_idx])
383
for offset in xrange(match_len):
384
left = child_lines[start_child+child_idx+offset]
385
right = right_lines[start_right+right_idx+offset]
386
if left[0] == right[0]:
387
# The annotations match, just return the left one
389
elif left[0] == revision_id:
390
# The left parent marked this as unmatched, so let the
391
# right parent claim it
394
# Left and Right both claim this line
395
if heads_provider is None:
396
output_append((revision_id, left[1]))
398
heads = heads_provider.heads((left[0], right[0]))
400
output_append((iter(heads).next(), left[1]))
402
# Both claim different origins, get a stable result.
403
# If the result is not stable, there is a risk a
404
# performance degradation as criss-cross merges will
405
# flip-flop the attribution.
406
if _break_annotation_tie is None:
408
_old_break_annotation_tie([left, right]))
410
output_append(_break_annotation_tie([left, right]))
411
last_child_idx = child_idx + match_len
414
def _reannotate_annotated(right_parent_lines, new_lines, new_revision_id,
415
annotated_lines, heads_provider):
416
"""Update the annotations for a node based on another parent.
418
:param right_parent_lines: A list of annotated lines for the right-hand
420
:param new_lines: The unannotated new lines.
421
:param new_revision_id: The revision_id to attribute to lines which are not
422
present in either parent.
423
:param annotated_lines: A list of annotated lines. This should be the
424
annotation of new_lines based on parents seen so far.
425
:param heads_provider: When parents disagree on the lineage of a line, we
426
need to check if one side supersedes the other.
428
if len(new_lines) != len(annotated_lines):
429
raise AssertionError("mismatched new_lines and annotated_lines")
430
# First compare the newly annotated lines with the right annotated lines.
431
# Lines which were not changed in left or right should match. This tends to
432
# be the bulk of the lines, and they will need no further processing.
434
lines_extend = lines.extend
435
last_right_idx = 0 # The line just after the last match from the right side
437
matching_left_and_right = _get_matching_blocks(right_parent_lines,
439
for right_idx, left_idx, match_len in matching_left_and_right:
440
# annotated lines from last_left_idx to left_idx did not match the
441
# lines from last_right_idx to right_idx, the raw lines should be
442
# compared to determine what annotations need to be updated
443
if last_right_idx == right_idx or last_left_idx == left_idx:
444
# One of the sides is empty, so this is a pure insertion
445
lines_extend(annotated_lines[last_left_idx:left_idx])
447
# We need to see if any of the unannotated lines match
448
_find_matching_unannotated_lines(lines,
449
new_lines, annotated_lines,
450
last_left_idx, left_idx,
452
last_right_idx, right_idx,
455
last_right_idx = right_idx + match_len
456
last_left_idx = left_idx + match_len
457
# If left and right agree on a range, just push that into the output
458
lines_extend(annotated_lines[left_idx:left_idx + match_len])
463
from bzrlib._annotator_pyx import Annotator
464
except ImportError, e:
465
osutils.failed_to_load_extension(e)
466
from bzrlib._annotator_py import Annotator