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
from __future__ import absolute_import
21
19
# TODO: Choice of more or less verbose formats:
23
21
# interposed: show more details between blocks of modified lines
25
23
# TODO: Show which revision caused a line to merge into the parent
27
25
# TODO: perhaps abbreviate timescales depending on how recent they are
28
# e.g. "3:12 Tue", "13 Oct", "Oct 2005", etc.
26
# e.g. "3:12 Tue", "13 Oct", "Oct 2005", etc.
33
from bzrlib.lazy_import import lazy_import
34
lazy_import(globals(), """
35
31
from bzrlib import (
44
37
from bzrlib.config import extract_email_address
45
from bzrlib.repository import _strip_NULL_ghosts
46
from bzrlib.revision import (
52
def annotate_file_tree(tree, file_id, to_file, verbose=False, full=False,
53
show_ids=False, branch=None):
54
"""Annotate file_id in a tree.
56
The tree should already be read_locked() when annotate_file_tree is called.
58
:param tree: The tree to look for revision numbers and history from.
40
def annotate_file(branch, rev_id, file_id, verbose=False, full=False,
41
to_file=None, show_ids=False):
42
"""Annotate file_id at revision rev_id in branch.
44
The branch should already be read_locked() when annotate_file is called.
46
:param branch: The branch to look for revision numbers and history from.
47
:param rev_id: The revision id to annotate.
59
48
:param file_id: The file_id to annotate.
60
:param to_file: The file to output the annotation to.
61
49
:param verbose: Show all details rather than truncating to ensure
62
50
reasonable text width.
63
51
:param full: XXXX Not sure what this does.
52
:param to_file: The file to output the annotation to; if None stdout is
64
54
:param show_ids: Show revision ids in the annotation output.
65
:param branch: Branch to use for revision revno lookups
69
56
if to_file is None:
70
57
to_file = sys.stdout
72
59
# Handle the show_ids case
73
annotations = list(tree.annotate_iter(file_id))
75
return _show_id_annotations(annotations, to_file, full)
77
if not getattr(tree, "get_revision_id", False):
78
# Create a virtual revision to represent the current tree state.
79
# Should get some more pending commit attributes, like pending tags,
81
current_rev = Revision(CURRENT_REVISION)
82
current_rev.parent_ids = tree.get_parent_ids()
84
current_rev.committer = branch.get_config_stack().get('email')
85
except errors.NoWhoami:
86
current_rev.committer = 'local user'
87
current_rev.message = "?"
88
current_rev.timestamp = round(time.time(), 3)
89
current_rev.timezone = osutils.local_time_offset()
92
annotation = list(_expand_annotations(annotations, branch,
94
_print_annotations(annotation, verbose, to_file, full)
97
def _print_annotations(annotation, verbose, to_file, full):
98
"""Print annotations to to_file.
100
:param to_file: The file to output the annotation to.
101
:param verbose: Show all details rather than truncating to ensure
102
reasonable text width.
103
: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))
105
75
if len(annotation) == 0:
106
76
max_origin_len = max_revno_len = max_revid_len = 0
143
def _show_id_annotations(annotations, to_file, full):
147
max_origin_len = max(len(origin) for origin, text in annotations)
148
for origin, text in annotations:
149
if full or last_rev_id != origin:
153
to_file.write('%*s | %s' % (max_origin_len, this, text))
158
def _expand_annotations(annotations, branch, current_rev=None):
159
"""Expand a file's annotations into command line UI ready tuples.
161
Each tuple includes detailed information, such as the author name, and date
162
string for the commit, rather than just the revision id.
164
:param annotations: The annotations to expand.
165
:param revision_id_to_revno: A map from id to revision numbers.
166
: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.
168
repository = branch.repository
169
if current_rev is not None:
170
# This can probably become a function on MutableTree, get_revno_map
171
# there, or something.
172
last_revision = current_rev.revision_id
173
# XXX: Partially Cloned from branch, uses the old_get_graph, eep.
174
# XXX: The main difficulty is that we need to inject a single new node
175
# (current_rev) into the graph before it gets numbered, etc.
176
# Once KnownGraph gets an 'add_node()' function, we can use
177
# VF.get_known_graph_ancestry().
178
graph = repository.get_graph()
179
revision_graph = dict(((key, value) for key, value in
180
graph.iter_ancestry(current_rev.parent_ids) if value is not None))
181
revision_graph = _strip_NULL_ghosts(revision_graph)
182
revision_graph[last_revision] = current_rev.parent_ids
183
merge_sorted_revisions = tsort.merge_sort(
188
revision_id_to_revno = dict((rev_id, revno)
189
for seq_num, rev_id, depth, revno, end_of_merge in
190
merge_sorted_revisions)
192
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)
193
127
last_origin = None
194
128
revision_ids = set(o for o, t in annotations)
196
if CURRENT_REVISION in revision_ids:
197
revision_id_to_revno[CURRENT_REVISION] = (
198
"%d?" % (branch.revno() + 1),)
199
revisions[CURRENT_REVISION] = current_rev
200
revision_ids = [o for o in revision_ids if
201
repository.has_revision(o)]
202
revisions.update((r.revision_id, r) for r in
203
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))
204
133
for origin, text in annotations:
205
134
text = text.rstrip('\r\n')
206
135
if origin == last_origin:
229
158
def reannotate(parents_lines, new_lines, new_revision_id,
230
_left_matching_blocks=None,
231
heads_provider=None):
159
_left_matching_blocks=None):
232
160
"""Create a new annotated version from new lines and parent annotations.
234
162
:param parents_lines: List of annotated lines for all parents
235
163
:param new_lines: The un-annotated new lines
236
164
:param new_revision_id: The revision-id to associate with new lines
237
165
(will often be CURRENT_REVISION)
238
166
:param left_matching_blocks: a hint about which areas are common
239
167
between the text and its left-hand-parent. The format is
240
the SequenceMatcher.get_matching_blocks format
241
(start_left, start_right, length_of_match).
242
:param heads_provider: An object which provides a .heads() call to resolve
243
if any revision ids are children of others.
244
If None, then any ancestry disputes will be resolved with
168
the SequenceMatcher.get_matching_blocks format.
247
170
if len(parents_lines) == 0:
248
lines = [(new_revision_id, line) for line in new_lines]
171
for line in new_lines:
172
yield new_revision_id, line
249
173
elif len(parents_lines) == 1:
250
lines = _reannotate(parents_lines[0], new_lines, new_revision_id,
251
_left_matching_blocks)
252
elif len(parents_lines) == 2:
253
left = _reannotate(parents_lines[0], new_lines, new_revision_id,
254
_left_matching_blocks)
255
lines = _reannotate_annotated(parents_lines[1], new_lines,
256
new_revision_id, left,
174
for data in _reannotate(parents_lines[0], new_lines, new_revision_id,
175
_left_matching_blocks):
259
reannotations = [_reannotate(parents_lines[0], new_lines,
260
new_revision_id, _left_matching_blocks)]
261
reannotations.extend(_reannotate(p, new_lines, new_revision_id)
262
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)]
264
181
for annos in zip(*reannotations):
265
182
origins = set(a for a, l in annos)
266
184
if len(origins) == 1:
267
# All the parents agree, so just return the first one
268
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
271
if len(origins) == 2 and new_revision_id in origins:
272
origins.remove(new_revision_id)
273
if len(origins) == 1:
274
lines.append((origins.pop(), line))
276
lines.append((new_revision_id, line))
189
yield new_revision_id, line
280
192
def _reannotate(parent_lines, new_lines, new_revision_id,
285
197
matcher = patiencediff.PatienceSequenceMatcher(None,
286
198
plain_parent_lines, new_lines)
287
199
matching_blocks = matcher.get_matching_blocks()
289
200
for i, j, n in matching_blocks:
290
201
for line in new_lines[new_cur:j]:
291
lines.append((new_revision_id, line))
292
lines.extend(parent_lines[i:i+n])
202
yield new_revision_id, line
203
for data in parent_lines[i:i+n]:
297
def _get_matching_blocks(old, new):
298
matcher = patiencediff.PatienceSequenceMatcher(None, old, new)
299
return matcher.get_matching_blocks()
302
_break_annotation_tie = None
304
def _old_break_annotation_tie(annotated_lines):
305
"""Chose an attribution between several possible ones.
307
:param annotated_lines: A list of tuples ((file_id, rev_id), line) where
308
the lines are identical but the revids different while no parent
309
relation exist between them
311
:return : The "winning" line. This must be one with a revid that
312
guarantees that further criss-cross merges will converge. Failing to
313
do so have performance implications.
315
# sort lexicographically so that we always get a stable result.
317
# TODO: while 'sort' is the easiest (and nearly the only possible solution)
318
# with the current implementation, chosing the oldest revision is known to
319
# provide better results (as in matching user expectations). The most
320
# common use case being manual cherry-pick from an already existing
322
return sorted(annotated_lines)[0]
325
def _find_matching_unannotated_lines(output_lines, plain_child_lines,
326
child_lines, start_child, end_child,
327
right_lines, start_right, end_right,
328
heads_provider, revision_id):
329
"""Find lines in plain_right_lines that match the existing lines.
331
:param output_lines: Append final annotated lines to this list
332
:param plain_child_lines: The unannotated new lines for the child text
333
:param child_lines: Lines for the child text which have been annotated
336
:param start_child: Position in plain_child_lines and child_lines to start
338
:param end_child: Last position in plain_child_lines and child_lines to
340
:param right_lines: The annotated lines for the whole text for the right
342
:param start_right: Position in right_lines to start the match
343
:param end_right: Last position in right_lines to search for a match
344
:param heads_provider: When parents disagree on the lineage of a line, we
345
need to check if one side supersedes the other
346
:param revision_id: The label to give if a line should be labeled 'tip'
348
output_extend = output_lines.extend
349
output_append = output_lines.append
350
# We need to see if any of the unannotated lines match
351
plain_right_subset = [l for a,l in right_lines[start_right:end_right]]
352
plain_child_subset = plain_child_lines[start_child:end_child]
353
match_blocks = _get_matching_blocks(plain_right_subset, plain_child_subset)
357
for right_idx, child_idx, match_len in match_blocks:
358
# All the lines that don't match are just passed along
359
if child_idx > last_child_idx:
360
output_extend(child_lines[start_child + last_child_idx
361
:start_child + child_idx])
362
for offset in xrange(match_len):
363
left = child_lines[start_child+child_idx+offset]
364
right = right_lines[start_right+right_idx+offset]
365
if left[0] == right[0]:
366
# The annotations match, just return the left one
368
elif left[0] == revision_id:
369
# The left parent marked this as unmatched, so let the
370
# right parent claim it
373
# Left and Right both claim this line
374
if heads_provider is None:
375
output_append((revision_id, left[1]))
377
heads = heads_provider.heads((left[0], right[0]))
379
output_append((iter(heads).next(), left[1]))
381
# Both claim different origins, get a stable result.
382
# If the result is not stable, there is a risk a
383
# performance degradation as criss-cross merges will
384
# flip-flop the attribution.
385
if _break_annotation_tie is None:
387
_old_break_annotation_tie([left, right]))
389
output_append(_break_annotation_tie([left, right]))
390
last_child_idx = child_idx + match_len
393
def _reannotate_annotated(right_parent_lines, new_lines, new_revision_id,
394
annotated_lines, heads_provider):
395
"""Update the annotations for a node based on another parent.
397
:param right_parent_lines: A list of annotated lines for the right-hand
399
:param new_lines: The unannotated new lines.
400
:param new_revision_id: The revision_id to attribute to lines which are not
401
present in either parent.
402
:param annotated_lines: A list of annotated lines. This should be the
403
annotation of new_lines based on parents seen so far.
404
:param heads_provider: When parents disagree on the lineage of a line, we
405
need to check if one side supersedes the other.
407
if len(new_lines) != len(annotated_lines):
408
raise AssertionError("mismatched new_lines and annotated_lines")
409
# First compare the newly annotated lines with the right annotated lines.
410
# Lines which were not changed in left or right should match. This tends to
411
# be the bulk of the lines, and they will need no further processing.
413
lines_extend = lines.extend
414
last_right_idx = 0 # The line just after the last match from the right side
416
matching_left_and_right = _get_matching_blocks(right_parent_lines,
418
for right_idx, left_idx, match_len in matching_left_and_right:
419
# annotated lines from last_left_idx to left_idx did not match the
420
# lines from last_right_idx to right_idx, the raw lines should be
421
# compared to determine what annotations need to be updated
422
if last_right_idx == right_idx or last_left_idx == left_idx:
423
# One of the sides is empty, so this is a pure insertion
424
lines_extend(annotated_lines[last_left_idx:left_idx])
426
# We need to see if any of the unannotated lines match
427
_find_matching_unannotated_lines(lines,
428
new_lines, annotated_lines,
429
last_left_idx, left_idx,
431
last_right_idx, right_idx,
434
last_right_idx = right_idx + match_len
435
last_left_idx = left_idx + match_len
436
# If left and right agree on a range, just push that into the output
437
lines_extend(annotated_lines[left_idx:left_idx + match_len])
442
from bzrlib._annotator_pyx import Annotator
443
except ImportError, e:
444
osutils.failed_to_load_extension(e)
445
from bzrlib._annotator_py import Annotator