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 .lazy_import import lazy_import
34
lazy_import(globals(), """
47
extract_email_address,
49
from .repository import _strip_NULL_ghosts
50
from .revision import (
56
def annotate_file_tree(tree, path, to_file, verbose=False, full=False,
57
show_ids=False, branch=None, file_id=None):
58
"""Annotate file_id in a tree.
60
The tree should already be read_locked() when annotate_file_tree is called.
62
:param tree: The tree to look for revision numbers and history from.
63
:param path: The path to annotate
64
:param to_file: The file to output the annotation to.
65
:param verbose: Show all details rather than truncating to ensure
66
reasonable text width.
67
:param full: XXXX Not sure what this does.
68
:param show_ids: Show revision ids in the annotation output.
69
:param file_id: The file_id to annotate (must match file path)
70
:param branch: Branch to use for revision revno lookups
37
from bzrlib.config import extract_email_address
40
def annotate_file(branch, rev_id, file_id, verbose=False, full=False,
41
to_file=None, show_ids=False):
74
42
if to_file is None:
75
43
to_file = sys.stdout
77
45
# Handle the show_ids case
78
annotations = list(tree.annotate_iter(path, file_id))
80
return _show_id_annotations(annotations, to_file, full)
82
if not getattr(tree, "get_revision_id", False):
83
# Create a virtual revision to represent the current tree state.
84
# Should get some more pending commit attributes, like pending tags,
86
current_rev = Revision(CURRENT_REVISION)
87
current_rev.parent_ids = tree.get_parent_ids()
89
current_rev.committer = branch.get_config_stack().get('email')
91
current_rev.committer = 'local user'
92
current_rev.message = "?"
93
current_rev.timestamp = round(time.time(), 3)
94
current_rev.timezone = osutils.local_time_offset()
97
annotation = list(_expand_annotations(annotations, branch,
99
_print_annotations(annotation, verbose, to_file, full)
102
def _print_annotations(annotation, verbose, to_file, full):
103
"""Print annotations to to_file.
105
:param to_file: The file to output the annotation to.
106
:param verbose: Show all details rather than truncating to ensure
107
reasonable text width.
108
:param full: XXXX Not sure what this does.
48
annotations = _annotations(branch.repository, file_id, rev_id)
49
max_origin_len = max(len(origin) for origin, text in annotations)
50
for origin, text in annotations:
51
if full or last_rev_id != origin:
55
to_file.write('%*s | %s' % (max_origin_len, this, text))
59
# Calculate the lengths of the various columns
60
annotation = list(_annotate_file(branch, rev_id, file_id))
110
61
if len(annotation) == 0:
111
62
max_origin_len = max_revno_len = max_revid_len = 0
129
82
anno = "%-*s %-7s " % (max_revno_len, revno_str, author[:7])
130
83
if anno.lstrip() == "" and full:
132
# GZ 2017-05-21: Writing both unicode annotation and bytes from file
133
# which the given to_file must cope with.
87
except UnicodeEncodeError:
88
# cmd_annotate should be passing in an 'exact' object, which means
89
# we have a direct handle to sys.stdout or equivalent. It may not
90
# be able to handle the exact Unicode characters, but 'annotate' is
91
# a user function (non-scripting), so shouldn't die because of
92
# unrepresentable annotation characters. So encode using 'replace',
93
# and write them again.
94
to_file.write(anno.encode(encoding, 'replace'))
135
95
to_file.write('| %s\n' % (text,))
139
def _show_id_annotations(annotations, to_file, full):
143
max_origin_len = max(len(origin) for origin, text in annotations)
144
for origin, text in annotations:
145
if full or last_rev_id != origin:
149
to_file.write('%*s | %s' % (max_origin_len, this, text))
154
def _expand_annotations(annotations, branch, current_rev=None):
155
"""Expand a file's annotations into command line UI ready tuples.
157
Each tuple includes detailed information, such as the author name, and date
158
string for the commit, rather than just the revision id.
160
:param annotations: The annotations to expand.
161
:param revision_id_to_revno: A map from id to revision numbers.
162
:param branch: A locked branch to query for revision details.
99
def _annotations(repo, file_id, rev_id):
100
"""Return the list of (origin,text) for a revision of a file in a repository."""
101
w = repo.weave_store.get_weave(file_id, repo.get_transaction())
102
return list(w.annotate_iter(rev_id))
105
def _annotate_file(branch, rev_id, file_id):
106
"""Yield the origins for each line of a file.
108
This includes detailed information, such as the author name, and
109
date string for the commit, rather than just the revision id.
164
repository = branch.repository
111
revision_id_to_revno = branch.get_revision_id_to_revno_map()
112
annotations = _annotations(branch.repository, file_id, rev_id)
165
114
revision_ids = set(o for o, t in annotations)
166
if current_rev is not None:
167
# This can probably become a function on MutableTree, get_revno_map
168
# there, or something.
169
last_revision = current_rev.revision_id
170
# XXX: Partially Cloned from branch, uses the old_get_graph, eep.
171
# XXX: The main difficulty is that we need to inject a single new node
172
# (current_rev) into the graph before it gets numbered, etc.
173
# Once KnownGraph gets an 'add_node()' function, we can use
174
# VF.get_known_graph_ancestry().
175
graph = repository.get_graph()
176
revision_graph = dict(((key, value) for key, value in
177
graph.iter_ancestry(current_rev.parent_ids) if value is not None))
178
revision_graph = _strip_NULL_ghosts(revision_graph)
179
revision_graph[last_revision] = current_rev.parent_ids
180
merge_sorted_revisions = tsort.merge_sort(
185
revision_id_to_revno = dict((rev_id, revno)
186
for seq_num, rev_id, depth, revno, end_of_merge in
187
merge_sorted_revisions)
189
# TODO(jelmer): Only look up the revision ids that we need (i.e. those
190
# in revision_ids). Possibly add a HPSS call that can look those up
192
revision_id_to_revno = branch.get_revision_id_to_revno_map()
195
if CURRENT_REVISION in revision_ids:
196
revision_id_to_revno[CURRENT_REVISION] = (
197
"%d?" % (branch.revno() + 1),)
198
revisions[CURRENT_REVISION] = current_rev
201
repository.iter_revisions(revision_ids)
202
if entry[1] is not None)
115
revision_ids = [o for o in revision_ids if
116
branch.repository.has_revision(o)]
117
revisions = dict((r.revision_id, r) for r in
118
branch.repository.get_revisions(revision_ids))
203
119
for origin, text in annotations:
204
text = text.rstrip(b'\r\n')
120
text = text.rstrip('\r\n')
205
121
if origin == last_origin:
206
(revno_str, author, date_str) = ('', '', '')
122
(revno_str, author, date_str) = ('','','')
208
124
last_origin = origin
209
125
if origin not in revisions:
210
(revno_str, author, date_str) = ('?', '?', '?')
126
(revno_str, author, date_str) = ('?','?','?')
212
128
revno_str = '.'.join(str(i) for i in
213
129
revision_id_to_revno[origin])
217
133
time.gmtime(rev.timestamp + tz))
218
134
# a lazy way to get something like the email address
219
135
# TODO: Get real email address
220
author = rev.get_apparent_authors()[0]
136
author = rev.get_apparent_author()
222
138
author = extract_email_address(author)
223
except NoEmailInUsername:
139
except errors.NoEmailInUsername:
224
140
pass # use the whole name
225
141
yield (revno_str, author, date_str, origin, text)
228
144
def reannotate(parents_lines, new_lines, new_revision_id,
229
_left_matching_blocks=None,
230
heads_provider=None):
145
_left_matching_blocks=None):
231
146
"""Create a new annotated version from new lines and parent annotations.
233
148
:param parents_lines: List of annotated lines for all parents
234
149
:param new_lines: The un-annotated new lines
235
150
:param new_revision_id: The revision-id to associate with new lines
236
151
(will often be CURRENT_REVISION)
237
152
:param left_matching_blocks: a hint about which areas are common
238
153
between the text and its left-hand-parent. The format is
239
the SequenceMatcher.get_matching_blocks format
240
(start_left, start_right, length_of_match).
241
:param heads_provider: An object which provides a .heads() call to resolve
242
if any revision ids are children of others.
243
If None, then any ancestry disputes will be resolved with
154
the SequenceMatcher.get_matching_blocks format.
246
156
if len(parents_lines) == 0:
247
lines = [(new_revision_id, line) for line in new_lines]
157
for line in new_lines:
158
yield new_revision_id, line
248
159
elif len(parents_lines) == 1:
249
lines = _reannotate(parents_lines[0], new_lines, new_revision_id,
250
_left_matching_blocks)
251
elif len(parents_lines) == 2:
252
left = _reannotate(parents_lines[0], new_lines, new_revision_id,
253
_left_matching_blocks)
254
lines = _reannotate_annotated(parents_lines[1], new_lines,
255
new_revision_id, left,
160
for data in _reannotate(parents_lines[0], new_lines, new_revision_id,
161
_left_matching_blocks):
258
reannotations = [_reannotate(parents_lines[0], new_lines,
259
new_revision_id, _left_matching_blocks)]
260
reannotations.extend(_reannotate(p, new_lines, new_revision_id)
261
for p in parents_lines[1:])
164
block_list = [_left_matching_blocks] + [None] * len(parents_lines)
165
reannotations = [list(_reannotate(p, new_lines, new_revision_id, b))
166
for p, b in zip(parents_lines, block_list)]
263
167
for annos in zip(*reannotations):
264
168
origins = set(a for a, l in annos)
265
170
if len(origins) == 1:
266
# All the parents agree, so just return the first one
267
lines.append(annos[0])
171
yield iter(origins).next(), line
172
elif len(origins) == 2 and new_revision_id in origins:
173
yield (x for x in origins if x != new_revision_id).next(), line
270
if len(origins) == 2 and new_revision_id in origins:
271
origins.remove(new_revision_id)
272
if len(origins) == 1:
273
lines.append((origins.pop(), line))
275
lines.append((new_revision_id, line))
175
yield new_revision_id, line
279
178
def _reannotate(parent_lines, new_lines, new_revision_id,
284
183
matcher = patiencediff.PatienceSequenceMatcher(None,
285
184
plain_parent_lines, new_lines)
286
185
matching_blocks = matcher.get_matching_blocks()
288
186
for i, j, n in matching_blocks:
289
187
for line in new_lines[new_cur:j]:
290
lines.append((new_revision_id, line))
291
lines.extend(parent_lines[i:i+n])
188
yield new_revision_id, line
189
for data in parent_lines[i:i+n]:
296
def _get_matching_blocks(old, new):
297
matcher = patiencediff.PatienceSequenceMatcher(None, old, new)
298
return matcher.get_matching_blocks()
301
_break_annotation_tie = None
303
def _old_break_annotation_tie(annotated_lines):
304
"""Chose an attribution between several possible ones.
306
:param annotated_lines: A list of tuples ((file_id, rev_id), line) where
307
the lines are identical but the revids different while no parent
308
relation exist between them
310
:return : The "winning" line. This must be one with a revid that
311
guarantees that further criss-cross merges will converge. Failing to
312
do so have performance implications.
314
# sort lexicographically so that we always get a stable result.
316
# TODO: while 'sort' is the easiest (and nearly the only possible solution)
317
# with the current implementation, chosing the oldest revision is known to
318
# provide better results (as in matching user expectations). The most
319
# common use case being manual cherry-pick from an already existing
321
return sorted(annotated_lines)[0]
324
def _find_matching_unannotated_lines(output_lines, plain_child_lines,
325
child_lines, start_child, end_child,
326
right_lines, start_right, end_right,
327
heads_provider, revision_id):
328
"""Find lines in plain_right_lines that match the existing lines.
330
:param output_lines: Append final annotated lines to this list
331
:param plain_child_lines: The unannotated new lines for the child text
332
:param child_lines: Lines for the child text which have been annotated
335
:param start_child: Position in plain_child_lines and child_lines to start
337
:param end_child: Last position in plain_child_lines and child_lines to
339
:param right_lines: The annotated lines for the whole text for the right
341
:param start_right: Position in right_lines to start the match
342
:param end_right: Last position in right_lines to search for a match
343
:param heads_provider: When parents disagree on the lineage of a line, we
344
need to check if one side supersedes the other
345
:param revision_id: The label to give if a line should be labeled 'tip'
347
output_extend = output_lines.extend
348
output_append = output_lines.append
349
# We need to see if any of the unannotated lines match
350
plain_right_subset = [l for a, l in right_lines[start_right:end_right]]
351
plain_child_subset = plain_child_lines[start_child:end_child]
352
match_blocks = _get_matching_blocks(plain_right_subset, plain_child_subset)
356
for right_idx, child_idx, match_len in match_blocks:
357
# All the lines that don't match are just passed along
358
if child_idx > last_child_idx:
359
output_extend(child_lines[start_child + last_child_idx
360
:start_child + child_idx])
361
for offset in range(match_len):
362
left = child_lines[start_child+child_idx+offset]
363
right = right_lines[start_right+right_idx+offset]
364
if left[0] == right[0]:
365
# The annotations match, just return the left one
367
elif left[0] == revision_id:
368
# The left parent marked this as unmatched, so let the
369
# right parent claim it
372
# Left and Right both claim this line
373
if heads_provider is None:
374
output_append((revision_id, left[1]))
376
heads = heads_provider.heads((left[0], right[0]))
378
output_append((next(iter(heads)), left[1]))
380
# Both claim different origins, get a stable result.
381
# If the result is not stable, there is a risk a
382
# performance degradation as criss-cross merges will
383
# flip-flop the attribution.
384
if _break_annotation_tie is None:
386
_old_break_annotation_tie([left, right]))
388
output_append(_break_annotation_tie([left, right]))
389
last_child_idx = child_idx + match_len
392
def _reannotate_annotated(right_parent_lines, new_lines, new_revision_id,
393
annotated_lines, heads_provider):
394
"""Update the annotations for a node based on another parent.
396
:param right_parent_lines: A list of annotated lines for the right-hand
398
:param new_lines: The unannotated new lines.
399
:param new_revision_id: The revision_id to attribute to lines which are not
400
present in either parent.
401
:param annotated_lines: A list of annotated lines. This should be the
402
annotation of new_lines based on parents seen so far.
403
:param heads_provider: When parents disagree on the lineage of a line, we
404
need to check if one side supersedes the other.
406
if len(new_lines) != len(annotated_lines):
407
raise AssertionError("mismatched new_lines and annotated_lines")
408
# First compare the newly annotated lines with the right annotated lines.
409
# Lines which were not changed in left or right should match. This tends to
410
# be the bulk of the lines, and they will need no further processing.
412
lines_extend = lines.extend
413
last_right_idx = 0 # The line just after the last match from the right side
415
matching_left_and_right = _get_matching_blocks(right_parent_lines,
417
for right_idx, left_idx, match_len in matching_left_and_right:
418
# annotated lines from last_left_idx to left_idx did not match the
419
# lines from last_right_idx to right_idx, the raw lines should be
420
# compared to determine what annotations need to be updated
421
if last_right_idx == right_idx or last_left_idx == left_idx:
422
# One of the sides is empty, so this is a pure insertion
423
lines_extend(annotated_lines[last_left_idx:left_idx])
425
# We need to see if any of the unannotated lines match
426
_find_matching_unannotated_lines(lines,
427
new_lines, annotated_lines,
428
last_left_idx, left_idx,
430
last_right_idx, right_idx,
433
last_right_idx = right_idx + match_len
434
last_left_idx = left_idx + match_len
435
# If left and right agree on a range, just push that into the output
436
lines_extend(annotated_lines[left_idx:left_idx + match_len])
441
from breezy._annotator_pyx import Annotator
442
except ImportError as e:
443
osutils.failed_to_load_extension(e)
444
from breezy._annotator_py import Annotator