« back to all changes in this revision
Viewing changes to loggerhead/history.py
- browse files at revision 128.13.48
- compare with another revision
- download diff
- download tarball
- view history from revision 128.13.48
- Committer: Martin Albisetti
- Date: 2008-07-22 00:40:03 UTC
- mfrom: (182 loggerhead.search_integration)
- mto: This revision was merged to the branch mainline in revision 187.
- Revision ID: argentina@gmail.com-20080722004003-lx1fp0tthpp6kl92
Merged from trunk, resolved billions of conflicts
- files added:
- loggerhead/controllers/search_ui.py
- files removed:
- loggerhead/static/javascript/custom.js
- files modified:
- NEWS
added
removed
loggerhead/history.py
35
35
import time
36
36
from StringIO import StringIO
37
37
38
from loggerhead import search
38
39
from loggerhead import util
39
from loggerhead.util import decorator
40
from loggerhead.wholehistory import compute_whole_history_data
40
41
41
42
import bzrlib
42
43
import bzrlib.branch
47
48
import bzrlib.tsort
48
49
import bzrlib.ui
49
50
50
51
with_branch_lock = util.with_lock('_lock', 'branch')
52
53
54
@decorator
55
def with_bzrlib_read_lock(unbound):
56
def bzrlib_read_locked(self, *args, **kw):
57
#self.log.debug('-> %r bzr lock', id(threading.currentThread()))
58
self._branch.repository.lock_read()
59
try:
60
return unbound(self, *args, **kw)
61
finally:
62
self._branch.repository.unlock()
63
#self.log.debug('<- %r bzr lock', id(threading.currentThread()))
64
return bzrlib_read_locked
65
66
67
51
# bzrlib's UIFactory is not thread-safe
68
52
uihack = threading.local()
69
53
107
91
out_chunk_list = []
108
92
for chunk in chunk_list:
109
93
line_list = []
110
wrap_char = '<wbr/>'
111
94
delete_list, insert_list = [], []
112
95
for line in chunk.diff:
113
# Add <wbr/> every X characters so we can wrap properly
114
wrap_line = re.findall(r'.{%d}|.+$' % 78, line.line)
115
wrap_lines = [util.html_clean(_line) for _line in wrap_line]
116
wrapped_line = wrap_char.join(wrap_lines)
117
118
96
if line.type == 'context':
119
97
if len(delete_list) or len(insert_list):
120
_process_side_by_side_buffers(line_list, delete_list,
121
insert_list)
98
_process_side_by_side_buffers(line_list, delete_list, insert_list)
122
99
delete_list, insert_list = [], []
123
line_list.append(util.Container(old_lineno=line.old_lineno,
124
new_lineno=line.new_lineno,
125
old_line=wrapped_line,
126
new_line=wrapped_line,
127
old_type=line.type,
128
new_type=line.type))
100
line_list.append(util.Container(old_lineno=line.old_lineno, new_lineno=line.new_lineno,
101
old_line=line.line, new_line=line.line,
102
old_type=line.type, new_type=line.type))
129
103
elif line.type == 'delete':
130
delete_list.append((line.old_lineno, wrapped_line, line.type))
104
delete_list.append((line.old_lineno, line.line, line.type))
131
105
elif line.type == 'insert':
132
insert_list.append((line.new_lineno, wrapped_line, line.type))
106
insert_list.append((line.new_lineno, line.line, line.type))
133
107
if len(delete_list) or len(insert_list):
134
108
_process_side_by_side_buffers(line_list, delete_list, insert_list)
135
109
out_chunk_list.append(util.Container(diff=line_list))
198
172
199
173
200
174
class History (object):
201
202
def __init__(self):
175
"""Decorate a branch to provide information for rendering.
176
177
History objects are expected to be short lived -- when serving a request
178
for a particular branch, open it, read-lock it, wrap a History object
179
around it, serve the request, throw the History object away, unlock the
180
branch and throw it away.
181
182
:ivar _file_change_cache: xx
183
"""
184
185
def __init__(self, branch, whole_history_data_cache):
186
assert branch.is_locked(), (
187
"Can only construct a History object with a read-locked branch.")
203
188
self._file_change_cache = None
204
self._lock = threading.RLock()
205
206
@classmethod
207
def from_branch(cls, branch):
208
z = time.time()
209
self = cls()
210
189
self._branch = branch
211
self._last_revid = self._branch.last_revision()
212
213
self.log = logging.getLogger('loggerhead.%s' % (self._branch.nick,))
214
215
graph = branch.repository.get_graph()
216
parent_map = dict(((key, value) for key, value in
217
graph.iter_ancestry([self._last_revid]) if value is not None))
218
219
self._revision_graph = self._strip_NULL_ghosts(parent_map)
220
self._full_history = []
221
self._revision_info = {}
222
self._revno_revid = {}
223
if bzrlib.revision.is_null(self._last_revid):
224
self._merge_sort = []
225
else:
226
self._merge_sort = bzrlib.tsort.merge_sort(
227
self._revision_graph, self._last_revid, generate_revno=True)
228
229
for (seq, revid, merge_depth, revno, end_of_merge) in self._merge_sort:
230
self._full_history.append(revid)
231
revno_str = '.'.join(str(n) for n in revno)
232
self._revno_revid[revno_str] = revid
233
self._revision_info[revid] = (
234
seq, revid, merge_depth, revno_str, end_of_merge)
235
236
# cache merge info
237
self._where_merged = {}
238
239
for revid in self._revision_graph.keys():
240
if self._revision_info[revid][2] == 0:
241
continue
242
for parent in self._revision_graph[revid]:
243
self._where_merged.setdefault(parent, set()).add(revid)
244
245
self.log.info('built revision graph cache: %r secs' % (time.time() - z,))
246
return self
247
248
@staticmethod
249
def _strip_NULL_ghosts(revision_graph):
250
"""
251
Copied over from bzrlib meant as a temporary workaround deprecated
252
methods.
253
"""
254
255
# Filter ghosts, and null:
256
if bzrlib.revision.NULL_REVISION in revision_graph:
257
del revision_graph[bzrlib.revision.NULL_REVISION]
258
for key, parents in revision_graph.items():
259
revision_graph[key] = tuple(parent for parent in parents if parent
260
in revision_graph)
261
return revision_graph
262
263
@classmethod
264
def from_folder(cls, path):
265
b = bzrlib.branch.Branch.open(path)
266
b.lock_read()
267
try:
268
return cls.from_branch(b)
269
finally:
270
b.unlock()
271
272
@with_branch_lock
273
def out_of_date(self):
274
# the branch may have been upgraded on disk, in which case we're stale.
275
newly_opened = bzrlib.branch.Branch.open(self._branch.base)
276
if self._branch.__class__ is not \
277
newly_opened.__class__:
278
return True
279
if self._branch.repository.__class__ is not \
280
newly_opened.repository.__class__:
281
return True
282
return self._branch.last_revision() != self._last_revid
190
self.log = logging.getLogger('loggerhead.%s' % (branch.nick,))
191
192
self.last_revid = branch.last_revision()
193
194
whole_history_data = whole_history_data_cache.get(self.last_revid)
195
if whole_history_data is None:
196
whole_history_data = compute_whole_history_data(branch)
197
whole_history_data_cache[self.last_revid] = whole_history_data
198
199
(self._revision_graph, self._full_history, self._revision_info,
200
self._revno_revid, self._merge_sort, self._where_merged
201
) = whole_history_data
283
202
284
203
def use_file_cache(self, cache):
285
204
self._file_change_cache = cache
288
207
def has_revisions(self):
289
208
return not bzrlib.revision.is_null(self.last_revid)
290
209
291
last_revid = property(lambda self: self._last_revid, None, None)
292
293
@with_branch_lock
294
210
def get_config(self):
295
211
return self._branch.get_config()
296
212
301
217
seq, revid, merge_depth, revno_str, end_of_merge = self._revision_info[revid]
302
218
return revno_str
303
219
304
def get_revision_history(self):
305
return self._full_history
306
307
220
def get_revids_from(self, revid_list, start_revid):
308
221
"""
309
222
Yield the mainline (wrt start_revid) revisions that merged each
331
244
return
332
245
revid = parents[0]
333
246
334
@with_branch_lock
335
247
def get_short_revision_history_by_fileid(self, file_id):
336
248
# wow. is this really the only way we can get this list? by
337
249
# man-handling the weave store directly? :-0
338
250
# FIXME: would be awesome if we could get, for a folder, the list of
339
251
# revisions where items within that folder changed.
340
w = self._branch.repository.weave_store.get_weave(file_id, self._branch.repository.get_transaction())
341
w_revids = w.versions()
342
revids = [r for r in self._full_history if r in w_revids]
343
return revids
252
possible_keys = [(file_id, revid) for revid in self._full_history]
253
existing_keys = self._branch.repository.texts.get_parent_map(possible_keys)
254
return [revid for _, revid in existing_keys.iterkeys()]
344
255
345
@with_branch_lock
346
256
def get_revision_history_since(self, revid_list, date):
347
257
# if a user asks for revisions starting at 01-sep, they mean inclusive,
348
258
# so start at midnight on 02-sep.
356
266
index = -index
357
267
return revid_list[index:]
358
268
359
@with_branch_lock
360
269
def get_search_revid_list(self, query, revid_list):
361
270
"""
362
271
given a "quick-search" query, try a few obvious possible meanings:
395
304
if date is not None:
396
305
if revid_list is None:
397
306
# if no limit to the query was given, search only the direct-parent path.
398
revid_list = list(self.get_revids_from(None, self._last_revid))
307
revid_list = list(self.get_revids_from(None, self.last_revid))
399
308
return self.get_revision_history_since(revid_list, date)
400
309
401
310
revno_re = re.compile(r'^[\d\.]+$')
411
320
if revid is None:
412
321
return revid
413
322
if revid == 'head:':
414
return self._last_revid
323
return self.last_revid
415
324
if self.revno_re.match(revid):
416
325
revid = self._revno_revid[revid]
417
326
return revid
418
327
419
@with_branch_lock
420
328
def get_file_view(self, revid, file_id):
421
329
"""
422
330
Given a revid and optional path, return a (revlist, revid) for
426
334
If file_id is None, the entire revision history is the list scope.
427
335
"""
428
336
if revid is None:
429
revid = self._last_revid
337
revid = self.last_revid
430
338
if file_id is not None:
431
339
# since revid is 'start_revid', possibly should start the path
432
340
# tracing from revid... FIXME
436
344
revlist = list(self.get_revids_from(None, revid))
437
345
return revlist
438
346
439
@with_branch_lock
440
347
def get_view(self, revid, start_revid, file_id, query=None):
441
348
"""
442
349
use the URL parameters (revid, start_revid, file_id, and query) to
462
369
contain vital context for future url navigation.
463
370
"""
464
371
if start_revid is None:
465
start_revid = self._last_revid
372
start_revid = self.last_revid
466
373
467
374
if query is None:
468
375
revid_list = self.get_file_view(start_revid, file_id)
480
387
revid_list = self.get_file_view(start_revid, file_id)
481
388
else:
482
389
revid_list = None
483
484
revid_list = self.get_search_revid_list(query, revid_list)
390
revid_list = search.search_revisions(self._branch, query)
485
391
if revid_list and len(revid_list) > 0:
486
392
if revid not in revid_list:
487
393
revid = revid_list[0]
488
394
return revid, start_revid, revid_list
489
395
else:
396
# XXX: This should return a message saying that the search could
397
# not be completed due to either missing the plugin or missing a
398
# search index.
490
399
return None, None, []
491
400
492
@with_branch_lock
493
401
def get_inventory(self, revid):
494
402
return self._branch.repository.get_revision_inventory(revid)
495
403
496
@with_branch_lock
497
404
def get_path(self, revid, file_id):
498
405
if (file_id is None) or (file_id == ''):
499
406
return ''
502
409
path = '/' + path
503
410
return path
504
411
505
@with_branch_lock
506
412
def get_file_id(self, revid, path):
507
413
if (len(path) > 0) and not path.startswith('/'):
508
414
path = '/' + path
581
487
else:
582
488
p.branch_nick = '(missing)'
583
489
584
@with_branch_lock
585
490
def get_changes(self, revid_list):
586
491
"""Return a list of changes objects for the given revids.
587
492
608
513
609
514
return changes
610
515
611
@with_branch_lock
612
@with_bzrlib_read_lock
613
516
def get_changes_uncached(self, revid_list):
614
517
# FIXME: deprecated method in getting a null revision
615
518
revid_list = filter(lambda revid: not bzrlib.revision.is_null(revid),
680
583
681
584
return [self.parse_delta(delta) for delta in delta_list]
682
585
683
@with_branch_lock
684
586
def get_file_changes(self, entries):
685
587
if self._file_change_cache is None:
686
588
return self.get_file_changes_uncached(entries)
693
595
for entry, changes in zip(entries, changes_list):
694
596
entry.changes = changes
695
597
696
@with_branch_lock
697
598
def get_change_with_diff(self, revid, compare_revid=None):
698
599
change = self.get_changes([revid])[0]
699
600
712
613
713
614
return change
714
615
715
@with_branch_lock
716
616
def get_file(self, file_id, revid):
717
617
"returns (path, filename, data)"
718
618
inv = self.get_inventory(revid)
786
686
old_lineno = lines[0]
787
687
new_lineno = lines[1]
788
688
elif line.startswith(' '):
789
chunk.diff.append(util.Container(old_lineno=old_lineno,
790
new_lineno=new_lineno,
791
type='context',
792
line=line[1:]))
689
chunk.diff.append(util.Container(old_lineno=old_lineno, new_lineno=new_lineno,
690
type='context', line=util.fixed_width(line[1:])))
793
691
old_lineno += 1
794
692
new_lineno += 1
795
693
elif line.startswith('+'):
796
chunk.diff.append(util.Container(old_lineno=None,
797
new_lineno=new_lineno,
798
type='insert', line=line[1:]))
694
chunk.diff.append(util.Container(old_lineno=None, new_lineno=new_lineno,
695
type='insert', line=util.fixed_width(line[1:])))
799
696
new_lineno += 1
800
697
elif line.startswith('-'):
801
chunk.diff.append(util.Container(old_lineno=old_lineno,
802
new_lineno=None,
803
type='delete', line=line[1:]))
698
chunk.diff.append(util.Container(old_lineno=old_lineno, new_lineno=None,
699
type='delete', line=util.fixed_width(line[1:])))
804
700
old_lineno += 1
805
701
else:
806
chunk.diff.append(util.Container(old_lineno=None,
807
new_lineno=None,
808
type='unknown',
809
line=repr(line)))
702
chunk.diff.append(util.Container(old_lineno=None, new_lineno=None,
703
type='unknown', line=util.fixed_width(repr(line))))
810
704
if chunk is not None:
811
705
chunks.append(chunk)
812
706
return chunks
851
745
for m in change.changes.modified:
852
746
m.sbs_chunks = _make_side_by_side(m.chunks)
853
747
854
@with_branch_lock
855
748
def get_filelist(self, inv, file_id, sort_type=None):
856
749
"""
857
750
return the list of all files (and their attributes) within a given
890
783
file_list.sort(key=lambda x: x.size)
891
784
elif sort_type == 'date':
892
785
file_list.sort(key=lambda x: x.change.date)
786
893
787
# Always sort by kind to get directories first
894
788
file_list.sort(key=lambda x: x.kind != 'directory')
895
789
903
797
904
798
_BADCHARS_RE = re.compile(ur'[\x00-\x08\x0b\x0e-\x1f]')
905
799
906
@with_branch_lock
907
800
def annotate_file(self, file_id, revid):
908
801
z = time.time()
909
802
lineno = 1
Loggerhead is a web-based interface for Breezy
Version: 2.0.1