1
# Copyright (C) 2010 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
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
17
from __future__ import absolute_import
21
from ...lazy_import import lazy_import
22
lazy_import(globals(), """
23
from fnmatch import fnmatch
25
from breezy._termcolor import color_string, FG
32
revision as _mod_revision,
39
from breezy.revisionspec import (
44
from breezy.sixish import (
48
_user_encoding = osutils.get_user_encoding()
51
class _RevisionNotLinear(Exception):
52
"""Raised when a revision is not on left-hand history."""
55
def _rev_on_mainline(rev_tuple):
56
"""returns True is rev tuple is on mainline"""
57
if len(rev_tuple) == 1:
59
return rev_tuple[1] == 0 and rev_tuple[2] == 0
62
# NOTE: _linear_view_revisions is basided on
63
# breezy.log._linear_view_revisions.
64
# This should probably be a common public API
65
def _linear_view_revisions(branch, start_rev_id, end_rev_id):
66
# requires that start is older than end
67
repo = branch.repository
68
graph = repo.get_graph()
69
for revision_id in graph.iter_lefthand_ancestry(
70
end_rev_id, (_mod_revision.NULL_REVISION, )):
71
revno = branch.revision_id_to_dotted_revno(revision_id)
72
revno_str = '.'.join(str(n) for n in revno)
73
if revision_id == start_rev_id:
74
yield revision_id, revno_str, 0
76
yield revision_id, revno_str, 0
79
# NOTE: _graph_view_revisions is copied from
80
# breezy.log._graph_view_revisions.
81
# This should probably be a common public API
82
def _graph_view_revisions(branch, start_rev_id, end_rev_id,
83
rebase_initial_depths=True):
84
"""Calculate revisions to view including merges, newest to oldest.
86
:param branch: the branch
87
:param start_rev_id: the lower revision-id
88
:param end_rev_id: the upper revision-id
89
:param rebase_initial_depth: should depths be rebased until a mainline
91
:return: An iterator of (revision_id, dotted_revno, merge_depth) tuples.
93
# requires that start is older than end
94
view_revisions = branch.iter_merge_sorted_revisions(
95
start_revision_id=end_rev_id, stop_revision_id=start_rev_id,
96
stop_rule="with-merges")
97
if not rebase_initial_depths:
98
for (rev_id, merge_depth, revno, end_of_merge
100
yield rev_id, '.'.join(map(str, revno)), merge_depth
102
# We're following a development line starting at a merged revision.
103
# We need to adjust depths down by the initial depth until we find
104
# a depth less than it. Then we use that depth as the adjustment.
105
# If and when we reach the mainline, depth adjustment ends.
106
depth_adjustment = None
107
for (rev_id, merge_depth, revno, end_of_merge
109
if depth_adjustment is None:
110
depth_adjustment = merge_depth
112
if merge_depth < depth_adjustment:
113
# From now on we reduce the depth adjustement, this can be
114
# surprising for users. The alternative requires two passes
115
# which breaks the fast display of the first revision
117
depth_adjustment = merge_depth
118
merge_depth -= depth_adjustment
119
yield rev_id, '.'.join(map(str, revno)), merge_depth
122
def compile_pattern(pattern, flags=0):
124
return re.compile(pattern, flags)
125
except re.error as e:
126
raise errors.BzrError("Invalid pattern: '%s'" % pattern)
130
def is_fixed_string(s):
131
if re.match("^([A-Za-z0-9_]|\\s)*$", s):
136
class _GrepDiffOutputter(object):
137
"""Precalculate formatting based on options given for diff grep.
140
def __init__(self, opts):
142
self.outf = opts.outf
144
if opts.fixed_string:
145
self._old = opts.pattern
146
self._new = color_string(opts.pattern, FG.BOLD_RED)
147
self.get_writer = self._get_writer_fixed_highlighted
149
flags = opts.patternc.flags
150
self._sub = re.compile(
151
opts.pattern.join(("((?:", ")+)")), flags).sub
152
self._highlight = color_string("\\1", FG.BOLD_RED)
153
self.get_writer = self._get_writer_regexp_highlighted
155
self.get_writer = self._get_writer_plain
157
def get_file_header_writer(self):
158
"""Get function for writing file headers"""
159
write = self.outf.write
160
eol_marker = self.opts.eol_marker
162
def _line_writer(line):
163
write(line + eol_marker)
165
def _line_writer_color(line):
166
write(FG.BOLD_MAGENTA + line + FG.NONE + eol_marker)
167
if self.opts.show_color:
168
return _line_writer_color
173
def get_revision_header_writer(self):
174
"""Get function for writing revno lines"""
175
write = self.outf.write
176
eol_marker = self.opts.eol_marker
178
def _line_writer(line):
179
write(line + eol_marker)
181
def _line_writer_color(line):
182
write(FG.BOLD_BLUE + line + FG.NONE + eol_marker)
183
if self.opts.show_color:
184
return _line_writer_color
189
def _get_writer_plain(self):
190
"""Get function for writing uncoloured output"""
191
write = self.outf.write
192
eol_marker = self.opts.eol_marker
194
def _line_writer(line):
195
write(line + eol_marker)
198
def _get_writer_regexp_highlighted(self):
199
"""Get function for writing output with regexp match highlighted"""
200
_line_writer = self._get_writer_plain()
201
sub, highlight = self._sub, self._highlight
203
def _line_writer_regexp_highlighted(line):
204
"""Write formatted line with matched pattern highlighted"""
205
return _line_writer(line=sub(highlight, line))
206
return _line_writer_regexp_highlighted
208
def _get_writer_fixed_highlighted(self):
209
"""Get function for writing output with search string highlighted"""
210
_line_writer = self._get_writer_plain()
211
old, new = self._old, self._new
213
def _line_writer_fixed_highlighted(line):
214
"""Write formatted line with string searched for highlighted"""
215
return _line_writer(line=line.replace(old, new))
216
return _line_writer_fixed_highlighted
220
wt, branch, relpath = \
221
controldir.ControlDir.open_containing_tree_or_branch('.')
222
with branch.lock_read():
224
start_rev = opts.revision[0]
226
# if no revision is sepcified for diff grep we grep all changesets.
227
opts.revision = [RevisionSpec.from_string('revno:1'),
228
RevisionSpec.from_string('last:1')]
229
start_rev = opts.revision[0]
230
start_revid = start_rev.as_revision_id(branch)
231
if start_revid == b'null:':
233
srevno_tuple = branch.revision_id_to_dotted_revno(start_revid)
234
if len(opts.revision) == 2:
235
end_rev = opts.revision[1]
236
end_revid = end_rev.as_revision_id(branch)
237
if end_revid is None:
238
end_revno, end_revid = branch.last_revision_info()
239
erevno_tuple = branch.revision_id_to_dotted_revno(end_revid)
241
grep_mainline = (_rev_on_mainline(srevno_tuple)
242
and _rev_on_mainline(erevno_tuple))
244
# ensure that we go in reverse order
245
if srevno_tuple > erevno_tuple:
246
srevno_tuple, erevno_tuple = erevno_tuple, srevno_tuple
247
start_revid, end_revid = end_revid, start_revid
249
# Optimization: Traversing the mainline in reverse order is much
250
# faster when we don't want to look at merged revs. We try this
251
# with _linear_view_revisions. If all revs are to be grepped we
252
# use the slower _graph_view_revisions
253
if opts.levels == 1 and grep_mainline:
254
given_revs = _linear_view_revisions(
255
branch, start_revid, end_revid)
257
given_revs = _graph_view_revisions(
258
branch, start_revid, end_revid)
260
# We do an optimization below. For grepping a specific revison
261
# We don't need to call _graph_view_revisions which is slow.
262
# We create the start_rev_tuple for only that specific revision.
263
# _graph_view_revisions is used only for revision range.
264
start_revno = '.'.join(map(str, srevno_tuple))
265
start_rev_tuple = (start_revid, start_revno, 0)
266
given_revs = [start_rev_tuple]
267
repo = branch.repository
268
diff_pattern = re.compile(
269
b"^[+\\-].*(" + opts.pattern.encode(_user_encoding) + b")")
270
file_pattern = re.compile(b"=== (modified|added|removed) file '.*'")
271
outputter = _GrepDiffOutputter(opts)
272
writeline = outputter.get_writer()
273
writerevno = outputter.get_revision_header_writer()
274
writefileheader = outputter.get_file_header_writer()
275
file_encoding = _user_encoding
276
for revid, revno, merge_depth in given_revs:
277
if opts.levels == 1 and merge_depth != 0:
278
# with level=1 show only top level
281
rev_spec = RevisionSpec_revid.from_string(
282
"revid:" + revid.decode('utf-8'))
283
new_rev = repo.get_revision(revid)
284
new_tree = rev_spec.as_tree(branch)
285
if len(new_rev.parent_ids) == 0:
286
ancestor_id = _mod_revision.NULL_REVISION
288
ancestor_id = new_rev.parent_ids[0]
289
old_tree = repo.revision_tree(ancestor_id)
291
diff.show_diff_trees(old_tree, new_tree, s,
292
old_label='', new_label='')
297
for line in text.splitlines():
298
if file_pattern.search(line):
301
elif diff_pattern.search(line):
303
writerevno("=== revno:%s ===" % (revno,))
304
display_revno = False
307
" %s" % (file_header.decode(file_encoding, 'replace'),))
309
line = line.decode(file_encoding, 'replace')
310
writeline(" %s" % (line,))
313
def versioned_grep(opts):
314
wt, branch, relpath = \
315
controldir.ControlDir.open_containing_tree_or_branch('.')
316
with branch.lock_read():
317
start_rev = opts.revision[0]
318
start_revid = start_rev.as_revision_id(branch)
319
if start_revid is None:
320
start_rev = RevisionSpec_revno.from_string("revno:1")
321
start_revid = start_rev.as_revision_id(branch)
322
srevno_tuple = branch.revision_id_to_dotted_revno(start_revid)
324
if len(opts.revision) == 2:
325
end_rev = opts.revision[1]
326
end_revid = end_rev.as_revision_id(branch)
327
if end_revid is None:
328
end_revno, end_revid = branch.last_revision_info()
329
erevno_tuple = branch.revision_id_to_dotted_revno(end_revid)
331
grep_mainline = (_rev_on_mainline(srevno_tuple)
332
and _rev_on_mainline(erevno_tuple))
334
# ensure that we go in reverse order
335
if srevno_tuple > erevno_tuple:
336
srevno_tuple, erevno_tuple = erevno_tuple, srevno_tuple
337
start_revid, end_revid = end_revid, start_revid
339
# Optimization: Traversing the mainline in reverse order is much
340
# faster when we don't want to look at merged revs. We try this
341
# with _linear_view_revisions. If all revs are to be grepped we
342
# use the slower _graph_view_revisions
343
if opts.levels == 1 and grep_mainline:
344
given_revs = _linear_view_revisions(
345
branch, start_revid, end_revid)
347
given_revs = _graph_view_revisions(
348
branch, start_revid, end_revid)
350
# We do an optimization below. For grepping a specific revison
351
# We don't need to call _graph_view_revisions which is slow.
352
# We create the start_rev_tuple for only that specific revision.
353
# _graph_view_revisions is used only for revision range.
354
start_revno = '.'.join(map(str, srevno_tuple))
355
start_rev_tuple = (start_revid, start_revno, 0)
356
given_revs = [start_rev_tuple]
358
# GZ 2010-06-02: Shouldn't be smuggling this on opts, but easy for now
359
opts.outputter = _Outputter(opts, use_cache=True)
361
for revid, revno, merge_depth in given_revs:
362
if opts.levels == 1 and merge_depth != 0:
363
# with level=1 show only top level
366
rev = RevisionSpec_revid.from_string(
367
"revid:" + revid.decode('utf-8'))
368
tree = rev.as_tree(branch)
369
for path in opts.path_list:
370
tree_path = osutils.pathjoin(relpath, path)
371
if not tree.has_filename(tree_path):
372
trace.warning("Skipped unknown file '%s'.", path)
375
if osutils.isdir(path):
377
dir_grep(tree, path, relpath, opts, revno, path_prefix)
380
tree, tree_path, '.', path, opts, revno)
383
def workingtree_grep(opts):
384
revno = opts.print_revno = None # for working tree set revno to None
386
tree, branch, relpath = \
387
controldir.ControlDir.open_containing_tree_or_branch('.')
389
msg = ('Cannot search working tree. Working tree not found.\n'
390
'To search for specific revision in history use the -r option.')
391
raise errors.BzrCommandError(msg)
393
# GZ 2010-06-02: Shouldn't be smuggling this on opts, but easy for now
394
opts.outputter = _Outputter(opts)
396
with tree.lock_read():
397
for path in opts.path_list:
398
if osutils.isdir(path):
400
dir_grep(tree, path, relpath, opts, revno, path_prefix)
402
with open(path, 'rb') as f:
403
_file_grep(f.read(), path, opts, revno)
406
def _skip_file(include, exclude, path):
407
if include and not _path_in_glob_list(path, include):
409
if exclude and _path_in_glob_list(path, exclude):
414
def dir_grep(tree, path, relpath, opts, revno, path_prefix):
415
# setup relpath to open files relative to cwd
418
rpath = osutils.pathjoin('..', relpath)
420
from_dir = osutils.pathjoin(relpath, path)
422
# start searching recursively from root
427
to_grep_append = to_grep.append
428
# GZ 2010-06-05: The cache dict used to be recycled every call to dir_grep
429
# and hits manually refilled. Could do this again if it was
430
# for a good reason, otherwise cache might want purging.
431
outputter = opts.outputter
432
for fp, fc, fkind, fid, entry in tree.list_files(include_root=False,
433
from_dir=from_dir, recursive=opts.recursive):
435
if _skip_file(opts.include, opts.exclude, fp):
438
if fc == 'V' and fkind == 'file':
439
tree_path = osutils.pathjoin(from_dir if from_dir else '', fp)
440
if revno is not None:
441
# If old result is valid, print results immediately.
442
# Otherwise, add file info to to_grep so that the
443
# loop later will get chunks and grep them
444
cache_id = tree.get_file_revision(tree_path)
445
if cache_id in outputter.cache:
446
# GZ 2010-06-05: Not really sure caching and re-outputting
447
# the old path is really the right thing,
448
# but it's what the old code seemed to do
449
outputter.write_cached_lines(cache_id, revno)
451
to_grep_append((tree_path, (fp, tree_path)))
453
# we are grepping working tree.
457
path_for_file = osutils.pathjoin(tree.basedir, from_dir, fp)
458
if opts.files_with_matches or opts.files_without_match:
459
# Optimize for wtree list-only as we don't need to read the
461
with open(path_for_file, 'rb', buffering=4096) as file:
462
_file_grep_list_only_wtree(file, fp, opts, path_prefix)
464
with open(path_for_file, 'rb') as f:
465
_file_grep(f.read(), fp, opts, revno, path_prefix)
467
if revno is not None: # grep versioned files
468
for (path, tree_path), chunks in tree.iter_files_bytes(to_grep):
469
path = _make_display_path(relpath, path)
470
_file_grep(b''.join(chunks), path, opts, revno, path_prefix,
471
tree.get_file_revision(tree_path))
474
def _make_display_path(relpath, path):
475
"""Return path string relative to user cwd.
477
Take tree's 'relpath' and user supplied 'path', and return path
478
that can be displayed to the user.
481
# update path so to display it w.r.t cwd
482
# handle windows slash separator
483
path = osutils.normpath(osutils.pathjoin(relpath, path))
484
path = path.replace('\\', '/')
485
path = path.replace(relpath + '/', '', 1)
489
def versioned_file_grep(tree, tree_path, relpath, path, opts, revno, path_prefix=None):
490
"""Create a file object for the specified id and pass it on to _file_grep.
493
path = _make_display_path(relpath, path)
494
file_text = tree.get_file_text(tree_path)
495
_file_grep(file_text, path, opts, revno, path_prefix)
498
def _path_in_glob_list(path, glob_list):
499
for glob in glob_list:
500
if fnmatch(path, glob):
505
def _file_grep_list_only_wtree(file, path, opts, path_prefix=None):
506
# test and skip binary files
507
if b'\x00' in file.read(1024):
509
trace.warning("Binary file '%s' skipped.", path)
512
file.seek(0) # search from beginning
515
if opts.fixed_string:
516
pattern = opts.pattern.encode(_user_encoding, 'replace')
521
else: # not fixed_string
523
if opts.patternc.search(line):
527
if (opts.files_with_matches and found) or \
528
(opts.files_without_match and not found):
529
if path_prefix and path_prefix != '.':
530
# user has passed a dir arg, show that as result prefix
531
path = osutils.pathjoin(path_prefix, path)
532
opts.outputter.get_writer(path, None, None)()
535
class _Outputter(object):
536
"""Precalculate formatting based on options given
538
The idea here is to do this work only once per run, and finally return a
539
function that will do the minimum amount possible for each match.
542
def __init__(self, opts, use_cache=False):
543
self.outf = opts.outf
545
# self.cache is used to cache results for dir grep based on fid.
546
# If the fid is does not change between results, it means that
547
# the result will be the same apart from revno. In such a case
548
# we avoid getting file chunks from repo and grepping. The result
549
# is just printed by replacing old revno with new one.
553
no_line = opts.files_with_matches or opts.files_without_match
557
self.get_writer = self._get_writer_plain
558
elif opts.fixed_string:
559
self._old = opts.pattern
560
self._new = color_string(opts.pattern, FG.BOLD_RED)
561
self.get_writer = self._get_writer_fixed_highlighted
563
flags = opts.patternc.flags
564
self._sub = re.compile(
565
opts.pattern.join(("((?:", ")+)")), flags).sub
566
self._highlight = color_string("\\1", FG.BOLD_RED)
567
self.get_writer = self._get_writer_regexp_highlighted
568
path_start = FG.MAGENTA
570
sep = color_string(':', FG.BOLD_CYAN)
571
rev_sep = color_string('~', FG.BOLD_YELLOW)
573
self.get_writer = self._get_writer_plain
574
path_start = path_end = ""
578
parts = [path_start, "%(path)s"]
580
parts.extend([rev_sep, "%(revno)s"])
581
self._format_initial = "".join(parts)
584
if not opts.print_revno:
585
parts.append(path_end)
588
parts.extend([sep, "%(lineno)s"])
589
parts.extend([sep, "%(line)s"])
590
parts.append(opts.eol_marker)
591
self._format_perline = "".join(parts)
593
def _get_writer_plain(self, path, revno, cache_id):
594
"""Get function for writing uncoloured output"""
595
per_line = self._format_perline
596
start = self._format_initial % {"path": path, "revno": revno}
597
write = self.outf.write
598
if self.cache is not None and cache_id is not None:
600
self.cache[cache_id] = path, result_list
601
add_to_cache = result_list.append
603
def _line_cache_and_writer(**kwargs):
604
"""Write formatted line and cache arguments"""
605
end = per_line % kwargs
608
return _line_cache_and_writer
610
def _line_writer(**kwargs):
611
"""Write formatted line from arguments given by underlying opts"""
612
write(start + per_line % kwargs)
615
def write_cached_lines(self, cache_id, revno):
616
"""Write cached results out again for new revision"""
617
cached_path, cached_matches = self.cache[cache_id]
618
start = self._format_initial % {"path": cached_path, "revno": revno}
619
write = self.outf.write
620
for end in cached_matches:
623
def _get_writer_regexp_highlighted(self, path, revno, cache_id):
624
"""Get function for writing output with regexp match highlighted"""
625
_line_writer = self._get_writer_plain(path, revno, cache_id)
626
sub, highlight = self._sub, self._highlight
628
def _line_writer_regexp_highlighted(line, **kwargs):
629
"""Write formatted line with matched pattern highlighted"""
630
return _line_writer(line=sub(highlight, line), **kwargs)
631
return _line_writer_regexp_highlighted
633
def _get_writer_fixed_highlighted(self, path, revno, cache_id):
634
"""Get function for writing output with search string highlighted"""
635
_line_writer = self._get_writer_plain(path, revno, cache_id)
636
old, new = self._old, self._new
638
def _line_writer_fixed_highlighted(line, **kwargs):
639
"""Write formatted line with string searched for highlighted"""
640
return _line_writer(line=line.replace(old, new), **kwargs)
641
return _line_writer_fixed_highlighted
644
def _file_grep(file_text, path, opts, revno, path_prefix=None, cache_id=None):
645
# test and skip binary files
646
if b'\x00' in file_text[:1024]:
648
trace.warning("Binary file '%s' skipped.", path)
651
if path_prefix and path_prefix != '.':
652
# user has passed a dir arg, show that as result prefix
653
path = osutils.pathjoin(path_prefix, path)
655
# GZ 2010-06-07: There's no actual guarentee the file contents will be in
656
# the user encoding, but we have to guess something and it
657
# is a reasonable default without a better mechanism.
658
file_encoding = _user_encoding
659
pattern = opts.pattern.encode(_user_encoding, 'replace')
661
writeline = opts.outputter.get_writer(path, revno, cache_id)
663
if opts.files_with_matches or opts.files_without_match:
664
if opts.fixed_string:
665
found = pattern in file_text
667
search = opts.patternc.search
668
if b"$" not in pattern:
669
found = search(file_text) is not None
671
for line in file_text.splitlines():
677
if (opts.files_with_matches and found) or \
678
(opts.files_without_match and not found):
680
elif opts.fixed_string:
681
# Fast path for no match, search through the entire file at once rather
682
# than a line at a time. <http://effbot.org/zone/stringlib.htm>
683
i = file_text.find(pattern)
686
b = file_text.rfind(b"\n", 0, i) + 1
688
start = file_text.count(b"\n", 0, b) + 1
689
file_text = file_text[b:]
691
for index, line in enumerate(file_text.splitlines()):
693
line = line.decode(file_encoding, 'replace')
694
writeline(lineno=index + start, line=line)
696
for line in file_text.splitlines():
698
line = line.decode(file_encoding, 'replace')
701
# Fast path on no match, the re module avoids bad behaviour in most
702
# standard cases, but perhaps could try and detect backtracking
703
# patterns here and avoid whole text search in those cases
704
search = opts.patternc.search
705
if b"$" not in pattern:
706
# GZ 2010-06-05: Grr, re.MULTILINE can't save us when searching
707
# through revisions as bazaar returns binary mode
708
# and trailing \r breaks $ as line ending match
709
m = search(file_text)
712
b = file_text.rfind(b"\n", 0, m.start()) + 1
714
start = file_text.count(b"\n", 0, b) + 1
715
file_text = file_text[b:]
719
for index, line in enumerate(file_text.splitlines()):
721
line = line.decode(file_encoding, 'replace')
722
writeline(lineno=index + start, line=line)
724
for line in file_text.splitlines():
726
line = line.decode(file_encoding, 'replace')
added
removed
breezy/plugins/grep/grep.py
