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
22
from ...lazy_import import lazy_import
23
lazy_import(globals(), """
24
from fnmatch import fnmatch
26
from breezy._termcolor import color_string, FG
28
from breezy.revisionspec import (
38
revision as _mod_revision,
45
from breezy.sixish import (
49
_user_encoding = osutils.get_user_encoding()
52
class _RevisionNotLinear(Exception):
53
"""Raised when a revision is not on left-hand history."""
56
def _rev_on_mainline(rev_tuple):
57
"""returns True is rev tuple is on mainline"""
58
if len(rev_tuple) == 1:
60
return rev_tuple[1] == 0 and rev_tuple[2] == 0
63
# NOTE: _linear_view_revisions is basided on
64
# breezy.log._linear_view_revisions.
65
# This should probably be a common public API
66
def _linear_view_revisions(branch, start_rev_id, end_rev_id):
67
# requires that start is older than end
68
repo = branch.repository
69
graph = repo.get_graph()
70
for revision_id in graph.iter_lefthand_ancestry(
71
end_rev_id, (_mod_revision.NULL_REVISION, )):
72
revno = branch.revision_id_to_dotted_revno(revision_id)
73
revno_str = '.'.join(str(n) for n in revno)
74
if revision_id == start_rev_id:
75
yield revision_id, revno_str, 0
77
yield revision_id, revno_str, 0
80
# NOTE: _graph_view_revisions is copied from
81
# breezy.log._graph_view_revisions.
82
# This should probably be a common public API
83
def _graph_view_revisions(branch, start_rev_id, end_rev_id,
84
rebase_initial_depths=True):
85
"""Calculate revisions to view including merges, newest to oldest.
87
:param branch: the branch
88
:param start_rev_id: the lower revision-id
89
:param end_rev_id: the upper revision-id
90
:param rebase_initial_depth: should depths be rebased until a mainline
92
:return: An iterator of (revision_id, dotted_revno, merge_depth) tuples.
94
# requires that start is older than end
95
view_revisions = branch.iter_merge_sorted_revisions(
96
start_revision_id=end_rev_id, stop_revision_id=start_rev_id,
97
stop_rule="with-merges")
98
if not rebase_initial_depths:
99
for (rev_id, merge_depth, revno, end_of_merge
101
yield rev_id, '.'.join(map(str, revno)), merge_depth
103
# We're following a development line starting at a merged revision.
104
# We need to adjust depths down by the initial depth until we find
105
# a depth less than it. Then we use that depth as the adjustment.
106
# If and when we reach the mainline, depth adjustment ends.
107
depth_adjustment = None
108
for (rev_id, merge_depth, revno, end_of_merge
110
if depth_adjustment is None:
111
depth_adjustment = merge_depth
113
if merge_depth < depth_adjustment:
114
# From now on we reduce the depth adjustement, this can be
115
# surprising for users. The alternative requires two passes
116
# which breaks the fast display of the first revision
118
depth_adjustment = merge_depth
119
merge_depth -= depth_adjustment
120
yield rev_id, '.'.join(map(str, revno)), merge_depth
123
def compile_pattern(pattern, flags=0):
126
# use python's re.compile as we need to catch re.error in case of bad pattern
127
lazy_regex.reset_compile()
128
patternc = re.compile(pattern, flags)
129
except re.error as e:
130
raise errors.BzrError("Invalid pattern: '%s'" % pattern)
134
def is_fixed_string(s):
135
if re.match("^([A-Za-z0-9_]|\s)*$", s):
140
class _GrepDiffOutputter(object):
141
"""Precalculate formatting based on options given for diff grep.
144
def __init__(self, opts):
146
self.outf = opts.outf
148
pat = opts.pattern.encode(_user_encoding, 'replace')
149
if opts.fixed_string:
151
self._new = color_string(pat, FG.BOLD_RED)
152
self.get_writer = self._get_writer_fixed_highlighted
154
flags = opts.patternc.flags
155
self._sub = re.compile(pat.join(("((?:",")+)")), flags).sub
156
self._highlight = color_string("\\1", FG.BOLD_RED)
157
self.get_writer = self._get_writer_regexp_highlighted
159
self.get_writer = self._get_writer_plain
161
def get_file_header_writer(self):
162
"""Get function for writing file headers"""
163
write = self.outf.write
164
eol_marker = self.opts.eol_marker
165
def _line_writer(line):
166
write(line + eol_marker)
167
def _line_writer_color(line):
168
write(FG.BOLD_MAGENTA + line + FG.NONE + eol_marker)
169
if self.opts.show_color:
170
return _line_writer_color
175
def get_revision_header_writer(self):
176
"""Get function for writing revno lines"""
177
write = self.outf.write
178
eol_marker = self.opts.eol_marker
179
def _line_writer(line):
180
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
193
def _line_writer(line):
194
write(line + eol_marker)
197
def _get_writer_regexp_highlighted(self):
198
"""Get function for writing output with regexp match highlighted"""
199
_line_writer = self._get_writer_plain()
200
sub, highlight = self._sub, self._highlight
201
def _line_writer_regexp_highlighted(line):
202
"""Write formatted line with matched pattern highlighted"""
203
return _line_writer(line=sub(highlight, line))
204
return _line_writer_regexp_highlighted
206
def _get_writer_fixed_highlighted(self):
207
"""Get function for writing output with search string highlighted"""
208
_line_writer = self._get_writer_plain()
209
old, new = self._old, self._new
210
def _line_writer_fixed_highlighted(line):
211
"""Write formatted line with string searched for highlighted"""
212
return _line_writer(line=line.replace(old, new))
213
return _line_writer_fixed_highlighted
217
wt, branch, relpath = \
218
controldir.ControlDir.open_containing_tree_or_branch('.')
222
start_rev = opts.revision[0]
224
# if no revision is sepcified for diff grep we grep all changesets.
225
opts.revision = [RevisionSpec.from_string('revno:1'),
226
RevisionSpec.from_string('last:1')]
227
start_rev = opts.revision[0]
228
start_revid = start_rev.as_revision_id(branch)
229
if start_revid == 'null:':
231
srevno_tuple = branch.revision_id_to_dotted_revno(start_revid)
232
if len(opts.revision) == 2:
233
end_rev = opts.revision[1]
234
end_revid = end_rev.as_revision_id(branch)
235
if end_revid is None:
236
end_revno, end_revid = branch.last_revision_info()
237
erevno_tuple = branch.revision_id_to_dotted_revno(end_revid)
239
grep_mainline = (_rev_on_mainline(srevno_tuple) and
240
_rev_on_mainline(erevno_tuple))
242
# ensure that we go in reverse order
243
if srevno_tuple > erevno_tuple:
244
srevno_tuple, erevno_tuple = erevno_tuple, srevno_tuple
245
start_revid, end_revid = end_revid, start_revid
247
# Optimization: Traversing the mainline in reverse order is much
248
# faster when we don't want to look at merged revs. We try this
249
# with _linear_view_revisions. If all revs are to be grepped we
250
# use the slower _graph_view_revisions
251
if opts.levels==1 and grep_mainline:
252
given_revs = _linear_view_revisions(branch, start_revid, end_revid)
254
given_revs = _graph_view_revisions(branch, start_revid, end_revid)
256
# We do an optimization below. For grepping a specific revison
257
# We don't need to call _graph_view_revisions which is slow.
258
# We create the start_rev_tuple for only that specific revision.
259
# _graph_view_revisions is used only for revision range.
260
start_revno = '.'.join(map(str, srevno_tuple))
261
start_rev_tuple = (start_revid, start_revno, 0)
262
given_revs = [start_rev_tuple]
263
repo = branch.repository
264
diff_pattern = re.compile("^[+\-].*(" + opts.pattern + ")")
265
file_pattern = re.compile("=== (modified|added|removed) file '.*'", re.UNICODE)
266
outputter = _GrepDiffOutputter(opts)
267
writeline = outputter.get_writer()
268
writerevno = outputter.get_revision_header_writer()
269
writefileheader = outputter.get_file_header_writer()
270
file_encoding = _user_encoding
271
for revid, revno, merge_depth in given_revs:
272
if opts.levels == 1 and merge_depth != 0:
273
# with level=1 show only top level
276
rev_spec = RevisionSpec_revid.from_string("revid:"+revid)
277
new_rev = repo.get_revision(revid)
278
new_tree = rev_spec.as_tree(branch)
279
if len(new_rev.parent_ids) == 0:
280
ancestor_id = _mod_revision.NULL_REVISION
282
ancestor_id = new_rev.parent_ids[0]
283
old_tree = repo.revision_tree(ancestor_id)
285
diff.show_diff_trees(old_tree, new_tree, s,
286
old_label='', new_label='')
291
for line in text.splitlines():
292
if file_pattern.search(line):
295
elif diff_pattern.search(line):
297
writerevno("=== revno:%s ===" % (revno,))
298
display_revno = False
300
writefileheader(" %s" % (file_header,))
302
line = line.decode(file_encoding, 'replace')
303
writeline(" %s" % (line,))
308
def versioned_grep(opts):
309
wt, branch, relpath = \
310
controldir.ControlDir.open_containing_tree_or_branch('.')
313
start_rev = opts.revision[0]
314
start_revid = start_rev.as_revision_id(branch)
315
if start_revid is None:
316
start_rev = RevisionSpec_revno.from_string("revno:1")
317
start_revid = start_rev.as_revision_id(branch)
318
srevno_tuple = branch.revision_id_to_dotted_revno(start_revid)
320
if len(opts.revision) == 2:
321
end_rev = opts.revision[1]
322
end_revid = end_rev.as_revision_id(branch)
323
if end_revid is None:
324
end_revno, end_revid = branch.last_revision_info()
325
erevno_tuple = branch.revision_id_to_dotted_revno(end_revid)
327
grep_mainline = (_rev_on_mainline(srevno_tuple) and
328
_rev_on_mainline(erevno_tuple))
330
# ensure that we go in reverse order
331
if srevno_tuple > erevno_tuple:
332
srevno_tuple, erevno_tuple = erevno_tuple, srevno_tuple
333
start_revid, end_revid = end_revid, start_revid
335
# Optimization: Traversing the mainline in reverse order is much
336
# faster when we don't want to look at merged revs. We try this
337
# with _linear_view_revisions. If all revs are to be grepped we
338
# use the slower _graph_view_revisions
339
if opts.levels == 1 and grep_mainline:
340
given_revs = _linear_view_revisions(branch, start_revid, end_revid)
342
given_revs = _graph_view_revisions(branch, start_revid, end_revid)
344
# We do an optimization below. For grepping a specific revison
345
# We don't need to call _graph_view_revisions which is slow.
346
# We create the start_rev_tuple for only that specific revision.
347
# _graph_view_revisions is used only for revision range.
348
start_revno = '.'.join(map(str, srevno_tuple))
349
start_rev_tuple = (start_revid, start_revno, 0)
350
given_revs = [start_rev_tuple]
352
# GZ 2010-06-02: Shouldn't be smuggling this on opts, but easy for now
353
opts.outputter = _Outputter(opts, use_cache=True)
355
for revid, revno, merge_depth in given_revs:
356
if opts.levels == 1 and merge_depth != 0:
357
# with level=1 show only top level
360
rev = RevisionSpec_revid.from_string("revid:"+revid)
361
tree = rev.as_tree(branch)
362
for path in opts.path_list:
363
path_for_id = osutils.pathjoin(relpath, path)
364
id = tree.path2id(path_for_id)
366
trace.warning("Skipped unknown file '%s'." % path)
369
if osutils.isdir(path):
371
dir_grep(tree, path, relpath, opts, revno, path_prefix)
373
versioned_file_grep(tree, id, '.', path, opts, revno)
378
def workingtree_grep(opts):
379
revno = opts.print_revno = None # for working tree set revno to None
381
tree, branch, relpath = \
382
controldir.ControlDir.open_containing_tree_or_branch('.')
384
msg = ('Cannot search working tree. Working tree not found.\n'
385
'To search for specific revision in history use the -r option.')
386
raise errors.BzrCommandError(msg)
388
# GZ 2010-06-02: Shouldn't be smuggling this on opts, but easy for now
389
opts.outputter = _Outputter(opts)
393
for path in opts.path_list:
394
if osutils.isdir(path):
396
dir_grep(tree, path, relpath, opts, revno, path_prefix)
398
_file_grep(open(path).read(), path, opts, revno)
403
def _skip_file(include, exclude, path):
404
if include and not _path_in_glob_list(path, include):
406
if exclude and _path_in_glob_list(path, exclude):
411
def dir_grep(tree, path, relpath, opts, revno, path_prefix):
412
# setup relpath to open files relative to cwd
415
rpath = osutils.pathjoin('..',relpath)
417
from_dir = osutils.pathjoin(relpath, path)
419
# start searching recursively from root
424
to_grep_append = to_grep.append
425
# GZ 2010-06-05: The cache dict used to be recycled every call to dir_grep
426
# and hits manually refilled. Could do this again if it was
427
# for a good reason, otherwise cache might want purging.
428
outputter = opts.outputter
429
for fp, fc, fkind, fid, entry in tree.list_files(include_root=False,
430
from_dir=from_dir, recursive=opts.recursive):
432
if _skip_file(opts.include, opts.exclude, fp):
435
if fc == 'V' and fkind == 'file':
437
# If old result is valid, print results immediately.
438
# Otherwise, add file info to to_grep so that the
439
# loop later will get chunks and grep them
440
cache_id = tree.get_file_revision(fid)
441
if cache_id in outputter.cache:
442
# GZ 2010-06-05: Not really sure caching and re-outputting
443
# the old path is really the right thing,
444
# but it's what the old code seemed to do
445
outputter.write_cached_lines(cache_id, revno)
447
to_grep_append((fid, (fp, fid)))
449
# we are grepping working tree.
453
path_for_file = osutils.pathjoin(tree.basedir, from_dir, fp)
454
if opts.files_with_matches or opts.files_without_match:
455
# Optimize for wtree list-only as we don't need to read the
457
file = open(path_for_file, 'r', buffering=4096)
458
_file_grep_list_only_wtree(file, fp, opts, path_prefix)
460
file_text = open(path_for_file, 'r').read()
461
_file_grep(file_text, fp, opts, revno, path_prefix)
463
if revno != None: # grep versioned files
464
for (path, fid), chunks in tree.iter_files_bytes(to_grep):
465
path = _make_display_path(relpath, path)
466
_file_grep(chunks[0], path, opts, revno, path_prefix,
467
tree.get_file_revision(fid, path))
470
def _make_display_path(relpath, path):
471
"""Return path string relative to user cwd.
473
Take tree's 'relpath' and user supplied 'path', and return path
474
that can be displayed to the user.
477
# update path so to display it w.r.t cwd
478
# handle windows slash separator
479
path = osutils.normpath(osutils.pathjoin(relpath, path))
480
path = path.replace('\\', '/')
481
path = path.replace(relpath + '/', '', 1)
485
def versioned_file_grep(tree, id, relpath, path, opts, revno, path_prefix = None):
486
"""Create a file object for the specified id and pass it on to _file_grep.
489
path = _make_display_path(relpath, path)
490
file_text = tree.get_file_text(id)
491
_file_grep(file_text, path, opts, revno, path_prefix)
494
def _path_in_glob_list(path, glob_list):
495
for glob in glob_list:
496
if fnmatch(path, glob):
501
def _file_grep_list_only_wtree(file, path, opts, path_prefix=None):
502
# test and skip binary files
503
if '\x00' in file.read(1024):
505
trace.warning("Binary file '%s' skipped." % path)
508
file.seek(0) # search from beginning
511
if opts.fixed_string:
512
pattern = opts.pattern.encode(_user_encoding, 'replace')
517
else: # not fixed_string
519
if opts.patternc.search(line):
523
if (opts.files_with_matches and found) or \
524
(opts.files_without_match and not found):
525
if path_prefix and path_prefix != '.':
526
# user has passed a dir arg, show that as result prefix
527
path = osutils.pathjoin(path_prefix, path)
528
opts.outputter.get_writer(path, None, None)()
531
class _Outputter(object):
532
"""Precalculate formatting based on options given
534
The idea here is to do this work only once per run, and finally return a
535
function that will do the minimum amount possible for each match.
537
def __init__(self, opts, use_cache=False):
538
self.outf = opts.outf
540
# self.cache is used to cache results for dir grep based on fid.
541
# If the fid is does not change between results, it means that
542
# the result will be the same apart from revno. In such a case
543
# we avoid getting file chunks from repo and grepping. The result
544
# is just printed by replacing old revno with new one.
548
no_line = opts.files_with_matches or opts.files_without_match
551
pat = opts.pattern.encode(_user_encoding, 'replace')
553
self.get_writer = self._get_writer_plain
554
elif opts.fixed_string:
556
self._new = color_string(pat, FG.BOLD_RED)
557
self.get_writer = self._get_writer_fixed_highlighted
559
flags = opts.patternc.flags
560
self._sub = re.compile(pat.join(("((?:",")+)")), flags).sub
561
self._highlight = color_string("\\1", FG.BOLD_RED)
562
self.get_writer = self._get_writer_regexp_highlighted
563
path_start = FG.MAGENTA
565
sep = color_string(':', FG.BOLD_CYAN)
566
rev_sep = color_string('~', FG.BOLD_YELLOW)
568
self.get_writer = self._get_writer_plain
569
path_start = path_end = ""
573
parts = [path_start, "%(path)s"]
575
parts.extend([rev_sep, "%(revno)s"])
576
self._format_initial = "".join(parts)
579
if not opts.print_revno:
580
parts.append(path_end)
583
parts.extend([sep, "%(lineno)s"])
584
parts.extend([sep, "%(line)s"])
585
parts.append(opts.eol_marker)
586
self._format_perline = "".join(parts)
588
def _get_writer_plain(self, path, revno, cache_id):
589
"""Get function for writing uncoloured output"""
590
per_line = self._format_perline
591
start = self._format_initial % {"path":path, "revno":revno}
592
write = self.outf.write
593
if self.cache is not None and cache_id is not None:
595
self.cache[cache_id] = path, result_list
596
add_to_cache = result_list.append
597
def _line_cache_and_writer(**kwargs):
598
"""Write formatted line and cache arguments"""
599
end = per_line % kwargs
602
return _line_cache_and_writer
603
def _line_writer(**kwargs):
604
"""Write formatted line from arguments given by underlying opts"""
605
write(start + per_line % kwargs)
608
def write_cached_lines(self, cache_id, revno):
609
"""Write cached results out again for new revision"""
610
cached_path, cached_matches = self.cache[cache_id]
611
start = self._format_initial % {"path":cached_path, "revno":revno}
612
write = self.outf.write
613
for end in cached_matches:
616
def _get_writer_regexp_highlighted(self, path, revno, cache_id):
617
"""Get function for writing output with regexp match highlighted"""
618
_line_writer = self._get_writer_plain(path, revno, cache_id)
619
sub, highlight = self._sub, self._highlight
620
def _line_writer_regexp_highlighted(line, **kwargs):
621
"""Write formatted line with matched pattern highlighted"""
622
return _line_writer(line=sub(highlight, line), **kwargs)
623
return _line_writer_regexp_highlighted
625
def _get_writer_fixed_highlighted(self, path, revno, cache_id):
626
"""Get function for writing output with search string highlighted"""
627
_line_writer = self._get_writer_plain(path, revno, cache_id)
628
old, new = self._old, self._new
629
def _line_writer_fixed_highlighted(line, **kwargs):
630
"""Write formatted line with string searched for highlighted"""
631
return _line_writer(line=line.replace(old, new), **kwargs)
632
return _line_writer_fixed_highlighted
635
def _file_grep(file_text, path, opts, revno, path_prefix=None, cache_id=None):
636
# test and skip binary files
637
if '\x00' in file_text[:1024]:
639
trace.warning("Binary file '%s' skipped." % path)
642
if path_prefix and path_prefix != '.':
643
# user has passed a dir arg, show that as result prefix
644
path = osutils.pathjoin(path_prefix, path)
646
# GZ 2010-06-07: There's no actual guarentee the file contents will be in
647
# the user encoding, but we have to guess something and it
648
# is a reasonable default without a better mechanism.
649
file_encoding = _user_encoding
650
pattern = opts.pattern.encode(_user_encoding, 'replace')
652
writeline = opts.outputter.get_writer(path, revno, cache_id)
654
if opts.files_with_matches or opts.files_without_match:
655
if opts.fixed_string:
656
found = pattern in file_text
658
search = opts.patternc.search
659
if "$" not in pattern:
660
found = search(file_text) is not None
662
for line in file_text.splitlines():
668
if (opts.files_with_matches and found) or \
669
(opts.files_without_match and not found):
671
elif opts.fixed_string:
672
# Fast path for no match, search through the entire file at once rather
673
# than a line at a time. <http://effbot.org/zone/stringlib.htm>
674
i = file_text.find(pattern)
677
b = file_text.rfind("\n", 0, i) + 1
679
start = file_text.count("\n", 0, b) + 1
680
file_text = file_text[b:]
682
for index, line in enumerate(file_text.splitlines()):
684
line = line.decode(file_encoding, 'replace')
685
writeline(lineno=index+start, line=line)
687
for line in file_text.splitlines():
689
line = line.decode(file_encoding, 'replace')
692
# Fast path on no match, the re module avoids bad behaviour in most
693
# standard cases, but perhaps could try and detect backtracking
694
# patterns here and avoid whole text search in those cases
695
search = opts.patternc.search
696
if "$" not in pattern:
697
# GZ 2010-06-05: Grr, re.MULTILINE can't save us when searching
698
# through revisions as bazaar returns binary mode
699
# and trailing \r breaks $ as line ending match
700
m = search(file_text)
703
b = file_text.rfind("\n", 0, m.start()) + 1
705
start = file_text.count("\n", 0, b) + 1
706
file_text = file_text[b:]
710
for index, line in enumerate(file_text.splitlines()):
712
line = line.decode(file_encoding, 'replace')
713
writeline(lineno=index+start, line=line)
715
for line in file_text.splitlines():
717
line = line.decode(file_encoding, 'replace')
added
removed
breezy/plugins/grep/grep.py
