« back to all changes in this revision
Viewing changes to tools/history2weaves.py
- browse files at revision 1219
- compare with another revision
- download diff
- download tarball
- view history from revision 1219
- Committer: Martin Pool
- Date: 2005-09-13 01:37:23 UTC
- Revision ID: mbp@sourcefrog.net-20050913013723-7e0026b48cbf08ff
- BROKEN: start refactoring fetch code to work well with weaves
- files removed:
- bzrlib/selftest/test_ancestry.py
- files renamed:
- bzrlib/selftest/test_weave.py => tools/testweave.py
- files modified:
- .bzrignore
added
removed
tools/history2weaves.py
1
1
#! /usr/bin/python
2
#
2
3
3
# Copyright (C) 2005 Canonical Ltd
4
#
4
5
5
# This program is free software; you can redistribute it and/or modify
6
6
# it under the terms of the GNU General Public License as published by
7
7
# the Free Software Foundation; either version 2 of the License, or
8
8
# (at your option) any later version.
9
#
9
10
10
# This program is distributed in the hope that it will be useful,
11
11
# but WITHOUT ANY WARRANTY; without even the implied warranty of
12
12
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13
13
# GNU General Public License for more details.
14
#
14
15
15
# You should have received a copy of the GNU General Public License
16
16
# along with this program; if not, write to the Free Software
17
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
18
18
19
19
"""Experiment in converting existing bzr branches to weaves."""
20
20
21
# To make this properly useful
22
#
23
# 1. assign text version ids, and put those text versions into
24
# the inventory as they're converted.
25
#
26
# 2. keep track of the previous version of each file, rather than
27
# just using the last one imported
28
#
29
# 3. assign entry versions when files are added, renamed or moved.
30
#
31
# 4. when merged-in versions are observed, walk down through them
32
# to discover everything, then commit bottom-up
33
#
34
# 5. track ancestry as things are merged in, and commit that in each
35
# revision
36
#
37
# Perhaps it's best to first walk the whole graph and make a plan for
38
# what should be imported in what order? Need a kind of topological
39
# sort of all revisions. (Or do we, can we just before doing a revision
40
# see that all its parents have either been converted or abandoned?)
41
42
if False:
43
try:
44
import psyco
45
psyco.full()
46
except ImportError:
47
pass
48
49
50
import tempfile
51
import hotshot, hotshot.stats
52
import sys
21
try:
22
import psyco
23
psyco.full()
24
except ImportError:
25
pass
26
27
53
28
import logging
54
import time
55
29
56
from bzrlib.branch import Branch, find_branch
30
import bzrlib.branch
57
31
from bzrlib.revfile import Revfile
58
32
from bzrlib.weave import Weave
59
33
from bzrlib.weavefile import read_weave, write_weave
60
34
from bzrlib.progress import ProgressBar
61
35
from bzrlib.atomicfile import AtomicFile
62
from bzrlib.xml4 import serializer_v4
63
from bzrlib.xml5 import serializer_v5
64
36
import bzrlib.trace
65
66
67
68
class Convert(object):
69
def __init__(self):
70
self.total_revs = 0
71
self.converted_revs = 0
72
self.text_count = 0
73
self.convert()
74
75
76
77
def convert(self):
78
bzrlib.trace.enable_default_logging()
79
self.pb = ProgressBar()
80
self.inv_weave = Weave('__inventory')
81
self.anc_weave = Weave('__ancestry')
82
83
last_text_sha = {}
84
85
# holds in-memory weaves for all files
86
text_weaves = {}
87
88
b = self.branch = Branch('.', relax_version_check=True)
89
90
revno = 1
91
rev_history = b.revision_history()
92
last_idx = None
93
inv_parents = []
94
95
# todo is a stack holding the revisions we still need to process;
96
# appending to it adds new highest-priority revisions
97
todo = rev_history[:]
98
todo.reverse()
99
self.total_revs = len(todo)
100
101
while todo:
102
self._convert_one_rev(todo.pop())
103
104
self.pb.clear()
105
print 'upgraded to weaves:'
106
print ' %6d revisions and inventories' % self.converted_revs
107
print ' %6d texts' % self.text_count
108
109
self._write_all_weaves()
110
111
112
def _write_all_weaves(self):
113
i = 0
114
return ############################################
115
# TODO: commit them all atomically at the end, not one by one
116
write_atomic_weave(self.inv_weave, 'weaves/inventory.weave')
117
write_atomic_weave(self.anc_weave, 'weaves/ancestry.weave')
118
for file_id, file_weave in text_weaves.items():
119
self.pb.update('writing weave', i, len(text_weaves))
120
write_atomic_weave(file_weave, 'weaves/%s.weave' % file_id)
121
i += 1
122
123
self.pb.clear()
124
125
126
def _convert_one_rev(self, rev_id):
127
self._bump_progress()
128
b = self.branch
129
rev_xml = b.revision_store[rev_id].read()
130
inv_xml = b.inventory_store[rev_id].read()
131
132
rev = serializer_v4.read_revision_from_string(rev_xml)
133
inv = serializer_v4.read_inventory_from_string(inv_xml)
134
135
return ##########################################
136
137
new_idx = self.inv_weave.add(rev_id, inv_parents, inv_xml)
37
import tempfile
38
import hotshot, hotshot.stats
39
import sys
40
41
def convert():
42
bzrlib.trace.enable_default_logging()
43
44
pb = ProgressBar()
45
46
inv_weave = Weave()
47
48
last_text_sha = {}
49
50
# holds in-memory weaves for all files
51
text_weaves = {}
52
53
b = bzrlib.branch.find_branch('.')
54
55
revno = 1
56
rev_history = b.revision_history()
57
last_idx = None
58
inv_parents = []
59
text_count = 0
60
61
for rev_id in rev_history:
62
pb.update('converting revision', revno, len(rev_history))
63
64
inv_xml = b.get_inventory_xml(rev_id).readlines()
65
66
new_idx = inv_weave.add(rev_id, inv_parents, inv_xml)
138
67
inv_parents = [new_idx]
139
68
140
69
tree = b.revision_tree(rev_id)
157
86
# revision then make a new weave; else find the old one
158
87
if file_id not in text_weaves:
159
88
text_weaves[file_id] = Weave()
160
89
161
90
w = text_weaves[file_id]
162
91
163
92
# base the new text version off whatever was last
173
102
text_count += 1
174
103
175
104
revno += 1
176
177
def _bump_progress(self):
178
self.converted_revs += 1
179
self.pb.update('converting revisions',
180
self.converted_revs,
181
self.total_revs)
105
106
pb.clear()
107
print '%6d revisions and inventories' % revno
108
print '%6d texts' % text_count
109
110
i = 0
111
# TODO: commit them all atomically at the end, not one by one
112
write_atomic_weave(inv_weave, 'weaves/inventory.weave')
113
for file_id, file_weave in text_weaves.items():
114
pb.update('writing weave', i, len(text_weaves))
115
write_atomic_weave(file_weave, 'weaves/%s.weave' % file_id)
116
i += 1
117
118
pb.clear()
182
119
183
120
184
121
def write_atomic_weave(weave, filename):
197
134
198
135
prof = hotshot.Profile(prof_f.name)
199
136
200
prof.runcall(Convert)
137
prof.runcall(convert)
201
138
prof.close()
202
139
203
140
stats = hotshot.stats.load(prof_f.name)
204
##stats.strip_dirs()
141
#stats.strip_dirs()
205
142
stats.sort_stats('time')
206
# XXX: Might like to write to stderr or the trace file instead but
207
# print_stats seems hardcoded to stdout
143
## XXX: Might like to write to stderr or the trace file instead but
144
## print_stats seems hardcoded to stdout
208
145
stats.print_stats(20)
209
146
210
147
211
148
if '-p' in sys.argv[1:]:
212
149
profile_convert()
213
150
else:
214
Convert()
151
convert()
215
152
Loggerhead is a web-based interface for Breezy
Version: 2.0.1