« back to all changes in this revision
Viewing changes to bzrlib/tsort.py
- browse files at revision 4629
- compare with another revision
- compare with revision 4629
- download diff
- download tarball
- view history from revision 4629
- Committer: Canonical.com Patch Queue Manager
- Date: 2009-08-19 18:04:49 UTC
- mfrom: (4593.5.43 1.19-known-graph-sorted)
- Revision ID: pqm@pqm.ubuntu.com-20090819180449-p5dibldf9pcp24n4
(jam) Add VersionedFiles.get_known_graph_ancestry and
KnownGraph.merge_sort()
KnownGraph.merge_sort()
- files renamed:
- bzrlib/tests/test_versionedfile.py => bzrlib/tests/per_versionedfile.py
- files modified:
- NEWS
added
removed
bzrlib/tsort.py
18
18
"""Topological sorting routines."""
19
19
20
20
21
from bzrlib import errors
22
import bzrlib.revision as _mod_revision
21
from bzrlib import (
22
errors,
23
graph as _mod_graph,
24
revision as _mod_revision,
25
)
23
26
24
27
25
28
__all__ = ["topo_sort", "TopoSorter", "merge_sort", "MergeSorter"]
30
33
31
34
graph -- sequence of pairs of node->parents_list.
32
35
33
The result is a list of node names, such that all parents come before
34
their children.
36
The result is a list of node names, such that all parents come before their
37
children.
35
38
36
39
node identifiers can be any hashable object, and are typically strings.
40
41
This function has the same purpose as the TopoSorter class, but uses a
42
different algorithm to sort the graph. That means that while both return a
43
list with parents before their child nodes, the exact ordering can be
44
different.
45
46
topo_sort is faster when the whole list is needed, while when iterating
47
over a part of the list, TopoSorter.iter_topo_order should be used.
37
48
"""
38
return TopoSorter(graph).sorted()
49
kg = _mod_graph.KnownGraph(dict(graph))
50
return kg.topo_sort()
39
51
40
52
41
53
class TopoSorter(object):
60
72
iteration or sorting may raise GraphCycleError if a cycle is present
61
73
in the graph.
62
74
"""
63
# a dict of the graph.
75
# store a dict of the graph.
64
76
self._graph = dict(graph)
65
self._visitable = set(self._graph)
66
### if debugging:
67
# self._original_graph = dict(graph)
68
69
# this is a stack storing the depth first search into the graph.
70
self._node_name_stack = []
71
# at each level of 'recursion' we have to check each parent. This
72
# stack stores the parents we have not yet checked for the node at the
73
# matching depth in _node_name_stack
74
self._pending_parents_stack = []
75
# this is a set of the completed nodes for fast checking whether a
76
# parent in a node we are processing on the stack has already been
77
# emitted and thus can be skipped.
78
self._completed_node_names = set()
79
77
80
78
def sorted(self):
81
79
"""Sort the graph and return as a list.
100
98
After finishing iteration the sorter is empty and you cannot continue
101
99
iteration.
102
100
"""
103
while self._graph:
101
graph = self._graph
102
visitable = set(graph)
103
104
# this is a stack storing the depth first search into the graph.
105
pending_node_stack = []
106
# at each level of 'recursion' we have to check each parent. This
107
# stack stores the parents we have not yet checked for the node at the
108
# matching depth in pending_node_stack
109
pending_parents_stack = []
110
111
# this is a set of the completed nodes for fast checking whether a
112
# parent in a node we are processing on the stack has already been
113
# emitted and thus can be skipped.
114
completed_node_names = set()
115
116
while graph:
104
117
# now pick a random node in the source graph, and transfer it to the
105
# top of the depth first search stack.
106
node_name, parents = self._graph.popitem()
107
self._push_node(node_name, parents)
108
while self._node_name_stack:
109
# loop until this call completes.
110
parents_to_visit = self._pending_parents_stack[-1]
111
# if all parents are done, the revision is done
118
# top of the depth first search stack of pending nodes.
119
node_name, parents = graph.popitem()
120
pending_node_stack.append(node_name)
121
pending_parents_stack.append(list(parents))
122
123
# loop until pending_node_stack is empty
124
while pending_node_stack:
125
parents_to_visit = pending_parents_stack[-1]
126
# if there are no parents left, the revision is done
112
127
if not parents_to_visit:
113
128
# append the revision to the topo sorted list
114
# all the nodes parents have been added to the output, now
115
# we can add it to the output.
116
yield self._pop_node()
129
# all the nodes parents have been added to the output,
130
# now we can add it to the output.
131
popped_node = pending_node_stack.pop()
132
pending_parents_stack.pop()
133
completed_node_names.add(popped_node)
134
yield popped_node
117
135
else:
118
while self._pending_parents_stack[-1]:
119
# recurse depth first into a single parent
120
next_node_name = self._pending_parents_stack[-1].pop()
121
if next_node_name in self._completed_node_names:
122
# this parent was completed by a child on the
123
# call stack. skip it.
124
continue
125
if next_node_name not in self._visitable:
126
continue
127
# otherwise transfer it from the source graph into the
128
# top of the current depth first search stack.
129
try:
130
parents = self._graph.pop(next_node_name)
131
except KeyError:
132
# if the next node is not in the source graph it has
133
# already been popped from it and placed into the
134
# current search stack (but not completed or we would
135
# have hit the continue 4 lines up.
136
# this indicates a cycle.
137
raise errors.GraphCycleError(self._node_name_stack)
138
self._push_node(next_node_name, parents)
139
# and do not continue processing parents until this 'call'
140
# has recursed.
141
break
142
143
def _push_node(self, node_name, parents):
144
"""Add node_name to the pending node stack.
145
146
Names in this stack will get emitted into the output as they are popped
147
off the stack.
148
"""
149
self._node_name_stack.append(node_name)
150
self._pending_parents_stack.append(list(parents))
151
152
def _pop_node(self):
153
"""Pop the top node off the stack
154
155
The node is appended to the sorted output.
156
"""
157
# we are returning from the flattened call frame:
158
# pop off the local variables
159
node_name = self._node_name_stack.pop()
160
self._pending_parents_stack.pop()
161
162
self._completed_node_names.add(node_name)
163
return node_name
136
# recurse depth first into a single parent
137
next_node_name = parents_to_visit.pop()
138
139
if next_node_name in completed_node_names:
140
# parent was already completed by a child, skip it.
141
continue
142
if next_node_name not in visitable:
143
# parent is not a node in the original graph, skip it.
144
continue
145
146
# transfer it along with its parents from the source graph
147
# into the top of the current depth first search stack.
148
try:
149
parents = graph.pop(next_node_name)
150
except KeyError:
151
# if the next node is not in the source graph it has
152
# already been popped from it and placed into the
153
# current search stack (but not completed or we would
154
# have hit the continue 6 lines up). this indicates a
155
# cycle.
156
raise errors.GraphCycleError(pending_node_stack)
157
pending_node_stack.append(next_node_name)
158
pending_parents_stack.append(list(parents))
164
159
165
160
166
161
def merge_sort(graph, branch_tip, mainline_revisions=None, generate_revno=False):
414
409
415
410
# seed the search with the tip of the branch
416
411
if (branch_tip is not None and
417
branch_tip != _mod_revision.NULL_REVISION):
412
branch_tip != _mod_revision.NULL_REVISION and
413
branch_tip != (_mod_revision.NULL_REVISION,)):
418
414
parents = self._graph.pop(branch_tip)
419
415
self._push_node(branch_tip, 0, parents)
420
416
571
567
# current search stack (but not completed or we would
572
568
# have hit the continue 4 lines up.
573
569
# this indicates a cycle.
574
raise errors.GraphCycleError(node_name_stack)
570
if next_node_name in self._original_graph:
571
raise errors.GraphCycleError(node_name_stack)
572
else:
573
# This is just a ghost parent, ignore it
574
continue
575
575
next_merge_depth = 0
576
576
if is_left_subtree:
577
577
# a new child branch from name_stack[-1]
673
673
else:
674
674
# no parents, use the root sequence
675
675
root_count = self._revno_to_branch_count.get(0, 0)
676
root_count = self._revno_to_branch_count.get(0, -1)
677
root_count += 1
676
678
if root_count:
677
679
revno = (0, root_count, 1)
678
680
else:
679
681
revno = (1,)
680
root_count += 1
681
682
self._revno_to_branch_count[0] = root_count
682
683
683
684
# store the revno for this node for future reference
Loggerhead is a web-based interface for Breezy
Version: 2.0.1