Show More
@@ -1,450 +1,454 b'' | |||||
1 | # branchmap.py - logic to computes, maintain and stores branchmap for local repo |
|
1 | # branchmap.py - logic to computes, maintain and stores branchmap for local repo | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from node import bin, hex, nullid, nullrev |
|
8 | from node import bin, hex, nullid, nullrev | |
9 | import encoding |
|
9 | import encoding | |
10 | import util |
|
10 | import util | |
11 | import time |
|
11 | import time | |
12 | from array import array |
|
12 | from array import array | |
13 | from struct import calcsize, pack, unpack |
|
13 | from struct import calcsize, pack, unpack | |
14 |
|
14 | |||
15 | def _filename(repo): |
|
15 | def _filename(repo): | |
16 | """name of a branchcache file for a given repo or repoview""" |
|
16 | """name of a branchcache file for a given repo or repoview""" | |
17 | filename = "cache/branch2" |
|
17 | filename = "cache/branch2" | |
18 | if repo.filtername: |
|
18 | if repo.filtername: | |
19 | filename = '%s-%s' % (filename, repo.filtername) |
|
19 | filename = '%s-%s' % (filename, repo.filtername) | |
20 | return filename |
|
20 | return filename | |
21 |
|
21 | |||
22 | def read(repo): |
|
22 | def read(repo): | |
23 | try: |
|
23 | try: | |
24 | f = repo.vfs(_filename(repo)) |
|
24 | f = repo.vfs(_filename(repo)) | |
25 | lines = f.read().split('\n') |
|
25 | lines = f.read().split('\n') | |
26 | f.close() |
|
26 | f.close() | |
27 | except (IOError, OSError): |
|
27 | except (IOError, OSError): | |
28 | return None |
|
28 | return None | |
29 |
|
29 | |||
30 | try: |
|
30 | try: | |
31 | cachekey = lines.pop(0).split(" ", 2) |
|
31 | cachekey = lines.pop(0).split(" ", 2) | |
32 | last, lrev = cachekey[:2] |
|
32 | last, lrev = cachekey[:2] | |
33 | last, lrev = bin(last), int(lrev) |
|
33 | last, lrev = bin(last), int(lrev) | |
34 | filteredhash = None |
|
34 | filteredhash = None | |
35 | if len(cachekey) > 2: |
|
35 | if len(cachekey) > 2: | |
36 | filteredhash = bin(cachekey[2]) |
|
36 | filteredhash = bin(cachekey[2]) | |
37 | partial = branchcache(tipnode=last, tiprev=lrev, |
|
37 | partial = branchcache(tipnode=last, tiprev=lrev, | |
38 | filteredhash=filteredhash) |
|
38 | filteredhash=filteredhash) | |
39 | if not partial.validfor(repo): |
|
39 | if not partial.validfor(repo): | |
40 | # invalidate the cache |
|
40 | # invalidate the cache | |
41 | raise ValueError('tip differs') |
|
41 | raise ValueError('tip differs') | |
42 | for l in lines: |
|
42 | for l in lines: | |
43 | if not l: |
|
43 | if not l: | |
44 | continue |
|
44 | continue | |
45 | node, state, label = l.split(" ", 2) |
|
45 | node, state, label = l.split(" ", 2) | |
46 | if state not in 'oc': |
|
46 | if state not in 'oc': | |
47 | raise ValueError('invalid branch state') |
|
47 | raise ValueError('invalid branch state') | |
48 | label = encoding.tolocal(label.strip()) |
|
48 | label = encoding.tolocal(label.strip()) | |
49 | if not node in repo: |
|
49 | if not node in repo: | |
50 | raise ValueError('node %s does not exist' % node) |
|
50 | raise ValueError('node %s does not exist' % node) | |
51 | node = bin(node) |
|
51 | node = bin(node) | |
52 | partial.setdefault(label, []).append(node) |
|
52 | partial.setdefault(label, []).append(node) | |
53 | if state == 'c': |
|
53 | if state == 'c': | |
54 | partial._closednodes.add(node) |
|
54 | partial._closednodes.add(node) | |
55 | except KeyboardInterrupt: |
|
55 | except KeyboardInterrupt: | |
56 | raise |
|
56 | raise | |
57 | except Exception, inst: |
|
57 | except Exception, inst: | |
58 | if repo.ui.debugflag: |
|
58 | if repo.ui.debugflag: | |
59 | msg = 'invalid branchheads cache' |
|
59 | msg = 'invalid branchheads cache' | |
60 | if repo.filtername is not None: |
|
60 | if repo.filtername is not None: | |
61 | msg += ' (%s)' % repo.filtername |
|
61 | msg += ' (%s)' % repo.filtername | |
62 | msg += ': %s\n' |
|
62 | msg += ': %s\n' | |
63 | repo.ui.debug(msg % inst) |
|
63 | repo.ui.debug(msg % inst) | |
64 | partial = None |
|
64 | partial = None | |
65 | return partial |
|
65 | return partial | |
66 |
|
66 | |||
67 | ### Nearest subset relation |
|
67 | ### Nearest subset relation | |
68 | # Nearest subset of filter X is a filter Y so that: |
|
68 | # Nearest subset of filter X is a filter Y so that: | |
69 | # * Y is included in X, |
|
69 | # * Y is included in X, | |
70 | # * X - Y is as small as possible. |
|
70 | # * X - Y is as small as possible. | |
71 | # This create and ordering used for branchmap purpose. |
|
71 | # This create and ordering used for branchmap purpose. | |
72 | # the ordering may be partial |
|
72 | # the ordering may be partial | |
73 | subsettable = {None: 'visible', |
|
73 | subsettable = {None: 'visible', | |
74 | 'visible': 'served', |
|
74 | 'visible': 'served', | |
75 | 'served': 'immutable', |
|
75 | 'served': 'immutable', | |
76 | 'immutable': 'base'} |
|
76 | 'immutable': 'base'} | |
77 |
|
77 | |||
78 | def updatecache(repo): |
|
78 | def updatecache(repo): | |
79 | cl = repo.changelog |
|
79 | cl = repo.changelog | |
80 | filtername = repo.filtername |
|
80 | filtername = repo.filtername | |
81 | partial = repo._branchcaches.get(filtername) |
|
81 | partial = repo._branchcaches.get(filtername) | |
82 |
|
82 | |||
83 | revs = [] |
|
83 | revs = [] | |
84 | if partial is None or not partial.validfor(repo): |
|
84 | if partial is None or not partial.validfor(repo): | |
85 | partial = read(repo) |
|
85 | partial = read(repo) | |
86 | if partial is None: |
|
86 | if partial is None: | |
87 | subsetname = subsettable.get(filtername) |
|
87 | subsetname = subsettable.get(filtername) | |
88 | if subsetname is None: |
|
88 | if subsetname is None: | |
89 | partial = branchcache() |
|
89 | partial = branchcache() | |
90 | else: |
|
90 | else: | |
91 | subset = repo.filtered(subsetname) |
|
91 | subset = repo.filtered(subsetname) | |
92 | partial = subset.branchmap().copy() |
|
92 | partial = subset.branchmap().copy() | |
93 | extrarevs = subset.changelog.filteredrevs - cl.filteredrevs |
|
93 | extrarevs = subset.changelog.filteredrevs - cl.filteredrevs | |
94 | revs.extend(r for r in extrarevs if r <= partial.tiprev) |
|
94 | revs.extend(r for r in extrarevs if r <= partial.tiprev) | |
95 | revs.extend(cl.revs(start=partial.tiprev + 1)) |
|
95 | revs.extend(cl.revs(start=partial.tiprev + 1)) | |
96 | if revs: |
|
96 | if revs: | |
97 | partial.update(repo, revs) |
|
97 | partial.update(repo, revs) | |
98 | partial.write(repo) |
|
98 | partial.write(repo) | |
99 |
|
99 | |||
100 | if repo._revbranchcache is not None: |
|
100 | if repo._revbranchcache is not None: | |
101 | repo._revbranchcache.write() |
|
101 | repo._revbranchcache.write() | |
102 |
|
102 | |||
103 | assert partial.validfor(repo), filtername |
|
103 | assert partial.validfor(repo), filtername | |
104 | repo._branchcaches[repo.filtername] = partial |
|
104 | repo._branchcaches[repo.filtername] = partial | |
105 |
|
105 | |||
106 | class branchcache(dict): |
|
106 | class branchcache(dict): | |
107 | """A dict like object that hold branches heads cache. |
|
107 | """A dict like object that hold branches heads cache. | |
108 |
|
108 | |||
109 | This cache is used to avoid costly computations to determine all the |
|
109 | This cache is used to avoid costly computations to determine all the | |
110 | branch heads of a repo. |
|
110 | branch heads of a repo. | |
111 |
|
111 | |||
112 | The cache is serialized on disk in the following format: |
|
112 | The cache is serialized on disk in the following format: | |
113 |
|
113 | |||
114 | <tip hex node> <tip rev number> [optional filtered repo hex hash] |
|
114 | <tip hex node> <tip rev number> [optional filtered repo hex hash] | |
115 | <branch head hex node> <open/closed state> <branch name> |
|
115 | <branch head hex node> <open/closed state> <branch name> | |
116 | <branch head hex node> <open/closed state> <branch name> |
|
116 | <branch head hex node> <open/closed state> <branch name> | |
117 | ... |
|
117 | ... | |
118 |
|
118 | |||
119 | The first line is used to check if the cache is still valid. If the |
|
119 | The first line is used to check if the cache is still valid. If the | |
120 | branch cache is for a filtered repo view, an optional third hash is |
|
120 | branch cache is for a filtered repo view, an optional third hash is | |
121 | included that hashes the hashes of all filtered revisions. |
|
121 | included that hashes the hashes of all filtered revisions. | |
122 |
|
122 | |||
123 | The open/closed state is represented by a single letter 'o' or 'c'. |
|
123 | The open/closed state is represented by a single letter 'o' or 'c'. | |
124 | This field can be used to avoid changelog reads when determining if a |
|
124 | This field can be used to avoid changelog reads when determining if a | |
125 | branch head closes a branch or not. |
|
125 | branch head closes a branch or not. | |
126 | """ |
|
126 | """ | |
127 |
|
127 | |||
128 | def __init__(self, entries=(), tipnode=nullid, tiprev=nullrev, |
|
128 | def __init__(self, entries=(), tipnode=nullid, tiprev=nullrev, | |
129 | filteredhash=None, closednodes=None): |
|
129 | filteredhash=None, closednodes=None): | |
130 | super(branchcache, self).__init__(entries) |
|
130 | super(branchcache, self).__init__(entries) | |
131 | self.tipnode = tipnode |
|
131 | self.tipnode = tipnode | |
132 | self.tiprev = tiprev |
|
132 | self.tiprev = tiprev | |
133 | self.filteredhash = filteredhash |
|
133 | self.filteredhash = filteredhash | |
134 | # closednodes is a set of nodes that close their branch. If the branch |
|
134 | # closednodes is a set of nodes that close their branch. If the branch | |
135 | # cache has been updated, it may contain nodes that are no longer |
|
135 | # cache has been updated, it may contain nodes that are no longer | |
136 | # heads. |
|
136 | # heads. | |
137 | if closednodes is None: |
|
137 | if closednodes is None: | |
138 | self._closednodes = set() |
|
138 | self._closednodes = set() | |
139 | else: |
|
139 | else: | |
140 | self._closednodes = closednodes |
|
140 | self._closednodes = closednodes | |
141 |
|
141 | |||
142 | def _hashfiltered(self, repo): |
|
142 | def _hashfiltered(self, repo): | |
143 | """build hash of revision filtered in the current cache |
|
143 | """build hash of revision filtered in the current cache | |
144 |
|
144 | |||
145 | Tracking tipnode and tiprev is not enough to ensure validity of the |
|
145 | Tracking tipnode and tiprev is not enough to ensure validity of the | |
146 | cache as they do not help to distinct cache that ignored various |
|
146 | cache as they do not help to distinct cache that ignored various | |
147 | revision bellow tiprev. |
|
147 | revision bellow tiprev. | |
148 |
|
148 | |||
149 | To detect such difference, we build a cache of all ignored revisions. |
|
149 | To detect such difference, we build a cache of all ignored revisions. | |
150 | """ |
|
150 | """ | |
151 | cl = repo.changelog |
|
151 | cl = repo.changelog | |
152 | if not cl.filteredrevs: |
|
152 | if not cl.filteredrevs: | |
153 | return None |
|
153 | return None | |
154 | key = None |
|
154 | key = None | |
155 | revs = sorted(r for r in cl.filteredrevs if r <= self.tiprev) |
|
155 | revs = sorted(r for r in cl.filteredrevs if r <= self.tiprev) | |
156 | if revs: |
|
156 | if revs: | |
157 | s = util.sha1() |
|
157 | s = util.sha1() | |
158 | for rev in revs: |
|
158 | for rev in revs: | |
159 | s.update('%s;' % rev) |
|
159 | s.update('%s;' % rev) | |
160 | key = s.digest() |
|
160 | key = s.digest() | |
161 | return key |
|
161 | return key | |
162 |
|
162 | |||
163 | def validfor(self, repo): |
|
163 | def validfor(self, repo): | |
164 | """Is the cache content valid regarding a repo |
|
164 | """Is the cache content valid regarding a repo | |
165 |
|
165 | |||
166 | - False when cached tipnode is unknown or if we detect a strip. |
|
166 | - False when cached tipnode is unknown or if we detect a strip. | |
167 | - True when cache is up to date or a subset of current repo.""" |
|
167 | - True when cache is up to date or a subset of current repo.""" | |
168 | try: |
|
168 | try: | |
169 | return ((self.tipnode == repo.changelog.node(self.tiprev)) |
|
169 | return ((self.tipnode == repo.changelog.node(self.tiprev)) | |
170 | and (self.filteredhash == self._hashfiltered(repo))) |
|
170 | and (self.filteredhash == self._hashfiltered(repo))) | |
171 | except IndexError: |
|
171 | except IndexError: | |
172 | return False |
|
172 | return False | |
173 |
|
173 | |||
174 | def _branchtip(self, heads): |
|
174 | def _branchtip(self, heads): | |
175 | '''Return tuple with last open head in heads and false, |
|
175 | '''Return tuple with last open head in heads and false, | |
176 | otherwise return last closed head and true.''' |
|
176 | otherwise return last closed head and true.''' | |
177 | tip = heads[-1] |
|
177 | tip = heads[-1] | |
178 | closed = True |
|
178 | closed = True | |
179 | for h in reversed(heads): |
|
179 | for h in reversed(heads): | |
180 | if h not in self._closednodes: |
|
180 | if h not in self._closednodes: | |
181 | tip = h |
|
181 | tip = h | |
182 | closed = False |
|
182 | closed = False | |
183 | break |
|
183 | break | |
184 | return tip, closed |
|
184 | return tip, closed | |
185 |
|
185 | |||
186 | def branchtip(self, branch): |
|
186 | def branchtip(self, branch): | |
187 | '''Return the tipmost open head on branch head, otherwise return the |
|
187 | '''Return the tipmost open head on branch head, otherwise return the | |
188 | tipmost closed head on branch. |
|
188 | tipmost closed head on branch. | |
189 | Raise KeyError for unknown branch.''' |
|
189 | Raise KeyError for unknown branch.''' | |
190 | return self._branchtip(self[branch])[0] |
|
190 | return self._branchtip(self[branch])[0] | |
191 |
|
191 | |||
192 | def branchheads(self, branch, closed=False): |
|
192 | def branchheads(self, branch, closed=False): | |
193 | heads = self[branch] |
|
193 | heads = self[branch] | |
194 | if not closed: |
|
194 | if not closed: | |
195 | heads = [h for h in heads if h not in self._closednodes] |
|
195 | heads = [h for h in heads if h not in self._closednodes] | |
196 | return heads |
|
196 | return heads | |
197 |
|
197 | |||
198 | def iterbranches(self): |
|
198 | def iterbranches(self): | |
199 | for bn, heads in self.iteritems(): |
|
199 | for bn, heads in self.iteritems(): | |
200 | yield (bn, heads) + self._branchtip(heads) |
|
200 | yield (bn, heads) + self._branchtip(heads) | |
201 |
|
201 | |||
202 | def copy(self): |
|
202 | def copy(self): | |
203 | """return an deep copy of the branchcache object""" |
|
203 | """return an deep copy of the branchcache object""" | |
204 | return branchcache(self, self.tipnode, self.tiprev, self.filteredhash, |
|
204 | return branchcache(self, self.tipnode, self.tiprev, self.filteredhash, | |
205 | self._closednodes) |
|
205 | self._closednodes) | |
206 |
|
206 | |||
207 | def write(self, repo): |
|
207 | def write(self, repo): | |
208 | try: |
|
208 | try: | |
209 | f = repo.vfs(_filename(repo), "w", atomictemp=True) |
|
209 | f = repo.vfs(_filename(repo), "w", atomictemp=True) | |
210 | cachekey = [hex(self.tipnode), str(self.tiprev)] |
|
210 | cachekey = [hex(self.tipnode), str(self.tiprev)] | |
211 | if self.filteredhash is not None: |
|
211 | if self.filteredhash is not None: | |
212 | cachekey.append(hex(self.filteredhash)) |
|
212 | cachekey.append(hex(self.filteredhash)) | |
213 | f.write(" ".join(cachekey) + '\n') |
|
213 | f.write(" ".join(cachekey) + '\n') | |
214 | nodecount = 0 |
|
214 | nodecount = 0 | |
215 | for label, nodes in sorted(self.iteritems()): |
|
215 | for label, nodes in sorted(self.iteritems()): | |
216 | for node in nodes: |
|
216 | for node in nodes: | |
217 | nodecount += 1 |
|
217 | nodecount += 1 | |
218 | if node in self._closednodes: |
|
218 | if node in self._closednodes: | |
219 | state = 'c' |
|
219 | state = 'c' | |
220 | else: |
|
220 | else: | |
221 | state = 'o' |
|
221 | state = 'o' | |
222 | f.write("%s %s %s\n" % (hex(node), state, |
|
222 | f.write("%s %s %s\n" % (hex(node), state, | |
223 | encoding.fromlocal(label))) |
|
223 | encoding.fromlocal(label))) | |
224 | f.close() |
|
224 | f.close() | |
225 | repo.ui.log('branchcache', |
|
225 | repo.ui.log('branchcache', | |
226 | 'wrote %s branch cache with %d labels and %d nodes\n', |
|
226 | 'wrote %s branch cache with %d labels and %d nodes\n', | |
227 | repo.filtername, len(self), nodecount) |
|
227 | repo.filtername, len(self), nodecount) | |
228 | except (IOError, OSError, util.Abort), inst: |
|
228 | except (IOError, OSError, util.Abort), inst: | |
229 | repo.ui.debug("couldn't write branch cache: %s\n" % inst) |
|
229 | repo.ui.debug("couldn't write branch cache: %s\n" % inst) | |
230 | # Abort may be raise by read only opener |
|
230 | # Abort may be raise by read only opener | |
231 | pass |
|
231 | pass | |
232 |
|
232 | |||
233 | def update(self, repo, revgen): |
|
233 | def update(self, repo, revgen): | |
234 | """Given a branchhead cache, self, that may have extra nodes or be |
|
234 | """Given a branchhead cache, self, that may have extra nodes or be | |
235 | missing heads, and a generator of nodes that are strictly a superset of |
|
235 | missing heads, and a generator of nodes that are strictly a superset of | |
236 | heads missing, this function updates self to be correct. |
|
236 | heads missing, this function updates self to be correct. | |
237 | """ |
|
237 | """ | |
238 | starttime = time.time() |
|
238 | starttime = time.time() | |
239 | cl = repo.changelog |
|
239 | cl = repo.changelog | |
240 | # collect new branch entries |
|
240 | # collect new branch entries | |
241 | newbranches = {} |
|
241 | newbranches = {} | |
242 | getbranchinfo = repo.revbranchcache().branchinfo |
|
242 | getbranchinfo = repo.revbranchcache().branchinfo | |
243 | for r in revgen: |
|
243 | for r in revgen: | |
244 | branch, closesbranch = getbranchinfo(r) |
|
244 | branch, closesbranch = getbranchinfo(r) | |
245 | newbranches.setdefault(branch, []).append(r) |
|
245 | newbranches.setdefault(branch, []).append(r) | |
246 | if closesbranch: |
|
246 | if closesbranch: | |
247 | self._closednodes.add(cl.node(r)) |
|
247 | self._closednodes.add(cl.node(r)) | |
248 |
|
248 | |||
249 | # fetch current topological heads to speed up filtering |
|
249 | # fetch current topological heads to speed up filtering | |
250 | topoheads = set(cl.headrevs()) |
|
250 | topoheads = set(cl.headrevs()) | |
251 |
|
251 | |||
252 | # if older branchheads are reachable from new ones, they aren't |
|
252 | # if older branchheads are reachable from new ones, they aren't | |
253 | # really branchheads. Note checking parents is insufficient: |
|
253 | # really branchheads. Note checking parents is insufficient: | |
254 | # 1 (branch a) -> 2 (branch b) -> 3 (branch a) |
|
254 | # 1 (branch a) -> 2 (branch b) -> 3 (branch a) | |
255 | for branch, newheadrevs in newbranches.iteritems(): |
|
255 | for branch, newheadrevs in newbranches.iteritems(): | |
256 | bheads = self.setdefault(branch, []) |
|
256 | bheads = self.setdefault(branch, []) | |
257 | bheadset = set(cl.rev(node) for node in bheads) |
|
257 | bheadset = set(cl.rev(node) for node in bheads) | |
258 |
|
258 | |||
259 | # This have been tested True on all internal usage of this function. |
|
259 | # This have been tested True on all internal usage of this function. | |
260 | # run it again in case of doubt |
|
260 | # run it again in case of doubt | |
261 | # assert not (set(bheadrevs) & set(newheadrevs)) |
|
261 | # assert not (set(bheadrevs) & set(newheadrevs)) | |
262 | newheadrevs.sort() |
|
262 | newheadrevs.sort() | |
263 | bheadset.update(newheadrevs) |
|
263 | bheadset.update(newheadrevs) | |
264 |
|
264 | |||
265 | # This prunes out two kinds of heads - heads that are superseded by |
|
265 | # This prunes out two kinds of heads - heads that are superseded by | |
266 | # a head in newheadrevs, and newheadrevs that are not heads because |
|
266 | # a head in newheadrevs, and newheadrevs that are not heads because | |
267 | # an existing head is their descendant. |
|
267 | # an existing head is their descendant. | |
268 | uncertain = bheadset - topoheads |
|
268 | uncertain = bheadset - topoheads | |
269 | if uncertain: |
|
269 | if uncertain: | |
270 | floorrev = min(uncertain) |
|
270 | floorrev = min(uncertain) | |
271 | ancestors = set(cl.ancestors(newheadrevs, floorrev)) |
|
271 | ancestors = set(cl.ancestors(newheadrevs, floorrev)) | |
272 | bheadset -= ancestors |
|
272 | bheadset -= ancestors | |
273 | bheadrevs = sorted(bheadset) |
|
273 | bheadrevs = sorted(bheadset) | |
274 | self[branch] = [cl.node(rev) for rev in bheadrevs] |
|
274 | self[branch] = [cl.node(rev) for rev in bheadrevs] | |
275 | tiprev = bheadrevs[-1] |
|
275 | tiprev = bheadrevs[-1] | |
276 | if tiprev > self.tiprev: |
|
276 | if tiprev > self.tiprev: | |
277 | self.tipnode = cl.node(tiprev) |
|
277 | self.tipnode = cl.node(tiprev) | |
278 | self.tiprev = tiprev |
|
278 | self.tiprev = tiprev | |
279 |
|
279 | |||
280 | if not self.validfor(repo): |
|
280 | if not self.validfor(repo): | |
281 | # cache key are not valid anymore |
|
281 | # cache key are not valid anymore | |
282 | self.tipnode = nullid |
|
282 | self.tipnode = nullid | |
283 | self.tiprev = nullrev |
|
283 | self.tiprev = nullrev | |
284 | for heads in self.values(): |
|
284 | for heads in self.values(): | |
285 | tiprev = max(cl.rev(node) for node in heads) |
|
285 | tiprev = max(cl.rev(node) for node in heads) | |
286 | if tiprev > self.tiprev: |
|
286 | if tiprev > self.tiprev: | |
287 | self.tipnode = cl.node(tiprev) |
|
287 | self.tipnode = cl.node(tiprev) | |
288 | self.tiprev = tiprev |
|
288 | self.tiprev = tiprev | |
289 | self.filteredhash = self._hashfiltered(repo) |
|
289 | self.filteredhash = self._hashfiltered(repo) | |
290 |
|
290 | |||
291 | duration = time.time() - starttime |
|
291 | duration = time.time() - starttime | |
292 | repo.ui.log('branchcache', 'updated %s branch cache in %.4f seconds\n', |
|
292 | repo.ui.log('branchcache', 'updated %s branch cache in %.4f seconds\n', | |
293 | repo.filtername, duration) |
|
293 | repo.filtername, duration) | |
294 |
|
294 | |||
295 | # Revision branch info cache |
|
295 | # Revision branch info cache | |
296 |
|
296 | |||
297 | _rbcversion = '-v1' |
|
297 | _rbcversion = '-v1' | |
298 | _rbcnames = 'cache/rbc-names' + _rbcversion |
|
298 | _rbcnames = 'cache/rbc-names' + _rbcversion | |
299 | _rbcrevs = 'cache/rbc-revs' + _rbcversion |
|
299 | _rbcrevs = 'cache/rbc-revs' + _rbcversion | |
300 | # [4 byte hash prefix][4 byte branch name number with sign bit indicating open] |
|
300 | # [4 byte hash prefix][4 byte branch name number with sign bit indicating open] | |
301 | _rbcrecfmt = '>4sI' |
|
301 | _rbcrecfmt = '>4sI' | |
302 | _rbcrecsize = calcsize(_rbcrecfmt) |
|
302 | _rbcrecsize = calcsize(_rbcrecfmt) | |
303 | _rbcnodelen = 4 |
|
303 | _rbcnodelen = 4 | |
304 | _rbcbranchidxmask = 0x7fffffff |
|
304 | _rbcbranchidxmask = 0x7fffffff | |
305 | _rbccloseflag = 0x80000000 |
|
305 | _rbccloseflag = 0x80000000 | |
306 |
|
306 | |||
307 | class revbranchcache(object): |
|
307 | class revbranchcache(object): | |
308 | """Persistent cache, mapping from revision number to branch name and close. |
|
308 | """Persistent cache, mapping from revision number to branch name and close. | |
309 | This is a low level cache, independent of filtering. |
|
309 | This is a low level cache, independent of filtering. | |
310 |
|
310 | |||
311 | Branch names are stored in rbc-names in internal encoding separated by 0. |
|
311 | Branch names are stored in rbc-names in internal encoding separated by 0. | |
312 | rbc-names is append-only, and each branch name is only stored once and will |
|
312 | rbc-names is append-only, and each branch name is only stored once and will | |
313 | thus have a unique index. |
|
313 | thus have a unique index. | |
314 |
|
314 | |||
315 | The branch info for each revision is stored in rbc-revs as constant size |
|
315 | The branch info for each revision is stored in rbc-revs as constant size | |
316 | records. The whole file is read into memory, but it is only 'parsed' on |
|
316 | records. The whole file is read into memory, but it is only 'parsed' on | |
317 | demand. The file is usually append-only but will be truncated if repo |
|
317 | demand. The file is usually append-only but will be truncated if repo | |
318 | modification is detected. |
|
318 | modification is detected. | |
319 | The record for each revision contains the first 4 bytes of the |
|
319 | The record for each revision contains the first 4 bytes of the | |
320 | corresponding node hash, and the record is only used if it still matches. |
|
320 | corresponding node hash, and the record is only used if it still matches. | |
321 | Even a completely trashed rbc-revs fill thus still give the right result |
|
321 | Even a completely trashed rbc-revs fill thus still give the right result | |
322 | while converging towards full recovery ... assuming no incorrectly matching |
|
322 | while converging towards full recovery ... assuming no incorrectly matching | |
323 | node hashes. |
|
323 | node hashes. | |
324 | The record also contains 4 bytes where 31 bits contains the index of the |
|
324 | The record also contains 4 bytes where 31 bits contains the index of the | |
325 | branch and the last bit indicate that it is a branch close commit. |
|
325 | branch and the last bit indicate that it is a branch close commit. | |
326 | The usage pattern for rbc-revs is thus somewhat similar to 00changelog.i |
|
326 | The usage pattern for rbc-revs is thus somewhat similar to 00changelog.i | |
327 | and will grow with it but be 1/8th of its size. |
|
327 | and will grow with it but be 1/8th of its size. | |
328 | """ |
|
328 | """ | |
329 |
|
329 | |||
330 | def __init__(self, repo, readonly=True): |
|
330 | def __init__(self, repo, readonly=True): | |
331 | assert repo.filtername is None |
|
331 | assert repo.filtername is None | |
332 | self._repo = repo |
|
332 | self._repo = repo | |
333 | self._names = [] # branch names in local encoding with static index |
|
333 | self._names = [] # branch names in local encoding with static index | |
334 | self._rbcrevs = array('c') # structs of type _rbcrecfmt |
|
334 | self._rbcrevs = array('c') # structs of type _rbcrecfmt | |
335 | self._rbcsnameslen = 0 |
|
335 | self._rbcsnameslen = 0 | |
336 | try: |
|
336 | try: | |
337 | bndata = repo.vfs.read(_rbcnames) |
|
337 | bndata = repo.vfs.read(_rbcnames) | |
338 | self._rbcsnameslen = len(bndata) # for verification before writing |
|
338 | self._rbcsnameslen = len(bndata) # for verification before writing | |
339 | self._names = [encoding.tolocal(bn) for bn in bndata.split('\0')] |
|
339 | self._names = [encoding.tolocal(bn) for bn in bndata.split('\0')] | |
340 | except (IOError, OSError), inst: |
|
340 | except (IOError, OSError), inst: | |
341 | repo.ui.debug("couldn't read revision branch cache names: %s\n" % |
|
341 | repo.ui.debug("couldn't read revision branch cache names: %s\n" % | |
342 | inst) |
|
342 | inst) | |
343 | if readonly: |
|
343 | if readonly: | |
344 | # don't try to use cache - fall back to the slow path |
|
344 | # don't try to use cache - fall back to the slow path | |
345 | self.branchinfo = self._branchinfo |
|
345 | self.branchinfo = self._branchinfo | |
346 |
|
346 | |||
347 | if self._names: |
|
347 | if self._names: | |
348 | try: |
|
348 | try: | |
349 | data = repo.vfs.read(_rbcrevs) |
|
349 | data = repo.vfs.read(_rbcrevs) | |
350 | self._rbcrevs.fromstring(data) |
|
350 | self._rbcrevs.fromstring(data) | |
351 | except (IOError, OSError), inst: |
|
351 | except (IOError, OSError), inst: | |
352 | repo.ui.debug("couldn't read revision branch cache: %s\n" % |
|
352 | repo.ui.debug("couldn't read revision branch cache: %s\n" % | |
353 | inst) |
|
353 | inst) | |
354 | # remember number of good records on disk |
|
354 | # remember number of good records on disk | |
355 | self._rbcrevslen = min(len(self._rbcrevs) // _rbcrecsize, |
|
355 | self._rbcrevslen = min(len(self._rbcrevs) // _rbcrecsize, | |
356 | len(repo.changelog)) |
|
356 | len(repo.changelog)) | |
357 | if self._rbcrevslen == 0: |
|
357 | if self._rbcrevslen == 0: | |
358 | self._names = [] |
|
358 | self._names = [] | |
359 | self._rbcnamescount = len(self._names) # number of good names on disk |
|
359 | self._rbcnamescount = len(self._names) # number of good names on disk | |
360 | self._namesreverse = dict((b, r) for r, b in enumerate(self._names)) |
|
360 | self._namesreverse = dict((b, r) for r, b in enumerate(self._names)) | |
361 |
|
361 | |||
362 | def branchinfo(self, rev): |
|
362 | def branchinfo(self, rev): | |
363 | """Return branch name and close flag for rev, using and updating |
|
363 | """Return branch name and close flag for rev, using and updating | |
364 | persistent cache.""" |
|
364 | persistent cache.""" | |
365 | changelog = self._repo.changelog |
|
365 | changelog = self._repo.changelog | |
366 | rbcrevidx = rev * _rbcrecsize |
|
366 | rbcrevidx = rev * _rbcrecsize | |
367 |
|
367 | |||
368 | # if requested rev is missing, add and populate all missing revs |
|
368 | # if requested rev is missing, add and populate all missing revs | |
369 | if len(self._rbcrevs) < rbcrevidx + _rbcrecsize: |
|
369 | if len(self._rbcrevs) < rbcrevidx + _rbcrecsize: | |
370 | first = len(self._rbcrevs) // _rbcrecsize |
|
370 | first = len(self._rbcrevs) // _rbcrecsize | |
371 | self._rbcrevs.extend('\0' * (len(changelog) * _rbcrecsize - |
|
371 | self._rbcrevs.extend('\0' * (len(changelog) * _rbcrecsize - | |
372 | len(self._rbcrevs))) |
|
372 | len(self._rbcrevs))) | |
373 | for r in xrange(first, len(changelog)): |
|
373 | for r in xrange(first, len(changelog)): | |
374 | self._branchinfo(r) |
|
374 | self._branchinfo(r) | |
375 |
|
375 | |||
376 | # fast path: extract data from cache, use it if node is matching |
|
376 | # fast path: extract data from cache, use it if node is matching | |
377 | reponode = changelog.node(rev)[:_rbcnodelen] |
|
377 | reponode = changelog.node(rev)[:_rbcnodelen] | |
378 | cachenode, branchidx = unpack( |
|
378 | cachenode, branchidx = unpack( | |
379 | _rbcrecfmt, buffer(self._rbcrevs, rbcrevidx, _rbcrecsize)) |
|
379 | _rbcrecfmt, buffer(self._rbcrevs, rbcrevidx, _rbcrecsize)) | |
380 | close = bool(branchidx & _rbccloseflag) |
|
380 | close = bool(branchidx & _rbccloseflag) | |
381 | if close: |
|
381 | if close: | |
382 | branchidx &= _rbcbranchidxmask |
|
382 | branchidx &= _rbcbranchidxmask | |
383 | if cachenode == reponode: |
|
383 | if cachenode == reponode: | |
384 | return self._names[branchidx], close |
|
384 | return self._names[branchidx], close | |
385 | # fall back to slow path and make sure it will be written to disk |
|
385 | # fall back to slow path and make sure it will be written to disk | |
386 | self._rbcrevslen = min(self._rbcrevslen, rev) |
|
386 | self._rbcrevslen = min(self._rbcrevslen, rev) | |
387 | return self._branchinfo(rev) |
|
387 | return self._branchinfo(rev) | |
388 |
|
388 | |||
389 | def _branchinfo(self, rev): |
|
389 | def _branchinfo(self, rev): | |
390 | """Retrieve branch info from changelog and update _rbcrevs""" |
|
390 | """Retrieve branch info from changelog and update _rbcrevs""" | |
391 | changelog = self._repo.changelog |
|
391 | changelog = self._repo.changelog | |
392 | b, close = changelog.branchinfo(rev) |
|
392 | b, close = changelog.branchinfo(rev) | |
393 | if b in self._namesreverse: |
|
393 | if b in self._namesreverse: | |
394 | branchidx = self._namesreverse[b] |
|
394 | branchidx = self._namesreverse[b] | |
395 | else: |
|
395 | else: | |
396 | branchidx = len(self._names) |
|
396 | branchidx = len(self._names) | |
397 | self._names.append(b) |
|
397 | self._names.append(b) | |
398 | self._namesreverse[b] = branchidx |
|
398 | self._namesreverse[b] = branchidx | |
399 | reponode = changelog.node(rev) |
|
399 | reponode = changelog.node(rev) | |
400 | if close: |
|
400 | if close: | |
401 | branchidx |= _rbccloseflag |
|
401 | branchidx |= _rbccloseflag | |
|
402 | self._setcachedata(rev, reponode, branchidx) | |||
|
403 | return b, close | |||
|
404 | ||||
|
405 | def _setcachedata(self, rev, node, branchidx): | |||
|
406 | """Writes the node's branch data to the in-memory cache data.""" | |||
402 | rbcrevidx = rev * _rbcrecsize |
|
407 | rbcrevidx = rev * _rbcrecsize | |
403 | rec = array('c') |
|
408 | rec = array('c') | |
404 |
rec.fromstring(pack(_rbcrecfmt, |
|
409 | rec.fromstring(pack(_rbcrecfmt, node, branchidx)) | |
405 | self._rbcrevs[rbcrevidx:rbcrevidx + _rbcrecsize] = rec |
|
410 | self._rbcrevs[rbcrevidx:rbcrevidx + _rbcrecsize] = rec | |
406 | return b, close |
|
|||
407 |
|
411 | |||
408 | def write(self): |
|
412 | def write(self): | |
409 | """Save branch cache if it is dirty.""" |
|
413 | """Save branch cache if it is dirty.""" | |
410 | repo = self._repo |
|
414 | repo = self._repo | |
411 | if self._rbcnamescount < len(self._names): |
|
415 | if self._rbcnamescount < len(self._names): | |
412 | try: |
|
416 | try: | |
413 | if self._rbcnamescount != 0: |
|
417 | if self._rbcnamescount != 0: | |
414 | f = repo.vfs.open(_rbcnames, 'ab') |
|
418 | f = repo.vfs.open(_rbcnames, 'ab') | |
415 | if f.tell() == self._rbcsnameslen: |
|
419 | if f.tell() == self._rbcsnameslen: | |
416 | f.write('\0') |
|
420 | f.write('\0') | |
417 | else: |
|
421 | else: | |
418 | f.close() |
|
422 | f.close() | |
419 | repo.ui.debug("%s changed - rewriting it\n" % _rbcnames) |
|
423 | repo.ui.debug("%s changed - rewriting it\n" % _rbcnames) | |
420 | self._rbcnamescount = 0 |
|
424 | self._rbcnamescount = 0 | |
421 | self._rbcrevslen = 0 |
|
425 | self._rbcrevslen = 0 | |
422 | if self._rbcnamescount == 0: |
|
426 | if self._rbcnamescount == 0: | |
423 | f = repo.vfs.open(_rbcnames, 'wb') |
|
427 | f = repo.vfs.open(_rbcnames, 'wb') | |
424 | f.write('\0'.join(encoding.fromlocal(b) |
|
428 | f.write('\0'.join(encoding.fromlocal(b) | |
425 | for b in self._names[self._rbcnamescount:])) |
|
429 | for b in self._names[self._rbcnamescount:])) | |
426 | self._rbcsnameslen = f.tell() |
|
430 | self._rbcsnameslen = f.tell() | |
427 | f.close() |
|
431 | f.close() | |
428 | except (IOError, OSError, util.Abort), inst: |
|
432 | except (IOError, OSError, util.Abort), inst: | |
429 | repo.ui.debug("couldn't write revision branch cache names: " |
|
433 | repo.ui.debug("couldn't write revision branch cache names: " | |
430 | "%s\n" % inst) |
|
434 | "%s\n" % inst) | |
431 | return |
|
435 | return | |
432 | self._rbcnamescount = len(self._names) |
|
436 | self._rbcnamescount = len(self._names) | |
433 |
|
437 | |||
434 | start = self._rbcrevslen * _rbcrecsize |
|
438 | start = self._rbcrevslen * _rbcrecsize | |
435 | if start != len(self._rbcrevs): |
|
439 | if start != len(self._rbcrevs): | |
436 | revs = min(len(repo.changelog), len(self._rbcrevs) // _rbcrecsize) |
|
440 | revs = min(len(repo.changelog), len(self._rbcrevs) // _rbcrecsize) | |
437 | try: |
|
441 | try: | |
438 | f = repo.vfs.open(_rbcrevs, 'ab') |
|
442 | f = repo.vfs.open(_rbcrevs, 'ab') | |
439 | if f.tell() != start: |
|
443 | if f.tell() != start: | |
440 | repo.ui.debug("truncating %s to %s\n" % (_rbcrevs, start)) |
|
444 | repo.ui.debug("truncating %s to %s\n" % (_rbcrevs, start)) | |
441 | f.seek(start) |
|
445 | f.seek(start) | |
442 | f.truncate() |
|
446 | f.truncate() | |
443 | end = revs * _rbcrecsize |
|
447 | end = revs * _rbcrecsize | |
444 | f.write(self._rbcrevs[start:end]) |
|
448 | f.write(self._rbcrevs[start:end]) | |
445 | f.close() |
|
449 | f.close() | |
446 | except (IOError, OSError, util.Abort), inst: |
|
450 | except (IOError, OSError, util.Abort), inst: | |
447 | repo.ui.debug("couldn't write revision branch cache: %s\n" % |
|
451 | repo.ui.debug("couldn't write revision branch cache: %s\n" % | |
448 | inst) |
|
452 | inst) | |
449 | return |
|
453 | return | |
450 | self._rbcrevslen = revs |
|
454 | self._rbcrevslen = revs |
General Comments 0
You need to be logged in to leave comments.
Login now