Show More
@@ -1,198 +1,198 b'' | |||
|
1 | 1 | # changelog.py - changelog class for mercurial |
|
2 | 2 | # |
|
3 | 3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
4 | 4 | # |
|
5 | 5 | # This software may be used and distributed according to the terms |
|
6 | 6 | # of the GNU General Public License, incorporated herein by reference. |
|
7 | 7 | |
|
8 | 8 | from revlog import * |
|
9 | 9 | from i18n import _ |
|
10 | 10 | import os, time, util |
|
11 | 11 | |
|
12 | 12 | def _string_escape(text): |
|
13 | 13 | """ |
|
14 | 14 | >>> d = {'nl': chr(10), 'bs': chr(92), 'cr': chr(13), 'nul': chr(0)} |
|
15 | 15 | >>> s = "ab%(nl)scd%(bs)s%(bs)sn%(nul)sab%(cr)scd%(bs)s%(nl)s" % d |
|
16 | 16 | >>> s |
|
17 | 17 | 'ab\\ncd\\\\\\\\n\\x00ab\\rcd\\\\\\n' |
|
18 | 18 | >>> res = _string_escape(s) |
|
19 | 19 | >>> s == _string_unescape(res) |
|
20 | 20 | True |
|
21 | 21 | """ |
|
22 | 22 | # subset of the string_escape codec |
|
23 | 23 | text = text.replace('\\', '\\\\').replace('\n', '\\n').replace('\r', '\\r') |
|
24 | 24 | return text.replace('\0', '\\0') |
|
25 | 25 | |
|
26 | 26 | def _string_unescape(text): |
|
27 | 27 | return text.decode('string_escape') |
|
28 | 28 | |
|
29 | 29 | class appender: |
|
30 | 30 | '''the changelog index must be update last on disk, so we use this class |
|
31 | 31 | to delay writes to it''' |
|
32 | 32 | def __init__(self, fp, buf): |
|
33 | 33 | self.data = buf |
|
34 | 34 | self.fp = fp |
|
35 | 35 | self.offset = fp.tell() |
|
36 | 36 | self.size = util.fstat(fp).st_size |
|
37 | 37 | |
|
38 | 38 | def end(self): |
|
39 | 39 | return self.size + len("".join(self.data)) |
|
40 | 40 | def tell(self): |
|
41 | 41 | return self.offset |
|
42 | 42 | def flush(self): |
|
43 | 43 | pass |
|
44 | 44 | def close(self): |
|
45 | 45 | self.fp.close() |
|
46 | 46 | |
|
47 | 47 | def seek(self, offset, whence=0): |
|
48 | 48 | '''virtual file offset spans real file and data''' |
|
49 | 49 | if whence == 0: |
|
50 | 50 | self.offset = offset |
|
51 | 51 | elif whence == 1: |
|
52 | 52 | self.offset += offset |
|
53 | 53 | elif whence == 2: |
|
54 | 54 | self.offset = self.end() + offset |
|
55 | 55 | if self.offset < self.size: |
|
56 | 56 | self.fp.seek(self.offset) |
|
57 | 57 | |
|
58 | 58 | def read(self, count=-1): |
|
59 | 59 | '''only trick here is reads that span real file and data''' |
|
60 | 60 | ret = "" |
|
61 | 61 | if self.offset < self.size: |
|
62 | 62 | s = self.fp.read(count) |
|
63 | 63 | ret = s |
|
64 | 64 | self.offset += len(s) |
|
65 | 65 | if count > 0: |
|
66 | 66 | count -= len(s) |
|
67 | 67 | if count != 0: |
|
68 | 68 | doff = self.offset - self.size |
|
69 | 69 | self.data.insert(0, "".join(self.data)) |
|
70 | 70 | del self.data[1:] |
|
71 | 71 | s = self.data[0][doff:doff+count] |
|
72 | 72 | self.offset += len(s) |
|
73 | 73 | ret += s |
|
74 | 74 | return ret |
|
75 | 75 | |
|
76 | 76 | def write(self, s): |
|
77 | self.data.append(s) | |
|
77 | self.data.append(str(s)) | |
|
78 | 78 | self.offset += len(s) |
|
79 | 79 | |
|
80 | 80 | class changelog(revlog): |
|
81 | 81 | def __init__(self, opener): |
|
82 | 82 | revlog.__init__(self, opener, "00changelog.i") |
|
83 | 83 | |
|
84 | 84 | def delayupdate(self): |
|
85 | 85 | "delay visibility of index updates to other readers" |
|
86 | 86 | self._realopener = self.opener |
|
87 | 87 | self.opener = self._delayopener |
|
88 | 88 | self._delaycount = self.count() |
|
89 | 89 | self._delaybuf = [] |
|
90 | 90 | self._delayname = None |
|
91 | 91 | |
|
92 | 92 | def finalize(self, tr): |
|
93 | 93 | "finalize index updates" |
|
94 | 94 | self.opener = self._realopener |
|
95 | 95 | # move redirected index data back into place |
|
96 | 96 | if self._delayname: |
|
97 | 97 | util.rename(self._delayname + ".a", self._delayname) |
|
98 | 98 | elif self._delaybuf: |
|
99 | 99 | fp = self.opener(self.indexfile, 'a') |
|
100 | 100 | fp.write("".join(self._delaybuf)) |
|
101 | 101 | fp.close() |
|
102 | 102 | del self._delaybuf |
|
103 | 103 | # split when we're done |
|
104 | 104 | self.checkinlinesize(tr) |
|
105 | 105 | |
|
106 | 106 | def _delayopener(self, name, mode='r'): |
|
107 | 107 | fp = self._realopener(name, mode) |
|
108 | 108 | # only divert the index |
|
109 | 109 | if not name == self.indexfile: |
|
110 | 110 | return fp |
|
111 | 111 | # if we're doing an initial clone, divert to another file |
|
112 | 112 | if self._delaycount == 0: |
|
113 | 113 | self._delayname = fp.name |
|
114 | 114 | return self._realopener(name + ".a", mode) |
|
115 | 115 | # otherwise, divert to memory |
|
116 | 116 | return appender(fp, self._delaybuf) |
|
117 | 117 | |
|
118 | 118 | def checkinlinesize(self, tr, fp=None): |
|
119 | 119 | if self.opener == self._delayopener: |
|
120 | 120 | return |
|
121 | 121 | return revlog.checkinlinesize(self, tr, fp) |
|
122 | 122 | |
|
123 | 123 | def decode_extra(self, text): |
|
124 | 124 | extra = {} |
|
125 | 125 | for l in text.split('\0'): |
|
126 | 126 | if not l: |
|
127 | 127 | continue |
|
128 | 128 | k, v = _string_unescape(l).split(':', 1) |
|
129 | 129 | extra[k] = v |
|
130 | 130 | return extra |
|
131 | 131 | |
|
132 | 132 | def encode_extra(self, d): |
|
133 | 133 | # keys must be sorted to produce a deterministic changelog entry |
|
134 | 134 | keys = d.keys() |
|
135 | 135 | keys.sort() |
|
136 | 136 | items = [_string_escape('%s:%s' % (k, d[k])) for k in keys] |
|
137 | 137 | return "\0".join(items) |
|
138 | 138 | |
|
139 | 139 | def extract(self, text): |
|
140 | 140 | """ |
|
141 | 141 | format used: |
|
142 | 142 | nodeid\n : manifest node in ascii |
|
143 | 143 | user\n : user, no \n or \r allowed |
|
144 | 144 | time tz extra\n : date (time is int or float, timezone is int) |
|
145 | 145 | : extra is metadatas, encoded and separated by '\0' |
|
146 | 146 | : older versions ignore it |
|
147 | 147 | files\n\n : files modified by the cset, no \n or \r allowed |
|
148 | 148 | (.*) : comment (free text, ideally utf-8) |
|
149 | 149 | |
|
150 | 150 | changelog v0 doesn't use extra |
|
151 | 151 | """ |
|
152 | 152 | if not text: |
|
153 | 153 | return (nullid, "", (0, 0), [], "", {'branch': 'default'}) |
|
154 | 154 | last = text.index("\n\n") |
|
155 | 155 | desc = util.tolocal(text[last + 2:]) |
|
156 | 156 | l = text[:last].split('\n') |
|
157 | 157 | manifest = bin(l[0]) |
|
158 | 158 | user = util.tolocal(l[1]) |
|
159 | 159 | |
|
160 | 160 | extra_data = l[2].split(' ', 2) |
|
161 | 161 | if len(extra_data) != 3: |
|
162 | 162 | time = float(extra_data.pop(0)) |
|
163 | 163 | try: |
|
164 | 164 | # various tools did silly things with the time zone field. |
|
165 | 165 | timezone = int(extra_data[0]) |
|
166 | 166 | except: |
|
167 | 167 | timezone = 0 |
|
168 | 168 | extra = {} |
|
169 | 169 | else: |
|
170 | 170 | time, timezone, extra = extra_data |
|
171 | 171 | time, timezone = float(time), int(timezone) |
|
172 | 172 | extra = self.decode_extra(extra) |
|
173 | 173 | if not extra.get('branch'): |
|
174 | 174 | extra['branch'] = 'default' |
|
175 | 175 | files = l[3:] |
|
176 | 176 | return (manifest, user, (time, timezone), files, desc, extra) |
|
177 | 177 | |
|
178 | 178 | def read(self, node): |
|
179 | 179 | return self.extract(self.revision(node)) |
|
180 | 180 | |
|
181 | 181 | def add(self, manifest, list, desc, transaction, p1=None, p2=None, |
|
182 | 182 | user=None, date=None, extra={}): |
|
183 | 183 | |
|
184 | 184 | user, desc = util.fromlocal(user), util.fromlocal(desc) |
|
185 | 185 | |
|
186 | 186 | if date: |
|
187 | 187 | parseddate = "%d %d" % util.parsedate(date) |
|
188 | 188 | else: |
|
189 | 189 | parseddate = "%d %d" % util.makedate() |
|
190 | 190 | if extra and extra.get("branch") in ("default", ""): |
|
191 | 191 | del extra["branch"] |
|
192 | 192 | if extra: |
|
193 | 193 | extra = self.encode_extra(extra) |
|
194 | 194 | parseddate = "%s %s" % (parseddate, extra) |
|
195 | 195 | list.sort() |
|
196 | 196 | l = [hex(manifest), user, parseddate] + list + ["", desc] |
|
197 | 197 | text = "\n".join(l) |
|
198 | 198 | return self.addrevision(text, transaction, self.count(), p1, p2) |
@@ -1,1291 +1,1291 b'' | |||
|
1 | 1 | """ |
|
2 | 2 | revlog.py - storage back-end for mercurial |
|
3 | 3 | |
|
4 | 4 | This provides efficient delta storage with O(1) retrieve and append |
|
5 | 5 | and O(changes) merge between branches |
|
6 | 6 | |
|
7 | 7 | Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
8 | 8 | |
|
9 | 9 | This software may be used and distributed according to the terms |
|
10 | 10 | of the GNU General Public License, incorporated herein by reference. |
|
11 | 11 | """ |
|
12 | 12 | |
|
13 | 13 | from node import * |
|
14 | 14 | from i18n import _ |
|
15 | 15 | import binascii, changegroup, errno, ancestor, mdiff, os |
|
16 | 16 | import sha, struct, util, zlib |
|
17 | 17 | |
|
18 | 18 | _pack = struct.pack |
|
19 | 19 | _unpack = struct.unpack |
|
20 | 20 | _compress = zlib.compress |
|
21 | 21 | _decompress = zlib.decompress |
|
22 | 22 | _sha = sha.new |
|
23 | 23 | |
|
24 | 24 | # revlog flags |
|
25 | 25 | REVLOGV0 = 0 |
|
26 | 26 | REVLOGNG = 1 |
|
27 | 27 | REVLOGNGINLINEDATA = (1 << 16) |
|
28 | 28 | REVLOG_DEFAULT_FLAGS = REVLOGNGINLINEDATA |
|
29 | 29 | REVLOG_DEFAULT_FORMAT = REVLOGNG |
|
30 | 30 | REVLOG_DEFAULT_VERSION = REVLOG_DEFAULT_FORMAT | REVLOG_DEFAULT_FLAGS |
|
31 | 31 | |
|
32 | 32 | class RevlogError(Exception): |
|
33 | 33 | pass |
|
34 | 34 | class LookupError(RevlogError): |
|
35 | 35 | pass |
|
36 | 36 | |
|
37 | 37 | def getoffset(q): |
|
38 | 38 | return int(q >> 16) |
|
39 | 39 | |
|
40 | 40 | def gettype(q): |
|
41 | 41 | return int(q & 0xFFFF) |
|
42 | 42 | |
|
43 | 43 | def offset_type(offset, type): |
|
44 | 44 | return long(long(offset) << 16 | type) |
|
45 | 45 | |
|
46 | 46 | def hash(text, p1, p2): |
|
47 | 47 | """generate a hash from the given text and its parent hashes |
|
48 | 48 | |
|
49 | 49 | This hash combines both the current file contents and its history |
|
50 | 50 | in a manner that makes it easy to distinguish nodes with the same |
|
51 | 51 | content in the revision graph. |
|
52 | 52 | """ |
|
53 | 53 | l = [p1, p2] |
|
54 | 54 | l.sort() |
|
55 | 55 | s = _sha(l[0]) |
|
56 | 56 | s.update(l[1]) |
|
57 | 57 | s.update(text) |
|
58 | 58 | return s.digest() |
|
59 | 59 | |
|
60 | 60 | def compress(text): |
|
61 | 61 | """ generate a possibly-compressed representation of text """ |
|
62 | 62 | if not text: |
|
63 | 63 | return ("", text) |
|
64 | 64 | if len(text) < 44: |
|
65 | 65 | if text[0] == '\0': |
|
66 | 66 | return ("", text) |
|
67 | 67 | return ('u', text) |
|
68 | 68 | bin = _compress(text) |
|
69 | 69 | if len(bin) > len(text): |
|
70 | 70 | if text[0] == '\0': |
|
71 | 71 | return ("", text) |
|
72 | 72 | return ('u', text) |
|
73 | 73 | return ("", bin) |
|
74 | 74 | |
|
75 | 75 | def decompress(bin): |
|
76 | 76 | """ decompress the given input """ |
|
77 | 77 | if not bin: |
|
78 | 78 | return bin |
|
79 | 79 | t = bin[0] |
|
80 | 80 | if t == '\0': |
|
81 | 81 | return bin |
|
82 | 82 | if t == 'x': |
|
83 | 83 | return _decompress(bin) |
|
84 | 84 | if t == 'u': |
|
85 | 85 | return bin[1:] |
|
86 | 86 | raise RevlogError(_("unknown compression type %r") % t) |
|
87 | 87 | |
|
88 | 88 | class lazyparser(object): |
|
89 | 89 | """ |
|
90 | 90 | this class avoids the need to parse the entirety of large indices |
|
91 | 91 | """ |
|
92 | 92 | |
|
93 | 93 | # lazyparser is not safe to use on windows if win32 extensions not |
|
94 | 94 | # available. it keeps file handle open, which make it not possible |
|
95 | 95 | # to break hardlinks on local cloned repos. |
|
96 | 96 | safe_to_use = os.name != 'nt' or (not util.is_win_9x() and |
|
97 | 97 | hasattr(util, 'win32api')) |
|
98 | 98 | |
|
99 | 99 | def __init__(self, dataf, size): |
|
100 | 100 | self.dataf = dataf |
|
101 | 101 | self.s = struct.calcsize(indexformatng) |
|
102 | 102 | self.datasize = size |
|
103 | 103 | self.l = size/self.s |
|
104 | 104 | self.index = [None] * self.l |
|
105 | 105 | self.map = {nullid: nullrev} |
|
106 | 106 | self.allmap = 0 |
|
107 | 107 | self.all = 0 |
|
108 | 108 | self.mapfind_count = 0 |
|
109 | 109 | |
|
110 | 110 | def loadmap(self): |
|
111 | 111 | """ |
|
112 | 112 | during a commit, we need to make sure the rev being added is |
|
113 | 113 | not a duplicate. This requires loading the entire index, |
|
114 | 114 | which is fairly slow. loadmap can load up just the node map, |
|
115 | 115 | which takes much less time. |
|
116 | 116 | """ |
|
117 | 117 | if self.allmap: |
|
118 | 118 | return |
|
119 | 119 | end = self.datasize |
|
120 | 120 | self.allmap = 1 |
|
121 | 121 | cur = 0 |
|
122 | 122 | count = 0 |
|
123 | 123 | blocksize = self.s * 256 |
|
124 | 124 | self.dataf.seek(0) |
|
125 | 125 | while cur < end: |
|
126 | 126 | data = self.dataf.read(blocksize) |
|
127 | 127 | off = 0 |
|
128 | 128 | for x in xrange(256): |
|
129 | 129 | n = data[off + ngshaoffset:off + ngshaoffset + 20] |
|
130 | 130 | self.map[n] = count |
|
131 | 131 | count += 1 |
|
132 | 132 | if count >= self.l: |
|
133 | 133 | break |
|
134 | 134 | off += self.s |
|
135 | 135 | cur += blocksize |
|
136 | 136 | |
|
137 | 137 | def loadblock(self, blockstart, blocksize, data=None): |
|
138 | 138 | if self.all: |
|
139 | 139 | return |
|
140 | 140 | if data is None: |
|
141 | 141 | self.dataf.seek(blockstart) |
|
142 | 142 | if blockstart + blocksize > self.datasize: |
|
143 | 143 | # the revlog may have grown since we've started running, |
|
144 | 144 | # but we don't have space in self.index for more entries. |
|
145 | 145 | # limit blocksize so that we don't get too much data. |
|
146 | 146 | blocksize = max(self.datasize - blockstart, 0) |
|
147 | 147 | data = self.dataf.read(blocksize) |
|
148 | 148 | lend = len(data) / self.s |
|
149 | 149 | i = blockstart / self.s |
|
150 | 150 | off = 0 |
|
151 | 151 | # lazyindex supports __delitem__ |
|
152 | 152 | if lend > len(self.index) - i: |
|
153 | 153 | lend = len(self.index) - i |
|
154 | 154 | for x in xrange(lend): |
|
155 | 155 | if self.index[i + x] == None: |
|
156 | 156 | b = data[off : off + self.s] |
|
157 | 157 | self.index[i + x] = b |
|
158 | 158 | n = b[ngshaoffset:ngshaoffset + 20] |
|
159 | 159 | self.map[n] = i + x |
|
160 | 160 | off += self.s |
|
161 | 161 | |
|
162 | 162 | def findnode(self, node): |
|
163 | 163 | """search backwards through the index file for a specific node""" |
|
164 | 164 | if self.allmap: |
|
165 | 165 | return None |
|
166 | 166 | |
|
167 | 167 | # hg log will cause many many searches for the manifest |
|
168 | 168 | # nodes. After we get called a few times, just load the whole |
|
169 | 169 | # thing. |
|
170 | 170 | if self.mapfind_count > 8: |
|
171 | 171 | self.loadmap() |
|
172 | 172 | if node in self.map: |
|
173 | 173 | return node |
|
174 | 174 | return None |
|
175 | 175 | self.mapfind_count += 1 |
|
176 | 176 | last = self.l - 1 |
|
177 | 177 | while self.index[last] != None: |
|
178 | 178 | if last == 0: |
|
179 | 179 | self.all = 1 |
|
180 | 180 | self.allmap = 1 |
|
181 | 181 | return None |
|
182 | 182 | last -= 1 |
|
183 | 183 | end = (last + 1) * self.s |
|
184 | 184 | blocksize = self.s * 256 |
|
185 | 185 | while end >= 0: |
|
186 | 186 | start = max(end - blocksize, 0) |
|
187 | 187 | self.dataf.seek(start) |
|
188 | 188 | data = self.dataf.read(end - start) |
|
189 | 189 | findend = end - start |
|
190 | 190 | while True: |
|
191 | 191 | # we're searching backwards, so we have to make sure |
|
192 | 192 | # we don't find a changeset where this node is a parent |
|
193 | 193 | off = data.find(node, 0, findend) |
|
194 | 194 | findend = off |
|
195 | 195 | if off >= 0: |
|
196 | 196 | i = off / self.s |
|
197 | 197 | off = i * self.s |
|
198 | 198 | n = data[off + ngshaoffset:off + ngshaoffset + 20] |
|
199 | 199 | if n == node: |
|
200 | 200 | self.map[n] = i + start / self.s |
|
201 | 201 | return node |
|
202 | 202 | else: |
|
203 | 203 | break |
|
204 | 204 | end -= blocksize |
|
205 | 205 | return None |
|
206 | 206 | |
|
207 | 207 | def loadindex(self, i=None, end=None): |
|
208 | 208 | if self.all: |
|
209 | 209 | return |
|
210 | 210 | all = False |
|
211 | 211 | if i == None: |
|
212 | 212 | blockstart = 0 |
|
213 | 213 | blocksize = (65536 / self.s) * self.s |
|
214 | 214 | end = self.datasize |
|
215 | 215 | all = True |
|
216 | 216 | else: |
|
217 | 217 | if end: |
|
218 | 218 | blockstart = i * self.s |
|
219 | 219 | end = end * self.s |
|
220 | 220 | blocksize = end - blockstart |
|
221 | 221 | else: |
|
222 | 222 | blockstart = (i & ~1023) * self.s |
|
223 | 223 | blocksize = self.s * 1024 |
|
224 | 224 | end = blockstart + blocksize |
|
225 | 225 | while blockstart < end: |
|
226 | 226 | self.loadblock(blockstart, blocksize) |
|
227 | 227 | blockstart += blocksize |
|
228 | 228 | if all: |
|
229 | 229 | self.all = True |
|
230 | 230 | |
|
231 | 231 | class lazyindex(object): |
|
232 | 232 | """a lazy version of the index array""" |
|
233 | 233 | def __init__(self, parser): |
|
234 | 234 | self.p = parser |
|
235 | 235 | def __len__(self): |
|
236 | 236 | return len(self.p.index) |
|
237 | 237 | def load(self, pos): |
|
238 | 238 | if pos < 0: |
|
239 | 239 | pos += len(self.p.index) |
|
240 | 240 | self.p.loadindex(pos) |
|
241 | 241 | return self.p.index[pos] |
|
242 | 242 | def __getitem__(self, pos): |
|
243 | 243 | return _unpack(indexformatng, self.p.index[pos] or self.load(pos)) |
|
244 | 244 | def __setitem__(self, pos, item): |
|
245 | 245 | self.p.index[pos] = _pack(indexformatng, *item) |
|
246 | 246 | def __delitem__(self, pos): |
|
247 | 247 | del self.p.index[pos] |
|
248 | 248 | def insert(self, pos, e): |
|
249 | 249 | self.p.index.insert(pos, _pack(indexformatng, *e)) |
|
250 | 250 | def append(self, e): |
|
251 | 251 | self.p.index.append(_pack(indexformatng, *e)) |
|
252 | 252 | |
|
253 | 253 | class lazymap(object): |
|
254 | 254 | """a lazy version of the node map""" |
|
255 | 255 | def __init__(self, parser): |
|
256 | 256 | self.p = parser |
|
257 | 257 | def load(self, key): |
|
258 | 258 | n = self.p.findnode(key) |
|
259 | 259 | if n == None: |
|
260 | 260 | raise KeyError(key) |
|
261 | 261 | def __contains__(self, key): |
|
262 | 262 | if key in self.p.map: |
|
263 | 263 | return True |
|
264 | 264 | self.p.loadmap() |
|
265 | 265 | return key in self.p.map |
|
266 | 266 | def __iter__(self): |
|
267 | 267 | yield nullid |
|
268 | 268 | for i in xrange(self.p.l): |
|
269 | 269 | ret = self.p.index[i] |
|
270 | 270 | if not ret: |
|
271 | 271 | self.p.loadindex(i) |
|
272 | 272 | ret = self.p.index[i] |
|
273 | 273 | if isinstance(ret, str): |
|
274 | 274 | ret = _unpack(indexformatng, ret) |
|
275 | 275 | yield ret[7] |
|
276 | 276 | def __getitem__(self, key): |
|
277 | 277 | try: |
|
278 | 278 | return self.p.map[key] |
|
279 | 279 | except KeyError: |
|
280 | 280 | try: |
|
281 | 281 | self.load(key) |
|
282 | 282 | return self.p.map[key] |
|
283 | 283 | except KeyError: |
|
284 | 284 | raise KeyError("node " + hex(key)) |
|
285 | 285 | def __setitem__(self, key, val): |
|
286 | 286 | self.p.map[key] = val |
|
287 | 287 | def __delitem__(self, key): |
|
288 | 288 | del self.p.map[key] |
|
289 | 289 | |
|
290 | 290 | indexformatv0 = ">4l20s20s20s" |
|
291 | 291 | v0shaoffset = 56 |
|
292 | 292 | |
|
293 | 293 | class revlogoldio(object): |
|
294 | 294 | def __init__(self): |
|
295 | 295 | self.size = struct.calcsize(indexformatv0) |
|
296 | 296 | |
|
297 | 297 | def parseindex(self, fp, inline): |
|
298 | 298 | s = self.size |
|
299 | 299 | index = [] |
|
300 | 300 | nodemap = {nullid: nullrev} |
|
301 | 301 | n = off = 0 |
|
302 | 302 | data = fp.read() |
|
303 | 303 | l = len(data) |
|
304 | 304 | while off + s <= l: |
|
305 | 305 | cur = data[off:off + s] |
|
306 | 306 | off += s |
|
307 | 307 | e = _unpack(indexformatv0, cur) |
|
308 | 308 | # transform to revlogv1 format |
|
309 | 309 | e2 = (offset_type(e[0], 0), e[1], -1, e[2], e[3], |
|
310 | 310 | nodemap[e[4]], nodemap[e[5]], e[6]) |
|
311 | 311 | index.append(e2) |
|
312 | 312 | nodemap[e[6]] = n |
|
313 | 313 | n += 1 |
|
314 | 314 | |
|
315 | 315 | return index, nodemap, None |
|
316 | 316 | |
|
317 | 317 | def packentry(self, entry, node, version, rev): |
|
318 | 318 | e2 = (getoffset(entry[0]), entry[1], entry[3], entry[4], |
|
319 | 319 | node(entry[5]), node(entry[6]), entry[7]) |
|
320 | 320 | return _pack(indexformatv0, *e2) |
|
321 | 321 | |
|
322 | 322 | # index ng: |
|
323 | 323 | # 6 bytes offset |
|
324 | 324 | # 2 bytes flags |
|
325 | 325 | # 4 bytes compressed length |
|
326 | 326 | # 4 bytes uncompressed length |
|
327 | 327 | # 4 bytes: base rev |
|
328 | 328 | # 4 bytes link rev |
|
329 | 329 | # 4 bytes parent 1 rev |
|
330 | 330 | # 4 bytes parent 2 rev |
|
331 | 331 | # 32 bytes: nodeid |
|
332 | 332 | indexformatng = ">Qiiiiii20s12x" |
|
333 | 333 | ngshaoffset = 32 |
|
334 | 334 | versionformat = ">I" |
|
335 | 335 | |
|
336 | 336 | class revlogio(object): |
|
337 | 337 | def __init__(self): |
|
338 | 338 | self.size = struct.calcsize(indexformatng) |
|
339 | 339 | |
|
340 | 340 | def parseindex(self, fp, inline): |
|
341 | 341 | try: |
|
342 | 342 | size = util.fstat(fp).st_size |
|
343 | 343 | except AttributeError: |
|
344 | 344 | size = 0 |
|
345 | 345 | |
|
346 | 346 | if lazyparser.safe_to_use and not inline and size > 1000000: |
|
347 | 347 | # big index, let's parse it on demand |
|
348 | 348 | parser = lazyparser(fp, size) |
|
349 | 349 | index = lazyindex(parser) |
|
350 | 350 | nodemap = lazymap(parser) |
|
351 | 351 | e = list(index[0]) |
|
352 | 352 | type = gettype(e[0]) |
|
353 | 353 | e[0] = offset_type(0, type) |
|
354 | 354 | index[0] = e |
|
355 | 355 | return index, nodemap, None |
|
356 | 356 | |
|
357 | 357 | s = self.size |
|
358 | 358 | cache = None |
|
359 | 359 | index = [] |
|
360 | 360 | nodemap = {nullid: nullrev} |
|
361 | 361 | n = off = 0 |
|
362 | 362 | # if we're not using lazymap, always read the whole index |
|
363 | 363 | data = fp.read() |
|
364 | 364 | l = len(data) - s |
|
365 | 365 | append = index.append |
|
366 | 366 | if inline: |
|
367 | 367 | cache = (0, data) |
|
368 | 368 | while off <= l: |
|
369 | 369 | e = _unpack(indexformatng, data[off:off + s]) |
|
370 | 370 | nodemap[e[7]] = n |
|
371 | 371 | append(e) |
|
372 | 372 | n += 1 |
|
373 | 373 | if e[1] < 0: |
|
374 | 374 | break |
|
375 | 375 | off += e[1] + s |
|
376 | 376 | else: |
|
377 | 377 | while off <= l: |
|
378 | 378 | e = _unpack(indexformatng, data[off:off + s]) |
|
379 | 379 | nodemap[e[7]] = n |
|
380 | 380 | append(e) |
|
381 | 381 | n += 1 |
|
382 | 382 | off += s |
|
383 | 383 | |
|
384 | 384 | e = list(index[0]) |
|
385 | 385 | type = gettype(e[0]) |
|
386 | 386 | e[0] = offset_type(0, type) |
|
387 | 387 | index[0] = e |
|
388 | 388 | |
|
389 | 389 | return index, nodemap, cache |
|
390 | 390 | |
|
391 | 391 | def packentry(self, entry, node, version, rev): |
|
392 | 392 | p = _pack(indexformatng, *entry) |
|
393 | 393 | if rev == 0: |
|
394 | 394 | p = _pack(versionformat, version) + p[4:] |
|
395 | 395 | return p |
|
396 | 396 | |
|
397 | 397 | class revlog(object): |
|
398 | 398 | """ |
|
399 | 399 | the underlying revision storage object |
|
400 | 400 | |
|
401 | 401 | A revlog consists of two parts, an index and the revision data. |
|
402 | 402 | |
|
403 | 403 | The index is a file with a fixed record size containing |
|
404 | 404 | information on each revision, includings its nodeid (hash), the |
|
405 | 405 | nodeids of its parents, the position and offset of its data within |
|
406 | 406 | the data file, and the revision it's based on. Finally, each entry |
|
407 | 407 | contains a linkrev entry that can serve as a pointer to external |
|
408 | 408 | data. |
|
409 | 409 | |
|
410 | 410 | The revision data itself is a linear collection of data chunks. |
|
411 | 411 | Each chunk represents a revision and is usually represented as a |
|
412 | 412 | delta against the previous chunk. To bound lookup time, runs of |
|
413 | 413 | deltas are limited to about 2 times the length of the original |
|
414 | 414 | version data. This makes retrieval of a version proportional to |
|
415 | 415 | its size, or O(1) relative to the number of revisions. |
|
416 | 416 | |
|
417 | 417 | Both pieces of the revlog are written to in an append-only |
|
418 | 418 | fashion, which means we never need to rewrite a file to insert or |
|
419 | 419 | remove data, and can use some simple techniques to avoid the need |
|
420 | 420 | for locking while reading. |
|
421 | 421 | """ |
|
422 | 422 | def __init__(self, opener, indexfile): |
|
423 | 423 | """ |
|
424 | 424 | create a revlog object |
|
425 | 425 | |
|
426 | 426 | opener is a function that abstracts the file opening operation |
|
427 | 427 | and can be used to implement COW semantics or the like. |
|
428 | 428 | """ |
|
429 | 429 | self.indexfile = indexfile |
|
430 | 430 | self.datafile = indexfile[:-2] + ".d" |
|
431 | 431 | self.opener = opener |
|
432 | 432 | self._cache = None |
|
433 | 433 | self._chunkcache = None |
|
434 | 434 | self.nodemap = {nullid: nullrev} |
|
435 | 435 | self.index = [] |
|
436 | 436 | |
|
437 | 437 | v = REVLOG_DEFAULT_VERSION |
|
438 | 438 | if hasattr(opener, "defversion"): |
|
439 | 439 | v = opener.defversion |
|
440 | 440 | if v & REVLOGNG: |
|
441 | 441 | v |= REVLOGNGINLINEDATA |
|
442 | 442 | |
|
443 | 443 | i = "" |
|
444 | 444 | try: |
|
445 | 445 | f = self.opener(self.indexfile) |
|
446 | 446 | i = f.read(4) |
|
447 | 447 | f.seek(0) |
|
448 | 448 | if len(i) > 0: |
|
449 | 449 | v = struct.unpack(versionformat, i)[0] |
|
450 | 450 | except IOError, inst: |
|
451 | 451 | if inst.errno != errno.ENOENT: |
|
452 | 452 | raise |
|
453 | 453 | |
|
454 | 454 | self.version = v |
|
455 | 455 | self._inline = v & REVLOGNGINLINEDATA |
|
456 | 456 | flags = v & ~0xFFFF |
|
457 | 457 | fmt = v & 0xFFFF |
|
458 | 458 | if fmt == REVLOGV0 and flags: |
|
459 | 459 | raise RevlogError(_("index %s unknown flags %#04x for format v0") |
|
460 | 460 | % (self.indexfile, flags >> 16)) |
|
461 | 461 | elif fmt == REVLOGNG and flags & ~REVLOGNGINLINEDATA: |
|
462 | 462 | raise RevlogError(_("index %s unknown flags %#04x for revlogng") |
|
463 | 463 | % (self.indexfile, flags >> 16)) |
|
464 | 464 | elif fmt > REVLOGNG: |
|
465 | 465 | raise RevlogError(_("index %s unknown format %d") |
|
466 | 466 | % (self.indexfile, fmt)) |
|
467 | 467 | |
|
468 | 468 | self._io = revlogio() |
|
469 | 469 | if self.version == REVLOGV0: |
|
470 | 470 | self._io = revlogoldio() |
|
471 | 471 | if i: |
|
472 | 472 | d = self._io.parseindex(f, self._inline) |
|
473 | 473 | self.index, self.nodemap, self._chunkcache = d |
|
474 | 474 | |
|
475 | 475 | # add the magic null revision at -1 |
|
476 | 476 | self.index.append((0, 0, 0, -1, -1, -1, -1, nullid)) |
|
477 | 477 | |
|
478 | 478 | def _loadindex(self, start, end): |
|
479 | 479 | """load a block of indexes all at once from the lazy parser""" |
|
480 | 480 | if isinstance(self.index, lazyindex): |
|
481 | 481 | self.index.p.loadindex(start, end) |
|
482 | 482 | |
|
483 | 483 | def _loadindexmap(self): |
|
484 | 484 | """loads both the map and the index from the lazy parser""" |
|
485 | 485 | if isinstance(self.index, lazyindex): |
|
486 | 486 | p = self.index.p |
|
487 | 487 | p.loadindex() |
|
488 | 488 | self.nodemap = p.map |
|
489 | 489 | |
|
490 | 490 | def _loadmap(self): |
|
491 | 491 | """loads the map from the lazy parser""" |
|
492 | 492 | if isinstance(self.nodemap, lazymap): |
|
493 | 493 | self.nodemap.p.loadmap() |
|
494 | 494 | self.nodemap = self.nodemap.p.map |
|
495 | 495 | |
|
496 | 496 | def tip(self): |
|
497 | 497 | return self.node(len(self.index) - 2) |
|
498 | 498 | def count(self): |
|
499 | 499 | return len(self.index) - 1 |
|
500 | 500 | |
|
501 | 501 | def rev(self, node): |
|
502 | 502 | try: |
|
503 | 503 | return self.nodemap[node] |
|
504 | 504 | except KeyError: |
|
505 | 505 | raise LookupError(_('%s: no node %s') % (self.indexfile, hex(node))) |
|
506 | 506 | def node(self, rev): |
|
507 | 507 | return self.index[rev][7] |
|
508 | 508 | def linkrev(self, node): |
|
509 | 509 | return self.index[self.rev(node)][4] |
|
510 | 510 | def parents(self, node): |
|
511 | 511 | d = self.index[self.rev(node)][5:7] |
|
512 | 512 | return (self.node(d[0]), self.node(d[1])) |
|
513 | 513 | def parentrevs(self, rev): |
|
514 | 514 | return self.index[rev][5:7] |
|
515 | 515 | def start(self, rev): |
|
516 | 516 | return int(self.index[rev][0] >> 16) |
|
517 | 517 | def end(self, rev): |
|
518 | 518 | return self.start(rev) + self.length(rev) |
|
519 | 519 | def length(self, rev): |
|
520 | 520 | return self.index[rev][1] |
|
521 | 521 | def base(self, rev): |
|
522 | 522 | return self.index[rev][3] |
|
523 | 523 | |
|
524 | 524 | def size(self, rev): |
|
525 | 525 | """return the length of the uncompressed text for a given revision""" |
|
526 | 526 | l = self.index[rev][2] |
|
527 | 527 | if l >= 0: |
|
528 | 528 | return l |
|
529 | 529 | |
|
530 | 530 | t = self.revision(self.node(rev)) |
|
531 | 531 | return len(t) |
|
532 | 532 | |
|
533 | 533 | # alternate implementation, The advantage to this code is it |
|
534 | 534 | # will be faster for a single revision. But, the results are not |
|
535 | 535 | # cached, so finding the size of every revision will be slower. |
|
536 | 536 | """ |
|
537 | 537 | if self.cache and self.cache[1] == rev: |
|
538 | 538 | return len(self.cache[2]) |
|
539 | 539 | |
|
540 | 540 | base = self.base(rev) |
|
541 | 541 | if self.cache and self.cache[1] >= base and self.cache[1] < rev: |
|
542 | 542 | base = self.cache[1] |
|
543 | 543 | text = self.cache[2] |
|
544 | 544 | else: |
|
545 | 545 | text = self.revision(self.node(base)) |
|
546 | 546 | |
|
547 | 547 | l = len(text) |
|
548 | 548 | for x in xrange(base + 1, rev + 1): |
|
549 | 549 | l = mdiff.patchedsize(l, self.chunk(x)) |
|
550 | 550 | return l |
|
551 | 551 | """ |
|
552 | 552 | |
|
553 | 553 | def reachable(self, node, stop=None): |
|
554 | 554 | """return a hash of all nodes ancestral to a given node, including |
|
555 | 555 | the node itself, stopping when stop is matched""" |
|
556 | 556 | reachable = {} |
|
557 | 557 | visit = [node] |
|
558 | 558 | reachable[node] = 1 |
|
559 | 559 | if stop: |
|
560 | 560 | stopn = self.rev(stop) |
|
561 | 561 | else: |
|
562 | 562 | stopn = 0 |
|
563 | 563 | while visit: |
|
564 | 564 | n = visit.pop(0) |
|
565 | 565 | if n == stop: |
|
566 | 566 | continue |
|
567 | 567 | if n == nullid: |
|
568 | 568 | continue |
|
569 | 569 | for p in self.parents(n): |
|
570 | 570 | if self.rev(p) < stopn: |
|
571 | 571 | continue |
|
572 | 572 | if p not in reachable: |
|
573 | 573 | reachable[p] = 1 |
|
574 | 574 | visit.append(p) |
|
575 | 575 | return reachable |
|
576 | 576 | |
|
577 | 577 | def nodesbetween(self, roots=None, heads=None): |
|
578 | 578 | """Return a tuple containing three elements. Elements 1 and 2 contain |
|
579 | 579 | a final list bases and heads after all the unreachable ones have been |
|
580 | 580 | pruned. Element 0 contains a topologically sorted list of all |
|
581 | 581 | |
|
582 | 582 | nodes that satisfy these constraints: |
|
583 | 583 | 1. All nodes must be descended from a node in roots (the nodes on |
|
584 | 584 | roots are considered descended from themselves). |
|
585 | 585 | 2. All nodes must also be ancestors of a node in heads (the nodes in |
|
586 | 586 | heads are considered to be their own ancestors). |
|
587 | 587 | |
|
588 | 588 | If roots is unspecified, nullid is assumed as the only root. |
|
589 | 589 | If heads is unspecified, it is taken to be the output of the |
|
590 | 590 | heads method (i.e. a list of all nodes in the repository that |
|
591 | 591 | have no children).""" |
|
592 | 592 | nonodes = ([], [], []) |
|
593 | 593 | if roots is not None: |
|
594 | 594 | roots = list(roots) |
|
595 | 595 | if not roots: |
|
596 | 596 | return nonodes |
|
597 | 597 | lowestrev = min([self.rev(n) for n in roots]) |
|
598 | 598 | else: |
|
599 | 599 | roots = [nullid] # Everybody's a descendent of nullid |
|
600 | 600 | lowestrev = nullrev |
|
601 | 601 | if (lowestrev == nullrev) and (heads is None): |
|
602 | 602 | # We want _all_ the nodes! |
|
603 | 603 | return ([self.node(r) for r in xrange(0, self.count())], |
|
604 | 604 | [nullid], list(self.heads())) |
|
605 | 605 | if heads is None: |
|
606 | 606 | # All nodes are ancestors, so the latest ancestor is the last |
|
607 | 607 | # node. |
|
608 | 608 | highestrev = self.count() - 1 |
|
609 | 609 | # Set ancestors to None to signal that every node is an ancestor. |
|
610 | 610 | ancestors = None |
|
611 | 611 | # Set heads to an empty dictionary for later discovery of heads |
|
612 | 612 | heads = {} |
|
613 | 613 | else: |
|
614 | 614 | heads = list(heads) |
|
615 | 615 | if not heads: |
|
616 | 616 | return nonodes |
|
617 | 617 | ancestors = {} |
|
618 | 618 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
619 | 619 | # Also, later we will be using it to filter out the heads we can't |
|
620 | 620 | # find from roots. |
|
621 | 621 | heads = dict.fromkeys(heads, 0) |
|
622 | 622 | # Start at the top and keep marking parents until we're done. |
|
623 | 623 | nodestotag = heads.keys() |
|
624 | 624 | # Remember where the top was so we can use it as a limit later. |
|
625 | 625 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
626 | 626 | while nodestotag: |
|
627 | 627 | # grab a node to tag |
|
628 | 628 | n = nodestotag.pop() |
|
629 | 629 | # Never tag nullid |
|
630 | 630 | if n == nullid: |
|
631 | 631 | continue |
|
632 | 632 | # A node's revision number represents its place in a |
|
633 | 633 | # topologically sorted list of nodes. |
|
634 | 634 | r = self.rev(n) |
|
635 | 635 | if r >= lowestrev: |
|
636 | 636 | if n not in ancestors: |
|
637 | 637 | # If we are possibly a descendent of one of the roots |
|
638 | 638 | # and we haven't already been marked as an ancestor |
|
639 | 639 | ancestors[n] = 1 # Mark as ancestor |
|
640 | 640 | # Add non-nullid parents to list of nodes to tag. |
|
641 | 641 | nodestotag.extend([p for p in self.parents(n) if |
|
642 | 642 | p != nullid]) |
|
643 | 643 | elif n in heads: # We've seen it before, is it a fake head? |
|
644 | 644 | # So it is, real heads should not be the ancestors of |
|
645 | 645 | # any other heads. |
|
646 | 646 | heads.pop(n) |
|
647 | 647 | if not ancestors: |
|
648 | 648 | return nonodes |
|
649 | 649 | # Now that we have our set of ancestors, we want to remove any |
|
650 | 650 | # roots that are not ancestors. |
|
651 | 651 | |
|
652 | 652 | # If one of the roots was nullid, everything is included anyway. |
|
653 | 653 | if lowestrev > nullrev: |
|
654 | 654 | # But, since we weren't, let's recompute the lowest rev to not |
|
655 | 655 | # include roots that aren't ancestors. |
|
656 | 656 | |
|
657 | 657 | # Filter out roots that aren't ancestors of heads |
|
658 | 658 | roots = [n for n in roots if n in ancestors] |
|
659 | 659 | # Recompute the lowest revision |
|
660 | 660 | if roots: |
|
661 | 661 | lowestrev = min([self.rev(n) for n in roots]) |
|
662 | 662 | else: |
|
663 | 663 | # No more roots? Return empty list |
|
664 | 664 | return nonodes |
|
665 | 665 | else: |
|
666 | 666 | # We are descending from nullid, and don't need to care about |
|
667 | 667 | # any other roots. |
|
668 | 668 | lowestrev = nullrev |
|
669 | 669 | roots = [nullid] |
|
670 | 670 | # Transform our roots list into a 'set' (i.e. a dictionary where the |
|
671 | 671 | # values don't matter. |
|
672 | 672 | descendents = dict.fromkeys(roots, 1) |
|
673 | 673 | # Also, keep the original roots so we can filter out roots that aren't |
|
674 | 674 | # 'real' roots (i.e. are descended from other roots). |
|
675 | 675 | roots = descendents.copy() |
|
676 | 676 | # Our topologically sorted list of output nodes. |
|
677 | 677 | orderedout = [] |
|
678 | 678 | # Don't start at nullid since we don't want nullid in our output list, |
|
679 | 679 | # and if nullid shows up in descedents, empty parents will look like |
|
680 | 680 | # they're descendents. |
|
681 | 681 | for r in xrange(max(lowestrev, 0), highestrev + 1): |
|
682 | 682 | n = self.node(r) |
|
683 | 683 | isdescendent = False |
|
684 | 684 | if lowestrev == nullrev: # Everybody is a descendent of nullid |
|
685 | 685 | isdescendent = True |
|
686 | 686 | elif n in descendents: |
|
687 | 687 | # n is already a descendent |
|
688 | 688 | isdescendent = True |
|
689 | 689 | # This check only needs to be done here because all the roots |
|
690 | 690 | # will start being marked is descendents before the loop. |
|
691 | 691 | if n in roots: |
|
692 | 692 | # If n was a root, check if it's a 'real' root. |
|
693 | 693 | p = tuple(self.parents(n)) |
|
694 | 694 | # If any of its parents are descendents, it's not a root. |
|
695 | 695 | if (p[0] in descendents) or (p[1] in descendents): |
|
696 | 696 | roots.pop(n) |
|
697 | 697 | else: |
|
698 | 698 | p = tuple(self.parents(n)) |
|
699 | 699 | # A node is a descendent if either of its parents are |
|
700 | 700 | # descendents. (We seeded the dependents list with the roots |
|
701 | 701 | # up there, remember?) |
|
702 | 702 | if (p[0] in descendents) or (p[1] in descendents): |
|
703 | 703 | descendents[n] = 1 |
|
704 | 704 | isdescendent = True |
|
705 | 705 | if isdescendent and ((ancestors is None) or (n in ancestors)): |
|
706 | 706 | # Only include nodes that are both descendents and ancestors. |
|
707 | 707 | orderedout.append(n) |
|
708 | 708 | if (ancestors is not None) and (n in heads): |
|
709 | 709 | # We're trying to figure out which heads are reachable |
|
710 | 710 | # from roots. |
|
711 | 711 | # Mark this head as having been reached |
|
712 | 712 | heads[n] = 1 |
|
713 | 713 | elif ancestors is None: |
|
714 | 714 | # Otherwise, we're trying to discover the heads. |
|
715 | 715 | # Assume this is a head because if it isn't, the next step |
|
716 | 716 | # will eventually remove it. |
|
717 | 717 | heads[n] = 1 |
|
718 | 718 | # But, obviously its parents aren't. |
|
719 | 719 | for p in self.parents(n): |
|
720 | 720 | heads.pop(p, None) |
|
721 | 721 | heads = [n for n in heads.iterkeys() if heads[n] != 0] |
|
722 | 722 | roots = roots.keys() |
|
723 | 723 | assert orderedout |
|
724 | 724 | assert roots |
|
725 | 725 | assert heads |
|
726 | 726 | return (orderedout, roots, heads) |
|
727 | 727 | |
|
728 | 728 | def heads(self, start=None, stop=None): |
|
729 | 729 | """return the list of all nodes that have no children |
|
730 | 730 | |
|
731 | 731 | if start is specified, only heads that are descendants of |
|
732 | 732 | start will be returned |
|
733 | 733 | if stop is specified, it will consider all the revs from stop |
|
734 | 734 | as if they had no children |
|
735 | 735 | """ |
|
736 | 736 | if start is None and stop is None: |
|
737 | 737 | count = self.count() |
|
738 | 738 | if not count: |
|
739 | 739 | return [nullid] |
|
740 | 740 | ishead = [1] * (count + 1) |
|
741 | 741 | index = self.index |
|
742 | 742 | for r in xrange(count): |
|
743 | 743 | e = index[r] |
|
744 | 744 | ishead[e[5]] = ishead[e[6]] = 0 |
|
745 | 745 | return [self.node(r) for r in xrange(count) if ishead[r]] |
|
746 | 746 | |
|
747 | 747 | if start is None: |
|
748 | 748 | start = nullid |
|
749 | 749 | if stop is None: |
|
750 | 750 | stop = [] |
|
751 | 751 | stoprevs = dict.fromkeys([self.rev(n) for n in stop]) |
|
752 | 752 | startrev = self.rev(start) |
|
753 | 753 | reachable = {startrev: 1} |
|
754 | 754 | heads = {startrev: 1} |
|
755 | 755 | |
|
756 | 756 | parentrevs = self.parentrevs |
|
757 | 757 | for r in xrange(startrev + 1, self.count()): |
|
758 | 758 | for p in parentrevs(r): |
|
759 | 759 | if p in reachable: |
|
760 | 760 | if r not in stoprevs: |
|
761 | 761 | reachable[r] = 1 |
|
762 | 762 | heads[r] = 1 |
|
763 | 763 | if p in heads and p not in stoprevs: |
|
764 | 764 | del heads[p] |
|
765 | 765 | |
|
766 | 766 | return [self.node(r) for r in heads] |
|
767 | 767 | |
|
768 | 768 | def children(self, node): |
|
769 | 769 | """find the children of a given node""" |
|
770 | 770 | c = [] |
|
771 | 771 | p = self.rev(node) |
|
772 | 772 | for r in range(p + 1, self.count()): |
|
773 | 773 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
774 | 774 | if prevs: |
|
775 | 775 | for pr in prevs: |
|
776 | 776 | if pr == p: |
|
777 | 777 | c.append(self.node(r)) |
|
778 | 778 | elif p == nullrev: |
|
779 | 779 | c.append(self.node(r)) |
|
780 | 780 | return c |
|
781 | 781 | |
|
782 | 782 | def _match(self, id): |
|
783 | 783 | if isinstance(id, (long, int)): |
|
784 | 784 | # rev |
|
785 | 785 | return self.node(id) |
|
786 | 786 | if len(id) == 20: |
|
787 | 787 | # possibly a binary node |
|
788 | 788 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
789 | 789 | try: |
|
790 | 790 | node = id |
|
791 | 791 | r = self.rev(node) # quick search the index |
|
792 | 792 | return node |
|
793 | 793 | except LookupError: |
|
794 | 794 | pass # may be partial hex id |
|
795 | 795 | try: |
|
796 | 796 | # str(rev) |
|
797 | 797 | rev = int(id) |
|
798 | 798 | if str(rev) != id: |
|
799 | 799 | raise ValueError |
|
800 | 800 | if rev < 0: |
|
801 | 801 | rev = self.count() + rev |
|
802 | 802 | if rev < 0 or rev >= self.count(): |
|
803 | 803 | raise ValueError |
|
804 | 804 | return self.node(rev) |
|
805 | 805 | except (ValueError, OverflowError): |
|
806 | 806 | pass |
|
807 | 807 | if len(id) == 40: |
|
808 | 808 | try: |
|
809 | 809 | # a full hex nodeid? |
|
810 | 810 | node = bin(id) |
|
811 | 811 | r = self.rev(node) |
|
812 | 812 | return node |
|
813 | 813 | except TypeError: |
|
814 | 814 | pass |
|
815 | 815 | |
|
816 | 816 | def _partialmatch(self, id): |
|
817 | 817 | if len(id) < 40: |
|
818 | 818 | try: |
|
819 | 819 | # hex(node)[:...] |
|
820 | 820 | bin_id = bin(id[:len(id) & ~1]) # grab an even number of digits |
|
821 | 821 | node = None |
|
822 | 822 | for n in self.nodemap: |
|
823 | 823 | if n.startswith(bin_id) and hex(n).startswith(id): |
|
824 | 824 | if node is not None: |
|
825 | 825 | raise LookupError(_("Ambiguous identifier")) |
|
826 | 826 | node = n |
|
827 | 827 | if node is not None: |
|
828 | 828 | return node |
|
829 | 829 | except TypeError: |
|
830 | 830 | pass |
|
831 | 831 | |
|
832 | 832 | def lookup(self, id): |
|
833 | 833 | """locate a node based on: |
|
834 | 834 | - revision number or str(revision number) |
|
835 | 835 | - nodeid or subset of hex nodeid |
|
836 | 836 | """ |
|
837 | 837 | n = self._match(id) |
|
838 | 838 | if n is not None: |
|
839 | 839 | return n |
|
840 | 840 | n = self._partialmatch(id) |
|
841 | 841 | if n: |
|
842 | 842 | return n |
|
843 | 843 | |
|
844 | 844 | raise LookupError(_("No match found")) |
|
845 | 845 | |
|
846 | 846 | def cmp(self, node, text): |
|
847 | 847 | """compare text with a given file revision""" |
|
848 | 848 | p1, p2 = self.parents(node) |
|
849 | 849 | return hash(text, p1, p2) != node |
|
850 | 850 | |
|
851 | 851 | def chunk(self, rev, df=None): |
|
852 | 852 | def loadcache(df): |
|
853 | 853 | if not df: |
|
854 | 854 | if self._inline: |
|
855 | 855 | df = self.opener(self.indexfile) |
|
856 | 856 | else: |
|
857 | 857 | df = self.opener(self.datafile) |
|
858 | 858 | df.seek(start) |
|
859 | 859 | self._chunkcache = (start, df.read(cache_length)) |
|
860 | 860 | |
|
861 | 861 | start, length = self.start(rev), self.length(rev) |
|
862 | 862 | if self._inline: |
|
863 | 863 | start += (rev + 1) * self._io.size |
|
864 | 864 | end = start + length |
|
865 | 865 | |
|
866 | 866 | offset = 0 |
|
867 | 867 | if not self._chunkcache: |
|
868 | 868 | cache_length = max(65536, length) |
|
869 | 869 | loadcache(df) |
|
870 | 870 | else: |
|
871 | 871 | cache_start = self._chunkcache[0] |
|
872 | 872 | cache_length = len(self._chunkcache[1]) |
|
873 | 873 | cache_end = cache_start + cache_length |
|
874 | 874 | if start >= cache_start and end <= cache_end: |
|
875 | 875 | # it is cached |
|
876 | 876 | offset = start - cache_start |
|
877 | 877 | else: |
|
878 | 878 | cache_length = max(65536, length) |
|
879 | 879 | loadcache(df) |
|
880 | 880 | |
|
881 | 881 | # avoid copying large chunks |
|
882 | 882 | c = self._chunkcache[1] |
|
883 | 883 | if cache_length != length: |
|
884 | 884 | c = c[offset:offset + length] |
|
885 | 885 | |
|
886 | 886 | return decompress(c) |
|
887 | 887 | |
|
888 | 888 | def delta(self, node): |
|
889 | 889 | """return or calculate a delta between a node and its predecessor""" |
|
890 | 890 | r = self.rev(node) |
|
891 | 891 | return self.revdiff(r - 1, r) |
|
892 | 892 | |
|
893 | 893 | def revdiff(self, rev1, rev2): |
|
894 | 894 | """return or calculate a delta between two revisions""" |
|
895 | 895 | if rev1 + 1 == rev2 and self.base(rev1) == self.base(rev2): |
|
896 | 896 | return self.chunk(rev2) |
|
897 | 897 | |
|
898 | 898 | return mdiff.textdiff(self.revision(self.node(rev1)), |
|
899 | 899 | self.revision(self.node(rev2))) |
|
900 | 900 | |
|
901 | 901 | def revision(self, node): |
|
902 | 902 | """return an uncompressed revision of a given""" |
|
903 | 903 | if node == nullid: |
|
904 | 904 | return "" |
|
905 | 905 | if self._cache and self._cache[0] == node: |
|
906 | return self._cache[2] | |
|
906 | return str(self._cache[2]) | |
|
907 | 907 | |
|
908 | 908 | # look up what we need to read |
|
909 | 909 | text = None |
|
910 | 910 | rev = self.rev(node) |
|
911 | 911 | base = self.base(rev) |
|
912 | 912 | |
|
913 | 913 | # check rev flags |
|
914 | 914 | if self.index[rev][0] & 0xFFFF: |
|
915 | 915 | raise RevlogError(_('incompatible revision flag %x') % |
|
916 | 916 | (self.index[rev][0] & 0xFFFF)) |
|
917 | 917 | |
|
918 | 918 | if self._inline: |
|
919 | 919 | # we probably have the whole chunk cached |
|
920 | 920 | df = None |
|
921 | 921 | else: |
|
922 | 922 | df = self.opener(self.datafile) |
|
923 | 923 | |
|
924 | 924 | # do we have useful data cached? |
|
925 | 925 | if self._cache and self._cache[1] >= base and self._cache[1] < rev: |
|
926 | 926 | base = self._cache[1] |
|
927 | text = self._cache[2] | |
|
927 | text = str(self._cache[2]) | |
|
928 | 928 | self._loadindex(base, rev + 1) |
|
929 | 929 | else: |
|
930 | 930 | self._loadindex(base, rev + 1) |
|
931 | 931 | text = self.chunk(base, df=df) |
|
932 | 932 | |
|
933 | 933 | bins = [self.chunk(r, df) for r in xrange(base + 1, rev + 1)] |
|
934 | 934 | text = mdiff.patches(text, bins) |
|
935 | 935 | p1, p2 = self.parents(node) |
|
936 | 936 | if node != hash(text, p1, p2): |
|
937 | 937 | raise RevlogError(_("integrity check failed on %s:%d") |
|
938 | 938 | % (self.datafile, rev)) |
|
939 | 939 | |
|
940 | 940 | self._cache = (node, rev, text) |
|
941 | 941 | return text |
|
942 | 942 | |
|
943 | 943 | def checkinlinesize(self, tr, fp=None): |
|
944 | 944 | if not self._inline: |
|
945 | 945 | return |
|
946 | 946 | if not fp: |
|
947 | 947 | fp = self.opener(self.indexfile, 'r') |
|
948 | 948 | fp.seek(0, 2) |
|
949 | 949 | size = fp.tell() |
|
950 | 950 | if size < 131072: |
|
951 | 951 | return |
|
952 | 952 | trinfo = tr.find(self.indexfile) |
|
953 | 953 | if trinfo == None: |
|
954 | 954 | raise RevlogError(_("%s not found in the transaction") |
|
955 | 955 | % self.indexfile) |
|
956 | 956 | |
|
957 | 957 | trindex = trinfo[2] |
|
958 | 958 | dataoff = self.start(trindex) |
|
959 | 959 | |
|
960 | 960 | tr.add(self.datafile, dataoff) |
|
961 | 961 | df = self.opener(self.datafile, 'w') |
|
962 | 962 | calc = self._io.size |
|
963 | 963 | for r in xrange(self.count()): |
|
964 | 964 | start = self.start(r) + (r + 1) * calc |
|
965 | 965 | length = self.length(r) |
|
966 | 966 | fp.seek(start) |
|
967 | 967 | d = fp.read(length) |
|
968 | 968 | df.write(d) |
|
969 | 969 | fp.close() |
|
970 | 970 | df.close() |
|
971 | 971 | fp = self.opener(self.indexfile, 'w', atomictemp=True) |
|
972 | 972 | self.version &= ~(REVLOGNGINLINEDATA) |
|
973 | 973 | self._inline = False |
|
974 | 974 | for i in xrange(self.count()): |
|
975 | 975 | e = self._io.packentry(self.index[i], self.node, self.version, i) |
|
976 | 976 | fp.write(e) |
|
977 | 977 | |
|
978 | 978 | # if we don't call rename, the temp file will never replace the |
|
979 | 979 | # real index |
|
980 | 980 | fp.rename() |
|
981 | 981 | |
|
982 | 982 | tr.replace(self.indexfile, trindex * calc) |
|
983 | 983 | self._chunkcache = None |
|
984 | 984 | |
|
985 | 985 | def addrevision(self, text, transaction, link, p1, p2, d=None): |
|
986 | 986 | """add a revision to the log |
|
987 | 987 | |
|
988 | 988 | text - the revision data to add |
|
989 | 989 | transaction - the transaction object used for rollback |
|
990 | 990 | link - the linkrev data to add |
|
991 | 991 | p1, p2 - the parent nodeids of the revision |
|
992 | 992 | d - an optional precomputed delta |
|
993 | 993 | """ |
|
994 | 994 | dfh = None |
|
995 | 995 | if not self._inline: |
|
996 | 996 | dfh = self.opener(self.datafile, "a") |
|
997 | 997 | ifh = self.opener(self.indexfile, "a+") |
|
998 | 998 | return self._addrevision(text, transaction, link, p1, p2, d, ifh, dfh) |
|
999 | 999 | |
|
1000 | 1000 | def _addrevision(self, text, transaction, link, p1, p2, d, ifh, dfh): |
|
1001 | 1001 | node = hash(text, p1, p2) |
|
1002 | 1002 | if node in self.nodemap: |
|
1003 | 1003 | return node |
|
1004 | 1004 | |
|
1005 | 1005 | curr = self.count() |
|
1006 | 1006 | prev = curr - 1 |
|
1007 | 1007 | base = self.base(prev) |
|
1008 | 1008 | offset = self.end(prev) |
|
1009 | 1009 | |
|
1010 | 1010 | if curr: |
|
1011 | 1011 | if not d: |
|
1012 | 1012 | ptext = self.revision(self.node(prev)) |
|
1013 | 1013 | d = mdiff.textdiff(ptext, text) |
|
1014 | 1014 | data = compress(d) |
|
1015 | 1015 | l = len(data[1]) + len(data[0]) |
|
1016 | 1016 | dist = l + offset - self.start(base) |
|
1017 | 1017 | |
|
1018 | 1018 | # full versions are inserted when the needed deltas |
|
1019 | 1019 | # become comparable to the uncompressed text |
|
1020 | 1020 | if not curr or dist > len(text) * 2: |
|
1021 | 1021 | data = compress(text) |
|
1022 | 1022 | l = len(data[1]) + len(data[0]) |
|
1023 | 1023 | base = curr |
|
1024 | 1024 | |
|
1025 | 1025 | e = (offset_type(offset, 0), l, len(text), |
|
1026 | 1026 | base, link, self.rev(p1), self.rev(p2), node) |
|
1027 | 1027 | self.index.insert(-1, e) |
|
1028 | 1028 | self.nodemap[node] = curr |
|
1029 | 1029 | |
|
1030 | 1030 | entry = self._io.packentry(e, self.node, self.version, curr) |
|
1031 | 1031 | if not self._inline: |
|
1032 | 1032 | transaction.add(self.datafile, offset) |
|
1033 | 1033 | transaction.add(self.indexfile, curr * len(entry)) |
|
1034 | 1034 | if data[0]: |
|
1035 | 1035 | dfh.write(data[0]) |
|
1036 | 1036 | dfh.write(data[1]) |
|
1037 | 1037 | dfh.flush() |
|
1038 | 1038 | ifh.write(entry) |
|
1039 | 1039 | else: |
|
1040 | 1040 | offset += curr * self._io.size |
|
1041 | 1041 | transaction.add(self.indexfile, offset, curr) |
|
1042 | 1042 | ifh.write(entry) |
|
1043 | 1043 | ifh.write(data[0]) |
|
1044 | 1044 | ifh.write(data[1]) |
|
1045 | 1045 | self.checkinlinesize(transaction, ifh) |
|
1046 | 1046 | |
|
1047 | 1047 | self._cache = (node, curr, text) |
|
1048 | 1048 | return node |
|
1049 | 1049 | |
|
1050 | 1050 | def ancestor(self, a, b): |
|
1051 | 1051 | """calculate the least common ancestor of nodes a and b""" |
|
1052 | 1052 | |
|
1053 | 1053 | def parents(rev): |
|
1054 | 1054 | return [p for p in self.parentrevs(rev) if p != nullrev] |
|
1055 | 1055 | |
|
1056 | 1056 | c = ancestor.ancestor(self.rev(a), self.rev(b), parents) |
|
1057 | 1057 | if c is None: |
|
1058 | 1058 | return nullid |
|
1059 | 1059 | |
|
1060 | 1060 | return self.node(c) |
|
1061 | 1061 | |
|
1062 | 1062 | def group(self, nodelist, lookup, infocollect=None): |
|
1063 | 1063 | """calculate a delta group |
|
1064 | 1064 | |
|
1065 | 1065 | Given a list of changeset revs, return a set of deltas and |
|
1066 | 1066 | metadata corresponding to nodes. the first delta is |
|
1067 | 1067 | parent(nodes[0]) -> nodes[0] the receiver is guaranteed to |
|
1068 | 1068 | have this parent as it has all history before these |
|
1069 | 1069 | changesets. parent is parent[0] |
|
1070 | 1070 | """ |
|
1071 | 1071 | revs = [self.rev(n) for n in nodelist] |
|
1072 | 1072 | |
|
1073 | 1073 | # if we don't have any revisions touched by these changesets, bail |
|
1074 | 1074 | if not revs: |
|
1075 | 1075 | yield changegroup.closechunk() |
|
1076 | 1076 | return |
|
1077 | 1077 | |
|
1078 | 1078 | # add the parent of the first rev |
|
1079 | 1079 | p = self.parents(self.node(revs[0]))[0] |
|
1080 | 1080 | revs.insert(0, self.rev(p)) |
|
1081 | 1081 | |
|
1082 | 1082 | # build deltas |
|
1083 | 1083 | for d in xrange(0, len(revs) - 1): |
|
1084 | 1084 | a, b = revs[d], revs[d + 1] |
|
1085 | 1085 | nb = self.node(b) |
|
1086 | 1086 | |
|
1087 | 1087 | if infocollect is not None: |
|
1088 | 1088 | infocollect(nb) |
|
1089 | 1089 | |
|
1090 | 1090 | p = self.parents(nb) |
|
1091 | 1091 | meta = nb + p[0] + p[1] + lookup(nb) |
|
1092 | 1092 | if a == -1: |
|
1093 | 1093 | d = self.revision(nb) |
|
1094 | 1094 | meta += mdiff.trivialdiffheader(len(d)) |
|
1095 | 1095 | else: |
|
1096 | 1096 | d = self.revdiff(a, b) |
|
1097 | 1097 | yield changegroup.chunkheader(len(meta) + len(d)) |
|
1098 | 1098 | yield meta |
|
1099 | 1099 | if len(d) > 2**20: |
|
1100 | 1100 | pos = 0 |
|
1101 | 1101 | while pos < len(d): |
|
1102 | 1102 | pos2 = pos + 2 ** 18 |
|
1103 | 1103 | yield d[pos:pos2] |
|
1104 | 1104 | pos = pos2 |
|
1105 | 1105 | else: |
|
1106 | 1106 | yield d |
|
1107 | 1107 | |
|
1108 | 1108 | yield changegroup.closechunk() |
|
1109 | 1109 | |
|
1110 | 1110 | def addgroup(self, revs, linkmapper, transaction, unique=0): |
|
1111 | 1111 | """ |
|
1112 | 1112 | add a delta group |
|
1113 | 1113 | |
|
1114 | 1114 | given a set of deltas, add them to the revision log. the |
|
1115 | 1115 | first delta is against its parent, which should be in our |
|
1116 | 1116 | log, the rest are against the previous delta. |
|
1117 | 1117 | """ |
|
1118 | 1118 | |
|
1119 | 1119 | #track the base of the current delta log |
|
1120 | 1120 | r = self.count() |
|
1121 | 1121 | t = r - 1 |
|
1122 | 1122 | node = None |
|
1123 | 1123 | |
|
1124 | 1124 | base = prev = nullrev |
|
1125 | 1125 | start = end = textlen = 0 |
|
1126 | 1126 | if r: |
|
1127 | 1127 | end = self.end(t) |
|
1128 | 1128 | |
|
1129 | 1129 | ifh = self.opener(self.indexfile, "a+") |
|
1130 | 1130 | isize = r * self._io.size |
|
1131 | 1131 | if self._inline: |
|
1132 | 1132 | transaction.add(self.indexfile, end + isize, r) |
|
1133 | 1133 | dfh = None |
|
1134 | 1134 | else: |
|
1135 | 1135 | transaction.add(self.indexfile, isize, r) |
|
1136 | 1136 | transaction.add(self.datafile, end) |
|
1137 | 1137 | dfh = self.opener(self.datafile, "a") |
|
1138 | 1138 | |
|
1139 | 1139 | # loop through our set of deltas |
|
1140 | 1140 | chain = None |
|
1141 | 1141 | for chunk in revs: |
|
1142 | 1142 | node, p1, p2, cs = struct.unpack("20s20s20s20s", chunk[:80]) |
|
1143 | 1143 | link = linkmapper(cs) |
|
1144 | 1144 | if node in self.nodemap: |
|
1145 | 1145 | # this can happen if two branches make the same change |
|
1146 | 1146 | # if unique: |
|
1147 | 1147 | # raise RevlogError(_("already have %s") % hex(node[:4])) |
|
1148 | 1148 | chain = node |
|
1149 | 1149 | continue |
|
1150 | 1150 | delta = buffer(chunk, 80) |
|
1151 | 1151 | del chunk |
|
1152 | 1152 | |
|
1153 | 1153 | for p in (p1, p2): |
|
1154 | 1154 | if not p in self.nodemap: |
|
1155 | 1155 | raise LookupError(_("unknown parent %s") % short(p)) |
|
1156 | 1156 | |
|
1157 | 1157 | if not chain: |
|
1158 | 1158 | # retrieve the parent revision of the delta chain |
|
1159 | 1159 | chain = p1 |
|
1160 | 1160 | if not chain in self.nodemap: |
|
1161 | 1161 | raise LookupError(_("unknown base %s") % short(chain[:4])) |
|
1162 | 1162 | |
|
1163 | 1163 | # full versions are inserted when the needed deltas become |
|
1164 | 1164 | # comparable to the uncompressed text or when the previous |
|
1165 | 1165 | # version is not the one we have a delta against. We use |
|
1166 | 1166 | # the size of the previous full rev as a proxy for the |
|
1167 | 1167 | # current size. |
|
1168 | 1168 | |
|
1169 | 1169 | if chain == prev: |
|
1170 | 1170 | cdelta = compress(delta) |
|
1171 | 1171 | cdeltalen = len(cdelta[0]) + len(cdelta[1]) |
|
1172 | 1172 | textlen = mdiff.patchedsize(textlen, delta) |
|
1173 | 1173 | |
|
1174 | 1174 | if chain != prev or (end - start + cdeltalen) > textlen * 2: |
|
1175 | 1175 | # flush our writes here so we can read it in revision |
|
1176 | 1176 | if dfh: |
|
1177 | 1177 | dfh.flush() |
|
1178 | 1178 | ifh.flush() |
|
1179 | 1179 | text = self.revision(chain) |
|
1180 | 1180 | if len(text) == 0: |
|
1181 | 1181 | # skip over trivial delta header |
|
1182 | 1182 | text = buffer(delta, 12) |
|
1183 | 1183 | else: |
|
1184 | 1184 | text = mdiff.patches(text, [delta]) |
|
1185 | 1185 | del delta |
|
1186 | 1186 | chk = self._addrevision(text, transaction, link, p1, p2, None, |
|
1187 | 1187 | ifh, dfh) |
|
1188 | 1188 | if not dfh and not self._inline: |
|
1189 | 1189 | # addrevision switched from inline to conventional |
|
1190 | 1190 | # reopen the index |
|
1191 | 1191 | dfh = self.opener(self.datafile, "a") |
|
1192 | 1192 | ifh = self.opener(self.indexfile, "a") |
|
1193 | 1193 | if chk != node: |
|
1194 | 1194 | raise RevlogError(_("consistency error adding group")) |
|
1195 | 1195 | textlen = len(text) |
|
1196 | 1196 | else: |
|
1197 | 1197 | e = (offset_type(end, 0), cdeltalen, textlen, base, |
|
1198 | 1198 | link, self.rev(p1), self.rev(p2), node) |
|
1199 | 1199 | self.index.insert(-1, e) |
|
1200 | 1200 | self.nodemap[node] = r |
|
1201 | 1201 | entry = self._io.packentry(e, self.node, self.version, r) |
|
1202 | 1202 | if self._inline: |
|
1203 | 1203 | ifh.write(entry) |
|
1204 | 1204 | ifh.write(cdelta[0]) |
|
1205 | 1205 | ifh.write(cdelta[1]) |
|
1206 | 1206 | self.checkinlinesize(transaction, ifh) |
|
1207 | 1207 | if not self._inline: |
|
1208 | 1208 | dfh = self.opener(self.datafile, "a") |
|
1209 | 1209 | ifh = self.opener(self.indexfile, "a") |
|
1210 | 1210 | else: |
|
1211 | 1211 | dfh.write(cdelta[0]) |
|
1212 | 1212 | dfh.write(cdelta[1]) |
|
1213 | 1213 | ifh.write(entry) |
|
1214 | 1214 | |
|
1215 | 1215 | t, r, chain, prev = r, r + 1, node, node |
|
1216 | 1216 | base = self.base(t) |
|
1217 | 1217 | start = self.start(base) |
|
1218 | 1218 | end = self.end(t) |
|
1219 | 1219 | |
|
1220 | 1220 | return node |
|
1221 | 1221 | |
|
1222 | 1222 | def strip(self, rev, minlink): |
|
1223 | 1223 | if self.count() == 0 or rev >= self.count(): |
|
1224 | 1224 | return |
|
1225 | 1225 | |
|
1226 | 1226 | if isinstance(self.index, lazyindex): |
|
1227 | 1227 | self._loadindexmap() |
|
1228 | 1228 | |
|
1229 | 1229 | # When stripping away a revision, we need to make sure it |
|
1230 | 1230 | # does not actually belong to an older changeset. |
|
1231 | 1231 | # The minlink parameter defines the oldest revision |
|
1232 | 1232 | # we're allowed to strip away. |
|
1233 | 1233 | while minlink > self.index[rev][4]: |
|
1234 | 1234 | rev += 1 |
|
1235 | 1235 | if rev >= self.count(): |
|
1236 | 1236 | return |
|
1237 | 1237 | |
|
1238 | 1238 | # first truncate the files on disk |
|
1239 | 1239 | end = self.start(rev) |
|
1240 | 1240 | if not self._inline: |
|
1241 | 1241 | df = self.opener(self.datafile, "a") |
|
1242 | 1242 | df.truncate(end) |
|
1243 | 1243 | end = rev * self._io.size |
|
1244 | 1244 | else: |
|
1245 | 1245 | end += rev * self._io.size |
|
1246 | 1246 | |
|
1247 | 1247 | indexf = self.opener(self.indexfile, "a") |
|
1248 | 1248 | indexf.truncate(end) |
|
1249 | 1249 | |
|
1250 | 1250 | # then reset internal state in memory to forget those revisions |
|
1251 | 1251 | self._cache = None |
|
1252 | 1252 | self._chunkcache = None |
|
1253 | 1253 | for x in xrange(rev, self.count()): |
|
1254 | 1254 | del self.nodemap[self.node(x)] |
|
1255 | 1255 | |
|
1256 | 1256 | del self.index[rev:-1] |
|
1257 | 1257 | |
|
1258 | 1258 | def checksize(self): |
|
1259 | 1259 | expected = 0 |
|
1260 | 1260 | if self.count(): |
|
1261 | 1261 | expected = max(0, self.end(self.count() - 1)) |
|
1262 | 1262 | |
|
1263 | 1263 | try: |
|
1264 | 1264 | f = self.opener(self.datafile) |
|
1265 | 1265 | f.seek(0, 2) |
|
1266 | 1266 | actual = f.tell() |
|
1267 | 1267 | dd = actual - expected |
|
1268 | 1268 | except IOError, inst: |
|
1269 | 1269 | if inst.errno != errno.ENOENT: |
|
1270 | 1270 | raise |
|
1271 | 1271 | dd = 0 |
|
1272 | 1272 | |
|
1273 | 1273 | try: |
|
1274 | 1274 | f = self.opener(self.indexfile) |
|
1275 | 1275 | f.seek(0, 2) |
|
1276 | 1276 | actual = f.tell() |
|
1277 | 1277 | s = self._io.size |
|
1278 | 1278 | i = max(0, actual / s) |
|
1279 | 1279 | di = actual - (i * s) |
|
1280 | 1280 | if self._inline: |
|
1281 | 1281 | databytes = 0 |
|
1282 | 1282 | for r in xrange(self.count()): |
|
1283 | 1283 | databytes += max(0, self.length(r)) |
|
1284 | 1284 | dd = 0 |
|
1285 | 1285 | di = actual - self.count() * s - databytes |
|
1286 | 1286 | except IOError, inst: |
|
1287 | 1287 | if inst.errno != errno.ENOENT: |
|
1288 | 1288 | raise |
|
1289 | 1289 | di = 0 |
|
1290 | 1290 | |
|
1291 | 1291 | return (dd, di) |
@@ -1,1690 +1,1690 b'' | |||
|
1 | 1 | """ |
|
2 | 2 | util.py - Mercurial utility functions and platform specfic implementations |
|
3 | 3 | |
|
4 | 4 | Copyright 2005 K. Thananchayan <thananck@yahoo.com> |
|
5 | 5 | Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
6 | 6 | Copyright 2006 Vadim Gelfer <vadim.gelfer@gmail.com> |
|
7 | 7 | |
|
8 | 8 | This software may be used and distributed according to the terms |
|
9 | 9 | of the GNU General Public License, incorporated herein by reference. |
|
10 | 10 | |
|
11 | 11 | This contains helper routines that are independent of the SCM core and hide |
|
12 | 12 | platform-specific details from the core. |
|
13 | 13 | """ |
|
14 | 14 | |
|
15 | 15 | from i18n import _ |
|
16 | 16 | import cStringIO, errno, getpass, popen2, re, shutil, sys, tempfile, strutil |
|
17 | 17 | import os, stat, threading, time, calendar, ConfigParser, locale, glob, osutil |
|
18 | 18 | |
|
19 | 19 | try: |
|
20 | 20 | set = set |
|
21 | 21 | frozenset = frozenset |
|
22 | 22 | except NameError: |
|
23 | 23 | from sets import Set as set, ImmutableSet as frozenset |
|
24 | 24 | |
|
25 | 25 | try: |
|
26 | 26 | _encoding = os.environ.get("HGENCODING") |
|
27 | 27 | if sys.platform == 'darwin' and not _encoding: |
|
28 | 28 | # On darwin, getpreferredencoding ignores the locale environment and |
|
29 | 29 | # always returns mac-roman. We override this if the environment is |
|
30 | 30 | # not C (has been customized by the user). |
|
31 | 31 | locale.setlocale(locale.LC_CTYPE, '') |
|
32 | 32 | _encoding = locale.getlocale()[1] |
|
33 | 33 | if not _encoding: |
|
34 | 34 | _encoding = locale.getpreferredencoding() or 'ascii' |
|
35 | 35 | except locale.Error: |
|
36 | 36 | _encoding = 'ascii' |
|
37 | 37 | _encodingmode = os.environ.get("HGENCODINGMODE", "strict") |
|
38 | 38 | _fallbackencoding = 'ISO-8859-1' |
|
39 | 39 | |
|
40 | 40 | def tolocal(s): |
|
41 | 41 | """ |
|
42 | 42 | Convert a string from internal UTF-8 to local encoding |
|
43 | 43 | |
|
44 | 44 | All internal strings should be UTF-8 but some repos before the |
|
45 | 45 | implementation of locale support may contain latin1 or possibly |
|
46 | 46 | other character sets. We attempt to decode everything strictly |
|
47 | 47 | using UTF-8, then Latin-1, and failing that, we use UTF-8 and |
|
48 | 48 | replace unknown characters. |
|
49 | 49 | """ |
|
50 | 50 | for e in ('UTF-8', _fallbackencoding): |
|
51 | 51 | try: |
|
52 | 52 | u = s.decode(e) # attempt strict decoding |
|
53 | 53 | return u.encode(_encoding, "replace") |
|
54 | 54 | except LookupError, k: |
|
55 | 55 | raise Abort(_("%s, please check your locale settings") % k) |
|
56 | 56 | except UnicodeDecodeError: |
|
57 | 57 | pass |
|
58 | 58 | u = s.decode("utf-8", "replace") # last ditch |
|
59 | 59 | return u.encode(_encoding, "replace") |
|
60 | 60 | |
|
61 | 61 | def fromlocal(s): |
|
62 | 62 | """ |
|
63 | 63 | Convert a string from the local character encoding to UTF-8 |
|
64 | 64 | |
|
65 | 65 | We attempt to decode strings using the encoding mode set by |
|
66 | 66 | HGENCODINGMODE, which defaults to 'strict'. In this mode, unknown |
|
67 | 67 | characters will cause an error message. Other modes include |
|
68 | 68 | 'replace', which replaces unknown characters with a special |
|
69 | 69 | Unicode character, and 'ignore', which drops the character. |
|
70 | 70 | """ |
|
71 | 71 | try: |
|
72 | 72 | return s.decode(_encoding, _encodingmode).encode("utf-8") |
|
73 | 73 | except UnicodeDecodeError, inst: |
|
74 | 74 | sub = s[max(0, inst.start-10):inst.start+10] |
|
75 | 75 | raise Abort("decoding near '%s': %s!" % (sub, inst)) |
|
76 | 76 | except LookupError, k: |
|
77 | 77 | raise Abort(_("%s, please check your locale settings") % k) |
|
78 | 78 | |
|
79 | 79 | def locallen(s): |
|
80 | 80 | """Find the length in characters of a local string""" |
|
81 | 81 | return len(s.decode(_encoding, "replace")) |
|
82 | 82 | |
|
83 | 83 | def localsub(s, a, b=None): |
|
84 | 84 | try: |
|
85 | 85 | u = s.decode(_encoding, _encodingmode) |
|
86 | 86 | if b is not None: |
|
87 | 87 | u = u[a:b] |
|
88 | 88 | else: |
|
89 | 89 | u = u[:a] |
|
90 | 90 | return u.encode(_encoding, _encodingmode) |
|
91 | 91 | except UnicodeDecodeError, inst: |
|
92 | 92 | sub = s[max(0, inst.start-10), inst.start+10] |
|
93 | 93 | raise Abort(_("decoding near '%s': %s!") % (sub, inst)) |
|
94 | 94 | |
|
95 | 95 | # used by parsedate |
|
96 | 96 | defaultdateformats = ( |
|
97 | 97 | '%Y-%m-%d %H:%M:%S', |
|
98 | 98 | '%Y-%m-%d %I:%M:%S%p', |
|
99 | 99 | '%Y-%m-%d %H:%M', |
|
100 | 100 | '%Y-%m-%d %I:%M%p', |
|
101 | 101 | '%Y-%m-%d', |
|
102 | 102 | '%m-%d', |
|
103 | 103 | '%m/%d', |
|
104 | 104 | '%m/%d/%y', |
|
105 | 105 | '%m/%d/%Y', |
|
106 | 106 | '%a %b %d %H:%M:%S %Y', |
|
107 | 107 | '%a %b %d %I:%M:%S%p %Y', |
|
108 | 108 | '%a, %d %b %Y %H:%M:%S', # GNU coreutils "/bin/date --rfc-2822" |
|
109 | 109 | '%b %d %H:%M:%S %Y', |
|
110 | 110 | '%b %d %I:%M:%S%p %Y', |
|
111 | 111 | '%b %d %H:%M:%S', |
|
112 | 112 | '%b %d %I:%M:%S%p', |
|
113 | 113 | '%b %d %H:%M', |
|
114 | 114 | '%b %d %I:%M%p', |
|
115 | 115 | '%b %d %Y', |
|
116 | 116 | '%b %d', |
|
117 | 117 | '%H:%M:%S', |
|
118 | 118 | '%I:%M:%SP', |
|
119 | 119 | '%H:%M', |
|
120 | 120 | '%I:%M%p', |
|
121 | 121 | ) |
|
122 | 122 | |
|
123 | 123 | extendeddateformats = defaultdateformats + ( |
|
124 | 124 | "%Y", |
|
125 | 125 | "%Y-%m", |
|
126 | 126 | "%b", |
|
127 | 127 | "%b %Y", |
|
128 | 128 | ) |
|
129 | 129 | |
|
130 | 130 | class SignalInterrupt(Exception): |
|
131 | 131 | """Exception raised on SIGTERM and SIGHUP.""" |
|
132 | 132 | |
|
133 | 133 | # differences from SafeConfigParser: |
|
134 | 134 | # - case-sensitive keys |
|
135 | 135 | # - allows values that are not strings (this means that you may not |
|
136 | 136 | # be able to save the configuration to a file) |
|
137 | 137 | class configparser(ConfigParser.SafeConfigParser): |
|
138 | 138 | def optionxform(self, optionstr): |
|
139 | 139 | return optionstr |
|
140 | 140 | |
|
141 | 141 | def set(self, section, option, value): |
|
142 | 142 | return ConfigParser.ConfigParser.set(self, section, option, value) |
|
143 | 143 | |
|
144 | 144 | def _interpolate(self, section, option, rawval, vars): |
|
145 | 145 | if not isinstance(rawval, basestring): |
|
146 | 146 | return rawval |
|
147 | 147 | return ConfigParser.SafeConfigParser._interpolate(self, section, |
|
148 | 148 | option, rawval, vars) |
|
149 | 149 | |
|
150 | 150 | def cachefunc(func): |
|
151 | 151 | '''cache the result of function calls''' |
|
152 | 152 | # XXX doesn't handle keywords args |
|
153 | 153 | cache = {} |
|
154 | 154 | if func.func_code.co_argcount == 1: |
|
155 | 155 | # we gain a small amount of time because |
|
156 | 156 | # we don't need to pack/unpack the list |
|
157 | 157 | def f(arg): |
|
158 | 158 | if arg not in cache: |
|
159 | 159 | cache[arg] = func(arg) |
|
160 | 160 | return cache[arg] |
|
161 | 161 | else: |
|
162 | 162 | def f(*args): |
|
163 | 163 | if args not in cache: |
|
164 | 164 | cache[args] = func(*args) |
|
165 | 165 | return cache[args] |
|
166 | 166 | |
|
167 | 167 | return f |
|
168 | 168 | |
|
169 | 169 | def pipefilter(s, cmd): |
|
170 | 170 | '''filter string S through command CMD, returning its output''' |
|
171 | 171 | (pin, pout) = os.popen2(cmd, 'b') |
|
172 | 172 | def writer(): |
|
173 | 173 | try: |
|
174 | 174 | pin.write(s) |
|
175 | 175 | pin.close() |
|
176 | 176 | except IOError, inst: |
|
177 | 177 | if inst.errno != errno.EPIPE: |
|
178 | 178 | raise |
|
179 | 179 | |
|
180 | 180 | # we should use select instead on UNIX, but this will work on most |
|
181 | 181 | # systems, including Windows |
|
182 | 182 | w = threading.Thread(target=writer) |
|
183 | 183 | w.start() |
|
184 | 184 | f = pout.read() |
|
185 | 185 | pout.close() |
|
186 | 186 | w.join() |
|
187 | 187 | return f |
|
188 | 188 | |
|
189 | 189 | def tempfilter(s, cmd): |
|
190 | 190 | '''filter string S through a pair of temporary files with CMD. |
|
191 | 191 | CMD is used as a template to create the real command to be run, |
|
192 | 192 | with the strings INFILE and OUTFILE replaced by the real names of |
|
193 | 193 | the temporary files generated.''' |
|
194 | 194 | inname, outname = None, None |
|
195 | 195 | try: |
|
196 | 196 | infd, inname = tempfile.mkstemp(prefix='hg-filter-in-') |
|
197 | 197 | fp = os.fdopen(infd, 'wb') |
|
198 | 198 | fp.write(s) |
|
199 | 199 | fp.close() |
|
200 | 200 | outfd, outname = tempfile.mkstemp(prefix='hg-filter-out-') |
|
201 | 201 | os.close(outfd) |
|
202 | 202 | cmd = cmd.replace('INFILE', inname) |
|
203 | 203 | cmd = cmd.replace('OUTFILE', outname) |
|
204 | 204 | code = os.system(cmd) |
|
205 | 205 | if sys.platform == 'OpenVMS' and code & 1: |
|
206 | 206 | code = 0 |
|
207 | 207 | if code: raise Abort(_("command '%s' failed: %s") % |
|
208 | 208 | (cmd, explain_exit(code))) |
|
209 | 209 | return open(outname, 'rb').read() |
|
210 | 210 | finally: |
|
211 | 211 | try: |
|
212 | 212 | if inname: os.unlink(inname) |
|
213 | 213 | except: pass |
|
214 | 214 | try: |
|
215 | 215 | if outname: os.unlink(outname) |
|
216 | 216 | except: pass |
|
217 | 217 | |
|
218 | 218 | filtertable = { |
|
219 | 219 | 'tempfile:': tempfilter, |
|
220 | 220 | 'pipe:': pipefilter, |
|
221 | 221 | } |
|
222 | 222 | |
|
223 | 223 | def filter(s, cmd): |
|
224 | 224 | "filter a string through a command that transforms its input to its output" |
|
225 | 225 | for name, fn in filtertable.iteritems(): |
|
226 | 226 | if cmd.startswith(name): |
|
227 | 227 | return fn(s, cmd[len(name):].lstrip()) |
|
228 | 228 | return pipefilter(s, cmd) |
|
229 | 229 | |
|
230 | 230 | def binary(s): |
|
231 | 231 | """return true if a string is binary data using diff's heuristic""" |
|
232 | 232 | if s and '\0' in s[:4096]: |
|
233 | 233 | return True |
|
234 | 234 | return False |
|
235 | 235 | |
|
236 | 236 | def unique(g): |
|
237 | 237 | """return the uniq elements of iterable g""" |
|
238 | 238 | seen = {} |
|
239 | 239 | l = [] |
|
240 | 240 | for f in g: |
|
241 | 241 | if f not in seen: |
|
242 | 242 | seen[f] = 1 |
|
243 | 243 | l.append(f) |
|
244 | 244 | return l |
|
245 | 245 | |
|
246 | 246 | class Abort(Exception): |
|
247 | 247 | """Raised if a command needs to print an error and exit.""" |
|
248 | 248 | |
|
249 | 249 | class UnexpectedOutput(Abort): |
|
250 | 250 | """Raised to print an error with part of output and exit.""" |
|
251 | 251 | |
|
252 | 252 | def always(fn): return True |
|
253 | 253 | def never(fn): return False |
|
254 | 254 | |
|
255 | 255 | def expand_glob(pats): |
|
256 | 256 | '''On Windows, expand the implicit globs in a list of patterns''' |
|
257 | 257 | if os.name != 'nt': |
|
258 | 258 | return list(pats) |
|
259 | 259 | ret = [] |
|
260 | 260 | for p in pats: |
|
261 | 261 | kind, name = patkind(p, None) |
|
262 | 262 | if kind is None: |
|
263 | 263 | globbed = glob.glob(name) |
|
264 | 264 | if globbed: |
|
265 | 265 | ret.extend(globbed) |
|
266 | 266 | continue |
|
267 | 267 | # if we couldn't expand the glob, just keep it around |
|
268 | 268 | ret.append(p) |
|
269 | 269 | return ret |
|
270 | 270 | |
|
271 | 271 | def patkind(name, dflt_pat='glob'): |
|
272 | 272 | """Split a string into an optional pattern kind prefix and the |
|
273 | 273 | actual pattern.""" |
|
274 | 274 | for prefix in 're', 'glob', 'path', 'relglob', 'relpath', 'relre': |
|
275 | 275 | if name.startswith(prefix + ':'): return name.split(':', 1) |
|
276 | 276 | return dflt_pat, name |
|
277 | 277 | |
|
278 | 278 | def globre(pat, head='^', tail='$'): |
|
279 | 279 | "convert a glob pattern into a regexp" |
|
280 | 280 | i, n = 0, len(pat) |
|
281 | 281 | res = '' |
|
282 | 282 | group = False |
|
283 | 283 | def peek(): return i < n and pat[i] |
|
284 | 284 | while i < n: |
|
285 | 285 | c = pat[i] |
|
286 | 286 | i = i+1 |
|
287 | 287 | if c == '*': |
|
288 | 288 | if peek() == '*': |
|
289 | 289 | i += 1 |
|
290 | 290 | res += '.*' |
|
291 | 291 | else: |
|
292 | 292 | res += '[^/]*' |
|
293 | 293 | elif c == '?': |
|
294 | 294 | res += '.' |
|
295 | 295 | elif c == '[': |
|
296 | 296 | j = i |
|
297 | 297 | if j < n and pat[j] in '!]': |
|
298 | 298 | j += 1 |
|
299 | 299 | while j < n and pat[j] != ']': |
|
300 | 300 | j += 1 |
|
301 | 301 | if j >= n: |
|
302 | 302 | res += '\\[' |
|
303 | 303 | else: |
|
304 | 304 | stuff = pat[i:j].replace('\\','\\\\') |
|
305 | 305 | i = j + 1 |
|
306 | 306 | if stuff[0] == '!': |
|
307 | 307 | stuff = '^' + stuff[1:] |
|
308 | 308 | elif stuff[0] == '^': |
|
309 | 309 | stuff = '\\' + stuff |
|
310 | 310 | res = '%s[%s]' % (res, stuff) |
|
311 | 311 | elif c == '{': |
|
312 | 312 | group = True |
|
313 | 313 | res += '(?:' |
|
314 | 314 | elif c == '}' and group: |
|
315 | 315 | res += ')' |
|
316 | 316 | group = False |
|
317 | 317 | elif c == ',' and group: |
|
318 | 318 | res += '|' |
|
319 | 319 | elif c == '\\': |
|
320 | 320 | p = peek() |
|
321 | 321 | if p: |
|
322 | 322 | i += 1 |
|
323 | 323 | res += re.escape(p) |
|
324 | 324 | else: |
|
325 | 325 | res += re.escape(c) |
|
326 | 326 | else: |
|
327 | 327 | res += re.escape(c) |
|
328 | 328 | return head + res + tail |
|
329 | 329 | |
|
330 | 330 | _globchars = {'[': 1, '{': 1, '*': 1, '?': 1} |
|
331 | 331 | |
|
332 | 332 | def pathto(root, n1, n2): |
|
333 | 333 | '''return the relative path from one place to another. |
|
334 | 334 | root should use os.sep to separate directories |
|
335 | 335 | n1 should use os.sep to separate directories |
|
336 | 336 | n2 should use "/" to separate directories |
|
337 | 337 | returns an os.sep-separated path. |
|
338 | 338 | |
|
339 | 339 | If n1 is a relative path, it's assumed it's |
|
340 | 340 | relative to root. |
|
341 | 341 | n2 should always be relative to root. |
|
342 | 342 | ''' |
|
343 | 343 | if not n1: return localpath(n2) |
|
344 | 344 | if os.path.isabs(n1): |
|
345 | 345 | if os.path.splitdrive(root)[0] != os.path.splitdrive(n1)[0]: |
|
346 | 346 | return os.path.join(root, localpath(n2)) |
|
347 | 347 | n2 = '/'.join((pconvert(root), n2)) |
|
348 | 348 | a, b = n1.split(os.sep), n2.split('/') |
|
349 | 349 | a.reverse() |
|
350 | 350 | b.reverse() |
|
351 | 351 | while a and b and a[-1] == b[-1]: |
|
352 | 352 | a.pop() |
|
353 | 353 | b.pop() |
|
354 | 354 | b.reverse() |
|
355 | 355 | return os.sep.join((['..'] * len(a)) + b) |
|
356 | 356 | |
|
357 | 357 | def canonpath(root, cwd, myname): |
|
358 | 358 | """return the canonical path of myname, given cwd and root""" |
|
359 | 359 | if root == os.sep: |
|
360 | 360 | rootsep = os.sep |
|
361 | 361 | elif root.endswith(os.sep): |
|
362 | 362 | rootsep = root |
|
363 | 363 | else: |
|
364 | 364 | rootsep = root + os.sep |
|
365 | 365 | name = myname |
|
366 | 366 | if not os.path.isabs(name): |
|
367 | 367 | name = os.path.join(root, cwd, name) |
|
368 | 368 | name = os.path.normpath(name) |
|
369 | 369 | audit_path = path_auditor(root) |
|
370 | 370 | if name != rootsep and name.startswith(rootsep): |
|
371 | 371 | name = name[len(rootsep):] |
|
372 | 372 | audit_path(name) |
|
373 | 373 | return pconvert(name) |
|
374 | 374 | elif name == root: |
|
375 | 375 | return '' |
|
376 | 376 | else: |
|
377 | 377 | # Determine whether `name' is in the hierarchy at or beneath `root', |
|
378 | 378 | # by iterating name=dirname(name) until that causes no change (can't |
|
379 | 379 | # check name == '/', because that doesn't work on windows). For each |
|
380 | 380 | # `name', compare dev/inode numbers. If they match, the list `rel' |
|
381 | 381 | # holds the reversed list of components making up the relative file |
|
382 | 382 | # name we want. |
|
383 | 383 | root_st = os.stat(root) |
|
384 | 384 | rel = [] |
|
385 | 385 | while True: |
|
386 | 386 | try: |
|
387 | 387 | name_st = os.stat(name) |
|
388 | 388 | except OSError: |
|
389 | 389 | break |
|
390 | 390 | if samestat(name_st, root_st): |
|
391 | 391 | if not rel: |
|
392 | 392 | # name was actually the same as root (maybe a symlink) |
|
393 | 393 | return '' |
|
394 | 394 | rel.reverse() |
|
395 | 395 | name = os.path.join(*rel) |
|
396 | 396 | audit_path(name) |
|
397 | 397 | return pconvert(name) |
|
398 | 398 | dirname, basename = os.path.split(name) |
|
399 | 399 | rel.append(basename) |
|
400 | 400 | if dirname == name: |
|
401 | 401 | break |
|
402 | 402 | name = dirname |
|
403 | 403 | |
|
404 | 404 | raise Abort('%s not under root' % myname) |
|
405 | 405 | |
|
406 | 406 | def matcher(canonroot, cwd='', names=[], inc=[], exc=[], src=None): |
|
407 | 407 | return _matcher(canonroot, cwd, names, inc, exc, 'glob', src) |
|
408 | 408 | |
|
409 | 409 | def cmdmatcher(canonroot, cwd='', names=[], inc=[], exc=[], src=None, |
|
410 | 410 | globbed=False, default=None): |
|
411 | 411 | default = default or 'relpath' |
|
412 | 412 | if default == 'relpath' and not globbed: |
|
413 | 413 | names = expand_glob(names) |
|
414 | 414 | return _matcher(canonroot, cwd, names, inc, exc, default, src) |
|
415 | 415 | |
|
416 | 416 | def _matcher(canonroot, cwd, names, inc, exc, dflt_pat, src): |
|
417 | 417 | """build a function to match a set of file patterns |
|
418 | 418 | |
|
419 | 419 | arguments: |
|
420 | 420 | canonroot - the canonical root of the tree you're matching against |
|
421 | 421 | cwd - the current working directory, if relevant |
|
422 | 422 | names - patterns to find |
|
423 | 423 | inc - patterns to include |
|
424 | 424 | exc - patterns to exclude |
|
425 | 425 | dflt_pat - if a pattern in names has no explicit type, assume this one |
|
426 | 426 | src - where these patterns came from (e.g. .hgignore) |
|
427 | 427 | |
|
428 | 428 | a pattern is one of: |
|
429 | 429 | 'glob:<glob>' - a glob relative to cwd |
|
430 | 430 | 're:<regexp>' - a regular expression |
|
431 | 431 | 'path:<path>' - a path relative to canonroot |
|
432 | 432 | 'relglob:<glob>' - an unrooted glob (*.c matches C files in all dirs) |
|
433 | 433 | 'relpath:<path>' - a path relative to cwd |
|
434 | 434 | 'relre:<regexp>' - a regexp that doesn't have to match the start of a name |
|
435 | 435 | '<something>' - one of the cases above, selected by the dflt_pat argument |
|
436 | 436 | |
|
437 | 437 | returns: |
|
438 | 438 | a 3-tuple containing |
|
439 | 439 | - list of roots (places where one should start a recursive walk of the fs); |
|
440 | 440 | this often matches the explicit non-pattern names passed in, but also |
|
441 | 441 | includes the initial part of glob: patterns that has no glob characters |
|
442 | 442 | - a bool match(filename) function |
|
443 | 443 | - a bool indicating if any patterns were passed in |
|
444 | 444 | """ |
|
445 | 445 | |
|
446 | 446 | # a common case: no patterns at all |
|
447 | 447 | if not names and not inc and not exc: |
|
448 | 448 | return [], always, False |
|
449 | 449 | |
|
450 | 450 | def contains_glob(name): |
|
451 | 451 | for c in name: |
|
452 | 452 | if c in _globchars: return True |
|
453 | 453 | return False |
|
454 | 454 | |
|
455 | 455 | def regex(kind, name, tail): |
|
456 | 456 | '''convert a pattern into a regular expression''' |
|
457 | 457 | if not name: |
|
458 | 458 | return '' |
|
459 | 459 | if kind == 're': |
|
460 | 460 | return name |
|
461 | 461 | elif kind == 'path': |
|
462 | 462 | return '^' + re.escape(name) + '(?:/|$)' |
|
463 | 463 | elif kind == 'relglob': |
|
464 | 464 | return globre(name, '(?:|.*/)', tail) |
|
465 | 465 | elif kind == 'relpath': |
|
466 | 466 | return re.escape(name) + '(?:/|$)' |
|
467 | 467 | elif kind == 'relre': |
|
468 | 468 | if name.startswith('^'): |
|
469 | 469 | return name |
|
470 | 470 | return '.*' + name |
|
471 | 471 | return globre(name, '', tail) |
|
472 | 472 | |
|
473 | 473 | def matchfn(pats, tail): |
|
474 | 474 | """build a matching function from a set of patterns""" |
|
475 | 475 | if not pats: |
|
476 | 476 | return |
|
477 | 477 | try: |
|
478 | 478 | pat = '(?:%s)' % '|'.join([regex(k, p, tail) for (k, p) in pats]) |
|
479 | 479 | return re.compile(pat).match |
|
480 | 480 | except OverflowError: |
|
481 | 481 | # We're using a Python with a tiny regex engine and we |
|
482 | 482 | # made it explode, so we'll divide the pattern list in two |
|
483 | 483 | # until it works |
|
484 | 484 | l = len(pats) |
|
485 | 485 | if l < 2: |
|
486 | 486 | raise |
|
487 | 487 | a, b = matchfn(pats[:l/2], tail), matchfn(pats[l/2:], tail) |
|
488 | 488 | return lambda s: a(s) or b(s) |
|
489 | 489 | except re.error: |
|
490 | 490 | for k, p in pats: |
|
491 | 491 | try: |
|
492 | 492 | re.compile('(?:%s)' % regex(k, p, tail)) |
|
493 | 493 | except re.error: |
|
494 | 494 | if src: |
|
495 | 495 | raise Abort("%s: invalid pattern (%s): %s" % |
|
496 | 496 | (src, k, p)) |
|
497 | 497 | else: |
|
498 | 498 | raise Abort("invalid pattern (%s): %s" % (k, p)) |
|
499 | 499 | raise Abort("invalid pattern") |
|
500 | 500 | |
|
501 | 501 | def globprefix(pat): |
|
502 | 502 | '''return the non-glob prefix of a path, e.g. foo/* -> foo''' |
|
503 | 503 | root = [] |
|
504 | 504 | for p in pat.split('/'): |
|
505 | 505 | if contains_glob(p): break |
|
506 | 506 | root.append(p) |
|
507 | 507 | return '/'.join(root) or '.' |
|
508 | 508 | |
|
509 | 509 | def normalizepats(names, default): |
|
510 | 510 | pats = [] |
|
511 | 511 | roots = [] |
|
512 | 512 | anypats = False |
|
513 | 513 | for kind, name in [patkind(p, default) for p in names]: |
|
514 | 514 | if kind in ('glob', 'relpath'): |
|
515 | 515 | name = canonpath(canonroot, cwd, name) |
|
516 | 516 | elif kind in ('relglob', 'path'): |
|
517 | 517 | name = normpath(name) |
|
518 | 518 | |
|
519 | 519 | pats.append((kind, name)) |
|
520 | 520 | |
|
521 | 521 | if kind in ('glob', 're', 'relglob', 'relre'): |
|
522 | 522 | anypats = True |
|
523 | 523 | |
|
524 | 524 | if kind == 'glob': |
|
525 | 525 | root = globprefix(name) |
|
526 | 526 | roots.append(root) |
|
527 | 527 | elif kind in ('relpath', 'path'): |
|
528 | 528 | roots.append(name or '.') |
|
529 | 529 | elif kind == 'relglob': |
|
530 | 530 | roots.append('.') |
|
531 | 531 | return roots, pats, anypats |
|
532 | 532 | |
|
533 | 533 | roots, pats, anypats = normalizepats(names, dflt_pat) |
|
534 | 534 | |
|
535 | 535 | patmatch = matchfn(pats, '$') or always |
|
536 | 536 | incmatch = always |
|
537 | 537 | if inc: |
|
538 | 538 | dummy, inckinds, dummy = normalizepats(inc, 'glob') |
|
539 | 539 | incmatch = matchfn(inckinds, '(?:/|$)') |
|
540 | 540 | excmatch = lambda fn: False |
|
541 | 541 | if exc: |
|
542 | 542 | dummy, exckinds, dummy = normalizepats(exc, 'glob') |
|
543 | 543 | excmatch = matchfn(exckinds, '(?:/|$)') |
|
544 | 544 | |
|
545 | 545 | if not names and inc and not exc: |
|
546 | 546 | # common case: hgignore patterns |
|
547 | 547 | match = incmatch |
|
548 | 548 | else: |
|
549 | 549 | match = lambda fn: incmatch(fn) and not excmatch(fn) and patmatch(fn) |
|
550 | 550 | |
|
551 | 551 | return (roots, match, (inc or exc or anypats) and True) |
|
552 | 552 | |
|
553 | 553 | _hgexecutable = None |
|
554 | 554 | |
|
555 | 555 | def hgexecutable(): |
|
556 | 556 | """return location of the 'hg' executable. |
|
557 | 557 | |
|
558 | 558 | Defaults to $HG or 'hg' in the search path. |
|
559 | 559 | """ |
|
560 | 560 | if _hgexecutable is None: |
|
561 | 561 | set_hgexecutable(os.environ.get('HG') or find_exe('hg', 'hg')) |
|
562 | 562 | return _hgexecutable |
|
563 | 563 | |
|
564 | 564 | def set_hgexecutable(path): |
|
565 | 565 | """set location of the 'hg' executable""" |
|
566 | 566 | global _hgexecutable |
|
567 | 567 | _hgexecutable = path |
|
568 | 568 | |
|
569 | 569 | def system(cmd, environ={}, cwd=None, onerr=None, errprefix=None): |
|
570 | 570 | '''enhanced shell command execution. |
|
571 | 571 | run with environment maybe modified, maybe in different dir. |
|
572 | 572 | |
|
573 | 573 | if command fails and onerr is None, return status. if ui object, |
|
574 | 574 | print error message and return status, else raise onerr object as |
|
575 | 575 | exception.''' |
|
576 | 576 | def py2shell(val): |
|
577 | 577 | 'convert python object into string that is useful to shell' |
|
578 | 578 | if val in (None, False): |
|
579 | 579 | return '0' |
|
580 | 580 | if val == True: |
|
581 | 581 | return '1' |
|
582 | 582 | return str(val) |
|
583 | 583 | oldenv = {} |
|
584 | 584 | for k in environ: |
|
585 | 585 | oldenv[k] = os.environ.get(k) |
|
586 | 586 | if cwd is not None: |
|
587 | 587 | oldcwd = os.getcwd() |
|
588 | 588 | origcmd = cmd |
|
589 | 589 | if os.name == 'nt': |
|
590 | 590 | cmd = '"%s"' % cmd |
|
591 | 591 | try: |
|
592 | 592 | for k, v in environ.iteritems(): |
|
593 | 593 | os.environ[k] = py2shell(v) |
|
594 | 594 | os.environ['HG'] = hgexecutable() |
|
595 | 595 | if cwd is not None and oldcwd != cwd: |
|
596 | 596 | os.chdir(cwd) |
|
597 | 597 | rc = os.system(cmd) |
|
598 | 598 | if sys.platform == 'OpenVMS' and rc & 1: |
|
599 | 599 | rc = 0 |
|
600 | 600 | if rc and onerr: |
|
601 | 601 | errmsg = '%s %s' % (os.path.basename(origcmd.split(None, 1)[0]), |
|
602 | 602 | explain_exit(rc)[0]) |
|
603 | 603 | if errprefix: |
|
604 | 604 | errmsg = '%s: %s' % (errprefix, errmsg) |
|
605 | 605 | try: |
|
606 | 606 | onerr.warn(errmsg + '\n') |
|
607 | 607 | except AttributeError: |
|
608 | 608 | raise onerr(errmsg) |
|
609 | 609 | return rc |
|
610 | 610 | finally: |
|
611 | 611 | for k, v in oldenv.iteritems(): |
|
612 | 612 | if v is None: |
|
613 | 613 | del os.environ[k] |
|
614 | 614 | else: |
|
615 | 615 | os.environ[k] = v |
|
616 | 616 | if cwd is not None and oldcwd != cwd: |
|
617 | 617 | os.chdir(oldcwd) |
|
618 | 618 | |
|
619 | 619 | # os.path.lexists is not available on python2.3 |
|
620 | 620 | def lexists(filename): |
|
621 | 621 | "test whether a file with this name exists. does not follow symlinks" |
|
622 | 622 | try: |
|
623 | 623 | os.lstat(filename) |
|
624 | 624 | except: |
|
625 | 625 | return False |
|
626 | 626 | return True |
|
627 | 627 | |
|
628 | 628 | def rename(src, dst): |
|
629 | 629 | """forcibly rename a file""" |
|
630 | 630 | try: |
|
631 | 631 | os.rename(src, dst) |
|
632 | 632 | except OSError, err: # FIXME: check err (EEXIST ?) |
|
633 | 633 | # on windows, rename to existing file is not allowed, so we |
|
634 | 634 | # must delete destination first. but if file is open, unlink |
|
635 | 635 | # schedules it for delete but does not delete it. rename |
|
636 | 636 | # happens immediately even for open files, so we create |
|
637 | 637 | # temporary file, delete it, rename destination to that name, |
|
638 | 638 | # then delete that. then rename is safe to do. |
|
639 | 639 | fd, temp = tempfile.mkstemp(dir=os.path.dirname(dst) or '.') |
|
640 | 640 | os.close(fd) |
|
641 | 641 | os.unlink(temp) |
|
642 | 642 | os.rename(dst, temp) |
|
643 | 643 | os.unlink(temp) |
|
644 | 644 | os.rename(src, dst) |
|
645 | 645 | |
|
646 | 646 | def unlink(f): |
|
647 | 647 | """unlink and remove the directory if it is empty""" |
|
648 | 648 | os.unlink(f) |
|
649 | 649 | # try removing directories that might now be empty |
|
650 | 650 | try: |
|
651 | 651 | os.removedirs(os.path.dirname(f)) |
|
652 | 652 | except OSError: |
|
653 | 653 | pass |
|
654 | 654 | |
|
655 | 655 | def copyfile(src, dest): |
|
656 | 656 | "copy a file, preserving mode" |
|
657 | 657 | if os.path.islink(src): |
|
658 | 658 | try: |
|
659 | 659 | os.unlink(dest) |
|
660 | 660 | except: |
|
661 | 661 | pass |
|
662 | 662 | os.symlink(os.readlink(src), dest) |
|
663 | 663 | else: |
|
664 | 664 | try: |
|
665 | 665 | shutil.copyfile(src, dest) |
|
666 | 666 | shutil.copymode(src, dest) |
|
667 | 667 | except shutil.Error, inst: |
|
668 | 668 | raise Abort(str(inst)) |
|
669 | 669 | |
|
670 | 670 | def copyfiles(src, dst, hardlink=None): |
|
671 | 671 | """Copy a directory tree using hardlinks if possible""" |
|
672 | 672 | |
|
673 | 673 | if hardlink is None: |
|
674 | 674 | hardlink = (os.stat(src).st_dev == |
|
675 | 675 | os.stat(os.path.dirname(dst)).st_dev) |
|
676 | 676 | |
|
677 | 677 | if os.path.isdir(src): |
|
678 | 678 | os.mkdir(dst) |
|
679 | 679 | for name, kind in osutil.listdir(src): |
|
680 | 680 | srcname = os.path.join(src, name) |
|
681 | 681 | dstname = os.path.join(dst, name) |
|
682 | 682 | copyfiles(srcname, dstname, hardlink) |
|
683 | 683 | else: |
|
684 | 684 | if hardlink: |
|
685 | 685 | try: |
|
686 | 686 | os_link(src, dst) |
|
687 | 687 | except (IOError, OSError): |
|
688 | 688 | hardlink = False |
|
689 | 689 | shutil.copy(src, dst) |
|
690 | 690 | else: |
|
691 | 691 | shutil.copy(src, dst) |
|
692 | 692 | |
|
693 | 693 | class path_auditor(object): |
|
694 | 694 | '''ensure that a filesystem path contains no banned components. |
|
695 | 695 | the following properties of a path are checked: |
|
696 | 696 | |
|
697 | 697 | - under top-level .hg |
|
698 | 698 | - starts at the root of a windows drive |
|
699 | 699 | - contains ".." |
|
700 | 700 | - traverses a symlink (e.g. a/symlink_here/b) |
|
701 | 701 | - inside a nested repository''' |
|
702 | 702 | |
|
703 | 703 | def __init__(self, root): |
|
704 | 704 | self.audited = set() |
|
705 | 705 | self.auditeddir = set() |
|
706 | 706 | self.root = root |
|
707 | 707 | |
|
708 | 708 | def __call__(self, path): |
|
709 | 709 | if path in self.audited: |
|
710 | 710 | return |
|
711 | 711 | normpath = os.path.normcase(path) |
|
712 | 712 | parts = normpath.split(os.sep) |
|
713 | 713 | if (os.path.splitdrive(path)[0] or parts[0] in ('.hg', '') |
|
714 | 714 | or os.pardir in parts): |
|
715 | 715 | raise Abort(_("path contains illegal component: %s") % path) |
|
716 | 716 | def check(prefix): |
|
717 | 717 | curpath = os.path.join(self.root, prefix) |
|
718 | 718 | try: |
|
719 | 719 | st = os.lstat(curpath) |
|
720 | 720 | except OSError, err: |
|
721 | 721 | # EINVAL can be raised as invalid path syntax under win32. |
|
722 | 722 | # They must be ignored for patterns can be checked too. |
|
723 | 723 | if err.errno not in (errno.ENOENT, errno.EINVAL): |
|
724 | 724 | raise |
|
725 | 725 | else: |
|
726 | 726 | if stat.S_ISLNK(st.st_mode): |
|
727 | 727 | raise Abort(_('path %r traverses symbolic link %r') % |
|
728 | 728 | (path, prefix)) |
|
729 | 729 | elif (stat.S_ISDIR(st.st_mode) and |
|
730 | 730 | os.path.isdir(os.path.join(curpath, '.hg'))): |
|
731 | 731 | raise Abort(_('path %r is inside repo %r') % |
|
732 | 732 | (path, prefix)) |
|
733 | 733 | |
|
734 | 734 | prefixes = [] |
|
735 | 735 | for c in strutil.rfindall(normpath, os.sep): |
|
736 | 736 | prefix = normpath[:c] |
|
737 | 737 | if prefix in self.auditeddir: |
|
738 | 738 | break |
|
739 | 739 | check(prefix) |
|
740 | 740 | prefixes.append(prefix) |
|
741 | 741 | |
|
742 | 742 | self.audited.add(path) |
|
743 | 743 | # only add prefixes to the cache after checking everything: we don't |
|
744 | 744 | # want to add "foo/bar/baz" before checking if there's a "foo/.hg" |
|
745 | 745 | self.auditeddir.update(prefixes) |
|
746 | 746 | |
|
747 | 747 | def _makelock_file(info, pathname): |
|
748 | 748 | ld = os.open(pathname, os.O_CREAT | os.O_WRONLY | os.O_EXCL) |
|
749 | 749 | os.write(ld, info) |
|
750 | 750 | os.close(ld) |
|
751 | 751 | |
|
752 | 752 | def _readlock_file(pathname): |
|
753 | 753 | return posixfile(pathname).read() |
|
754 | 754 | |
|
755 | 755 | def nlinks(pathname): |
|
756 | 756 | """Return number of hardlinks for the given file.""" |
|
757 | 757 | return os.lstat(pathname).st_nlink |
|
758 | 758 | |
|
759 | 759 | if hasattr(os, 'link'): |
|
760 | 760 | os_link = os.link |
|
761 | 761 | else: |
|
762 | 762 | def os_link(src, dst): |
|
763 | 763 | raise OSError(0, _("Hardlinks not supported")) |
|
764 | 764 | |
|
765 | 765 | def fstat(fp): |
|
766 | 766 | '''stat file object that may not have fileno method.''' |
|
767 | 767 | try: |
|
768 | 768 | return os.fstat(fp.fileno()) |
|
769 | 769 | except AttributeError: |
|
770 | 770 | return os.stat(fp.name) |
|
771 | 771 | |
|
772 | 772 | posixfile = file |
|
773 | 773 | |
|
774 | 774 | def is_win_9x(): |
|
775 | 775 | '''return true if run on windows 95, 98 or me.''' |
|
776 | 776 | try: |
|
777 | 777 | return sys.getwindowsversion()[3] == 1 |
|
778 | 778 | except AttributeError: |
|
779 | 779 | return os.name == 'nt' and 'command' in os.environ.get('comspec', '') |
|
780 | 780 | |
|
781 | 781 | getuser_fallback = None |
|
782 | 782 | |
|
783 | 783 | def getuser(): |
|
784 | 784 | '''return name of current user''' |
|
785 | 785 | try: |
|
786 | 786 | return getpass.getuser() |
|
787 | 787 | except ImportError: |
|
788 | 788 | # import of pwd will fail on windows - try fallback |
|
789 | 789 | if getuser_fallback: |
|
790 | 790 | return getuser_fallback() |
|
791 | 791 | # raised if win32api not available |
|
792 | 792 | raise Abort(_('user name not available - set USERNAME ' |
|
793 | 793 | 'environment variable')) |
|
794 | 794 | |
|
795 | 795 | def username(uid=None): |
|
796 | 796 | """Return the name of the user with the given uid. |
|
797 | 797 | |
|
798 | 798 | If uid is None, return the name of the current user.""" |
|
799 | 799 | try: |
|
800 | 800 | import pwd |
|
801 | 801 | if uid is None: |
|
802 | 802 | uid = os.getuid() |
|
803 | 803 | try: |
|
804 | 804 | return pwd.getpwuid(uid)[0] |
|
805 | 805 | except KeyError: |
|
806 | 806 | return str(uid) |
|
807 | 807 | except ImportError: |
|
808 | 808 | return None |
|
809 | 809 | |
|
810 | 810 | def groupname(gid=None): |
|
811 | 811 | """Return the name of the group with the given gid. |
|
812 | 812 | |
|
813 | 813 | If gid is None, return the name of the current group.""" |
|
814 | 814 | try: |
|
815 | 815 | import grp |
|
816 | 816 | if gid is None: |
|
817 | 817 | gid = os.getgid() |
|
818 | 818 | try: |
|
819 | 819 | return grp.getgrgid(gid)[0] |
|
820 | 820 | except KeyError: |
|
821 | 821 | return str(gid) |
|
822 | 822 | except ImportError: |
|
823 | 823 | return None |
|
824 | 824 | |
|
825 | 825 | # File system features |
|
826 | 826 | |
|
827 | 827 | def checkfolding(path): |
|
828 | 828 | """ |
|
829 | 829 | Check whether the given path is on a case-sensitive filesystem |
|
830 | 830 | |
|
831 | 831 | Requires a path (like /foo/.hg) ending with a foldable final |
|
832 | 832 | directory component. |
|
833 | 833 | """ |
|
834 | 834 | s1 = os.stat(path) |
|
835 | 835 | d, b = os.path.split(path) |
|
836 | 836 | p2 = os.path.join(d, b.upper()) |
|
837 | 837 | if path == p2: |
|
838 | 838 | p2 = os.path.join(d, b.lower()) |
|
839 | 839 | try: |
|
840 | 840 | s2 = os.stat(p2) |
|
841 | 841 | if s2 == s1: |
|
842 | 842 | return False |
|
843 | 843 | return True |
|
844 | 844 | except: |
|
845 | 845 | return True |
|
846 | 846 | |
|
847 | 847 | def checkexec(path): |
|
848 | 848 | """ |
|
849 | 849 | Check whether the given path is on a filesystem with UNIX-like exec flags |
|
850 | 850 | |
|
851 | 851 | Requires a directory (like /foo/.hg) |
|
852 | 852 | """ |
|
853 | 853 | try: |
|
854 | 854 | EXECFLAGS = stat.S_IXUSR | stat.S_IXGRP | stat.S_IXOTH |
|
855 | 855 | fh, fn = tempfile.mkstemp("", "", path) |
|
856 | 856 | os.close(fh) |
|
857 | 857 | m = os.stat(fn).st_mode |
|
858 | 858 | # VFAT on Linux can flip mode but it doesn't persist a FS remount. |
|
859 | 859 | # frequently we can detect it if files are created with exec bit on. |
|
860 | 860 | new_file_has_exec = m & EXECFLAGS |
|
861 | 861 | os.chmod(fn, m ^ EXECFLAGS) |
|
862 | 862 | exec_flags_cannot_flip = (os.stat(fn).st_mode == m) |
|
863 | 863 | os.unlink(fn) |
|
864 | 864 | except (IOError,OSError): |
|
865 | 865 | # we don't care, the user probably won't be able to commit anyway |
|
866 | 866 | return False |
|
867 | 867 | return not (new_file_has_exec or exec_flags_cannot_flip) |
|
868 | 868 | |
|
869 | 869 | def execfunc(path, fallback): |
|
870 | 870 | '''return an is_exec() function with default to fallback''' |
|
871 | 871 | if checkexec(path): |
|
872 | 872 | return lambda x: is_exec(os.path.join(path, x)) |
|
873 | 873 | return fallback |
|
874 | 874 | |
|
875 | 875 | def checklink(path): |
|
876 | 876 | """check whether the given path is on a symlink-capable filesystem""" |
|
877 | 877 | # mktemp is not racy because symlink creation will fail if the |
|
878 | 878 | # file already exists |
|
879 | 879 | name = tempfile.mktemp(dir=path) |
|
880 | 880 | try: |
|
881 | 881 | os.symlink(".", name) |
|
882 | 882 | os.unlink(name) |
|
883 | 883 | return True |
|
884 | 884 | except (OSError, AttributeError): |
|
885 | 885 | return False |
|
886 | 886 | |
|
887 | 887 | def linkfunc(path, fallback): |
|
888 | 888 | '''return an is_link() function with default to fallback''' |
|
889 | 889 | if checklink(path): |
|
890 | 890 | return lambda x: os.path.islink(os.path.join(path, x)) |
|
891 | 891 | return fallback |
|
892 | 892 | |
|
893 | 893 | _umask = os.umask(0) |
|
894 | 894 | os.umask(_umask) |
|
895 | 895 | |
|
896 | 896 | def needbinarypatch(): |
|
897 | 897 | """return True if patches should be applied in binary mode by default.""" |
|
898 | 898 | return os.name == 'nt' |
|
899 | 899 | |
|
900 | 900 | # Platform specific variants |
|
901 | 901 | if os.name == 'nt': |
|
902 | 902 | import msvcrt |
|
903 | 903 | nulldev = 'NUL:' |
|
904 | 904 | |
|
905 | 905 | class winstdout: |
|
906 | 906 | '''stdout on windows misbehaves if sent through a pipe''' |
|
907 | 907 | |
|
908 | 908 | def __init__(self, fp): |
|
909 | 909 | self.fp = fp |
|
910 | 910 | |
|
911 | 911 | def __getattr__(self, key): |
|
912 | 912 | return getattr(self.fp, key) |
|
913 | 913 | |
|
914 | 914 | def close(self): |
|
915 | 915 | try: |
|
916 | 916 | self.fp.close() |
|
917 | 917 | except: pass |
|
918 | 918 | |
|
919 | 919 | def write(self, s): |
|
920 | 920 | try: |
|
921 | 921 | return self.fp.write(s) |
|
922 | 922 | except IOError, inst: |
|
923 | 923 | if inst.errno != 0: raise |
|
924 | 924 | self.close() |
|
925 | 925 | raise IOError(errno.EPIPE, 'Broken pipe') |
|
926 | 926 | |
|
927 | 927 | def flush(self): |
|
928 | 928 | try: |
|
929 | 929 | return self.fp.flush() |
|
930 | 930 | except IOError, inst: |
|
931 | 931 | if inst.errno != errno.EINVAL: raise |
|
932 | 932 | self.close() |
|
933 | 933 | raise IOError(errno.EPIPE, 'Broken pipe') |
|
934 | 934 | |
|
935 | 935 | sys.stdout = winstdout(sys.stdout) |
|
936 | 936 | |
|
937 | 937 | def system_rcpath(): |
|
938 | 938 | try: |
|
939 | 939 | return system_rcpath_win32() |
|
940 | 940 | except: |
|
941 | 941 | return [r'c:\mercurial\mercurial.ini'] |
|
942 | 942 | |
|
943 | 943 | def user_rcpath(): |
|
944 | 944 | '''return os-specific hgrc search path to the user dir''' |
|
945 | 945 | try: |
|
946 | 946 | userrc = user_rcpath_win32() |
|
947 | 947 | except: |
|
948 | 948 | userrc = os.path.join(os.path.expanduser('~'), 'mercurial.ini') |
|
949 | 949 | path = [userrc] |
|
950 | 950 | userprofile = os.environ.get('USERPROFILE') |
|
951 | 951 | if userprofile: |
|
952 | 952 | path.append(os.path.join(userprofile, 'mercurial.ini')) |
|
953 | 953 | return path |
|
954 | 954 | |
|
955 | 955 | def parse_patch_output(output_line): |
|
956 | 956 | """parses the output produced by patch and returns the file name""" |
|
957 | 957 | pf = output_line[14:] |
|
958 | 958 | if pf[0] == '`': |
|
959 | 959 | pf = pf[1:-1] # Remove the quotes |
|
960 | 960 | return pf |
|
961 | 961 | |
|
962 | 962 | def testpid(pid): |
|
963 | 963 | '''return False if pid dead, True if running or not known''' |
|
964 | 964 | return True |
|
965 | 965 | |
|
966 | 966 | def set_exec(f, mode): |
|
967 | 967 | pass |
|
968 | 968 | |
|
969 | 969 | def set_link(f, mode): |
|
970 | 970 | pass |
|
971 | 971 | |
|
972 | 972 | def set_binary(fd): |
|
973 | 973 | msvcrt.setmode(fd.fileno(), os.O_BINARY) |
|
974 | 974 | |
|
975 | 975 | def pconvert(path): |
|
976 | 976 | return path.replace("\\", "/") |
|
977 | 977 | |
|
978 | 978 | def localpath(path): |
|
979 | 979 | return path.replace('/', '\\') |
|
980 | 980 | |
|
981 | 981 | def normpath(path): |
|
982 | 982 | return pconvert(os.path.normpath(path)) |
|
983 | 983 | |
|
984 | 984 | makelock = _makelock_file |
|
985 | 985 | readlock = _readlock_file |
|
986 | 986 | |
|
987 | 987 | def samestat(s1, s2): |
|
988 | 988 | return False |
|
989 | 989 | |
|
990 | 990 | # A sequence of backslashes is special iff it precedes a double quote: |
|
991 | 991 | # - if there's an even number of backslashes, the double quote is not |
|
992 | 992 | # quoted (i.e. it ends the quoted region) |
|
993 | 993 | # - if there's an odd number of backslashes, the double quote is quoted |
|
994 | 994 | # - in both cases, every pair of backslashes is unquoted into a single |
|
995 | 995 | # backslash |
|
996 | 996 | # (See http://msdn2.microsoft.com/en-us/library/a1y7w461.aspx ) |
|
997 | 997 | # So, to quote a string, we must surround it in double quotes, double |
|
998 | 998 | # the number of backslashes that preceed double quotes and add another |
|
999 | 999 | # backslash before every double quote (being careful with the double |
|
1000 | 1000 | # quote we've appended to the end) |
|
1001 | 1001 | _quotere = None |
|
1002 | 1002 | def shellquote(s): |
|
1003 | 1003 | global _quotere |
|
1004 | 1004 | if _quotere is None: |
|
1005 | 1005 | _quotere = re.compile(r'(\\*)("|\\$)') |
|
1006 | 1006 | return '"%s"' % _quotere.sub(r'\1\1\\\2', s) |
|
1007 | 1007 | |
|
1008 | 1008 | def quotecommand(cmd): |
|
1009 | 1009 | """Build a command string suitable for os.popen* calls.""" |
|
1010 | 1010 | # The extra quotes are needed because popen* runs the command |
|
1011 | 1011 | # through the current COMSPEC. cmd.exe suppress enclosing quotes. |
|
1012 | 1012 | return '"' + cmd + '"' |
|
1013 | 1013 | |
|
1014 | 1014 | def explain_exit(code): |
|
1015 | 1015 | return _("exited with status %d") % code, code |
|
1016 | 1016 | |
|
1017 | 1017 | # if you change this stub into a real check, please try to implement the |
|
1018 | 1018 | # username and groupname functions above, too. |
|
1019 | 1019 | def isowner(fp, st=None): |
|
1020 | 1020 | return True |
|
1021 | 1021 | |
|
1022 | 1022 | def find_in_path(name, path, default=None): |
|
1023 | 1023 | '''find name in search path. path can be string (will be split |
|
1024 | 1024 | with os.pathsep), or iterable thing that returns strings. if name |
|
1025 | 1025 | found, return path to name. else return default. name is looked up |
|
1026 | 1026 | using cmd.exe rules, using PATHEXT.''' |
|
1027 | 1027 | if isinstance(path, str): |
|
1028 | 1028 | path = path.split(os.pathsep) |
|
1029 | 1029 | |
|
1030 | 1030 | pathext = os.environ.get('PATHEXT', '.COM;.EXE;.BAT;.CMD') |
|
1031 | 1031 | pathext = pathext.lower().split(os.pathsep) |
|
1032 | 1032 | isexec = os.path.splitext(name)[1].lower() in pathext |
|
1033 | 1033 | |
|
1034 | 1034 | for p in path: |
|
1035 | 1035 | p_name = os.path.join(p, name) |
|
1036 | 1036 | |
|
1037 | 1037 | if isexec and os.path.exists(p_name): |
|
1038 | 1038 | return p_name |
|
1039 | 1039 | |
|
1040 | 1040 | for ext in pathext: |
|
1041 | 1041 | p_name_ext = p_name + ext |
|
1042 | 1042 | if os.path.exists(p_name_ext): |
|
1043 | 1043 | return p_name_ext |
|
1044 | 1044 | return default |
|
1045 | 1045 | |
|
1046 | 1046 | def set_signal_handler(): |
|
1047 | 1047 | try: |
|
1048 | 1048 | set_signal_handler_win32() |
|
1049 | 1049 | except NameError: |
|
1050 | 1050 | pass |
|
1051 | 1051 | |
|
1052 | 1052 | try: |
|
1053 | 1053 | # override functions with win32 versions if possible |
|
1054 | 1054 | from util_win32 import * |
|
1055 | 1055 | if not is_win_9x(): |
|
1056 | 1056 | posixfile = posixfile_nt |
|
1057 | 1057 | except ImportError: |
|
1058 | 1058 | pass |
|
1059 | 1059 | |
|
1060 | 1060 | else: |
|
1061 | 1061 | nulldev = '/dev/null' |
|
1062 | 1062 | |
|
1063 | 1063 | def rcfiles(path): |
|
1064 | 1064 | rcs = [os.path.join(path, 'hgrc')] |
|
1065 | 1065 | rcdir = os.path.join(path, 'hgrc.d') |
|
1066 | 1066 | try: |
|
1067 | 1067 | rcs.extend([os.path.join(rcdir, f) |
|
1068 | 1068 | for f, kind in osutil.listdir(rcdir) |
|
1069 | 1069 | if f.endswith(".rc")]) |
|
1070 | 1070 | except OSError: |
|
1071 | 1071 | pass |
|
1072 | 1072 | return rcs |
|
1073 | 1073 | |
|
1074 | 1074 | def system_rcpath(): |
|
1075 | 1075 | path = [] |
|
1076 | 1076 | # old mod_python does not set sys.argv |
|
1077 | 1077 | if len(getattr(sys, 'argv', [])) > 0: |
|
1078 | 1078 | path.extend(rcfiles(os.path.dirname(sys.argv[0]) + |
|
1079 | 1079 | '/../etc/mercurial')) |
|
1080 | 1080 | path.extend(rcfiles('/etc/mercurial')) |
|
1081 | 1081 | return path |
|
1082 | 1082 | |
|
1083 | 1083 | def user_rcpath(): |
|
1084 | 1084 | return [os.path.expanduser('~/.hgrc')] |
|
1085 | 1085 | |
|
1086 | 1086 | def parse_patch_output(output_line): |
|
1087 | 1087 | """parses the output produced by patch and returns the file name""" |
|
1088 | 1088 | pf = output_line[14:] |
|
1089 | 1089 | if os.sys.platform == 'OpenVMS': |
|
1090 | 1090 | if pf[0] == '`': |
|
1091 | 1091 | pf = pf[1:-1] # Remove the quotes |
|
1092 | 1092 | else: |
|
1093 | 1093 | if pf.startswith("'") and pf.endswith("'") and " " in pf: |
|
1094 | 1094 | pf = pf[1:-1] # Remove the quotes |
|
1095 | 1095 | return pf |
|
1096 | 1096 | |
|
1097 | 1097 | def is_exec(f): |
|
1098 | 1098 | """check whether a file is executable""" |
|
1099 | 1099 | return (os.lstat(f).st_mode & 0100 != 0) |
|
1100 | 1100 | |
|
1101 | 1101 | def set_exec(f, mode): |
|
1102 | 1102 | s = os.lstat(f).st_mode |
|
1103 | 1103 | if stat.S_ISLNK(s) or (s & 0100 != 0) == mode: |
|
1104 | 1104 | return |
|
1105 | 1105 | if mode: |
|
1106 | 1106 | # Turn on +x for every +r bit when making a file executable |
|
1107 | 1107 | # and obey umask. |
|
1108 | 1108 | os.chmod(f, s | (s & 0444) >> 2 & ~_umask) |
|
1109 | 1109 | else: |
|
1110 | 1110 | os.chmod(f, s & 0666) |
|
1111 | 1111 | |
|
1112 | 1112 | def set_link(f, mode): |
|
1113 | 1113 | """make a file a symbolic link/regular file |
|
1114 | 1114 | |
|
1115 | 1115 | if a file is changed to a link, its contents become the link data |
|
1116 | 1116 | if a link is changed to a file, its link data become its contents |
|
1117 | 1117 | """ |
|
1118 | 1118 | |
|
1119 | 1119 | m = os.path.islink(f) |
|
1120 | 1120 | if m == bool(mode): |
|
1121 | 1121 | return |
|
1122 | 1122 | |
|
1123 | 1123 | if mode: # switch file to link |
|
1124 | 1124 | data = file(f).read() |
|
1125 | 1125 | os.unlink(f) |
|
1126 | 1126 | os.symlink(data, f) |
|
1127 | 1127 | else: |
|
1128 | 1128 | data = os.readlink(f) |
|
1129 | 1129 | os.unlink(f) |
|
1130 | 1130 | file(f, "w").write(data) |
|
1131 | 1131 | |
|
1132 | 1132 | def set_binary(fd): |
|
1133 | 1133 | pass |
|
1134 | 1134 | |
|
1135 | 1135 | def pconvert(path): |
|
1136 | 1136 | return path |
|
1137 | 1137 | |
|
1138 | 1138 | def localpath(path): |
|
1139 | 1139 | return path |
|
1140 | 1140 | |
|
1141 | 1141 | normpath = os.path.normpath |
|
1142 | 1142 | samestat = os.path.samestat |
|
1143 | 1143 | |
|
1144 | 1144 | def makelock(info, pathname): |
|
1145 | 1145 | try: |
|
1146 | 1146 | os.symlink(info, pathname) |
|
1147 | 1147 | except OSError, why: |
|
1148 | 1148 | if why.errno == errno.EEXIST: |
|
1149 | 1149 | raise |
|
1150 | 1150 | else: |
|
1151 | 1151 | _makelock_file(info, pathname) |
|
1152 | 1152 | |
|
1153 | 1153 | def readlock(pathname): |
|
1154 | 1154 | try: |
|
1155 | 1155 | return os.readlink(pathname) |
|
1156 | 1156 | except OSError, why: |
|
1157 | 1157 | if why.errno in (errno.EINVAL, errno.ENOSYS): |
|
1158 | 1158 | return _readlock_file(pathname) |
|
1159 | 1159 | else: |
|
1160 | 1160 | raise |
|
1161 | 1161 | |
|
1162 | 1162 | def shellquote(s): |
|
1163 | 1163 | if os.sys.platform == 'OpenVMS': |
|
1164 | 1164 | return '"%s"' % s |
|
1165 | 1165 | else: |
|
1166 | 1166 | return "'%s'" % s.replace("'", "'\\''") |
|
1167 | 1167 | |
|
1168 | 1168 | def quotecommand(cmd): |
|
1169 | 1169 | return cmd |
|
1170 | 1170 | |
|
1171 | 1171 | def testpid(pid): |
|
1172 | 1172 | '''return False if pid dead, True if running or not sure''' |
|
1173 | 1173 | if os.sys.platform == 'OpenVMS': |
|
1174 | 1174 | return True |
|
1175 | 1175 | try: |
|
1176 | 1176 | os.kill(pid, 0) |
|
1177 | 1177 | return True |
|
1178 | 1178 | except OSError, inst: |
|
1179 | 1179 | return inst.errno != errno.ESRCH |
|
1180 | 1180 | |
|
1181 | 1181 | def explain_exit(code): |
|
1182 | 1182 | """return a 2-tuple (desc, code) describing a process's status""" |
|
1183 | 1183 | if os.WIFEXITED(code): |
|
1184 | 1184 | val = os.WEXITSTATUS(code) |
|
1185 | 1185 | return _("exited with status %d") % val, val |
|
1186 | 1186 | elif os.WIFSIGNALED(code): |
|
1187 | 1187 | val = os.WTERMSIG(code) |
|
1188 | 1188 | return _("killed by signal %d") % val, val |
|
1189 | 1189 | elif os.WIFSTOPPED(code): |
|
1190 | 1190 | val = os.WSTOPSIG(code) |
|
1191 | 1191 | return _("stopped by signal %d") % val, val |
|
1192 | 1192 | raise ValueError(_("invalid exit code")) |
|
1193 | 1193 | |
|
1194 | 1194 | def isowner(fp, st=None): |
|
1195 | 1195 | """Return True if the file object f belongs to the current user. |
|
1196 | 1196 | |
|
1197 | 1197 | The return value of a util.fstat(f) may be passed as the st argument. |
|
1198 | 1198 | """ |
|
1199 | 1199 | if st is None: |
|
1200 | 1200 | st = fstat(fp) |
|
1201 | 1201 | return st.st_uid == os.getuid() |
|
1202 | 1202 | |
|
1203 | 1203 | def find_in_path(name, path, default=None): |
|
1204 | 1204 | '''find name in search path. path can be string (will be split |
|
1205 | 1205 | with os.pathsep), or iterable thing that returns strings. if name |
|
1206 | 1206 | found, return path to name. else return default.''' |
|
1207 | 1207 | if isinstance(path, str): |
|
1208 | 1208 | path = path.split(os.pathsep) |
|
1209 | 1209 | for p in path: |
|
1210 | 1210 | p_name = os.path.join(p, name) |
|
1211 | 1211 | if os.path.exists(p_name): |
|
1212 | 1212 | return p_name |
|
1213 | 1213 | return default |
|
1214 | 1214 | |
|
1215 | 1215 | def set_signal_handler(): |
|
1216 | 1216 | pass |
|
1217 | 1217 | |
|
1218 | 1218 | def find_exe(name, default=None): |
|
1219 | 1219 | '''find path of an executable. |
|
1220 | 1220 | if name contains a path component, return it as is. otherwise, |
|
1221 | 1221 | use normal executable search path.''' |
|
1222 | 1222 | |
|
1223 | 1223 | if os.sep in name or sys.platform == 'OpenVMS': |
|
1224 | 1224 | # don't check the executable bit. if the file isn't |
|
1225 | 1225 | # executable, whoever tries to actually run it will give a |
|
1226 | 1226 | # much more useful error message. |
|
1227 | 1227 | return name |
|
1228 | 1228 | return find_in_path(name, os.environ.get('PATH', ''), default=default) |
|
1229 | 1229 | |
|
1230 | 1230 | def _buildencodefun(): |
|
1231 | 1231 | e = '_' |
|
1232 | 1232 | win_reserved = [ord(x) for x in '\\:*?"<>|'] |
|
1233 | 1233 | cmap = dict([ (chr(x), chr(x)) for x in xrange(127) ]) |
|
1234 | 1234 | for x in (range(32) + range(126, 256) + win_reserved): |
|
1235 | 1235 | cmap[chr(x)] = "~%02x" % x |
|
1236 | 1236 | for x in range(ord("A"), ord("Z")+1) + [ord(e)]: |
|
1237 | 1237 | cmap[chr(x)] = e + chr(x).lower() |
|
1238 | 1238 | dmap = {} |
|
1239 | 1239 | for k, v in cmap.iteritems(): |
|
1240 | 1240 | dmap[v] = k |
|
1241 | 1241 | def decode(s): |
|
1242 | 1242 | i = 0 |
|
1243 | 1243 | while i < len(s): |
|
1244 | 1244 | for l in xrange(1, 4): |
|
1245 | 1245 | try: |
|
1246 | 1246 | yield dmap[s[i:i+l]] |
|
1247 | 1247 | i += l |
|
1248 | 1248 | break |
|
1249 | 1249 | except KeyError: |
|
1250 | 1250 | pass |
|
1251 | 1251 | else: |
|
1252 | 1252 | raise KeyError |
|
1253 | 1253 | return (lambda s: "".join([cmap[c] for c in s]), |
|
1254 | 1254 | lambda s: "".join(list(decode(s)))) |
|
1255 | 1255 | |
|
1256 | 1256 | encodefilename, decodefilename = _buildencodefun() |
|
1257 | 1257 | |
|
1258 | 1258 | def encodedopener(openerfn, fn): |
|
1259 | 1259 | def o(path, *args, **kw): |
|
1260 | 1260 | return openerfn(fn(path), *args, **kw) |
|
1261 | 1261 | return o |
|
1262 | 1262 | |
|
1263 | 1263 | def mktempcopy(name, emptyok=False): |
|
1264 | 1264 | """Create a temporary file with the same contents from name |
|
1265 | 1265 | |
|
1266 | 1266 | The permission bits are copied from the original file. |
|
1267 | 1267 | |
|
1268 | 1268 | If the temporary file is going to be truncated immediately, you |
|
1269 | 1269 | can use emptyok=True as an optimization. |
|
1270 | 1270 | |
|
1271 | 1271 | Returns the name of the temporary file. |
|
1272 | 1272 | """ |
|
1273 | 1273 | d, fn = os.path.split(name) |
|
1274 | 1274 | fd, temp = tempfile.mkstemp(prefix='.%s-' % fn, dir=d) |
|
1275 | 1275 | os.close(fd) |
|
1276 | 1276 | # Temporary files are created with mode 0600, which is usually not |
|
1277 | 1277 | # what we want. If the original file already exists, just copy |
|
1278 | 1278 | # its mode. Otherwise, manually obey umask. |
|
1279 | 1279 | try: |
|
1280 | 1280 | st_mode = os.lstat(name).st_mode |
|
1281 | 1281 | except OSError, inst: |
|
1282 | 1282 | if inst.errno != errno.ENOENT: |
|
1283 | 1283 | raise |
|
1284 | 1284 | st_mode = 0666 & ~_umask |
|
1285 | 1285 | os.chmod(temp, st_mode) |
|
1286 | 1286 | if emptyok: |
|
1287 | 1287 | return temp |
|
1288 | 1288 | try: |
|
1289 | 1289 | try: |
|
1290 | 1290 | ifp = posixfile(name, "rb") |
|
1291 | 1291 | except IOError, inst: |
|
1292 | 1292 | if inst.errno == errno.ENOENT: |
|
1293 | 1293 | return temp |
|
1294 | 1294 | if not getattr(inst, 'filename', None): |
|
1295 | 1295 | inst.filename = name |
|
1296 | 1296 | raise |
|
1297 | 1297 | ofp = posixfile(temp, "wb") |
|
1298 | 1298 | for chunk in filechunkiter(ifp): |
|
1299 | 1299 | ofp.write(chunk) |
|
1300 | 1300 | ifp.close() |
|
1301 | 1301 | ofp.close() |
|
1302 | 1302 | except: |
|
1303 | 1303 | try: os.unlink(temp) |
|
1304 | 1304 | except: pass |
|
1305 | 1305 | raise |
|
1306 | 1306 | return temp |
|
1307 | 1307 | |
|
1308 | 1308 | class atomictempfile(posixfile): |
|
1309 | 1309 | """file-like object that atomically updates a file |
|
1310 | 1310 | |
|
1311 | 1311 | All writes will be redirected to a temporary copy of the original |
|
1312 | 1312 | file. When rename is called, the copy is renamed to the original |
|
1313 | 1313 | name, making the changes visible. |
|
1314 | 1314 | """ |
|
1315 | 1315 | def __init__(self, name, mode): |
|
1316 | 1316 | self.__name = name |
|
1317 | 1317 | self.temp = mktempcopy(name, emptyok=('w' in mode)) |
|
1318 | 1318 | posixfile.__init__(self, self.temp, mode) |
|
1319 | 1319 | |
|
1320 | 1320 | def rename(self): |
|
1321 | 1321 | if not self.closed: |
|
1322 | 1322 | posixfile.close(self) |
|
1323 | 1323 | rename(self.temp, localpath(self.__name)) |
|
1324 | 1324 | |
|
1325 | 1325 | def __del__(self): |
|
1326 | 1326 | if not self.closed: |
|
1327 | 1327 | try: |
|
1328 | 1328 | os.unlink(self.temp) |
|
1329 | 1329 | except: pass |
|
1330 | 1330 | posixfile.close(self) |
|
1331 | 1331 | |
|
1332 | 1332 | class opener(object): |
|
1333 | 1333 | """Open files relative to a base directory |
|
1334 | 1334 | |
|
1335 | 1335 | This class is used to hide the details of COW semantics and |
|
1336 | 1336 | remote file access from higher level code. |
|
1337 | 1337 | """ |
|
1338 | 1338 | def __init__(self, base, audit=True): |
|
1339 | 1339 | self.base = base |
|
1340 | 1340 | if audit: |
|
1341 | 1341 | self.audit_path = path_auditor(base) |
|
1342 | 1342 | else: |
|
1343 | 1343 | self.audit_path = always |
|
1344 | 1344 | |
|
1345 | 1345 | def __getattr__(self, name): |
|
1346 | 1346 | if name == '_can_symlink': |
|
1347 | 1347 | self._can_symlink = checklink(self.base) |
|
1348 | 1348 | return self._can_symlink |
|
1349 | 1349 | raise AttributeError(name) |
|
1350 | 1350 | |
|
1351 | 1351 | def __call__(self, path, mode="r", text=False, atomictemp=False): |
|
1352 | 1352 | self.audit_path(path) |
|
1353 | 1353 | f = os.path.join(self.base, path) |
|
1354 | 1354 | |
|
1355 | 1355 | if not text and "b" not in mode: |
|
1356 | 1356 | mode += "b" # for that other OS |
|
1357 | 1357 | |
|
1358 | 1358 | if mode[0] != "r": |
|
1359 | 1359 | try: |
|
1360 | 1360 | nlink = nlinks(f) |
|
1361 | 1361 | except OSError: |
|
1362 | 1362 | nlink = 0 |
|
1363 | 1363 | d = os.path.dirname(f) |
|
1364 | 1364 | if not os.path.isdir(d): |
|
1365 | 1365 | os.makedirs(d) |
|
1366 | 1366 | if atomictemp: |
|
1367 | 1367 | return atomictempfile(f, mode) |
|
1368 | 1368 | if nlink > 1: |
|
1369 | 1369 | rename(mktempcopy(f), f) |
|
1370 | 1370 | return posixfile(f, mode) |
|
1371 | 1371 | |
|
1372 | 1372 | def symlink(self, src, dst): |
|
1373 | 1373 | self.audit_path(dst) |
|
1374 | 1374 | linkname = os.path.join(self.base, dst) |
|
1375 | 1375 | try: |
|
1376 | 1376 | os.unlink(linkname) |
|
1377 | 1377 | except OSError: |
|
1378 | 1378 | pass |
|
1379 | 1379 | |
|
1380 | 1380 | dirname = os.path.dirname(linkname) |
|
1381 | 1381 | if not os.path.exists(dirname): |
|
1382 | 1382 | os.makedirs(dirname) |
|
1383 | 1383 | |
|
1384 | 1384 | if self._can_symlink: |
|
1385 | 1385 | try: |
|
1386 | 1386 | os.symlink(src, linkname) |
|
1387 | 1387 | except OSError, err: |
|
1388 | 1388 | raise OSError(err.errno, _('could not symlink to %r: %s') % |
|
1389 | 1389 | (src, err.strerror), linkname) |
|
1390 | 1390 | else: |
|
1391 | 1391 | f = self(dst, "w") |
|
1392 | 1392 | f.write(src) |
|
1393 | 1393 | f.close() |
|
1394 | 1394 | |
|
1395 | 1395 | class chunkbuffer(object): |
|
1396 | 1396 | """Allow arbitrary sized chunks of data to be efficiently read from an |
|
1397 | 1397 | iterator over chunks of arbitrary size.""" |
|
1398 | 1398 | |
|
1399 | 1399 | def __init__(self, in_iter): |
|
1400 | 1400 | """in_iter is the iterator that's iterating over the input chunks. |
|
1401 | 1401 | targetsize is how big a buffer to try to maintain.""" |
|
1402 | 1402 | self.iter = iter(in_iter) |
|
1403 | 1403 | self.buf = '' |
|
1404 | 1404 | self.targetsize = 2**16 |
|
1405 | 1405 | |
|
1406 | 1406 | def read(self, l): |
|
1407 | 1407 | """Read L bytes of data from the iterator of chunks of data. |
|
1408 | 1408 | Returns less than L bytes if the iterator runs dry.""" |
|
1409 | 1409 | if l > len(self.buf) and self.iter: |
|
1410 | 1410 | # Clamp to a multiple of self.targetsize |
|
1411 | 1411 | targetsize = max(l, self.targetsize) |
|
1412 | 1412 | collector = cStringIO.StringIO() |
|
1413 | 1413 | collector.write(self.buf) |
|
1414 | 1414 | collected = len(self.buf) |
|
1415 | 1415 | for chunk in self.iter: |
|
1416 | 1416 | collector.write(chunk) |
|
1417 | 1417 | collected += len(chunk) |
|
1418 | 1418 | if collected >= targetsize: |
|
1419 | 1419 | break |
|
1420 | 1420 | if collected < targetsize: |
|
1421 | 1421 | self.iter = False |
|
1422 | 1422 | self.buf = collector.getvalue() |
|
1423 | 1423 | if len(self.buf) == l: |
|
1424 | s, self.buf = self.buf, '' | |
|
1424 | s, self.buf = str(self.buf), '' | |
|
1425 | 1425 | else: |
|
1426 | 1426 | s, self.buf = self.buf[:l], buffer(self.buf, l) |
|
1427 | 1427 | return s |
|
1428 | 1428 | |
|
1429 | 1429 | def filechunkiter(f, size=65536, limit=None): |
|
1430 | 1430 | """Create a generator that produces the data in the file size |
|
1431 | 1431 | (default 65536) bytes at a time, up to optional limit (default is |
|
1432 | 1432 | to read all data). Chunks may be less than size bytes if the |
|
1433 | 1433 | chunk is the last chunk in the file, or the file is a socket or |
|
1434 | 1434 | some other type of file that sometimes reads less data than is |
|
1435 | 1435 | requested.""" |
|
1436 | 1436 | assert size >= 0 |
|
1437 | 1437 | assert limit is None or limit >= 0 |
|
1438 | 1438 | while True: |
|
1439 | 1439 | if limit is None: nbytes = size |
|
1440 | 1440 | else: nbytes = min(limit, size) |
|
1441 | 1441 | s = nbytes and f.read(nbytes) |
|
1442 | 1442 | if not s: break |
|
1443 | 1443 | if limit: limit -= len(s) |
|
1444 | 1444 | yield s |
|
1445 | 1445 | |
|
1446 | 1446 | def makedate(): |
|
1447 | 1447 | lt = time.localtime() |
|
1448 | 1448 | if lt[8] == 1 and time.daylight: |
|
1449 | 1449 | tz = time.altzone |
|
1450 | 1450 | else: |
|
1451 | 1451 | tz = time.timezone |
|
1452 | 1452 | return time.mktime(lt), tz |
|
1453 | 1453 | |
|
1454 | 1454 | def datestr(date=None, format='%a %b %d %H:%M:%S %Y', timezone=True, timezone_format=" %+03d%02d"): |
|
1455 | 1455 | """represent a (unixtime, offset) tuple as a localized time. |
|
1456 | 1456 | unixtime is seconds since the epoch, and offset is the time zone's |
|
1457 | 1457 | number of seconds away from UTC. if timezone is false, do not |
|
1458 | 1458 | append time zone to string.""" |
|
1459 | 1459 | t, tz = date or makedate() |
|
1460 | 1460 | s = time.strftime(format, time.gmtime(float(t) - tz)) |
|
1461 | 1461 | if timezone: |
|
1462 | 1462 | s += timezone_format % (-tz / 3600, ((-tz % 3600) / 60)) |
|
1463 | 1463 | return s |
|
1464 | 1464 | |
|
1465 | 1465 | def strdate(string, format, defaults=[]): |
|
1466 | 1466 | """parse a localized time string and return a (unixtime, offset) tuple. |
|
1467 | 1467 | if the string cannot be parsed, ValueError is raised.""" |
|
1468 | 1468 | def timezone(string): |
|
1469 | 1469 | tz = string.split()[-1] |
|
1470 | 1470 | if tz[0] in "+-" and len(tz) == 5 and tz[1:].isdigit(): |
|
1471 | 1471 | tz = int(tz) |
|
1472 | 1472 | offset = - 3600 * (tz / 100) - 60 * (tz % 100) |
|
1473 | 1473 | return offset |
|
1474 | 1474 | if tz == "GMT" or tz == "UTC": |
|
1475 | 1475 | return 0 |
|
1476 | 1476 | return None |
|
1477 | 1477 | |
|
1478 | 1478 | # NOTE: unixtime = localunixtime + offset |
|
1479 | 1479 | offset, date = timezone(string), string |
|
1480 | 1480 | if offset != None: |
|
1481 | 1481 | date = " ".join(string.split()[:-1]) |
|
1482 | 1482 | |
|
1483 | 1483 | # add missing elements from defaults |
|
1484 | 1484 | for part in defaults: |
|
1485 | 1485 | found = [True for p in part if ("%"+p) in format] |
|
1486 | 1486 | if not found: |
|
1487 | 1487 | date += "@" + defaults[part] |
|
1488 | 1488 | format += "@%" + part[0] |
|
1489 | 1489 | |
|
1490 | 1490 | timetuple = time.strptime(date, format) |
|
1491 | 1491 | localunixtime = int(calendar.timegm(timetuple)) |
|
1492 | 1492 | if offset is None: |
|
1493 | 1493 | # local timezone |
|
1494 | 1494 | unixtime = int(time.mktime(timetuple)) |
|
1495 | 1495 | offset = unixtime - localunixtime |
|
1496 | 1496 | else: |
|
1497 | 1497 | unixtime = localunixtime + offset |
|
1498 | 1498 | return unixtime, offset |
|
1499 | 1499 | |
|
1500 | 1500 | def parsedate(string, formats=None, defaults=None): |
|
1501 | 1501 | """parse a localized time string and return a (unixtime, offset) tuple. |
|
1502 | 1502 | The date may be a "unixtime offset" string or in one of the specified |
|
1503 | 1503 | formats.""" |
|
1504 | 1504 | if not string: |
|
1505 | 1505 | return 0, 0 |
|
1506 | 1506 | if not formats: |
|
1507 | 1507 | formats = defaultdateformats |
|
1508 | 1508 | string = string.strip() |
|
1509 | 1509 | try: |
|
1510 | 1510 | when, offset = map(int, string.split(' ')) |
|
1511 | 1511 | except ValueError: |
|
1512 | 1512 | # fill out defaults |
|
1513 | 1513 | if not defaults: |
|
1514 | 1514 | defaults = {} |
|
1515 | 1515 | now = makedate() |
|
1516 | 1516 | for part in "d mb yY HI M S".split(): |
|
1517 | 1517 | if part not in defaults: |
|
1518 | 1518 | if part[0] in "HMS": |
|
1519 | 1519 | defaults[part] = "00" |
|
1520 | 1520 | elif part[0] in "dm": |
|
1521 | 1521 | defaults[part] = "1" |
|
1522 | 1522 | else: |
|
1523 | 1523 | defaults[part] = datestr(now, "%" + part[0], False) |
|
1524 | 1524 | |
|
1525 | 1525 | for format in formats: |
|
1526 | 1526 | try: |
|
1527 | 1527 | when, offset = strdate(string, format, defaults) |
|
1528 | 1528 | except ValueError: |
|
1529 | 1529 | pass |
|
1530 | 1530 | else: |
|
1531 | 1531 | break |
|
1532 | 1532 | else: |
|
1533 | 1533 | raise Abort(_('invalid date: %r ') % string) |
|
1534 | 1534 | # validate explicit (probably user-specified) date and |
|
1535 | 1535 | # time zone offset. values must fit in signed 32 bits for |
|
1536 | 1536 | # current 32-bit linux runtimes. timezones go from UTC-12 |
|
1537 | 1537 | # to UTC+14 |
|
1538 | 1538 | if abs(when) > 0x7fffffff: |
|
1539 | 1539 | raise Abort(_('date exceeds 32 bits: %d') % when) |
|
1540 | 1540 | if offset < -50400 or offset > 43200: |
|
1541 | 1541 | raise Abort(_('impossible time zone offset: %d') % offset) |
|
1542 | 1542 | return when, offset |
|
1543 | 1543 | |
|
1544 | 1544 | def matchdate(date): |
|
1545 | 1545 | """Return a function that matches a given date match specifier |
|
1546 | 1546 | |
|
1547 | 1547 | Formats include: |
|
1548 | 1548 | |
|
1549 | 1549 | '{date}' match a given date to the accuracy provided |
|
1550 | 1550 | |
|
1551 | 1551 | '<{date}' on or before a given date |
|
1552 | 1552 | |
|
1553 | 1553 | '>{date}' on or after a given date |
|
1554 | 1554 | |
|
1555 | 1555 | """ |
|
1556 | 1556 | |
|
1557 | 1557 | def lower(date): |
|
1558 | 1558 | return parsedate(date, extendeddateformats)[0] |
|
1559 | 1559 | |
|
1560 | 1560 | def upper(date): |
|
1561 | 1561 | d = dict(mb="12", HI="23", M="59", S="59") |
|
1562 | 1562 | for days in "31 30 29".split(): |
|
1563 | 1563 | try: |
|
1564 | 1564 | d["d"] = days |
|
1565 | 1565 | return parsedate(date, extendeddateformats, d)[0] |
|
1566 | 1566 | except: |
|
1567 | 1567 | pass |
|
1568 | 1568 | d["d"] = "28" |
|
1569 | 1569 | return parsedate(date, extendeddateformats, d)[0] |
|
1570 | 1570 | |
|
1571 | 1571 | if date[0] == "<": |
|
1572 | 1572 | when = upper(date[1:]) |
|
1573 | 1573 | return lambda x: x <= when |
|
1574 | 1574 | elif date[0] == ">": |
|
1575 | 1575 | when = lower(date[1:]) |
|
1576 | 1576 | return lambda x: x >= when |
|
1577 | 1577 | elif date[0] == "-": |
|
1578 | 1578 | try: |
|
1579 | 1579 | days = int(date[1:]) |
|
1580 | 1580 | except ValueError: |
|
1581 | 1581 | raise Abort(_("invalid day spec: %s") % date[1:]) |
|
1582 | 1582 | when = makedate()[0] - days * 3600 * 24 |
|
1583 | 1583 | return lambda x: x >= when |
|
1584 | 1584 | elif " to " in date: |
|
1585 | 1585 | a, b = date.split(" to ") |
|
1586 | 1586 | start, stop = lower(a), upper(b) |
|
1587 | 1587 | return lambda x: x >= start and x <= stop |
|
1588 | 1588 | else: |
|
1589 | 1589 | start, stop = lower(date), upper(date) |
|
1590 | 1590 | return lambda x: x >= start and x <= stop |
|
1591 | 1591 | |
|
1592 | 1592 | def shortuser(user): |
|
1593 | 1593 | """Return a short representation of a user name or email address.""" |
|
1594 | 1594 | f = user.find('@') |
|
1595 | 1595 | if f >= 0: |
|
1596 | 1596 | user = user[:f] |
|
1597 | 1597 | f = user.find('<') |
|
1598 | 1598 | if f >= 0: |
|
1599 | 1599 | user = user[f+1:] |
|
1600 | 1600 | f = user.find(' ') |
|
1601 | 1601 | if f >= 0: |
|
1602 | 1602 | user = user[:f] |
|
1603 | 1603 | f = user.find('.') |
|
1604 | 1604 | if f >= 0: |
|
1605 | 1605 | user = user[:f] |
|
1606 | 1606 | return user |
|
1607 | 1607 | |
|
1608 | 1608 | def ellipsis(text, maxlength=400): |
|
1609 | 1609 | """Trim string to at most maxlength (default: 400) characters.""" |
|
1610 | 1610 | if len(text) <= maxlength: |
|
1611 | 1611 | return text |
|
1612 | 1612 | else: |
|
1613 | 1613 | return "%s..." % (text[:maxlength-3]) |
|
1614 | 1614 | |
|
1615 | 1615 | def walkrepos(path): |
|
1616 | 1616 | '''yield every hg repository under path, recursively.''' |
|
1617 | 1617 | def errhandler(err): |
|
1618 | 1618 | if err.filename == path: |
|
1619 | 1619 | raise err |
|
1620 | 1620 | |
|
1621 | 1621 | for root, dirs, files in os.walk(path, onerror=errhandler): |
|
1622 | 1622 | for d in dirs: |
|
1623 | 1623 | if d == '.hg': |
|
1624 | 1624 | yield root |
|
1625 | 1625 | dirs[:] = [] |
|
1626 | 1626 | break |
|
1627 | 1627 | |
|
1628 | 1628 | _rcpath = None |
|
1629 | 1629 | |
|
1630 | 1630 | def os_rcpath(): |
|
1631 | 1631 | '''return default os-specific hgrc search path''' |
|
1632 | 1632 | path = system_rcpath() |
|
1633 | 1633 | path.extend(user_rcpath()) |
|
1634 | 1634 | path = [os.path.normpath(f) for f in path] |
|
1635 | 1635 | return path |
|
1636 | 1636 | |
|
1637 | 1637 | def rcpath(): |
|
1638 | 1638 | '''return hgrc search path. if env var HGRCPATH is set, use it. |
|
1639 | 1639 | for each item in path, if directory, use files ending in .rc, |
|
1640 | 1640 | else use item. |
|
1641 | 1641 | make HGRCPATH empty to only look in .hg/hgrc of current repo. |
|
1642 | 1642 | if no HGRCPATH, use default os-specific path.''' |
|
1643 | 1643 | global _rcpath |
|
1644 | 1644 | if _rcpath is None: |
|
1645 | 1645 | if 'HGRCPATH' in os.environ: |
|
1646 | 1646 | _rcpath = [] |
|
1647 | 1647 | for p in os.environ['HGRCPATH'].split(os.pathsep): |
|
1648 | 1648 | if not p: continue |
|
1649 | 1649 | if os.path.isdir(p): |
|
1650 | 1650 | for f, kind in osutil.listdir(p): |
|
1651 | 1651 | if f.endswith('.rc'): |
|
1652 | 1652 | _rcpath.append(os.path.join(p, f)) |
|
1653 | 1653 | else: |
|
1654 | 1654 | _rcpath.append(p) |
|
1655 | 1655 | else: |
|
1656 | 1656 | _rcpath = os_rcpath() |
|
1657 | 1657 | return _rcpath |
|
1658 | 1658 | |
|
1659 | 1659 | def bytecount(nbytes): |
|
1660 | 1660 | '''return byte count formatted as readable string, with units''' |
|
1661 | 1661 | |
|
1662 | 1662 | units = ( |
|
1663 | 1663 | (100, 1<<30, _('%.0f GB')), |
|
1664 | 1664 | (10, 1<<30, _('%.1f GB')), |
|
1665 | 1665 | (1, 1<<30, _('%.2f GB')), |
|
1666 | 1666 | (100, 1<<20, _('%.0f MB')), |
|
1667 | 1667 | (10, 1<<20, _('%.1f MB')), |
|
1668 | 1668 | (1, 1<<20, _('%.2f MB')), |
|
1669 | 1669 | (100, 1<<10, _('%.0f KB')), |
|
1670 | 1670 | (10, 1<<10, _('%.1f KB')), |
|
1671 | 1671 | (1, 1<<10, _('%.2f KB')), |
|
1672 | 1672 | (1, 1, _('%.0f bytes')), |
|
1673 | 1673 | ) |
|
1674 | 1674 | |
|
1675 | 1675 | for multiplier, divisor, format in units: |
|
1676 | 1676 | if nbytes >= divisor * multiplier: |
|
1677 | 1677 | return format % (nbytes / float(divisor)) |
|
1678 | 1678 | return units[-1][2] % nbytes |
|
1679 | 1679 | |
|
1680 | 1680 | def drop_scheme(scheme, path): |
|
1681 | 1681 | sc = scheme + ':' |
|
1682 | 1682 | if path.startswith(sc): |
|
1683 | 1683 | path = path[len(sc):] |
|
1684 | 1684 | if path.startswith('//'): |
|
1685 | 1685 | path = path[2:] |
|
1686 | 1686 | return path |
|
1687 | 1687 | |
|
1688 | 1688 | def uirepr(s): |
|
1689 | 1689 | # Avoid double backslash in Windows path repr() |
|
1690 | 1690 | return repr(s).replace('\\\\', '\\') |
General Comments 0
You need to be logged in to leave comments.
Login now