Show More
@@ -1,298 +1,296 b'' | |||
|
1 | """ | |
|
2 | bundlerepo.py - repository class for viewing uncompressed bundles | |
|
3 | ||
|
4 | This provides a read-only repository interface to bundles as if | |
|
5 | they were part of the actual repository. | |
|
6 | ||
|
7 | Copyright 2006, 2007 Benoit Boissinot <bboissin@gmail.com> | |
|
8 | ||
|
9 | This software may be used and distributed according to the terms of the | |
|
10 | GNU General Public License version 2, incorporated herein by reference. | |
|
11 | """ | |
|
1 | # bundlerepo.py - repository class for viewing uncompressed bundles | |
|
2 | # | |
|
3 | # This provides a read-only repository interface to bundles as if | |
|
4 | # they were part of the actual repository. | |
|
5 | # | |
|
6 | # Copyright 2006, 2007 Benoit Boissinot <bboissin@gmail.com> | |
|
7 | # | |
|
8 | # This software may be used and distributed according to the terms of the | |
|
9 | # GNU General Public License version 2, incorporated herein by reference. | |
|
12 | 10 | |
|
13 | 11 | from node import nullid |
|
14 | 12 | from i18n import _ |
|
15 | 13 | import changegroup, util, os, struct, bz2, zlib, tempfile, shutil, mdiff |
|
16 | 14 | import localrepo, changelog, manifest, filelog, revlog, error |
|
17 | 15 | |
|
18 | 16 | class bundlerevlog(revlog.revlog): |
|
19 | 17 | def __init__(self, opener, indexfile, bundlefile, |
|
20 | 18 | linkmapper=None): |
|
21 | 19 | # How it works: |
|
22 | 20 | # to retrieve a revision, we need to know the offset of |
|
23 | 21 | # the revision in the bundlefile (an opened file). |
|
24 | 22 | # |
|
25 | 23 | # We store this offset in the index (start), to differentiate a |
|
26 | 24 | # rev in the bundle and from a rev in the revlog, we check |
|
27 | 25 | # len(index[r]). If the tuple is bigger than 7, it is a bundle |
|
28 | 26 | # (it is bigger since we store the node to which the delta is) |
|
29 | 27 | # |
|
30 | 28 | revlog.revlog.__init__(self, opener, indexfile) |
|
31 | 29 | self.bundlefile = bundlefile |
|
32 | 30 | self.basemap = {} |
|
33 | 31 | def chunkpositer(): |
|
34 | 32 | for chunk in changegroup.chunkiter(bundlefile): |
|
35 | 33 | pos = bundlefile.tell() |
|
36 | 34 | yield chunk, pos - len(chunk) |
|
37 | 35 | n = len(self) |
|
38 | 36 | prev = None |
|
39 | 37 | for chunk, start in chunkpositer(): |
|
40 | 38 | size = len(chunk) |
|
41 | 39 | if size < 80: |
|
42 | 40 | raise util.Abort(_("invalid changegroup")) |
|
43 | 41 | start += 80 |
|
44 | 42 | size -= 80 |
|
45 | 43 | node, p1, p2, cs = struct.unpack("20s20s20s20s", chunk[:80]) |
|
46 | 44 | if node in self.nodemap: |
|
47 | 45 | prev = node |
|
48 | 46 | continue |
|
49 | 47 | for p in (p1, p2): |
|
50 | 48 | if not p in self.nodemap: |
|
51 | 49 | raise error.LookupError(p1, self.indexfile, |
|
52 | 50 | _("unknown parent")) |
|
53 | 51 | if linkmapper is None: |
|
54 | 52 | link = n |
|
55 | 53 | else: |
|
56 | 54 | link = linkmapper(cs) |
|
57 | 55 | |
|
58 | 56 | if not prev: |
|
59 | 57 | prev = p1 |
|
60 | 58 | # start, size, full unc. size, base (unused), link, p1, p2, node |
|
61 | 59 | e = (revlog.offset_type(start, 0), size, -1, -1, link, |
|
62 | 60 | self.rev(p1), self.rev(p2), node) |
|
63 | 61 | self.basemap[n] = prev |
|
64 | 62 | self.index.insert(-1, e) |
|
65 | 63 | self.nodemap[node] = n |
|
66 | 64 | prev = node |
|
67 | 65 | n += 1 |
|
68 | 66 | |
|
69 | 67 | def bundle(self, rev): |
|
70 | 68 | """is rev from the bundle""" |
|
71 | 69 | if rev < 0: |
|
72 | 70 | return False |
|
73 | 71 | return rev in self.basemap |
|
74 | 72 | def bundlebase(self, rev): return self.basemap[rev] |
|
75 | 73 | def chunk(self, rev, df=None, cachelen=4096): |
|
76 | 74 | # Warning: in case of bundle, the diff is against bundlebase, |
|
77 | 75 | # not against rev - 1 |
|
78 | 76 | # XXX: could use some caching |
|
79 | 77 | if not self.bundle(rev): |
|
80 | 78 | return revlog.revlog.chunk(self, rev, df) |
|
81 | 79 | self.bundlefile.seek(self.start(rev)) |
|
82 | 80 | return self.bundlefile.read(self.length(rev)) |
|
83 | 81 | |
|
84 | 82 | def revdiff(self, rev1, rev2): |
|
85 | 83 | """return or calculate a delta between two revisions""" |
|
86 | 84 | if self.bundle(rev1) and self.bundle(rev2): |
|
87 | 85 | # hot path for bundle |
|
88 | 86 | revb = self.rev(self.bundlebase(rev2)) |
|
89 | 87 | if revb == rev1: |
|
90 | 88 | return self.chunk(rev2) |
|
91 | 89 | elif not self.bundle(rev1) and not self.bundle(rev2): |
|
92 | 90 | return revlog.revlog.revdiff(self, rev1, rev2) |
|
93 | 91 | |
|
94 | 92 | return mdiff.textdiff(self.revision(self.node(rev1)), |
|
95 | 93 | self.revision(self.node(rev2))) |
|
96 | 94 | |
|
97 | 95 | def revision(self, node): |
|
98 | 96 | """return an uncompressed revision of a given""" |
|
99 | 97 | if node == nullid: return "" |
|
100 | 98 | |
|
101 | 99 | text = None |
|
102 | 100 | chain = [] |
|
103 | 101 | iter_node = node |
|
104 | 102 | rev = self.rev(iter_node) |
|
105 | 103 | # reconstruct the revision if it is from a changegroup |
|
106 | 104 | while self.bundle(rev): |
|
107 | 105 | if self._cache and self._cache[0] == iter_node: |
|
108 | 106 | text = self._cache[2] |
|
109 | 107 | break |
|
110 | 108 | chain.append(rev) |
|
111 | 109 | iter_node = self.bundlebase(rev) |
|
112 | 110 | rev = self.rev(iter_node) |
|
113 | 111 | if text is None: |
|
114 | 112 | text = revlog.revlog.revision(self, iter_node) |
|
115 | 113 | |
|
116 | 114 | while chain: |
|
117 | 115 | delta = self.chunk(chain.pop()) |
|
118 | 116 | text = mdiff.patches(text, [delta]) |
|
119 | 117 | |
|
120 | 118 | p1, p2 = self.parents(node) |
|
121 | 119 | if node != revlog.hash(text, p1, p2): |
|
122 | 120 | raise error.RevlogError(_("integrity check failed on %s:%d") |
|
123 | 121 | % (self.datafile, self.rev(node))) |
|
124 | 122 | |
|
125 | 123 | self._cache = (node, self.rev(node), text) |
|
126 | 124 | return text |
|
127 | 125 | |
|
128 | 126 | def addrevision(self, text, transaction, link, p1=None, p2=None, d=None): |
|
129 | 127 | raise NotImplementedError |
|
130 | 128 | def addgroup(self, revs, linkmapper, transaction): |
|
131 | 129 | raise NotImplementedError |
|
132 | 130 | def strip(self, rev, minlink): |
|
133 | 131 | raise NotImplementedError |
|
134 | 132 | def checksize(self): |
|
135 | 133 | raise NotImplementedError |
|
136 | 134 | |
|
137 | 135 | class bundlechangelog(bundlerevlog, changelog.changelog): |
|
138 | 136 | def __init__(self, opener, bundlefile): |
|
139 | 137 | changelog.changelog.__init__(self, opener) |
|
140 | 138 | bundlerevlog.__init__(self, opener, self.indexfile, bundlefile) |
|
141 | 139 | |
|
142 | 140 | class bundlemanifest(bundlerevlog, manifest.manifest): |
|
143 | 141 | def __init__(self, opener, bundlefile, linkmapper): |
|
144 | 142 | manifest.manifest.__init__(self, opener) |
|
145 | 143 | bundlerevlog.__init__(self, opener, self.indexfile, bundlefile, |
|
146 | 144 | linkmapper) |
|
147 | 145 | |
|
148 | 146 | class bundlefilelog(bundlerevlog, filelog.filelog): |
|
149 | 147 | def __init__(self, opener, path, bundlefile, linkmapper): |
|
150 | 148 | filelog.filelog.__init__(self, opener, path) |
|
151 | 149 | bundlerevlog.__init__(self, opener, self.indexfile, bundlefile, |
|
152 | 150 | linkmapper) |
|
153 | 151 | |
|
154 | 152 | class bundlerepository(localrepo.localrepository): |
|
155 | 153 | def __init__(self, ui, path, bundlename): |
|
156 | 154 | self._tempparent = None |
|
157 | 155 | try: |
|
158 | 156 | localrepo.localrepository.__init__(self, ui, path) |
|
159 | 157 | except error.RepoError: |
|
160 | 158 | self._tempparent = tempfile.mkdtemp() |
|
161 | 159 | localrepo.instance(ui,self._tempparent,1) |
|
162 | 160 | localrepo.localrepository.__init__(self, ui, self._tempparent) |
|
163 | 161 | |
|
164 | 162 | if path: |
|
165 | 163 | self._url = 'bundle:' + path + '+' + bundlename |
|
166 | 164 | else: |
|
167 | 165 | self._url = 'bundle:' + bundlename |
|
168 | 166 | |
|
169 | 167 | self.tempfile = None |
|
170 | 168 | self.bundlefile = open(bundlename, "rb") |
|
171 | 169 | header = self.bundlefile.read(6) |
|
172 | 170 | if not header.startswith("HG"): |
|
173 | 171 | raise util.Abort(_("%s: not a Mercurial bundle file") % bundlename) |
|
174 | 172 | elif not header.startswith("HG10"): |
|
175 | 173 | raise util.Abort(_("%s: unknown bundle version") % bundlename) |
|
176 | 174 | elif (header == "HG10BZ") or (header == "HG10GZ"): |
|
177 | 175 | fdtemp, temp = tempfile.mkstemp(prefix="hg-bundle-", |
|
178 | 176 | suffix=".hg10un", dir=self.path) |
|
179 | 177 | self.tempfile = temp |
|
180 | 178 | fptemp = os.fdopen(fdtemp, 'wb') |
|
181 | 179 | def generator(f): |
|
182 | 180 | if header == "HG10BZ": |
|
183 | 181 | zd = bz2.BZ2Decompressor() |
|
184 | 182 | zd.decompress("BZ") |
|
185 | 183 | elif header == "HG10GZ": |
|
186 | 184 | zd = zlib.decompressobj() |
|
187 | 185 | for chunk in f: |
|
188 | 186 | yield zd.decompress(chunk) |
|
189 | 187 | gen = generator(util.filechunkiter(self.bundlefile, 4096)) |
|
190 | 188 | |
|
191 | 189 | try: |
|
192 | 190 | fptemp.write("HG10UN") |
|
193 | 191 | for chunk in gen: |
|
194 | 192 | fptemp.write(chunk) |
|
195 | 193 | finally: |
|
196 | 194 | fptemp.close() |
|
197 | 195 | self.bundlefile.close() |
|
198 | 196 | |
|
199 | 197 | self.bundlefile = open(self.tempfile, "rb") |
|
200 | 198 | # seek right after the header |
|
201 | 199 | self.bundlefile.seek(6) |
|
202 | 200 | elif header == "HG10UN": |
|
203 | 201 | # nothing to do |
|
204 | 202 | pass |
|
205 | 203 | else: |
|
206 | 204 | raise util.Abort(_("%s: unknown bundle compression type") |
|
207 | 205 | % bundlename) |
|
208 | 206 | # dict with the mapping 'filename' -> position in the bundle |
|
209 | 207 | self.bundlefilespos = {} |
|
210 | 208 | |
|
211 | 209 | def __getattr__(self, name): |
|
212 | 210 | if name == 'changelog': |
|
213 | 211 | self.changelog = bundlechangelog(self.sopener, self.bundlefile) |
|
214 | 212 | self.manstart = self.bundlefile.tell() |
|
215 | 213 | return self.changelog |
|
216 | 214 | elif name == 'manifest': |
|
217 | 215 | self.bundlefile.seek(self.manstart) |
|
218 | 216 | self.manifest = bundlemanifest(self.sopener, self.bundlefile, |
|
219 | 217 | self.changelog.rev) |
|
220 | 218 | self.filestart = self.bundlefile.tell() |
|
221 | 219 | return self.manifest |
|
222 | 220 | elif name == 'manstart': |
|
223 | 221 | self.changelog |
|
224 | 222 | return self.manstart |
|
225 | 223 | elif name == 'filestart': |
|
226 | 224 | self.manifest |
|
227 | 225 | return self.filestart |
|
228 | 226 | else: |
|
229 | 227 | return localrepo.localrepository.__getattr__(self, name) |
|
230 | 228 | |
|
231 | 229 | def url(self): |
|
232 | 230 | return self._url |
|
233 | 231 | |
|
234 | 232 | def file(self, f): |
|
235 | 233 | if not self.bundlefilespos: |
|
236 | 234 | self.bundlefile.seek(self.filestart) |
|
237 | 235 | while 1: |
|
238 | 236 | chunk = changegroup.getchunk(self.bundlefile) |
|
239 | 237 | if not chunk: |
|
240 | 238 | break |
|
241 | 239 | self.bundlefilespos[chunk] = self.bundlefile.tell() |
|
242 | 240 | for c in changegroup.chunkiter(self.bundlefile): |
|
243 | 241 | pass |
|
244 | 242 | |
|
245 | 243 | if f[0] == '/': |
|
246 | 244 | f = f[1:] |
|
247 | 245 | if f in self.bundlefilespos: |
|
248 | 246 | self.bundlefile.seek(self.bundlefilespos[f]) |
|
249 | 247 | return bundlefilelog(self.sopener, f, self.bundlefile, |
|
250 | 248 | self.changelog.rev) |
|
251 | 249 | else: |
|
252 | 250 | return filelog.filelog(self.sopener, f) |
|
253 | 251 | |
|
254 | 252 | def close(self): |
|
255 | 253 | """Close assigned bundle file immediately.""" |
|
256 | 254 | self.bundlefile.close() |
|
257 | 255 | |
|
258 | 256 | def __del__(self): |
|
259 | 257 | bundlefile = getattr(self, 'bundlefile', None) |
|
260 | 258 | if bundlefile and not bundlefile.closed: |
|
261 | 259 | bundlefile.close() |
|
262 | 260 | tempfile = getattr(self, 'tempfile', None) |
|
263 | 261 | if tempfile is not None: |
|
264 | 262 | os.unlink(tempfile) |
|
265 | 263 | if self._tempparent: |
|
266 | 264 | shutil.rmtree(self._tempparent, True) |
|
267 | 265 | |
|
268 | 266 | def cancopy(self): |
|
269 | 267 | return False |
|
270 | 268 | |
|
271 | 269 | def getcwd(self): |
|
272 | 270 | return os.getcwd() # always outside the repo |
|
273 | 271 | |
|
274 | 272 | def instance(ui, path, create): |
|
275 | 273 | if create: |
|
276 | 274 | raise util.Abort(_('cannot create new bundle repository')) |
|
277 | 275 | parentpath = ui.config("bundle", "mainreporoot", "") |
|
278 | 276 | if parentpath: |
|
279 | 277 | # Try to make the full path relative so we get a nice, short URL. |
|
280 | 278 | # In particular, we don't want temp dir names in test outputs. |
|
281 | 279 | cwd = os.getcwd() |
|
282 | 280 | if parentpath == cwd: |
|
283 | 281 | parentpath = '' |
|
284 | 282 | else: |
|
285 | 283 | cwd = os.path.join(cwd,'') |
|
286 | 284 | if parentpath.startswith(cwd): |
|
287 | 285 | parentpath = parentpath[len(cwd):] |
|
288 | 286 | path = util.drop_scheme('file', path) |
|
289 | 287 | if path.startswith('bundle:'): |
|
290 | 288 | path = util.drop_scheme('bundle', path) |
|
291 | 289 | s = path.split("+", 1) |
|
292 | 290 | if len(s) == 1: |
|
293 | 291 | repopath, bundlename = parentpath, s[0] |
|
294 | 292 | else: |
|
295 | 293 | repopath, bundlename = s |
|
296 | 294 | else: |
|
297 | 295 | repopath, bundlename = parentpath, path |
|
298 | 296 | return bundlerepository(ui, repopath, bundlename) |
@@ -1,141 +1,139 b'' | |||
|
1 | """ | |
|
2 | changegroup.py - Mercurial changegroup manipulation functions | |
|
3 | ||
|
4 | Copyright 2006 Matt Mackall <mpm@selenic.com> | |
|
5 | ||
|
6 | This software may be used and distributed according to the terms of the | |
|
7 | GNU General Public License version 2, incorporated herein by reference. | |
|
8 | """ | |
|
1 | # changegroup.py - Mercurial changegroup manipulation functions | |
|
2 | # | |
|
3 | # Copyright 2006 Matt Mackall <mpm@selenic.com> | |
|
4 | # | |
|
5 | # This software may be used and distributed according to the terms of the | |
|
6 | # GNU General Public License version 2, incorporated herein by reference. | |
|
9 | 7 | |
|
10 | 8 | from i18n import _ |
|
11 | 9 | import struct, os, bz2, zlib, util, tempfile |
|
12 | 10 | |
|
13 | 11 | def getchunk(source): |
|
14 | 12 | """get a chunk from a changegroup""" |
|
15 | 13 | d = source.read(4) |
|
16 | 14 | if not d: |
|
17 | 15 | return "" |
|
18 | 16 | l = struct.unpack(">l", d)[0] |
|
19 | 17 | if l <= 4: |
|
20 | 18 | return "" |
|
21 | 19 | d = source.read(l - 4) |
|
22 | 20 | if len(d) < l - 4: |
|
23 | 21 | raise util.Abort(_("premature EOF reading chunk" |
|
24 | 22 | " (got %d bytes, expected %d)") |
|
25 | 23 | % (len(d), l - 4)) |
|
26 | 24 | return d |
|
27 | 25 | |
|
28 | 26 | def chunkiter(source): |
|
29 | 27 | """iterate through the chunks in source""" |
|
30 | 28 | while 1: |
|
31 | 29 | c = getchunk(source) |
|
32 | 30 | if not c: |
|
33 | 31 | break |
|
34 | 32 | yield c |
|
35 | 33 | |
|
36 | 34 | def chunkheader(length): |
|
37 | 35 | """build a changegroup chunk header""" |
|
38 | 36 | return struct.pack(">l", length + 4) |
|
39 | 37 | |
|
40 | 38 | def closechunk(): |
|
41 | 39 | return struct.pack(">l", 0) |
|
42 | 40 | |
|
43 | 41 | class nocompress(object): |
|
44 | 42 | def compress(self, x): |
|
45 | 43 | return x |
|
46 | 44 | def flush(self): |
|
47 | 45 | return "" |
|
48 | 46 | |
|
49 | 47 | bundletypes = { |
|
50 | 48 | "": ("", nocompress), |
|
51 | 49 | "HG10UN": ("HG10UN", nocompress), |
|
52 | 50 | "HG10BZ": ("HG10", lambda: bz2.BZ2Compressor()), |
|
53 | 51 | "HG10GZ": ("HG10GZ", lambda: zlib.compressobj()), |
|
54 | 52 | } |
|
55 | 53 | |
|
56 | 54 | # hgweb uses this list to communicate it's preferred type |
|
57 | 55 | bundlepriority = ['HG10GZ', 'HG10BZ', 'HG10UN'] |
|
58 | 56 | |
|
59 | 57 | def writebundle(cg, filename, bundletype): |
|
60 | 58 | """Write a bundle file and return its filename. |
|
61 | 59 | |
|
62 | 60 | Existing files will not be overwritten. |
|
63 | 61 | If no filename is specified, a temporary file is created. |
|
64 | 62 | bz2 compression can be turned off. |
|
65 | 63 | The bundle file will be deleted in case of errors. |
|
66 | 64 | """ |
|
67 | 65 | |
|
68 | 66 | fh = None |
|
69 | 67 | cleanup = None |
|
70 | 68 | try: |
|
71 | 69 | if filename: |
|
72 | 70 | fh = open(filename, "wb") |
|
73 | 71 | else: |
|
74 | 72 | fd, filename = tempfile.mkstemp(prefix="hg-bundle-", suffix=".hg") |
|
75 | 73 | fh = os.fdopen(fd, "wb") |
|
76 | 74 | cleanup = filename |
|
77 | 75 | |
|
78 | 76 | header, compressor = bundletypes[bundletype] |
|
79 | 77 | fh.write(header) |
|
80 | 78 | z = compressor() |
|
81 | 79 | |
|
82 | 80 | # parse the changegroup data, otherwise we will block |
|
83 | 81 | # in case of sshrepo because we don't know the end of the stream |
|
84 | 82 | |
|
85 | 83 | # an empty chunkiter is the end of the changegroup |
|
86 | 84 | # a changegroup has at least 2 chunkiters (changelog and manifest). |
|
87 | 85 | # after that, an empty chunkiter is the end of the changegroup |
|
88 | 86 | empty = False |
|
89 | 87 | count = 0 |
|
90 | 88 | while not empty or count <= 2: |
|
91 | 89 | empty = True |
|
92 | 90 | count += 1 |
|
93 | 91 | for chunk in chunkiter(cg): |
|
94 | 92 | empty = False |
|
95 | 93 | fh.write(z.compress(chunkheader(len(chunk)))) |
|
96 | 94 | pos = 0 |
|
97 | 95 | while pos < len(chunk): |
|
98 | 96 | next = pos + 2**20 |
|
99 | 97 | fh.write(z.compress(chunk[pos:next])) |
|
100 | 98 | pos = next |
|
101 | 99 | fh.write(z.compress(closechunk())) |
|
102 | 100 | fh.write(z.flush()) |
|
103 | 101 | cleanup = None |
|
104 | 102 | return filename |
|
105 | 103 | finally: |
|
106 | 104 | if fh is not None: |
|
107 | 105 | fh.close() |
|
108 | 106 | if cleanup is not None: |
|
109 | 107 | os.unlink(cleanup) |
|
110 | 108 | |
|
111 | 109 | def unbundle(header, fh): |
|
112 | 110 | if header == 'HG10UN': |
|
113 | 111 | return fh |
|
114 | 112 | elif not header.startswith('HG'): |
|
115 | 113 | # old client with uncompressed bundle |
|
116 | 114 | def generator(f): |
|
117 | 115 | yield header |
|
118 | 116 | for chunk in f: |
|
119 | 117 | yield chunk |
|
120 | 118 | elif header == 'HG10GZ': |
|
121 | 119 | def generator(f): |
|
122 | 120 | zd = zlib.decompressobj() |
|
123 | 121 | for chunk in f: |
|
124 | 122 | yield zd.decompress(chunk) |
|
125 | 123 | elif header == 'HG10BZ': |
|
126 | 124 | def generator(f): |
|
127 | 125 | zd = bz2.BZ2Decompressor() |
|
128 | 126 | zd.decompress("BZ") |
|
129 | 127 | for chunk in util.filechunkiter(f, 4096): |
|
130 | 128 | yield zd.decompress(chunk) |
|
131 | 129 | return util.chunkbuffer(generator(fh)) |
|
132 | 130 | |
|
133 | 131 | def readbundle(fh, fname): |
|
134 | 132 | header = fh.read(6) |
|
135 | 133 | if not header.startswith('HG'): |
|
136 | 134 | raise util.Abort(_('%s: not a Mercurial bundle file') % fname) |
|
137 | 135 | if not header.startswith('HG10'): |
|
138 | 136 | raise util.Abort(_('%s: unknown bundle version') % fname) |
|
139 | 137 | elif header not in bundletypes: |
|
140 | 138 | raise util.Abort(_('%s: unknown bundle compression type') % fname) |
|
141 | 139 | return unbundle(header, fh) |
@@ -1,585 +1,583 b'' | |||
|
1 | """ | |
|
2 | dirstate.py - working directory tracking for mercurial | |
|
3 | ||
|
4 | Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
|
5 | ||
|
6 | This software may be used and distributed according to the terms of the | |
|
7 | GNU General Public License version 2, incorporated herein by reference. | |
|
8 | """ | |
|
1 | # dirstate.py - working directory tracking for mercurial | |
|
2 | # | |
|
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
|
4 | # | |
|
5 | # This software may be used and distributed according to the terms of the | |
|
6 | # GNU General Public License version 2, incorporated herein by reference. | |
|
9 | 7 | |
|
10 | 8 | from node import nullid |
|
11 | 9 | from i18n import _ |
|
12 | 10 | import struct, os, stat, util, errno, ignore |
|
13 | 11 | import cStringIO, osutil, sys, parsers |
|
14 | 12 | |
|
15 | 13 | _unknown = ('?', 0, 0, 0) |
|
16 | 14 | _format = ">cllll" |
|
17 | 15 | |
|
18 | 16 | def _finddirs(path): |
|
19 | 17 | pos = path.rfind('/') |
|
20 | 18 | while pos != -1: |
|
21 | 19 | yield path[:pos] |
|
22 | 20 | pos = path.rfind('/', 0, pos) |
|
23 | 21 | |
|
24 | 22 | def _incdirs(dirs, path): |
|
25 | 23 | for base in _finddirs(path): |
|
26 | 24 | if base in dirs: |
|
27 | 25 | dirs[base] += 1 |
|
28 | 26 | return |
|
29 | 27 | dirs[base] = 1 |
|
30 | 28 | |
|
31 | 29 | def _decdirs(dirs, path): |
|
32 | 30 | for base in _finddirs(path): |
|
33 | 31 | if dirs[base] > 1: |
|
34 | 32 | dirs[base] -= 1 |
|
35 | 33 | return |
|
36 | 34 | del dirs[base] |
|
37 | 35 | |
|
38 | 36 | class dirstate(object): |
|
39 | 37 | |
|
40 | 38 | def __init__(self, opener, ui, root): |
|
41 | 39 | self._opener = opener |
|
42 | 40 | self._root = root |
|
43 | 41 | self._rootdir = os.path.join(root, '') |
|
44 | 42 | self._dirty = False |
|
45 | 43 | self._dirtypl = False |
|
46 | 44 | self._ui = ui |
|
47 | 45 | |
|
48 | 46 | def __getattr__(self, name): |
|
49 | 47 | if name == '_map': |
|
50 | 48 | self._read() |
|
51 | 49 | return self._map |
|
52 | 50 | elif name == '_copymap': |
|
53 | 51 | self._read() |
|
54 | 52 | return self._copymap |
|
55 | 53 | elif name == '_foldmap': |
|
56 | 54 | _foldmap = {} |
|
57 | 55 | for name in self._map: |
|
58 | 56 | norm = os.path.normcase(name) |
|
59 | 57 | _foldmap[norm] = name |
|
60 | 58 | self._foldmap = _foldmap |
|
61 | 59 | return self._foldmap |
|
62 | 60 | elif name == '_branch': |
|
63 | 61 | try: |
|
64 | 62 | self._branch = (self._opener("branch").read().strip() |
|
65 | 63 | or "default") |
|
66 | 64 | except IOError: |
|
67 | 65 | self._branch = "default" |
|
68 | 66 | return self._branch |
|
69 | 67 | elif name == '_pl': |
|
70 | 68 | self._pl = [nullid, nullid] |
|
71 | 69 | try: |
|
72 | 70 | st = self._opener("dirstate").read(40) |
|
73 | 71 | if len(st) == 40: |
|
74 | 72 | self._pl = st[:20], st[20:40] |
|
75 | 73 | except IOError, err: |
|
76 | 74 | if err.errno != errno.ENOENT: raise |
|
77 | 75 | return self._pl |
|
78 | 76 | elif name == '_dirs': |
|
79 | 77 | dirs = {} |
|
80 | 78 | for f,s in self._map.iteritems(): |
|
81 | 79 | if s[0] != 'r': |
|
82 | 80 | _incdirs(dirs, f) |
|
83 | 81 | self._dirs = dirs |
|
84 | 82 | return self._dirs |
|
85 | 83 | elif name == '_ignore': |
|
86 | 84 | files = [self._join('.hgignore')] |
|
87 | 85 | for name, path in self._ui.configitems("ui"): |
|
88 | 86 | if name == 'ignore' or name.startswith('ignore.'): |
|
89 | 87 | files.append(os.path.expanduser(path)) |
|
90 | 88 | self._ignore = ignore.ignore(self._root, files, self._ui.warn) |
|
91 | 89 | return self._ignore |
|
92 | 90 | elif name == '_slash': |
|
93 | 91 | self._slash = self._ui.configbool('ui', 'slash') and os.sep != '/' |
|
94 | 92 | return self._slash |
|
95 | 93 | elif name == '_checklink': |
|
96 | 94 | self._checklink = util.checklink(self._root) |
|
97 | 95 | return self._checklink |
|
98 | 96 | elif name == '_checkexec': |
|
99 | 97 | self._checkexec = util.checkexec(self._root) |
|
100 | 98 | return self._checkexec |
|
101 | 99 | elif name == '_checkcase': |
|
102 | 100 | self._checkcase = not util.checkcase(self._join('.hg')) |
|
103 | 101 | return self._checkcase |
|
104 | 102 | elif name == 'normalize': |
|
105 | 103 | if self._checkcase: |
|
106 | 104 | self.normalize = self._normalize |
|
107 | 105 | else: |
|
108 | 106 | self.normalize = lambda x, y=False: x |
|
109 | 107 | return self.normalize |
|
110 | 108 | else: |
|
111 | 109 | raise AttributeError(name) |
|
112 | 110 | |
|
113 | 111 | def _join(self, f): |
|
114 | 112 | # much faster than os.path.join() |
|
115 | 113 | # it's safe because f is always a relative path |
|
116 | 114 | return self._rootdir + f |
|
117 | 115 | |
|
118 | 116 | def flagfunc(self, fallback): |
|
119 | 117 | if self._checklink: |
|
120 | 118 | if self._checkexec: |
|
121 | 119 | def f(x): |
|
122 | 120 | p = self._join(x) |
|
123 | 121 | if os.path.islink(p): |
|
124 | 122 | return 'l' |
|
125 | 123 | if util.is_exec(p): |
|
126 | 124 | return 'x' |
|
127 | 125 | return '' |
|
128 | 126 | return f |
|
129 | 127 | def f(x): |
|
130 | 128 | if os.path.islink(self._join(x)): |
|
131 | 129 | return 'l' |
|
132 | 130 | if 'x' in fallback(x): |
|
133 | 131 | return 'x' |
|
134 | 132 | return '' |
|
135 | 133 | return f |
|
136 | 134 | if self._checkexec: |
|
137 | 135 | def f(x): |
|
138 | 136 | if 'l' in fallback(x): |
|
139 | 137 | return 'l' |
|
140 | 138 | if util.is_exec(self._join(x)): |
|
141 | 139 | return 'x' |
|
142 | 140 | return '' |
|
143 | 141 | return f |
|
144 | 142 | return fallback |
|
145 | 143 | |
|
146 | 144 | def getcwd(self): |
|
147 | 145 | cwd = os.getcwd() |
|
148 | 146 | if cwd == self._root: return '' |
|
149 | 147 | # self._root ends with a path separator if self._root is '/' or 'C:\' |
|
150 | 148 | rootsep = self._root |
|
151 | 149 | if not util.endswithsep(rootsep): |
|
152 | 150 | rootsep += os.sep |
|
153 | 151 | if cwd.startswith(rootsep): |
|
154 | 152 | return cwd[len(rootsep):] |
|
155 | 153 | else: |
|
156 | 154 | # we're outside the repo. return an absolute path. |
|
157 | 155 | return cwd |
|
158 | 156 | |
|
159 | 157 | def pathto(self, f, cwd=None): |
|
160 | 158 | if cwd is None: |
|
161 | 159 | cwd = self.getcwd() |
|
162 | 160 | path = util.pathto(self._root, cwd, f) |
|
163 | 161 | if self._slash: |
|
164 | 162 | return util.normpath(path) |
|
165 | 163 | return path |
|
166 | 164 | |
|
167 | 165 | def __getitem__(self, key): |
|
168 | 166 | ''' current states: |
|
169 | 167 | n normal |
|
170 | 168 | m needs merging |
|
171 | 169 | r marked for removal |
|
172 | 170 | a marked for addition |
|
173 | 171 | ? not tracked''' |
|
174 | 172 | return self._map.get(key, ("?",))[0] |
|
175 | 173 | |
|
176 | 174 | def __contains__(self, key): |
|
177 | 175 | return key in self._map |
|
178 | 176 | |
|
179 | 177 | def __iter__(self): |
|
180 | 178 | for x in sorted(self._map): |
|
181 | 179 | yield x |
|
182 | 180 | |
|
183 | 181 | def parents(self): |
|
184 | 182 | return self._pl |
|
185 | 183 | |
|
186 | 184 | def branch(self): |
|
187 | 185 | return self._branch |
|
188 | 186 | |
|
189 | 187 | def setparents(self, p1, p2=nullid): |
|
190 | 188 | self._dirty = self._dirtypl = True |
|
191 | 189 | self._pl = p1, p2 |
|
192 | 190 | |
|
193 | 191 | def setbranch(self, branch): |
|
194 | 192 | self._branch = branch |
|
195 | 193 | self._opener("branch", "w").write(branch + '\n') |
|
196 | 194 | |
|
197 | 195 | def _read(self): |
|
198 | 196 | self._map = {} |
|
199 | 197 | self._copymap = {} |
|
200 | 198 | try: |
|
201 | 199 | st = self._opener("dirstate").read() |
|
202 | 200 | except IOError, err: |
|
203 | 201 | if err.errno != errno.ENOENT: raise |
|
204 | 202 | return |
|
205 | 203 | if not st: |
|
206 | 204 | return |
|
207 | 205 | |
|
208 | 206 | p = parsers.parse_dirstate(self._map, self._copymap, st) |
|
209 | 207 | if not self._dirtypl: |
|
210 | 208 | self._pl = p |
|
211 | 209 | |
|
212 | 210 | def invalidate(self): |
|
213 | 211 | for a in "_map _copymap _foldmap _branch _pl _dirs _ignore".split(): |
|
214 | 212 | if a in self.__dict__: |
|
215 | 213 | delattr(self, a) |
|
216 | 214 | self._dirty = False |
|
217 | 215 | |
|
218 | 216 | def copy(self, source, dest): |
|
219 | 217 | """Mark dest as a copy of source. Unmark dest if source is None. |
|
220 | 218 | """ |
|
221 | 219 | if source == dest: |
|
222 | 220 | return |
|
223 | 221 | self._dirty = True |
|
224 | 222 | if source is not None: |
|
225 | 223 | self._copymap[dest] = source |
|
226 | 224 | elif dest in self._copymap: |
|
227 | 225 | del self._copymap[dest] |
|
228 | 226 | |
|
229 | 227 | def copied(self, file): |
|
230 | 228 | return self._copymap.get(file, None) |
|
231 | 229 | |
|
232 | 230 | def copies(self): |
|
233 | 231 | return self._copymap |
|
234 | 232 | |
|
235 | 233 | def _droppath(self, f): |
|
236 | 234 | if self[f] not in "?r" and "_dirs" in self.__dict__: |
|
237 | 235 | _decdirs(self._dirs, f) |
|
238 | 236 | |
|
239 | 237 | def _addpath(self, f, check=False): |
|
240 | 238 | oldstate = self[f] |
|
241 | 239 | if check or oldstate == "r": |
|
242 | 240 | if '\r' in f or '\n' in f: |
|
243 | 241 | raise util.Abort( |
|
244 | 242 | _("'\\n' and '\\r' disallowed in filenames: %r") % f) |
|
245 | 243 | if f in self._dirs: |
|
246 | 244 | raise util.Abort(_('directory %r already in dirstate') % f) |
|
247 | 245 | # shadows |
|
248 | 246 | for d in _finddirs(f): |
|
249 | 247 | if d in self._dirs: |
|
250 | 248 | break |
|
251 | 249 | if d in self._map and self[d] != 'r': |
|
252 | 250 | raise util.Abort( |
|
253 | 251 | _('file %r in dirstate clashes with %r') % (d, f)) |
|
254 | 252 | if oldstate in "?r" and "_dirs" in self.__dict__: |
|
255 | 253 | _incdirs(self._dirs, f) |
|
256 | 254 | |
|
257 | 255 | def normal(self, f): |
|
258 | 256 | 'mark a file normal and clean' |
|
259 | 257 | self._dirty = True |
|
260 | 258 | self._addpath(f) |
|
261 | 259 | s = os.lstat(self._join(f)) |
|
262 | 260 | self._map[f] = ('n', s.st_mode, s.st_size, int(s.st_mtime)) |
|
263 | 261 | if f in self._copymap: |
|
264 | 262 | del self._copymap[f] |
|
265 | 263 | |
|
266 | 264 | def normallookup(self, f): |
|
267 | 265 | 'mark a file normal, but possibly dirty' |
|
268 | 266 | if self._pl[1] != nullid and f in self._map: |
|
269 | 267 | # if there is a merge going on and the file was either |
|
270 | 268 | # in state 'm' or dirty before being removed, restore that state. |
|
271 | 269 | entry = self._map[f] |
|
272 | 270 | if entry[0] == 'r' and entry[2] in (-1, -2): |
|
273 | 271 | source = self._copymap.get(f) |
|
274 | 272 | if entry[2] == -1: |
|
275 | 273 | self.merge(f) |
|
276 | 274 | elif entry[2] == -2: |
|
277 | 275 | self.normaldirty(f) |
|
278 | 276 | if source: |
|
279 | 277 | self.copy(source, f) |
|
280 | 278 | return |
|
281 | 279 | if entry[0] == 'm' or entry[0] == 'n' and entry[2] == -2: |
|
282 | 280 | return |
|
283 | 281 | self._dirty = True |
|
284 | 282 | self._addpath(f) |
|
285 | 283 | self._map[f] = ('n', 0, -1, -1) |
|
286 | 284 | if f in self._copymap: |
|
287 | 285 | del self._copymap[f] |
|
288 | 286 | |
|
289 | 287 | def normaldirty(self, f): |
|
290 | 288 | 'mark a file normal, but dirty' |
|
291 | 289 | self._dirty = True |
|
292 | 290 | self._addpath(f) |
|
293 | 291 | self._map[f] = ('n', 0, -2, -1) |
|
294 | 292 | if f in self._copymap: |
|
295 | 293 | del self._copymap[f] |
|
296 | 294 | |
|
297 | 295 | def add(self, f): |
|
298 | 296 | 'mark a file added' |
|
299 | 297 | self._dirty = True |
|
300 | 298 | self._addpath(f, True) |
|
301 | 299 | self._map[f] = ('a', 0, -1, -1) |
|
302 | 300 | if f in self._copymap: |
|
303 | 301 | del self._copymap[f] |
|
304 | 302 | |
|
305 | 303 | def remove(self, f): |
|
306 | 304 | 'mark a file removed' |
|
307 | 305 | self._dirty = True |
|
308 | 306 | self._droppath(f) |
|
309 | 307 | size = 0 |
|
310 | 308 | if self._pl[1] != nullid and f in self._map: |
|
311 | 309 | entry = self._map[f] |
|
312 | 310 | if entry[0] == 'm': |
|
313 | 311 | size = -1 |
|
314 | 312 | elif entry[0] == 'n' and entry[2] == -2: |
|
315 | 313 | size = -2 |
|
316 | 314 | self._map[f] = ('r', 0, size, 0) |
|
317 | 315 | if size == 0 and f in self._copymap: |
|
318 | 316 | del self._copymap[f] |
|
319 | 317 | |
|
320 | 318 | def merge(self, f): |
|
321 | 319 | 'mark a file merged' |
|
322 | 320 | self._dirty = True |
|
323 | 321 | s = os.lstat(self._join(f)) |
|
324 | 322 | self._addpath(f) |
|
325 | 323 | self._map[f] = ('m', s.st_mode, s.st_size, int(s.st_mtime)) |
|
326 | 324 | if f in self._copymap: |
|
327 | 325 | del self._copymap[f] |
|
328 | 326 | |
|
329 | 327 | def forget(self, f): |
|
330 | 328 | 'forget a file' |
|
331 | 329 | self._dirty = True |
|
332 | 330 | try: |
|
333 | 331 | self._droppath(f) |
|
334 | 332 | del self._map[f] |
|
335 | 333 | except KeyError: |
|
336 | 334 | self._ui.warn(_("not in dirstate: %s\n") % f) |
|
337 | 335 | |
|
338 | 336 | def _normalize(self, path, knownpath=False): |
|
339 | 337 | norm_path = os.path.normcase(path) |
|
340 | 338 | fold_path = self._foldmap.get(norm_path, None) |
|
341 | 339 | if fold_path is None: |
|
342 | 340 | if knownpath or not os.path.exists(os.path.join(self._root, path)): |
|
343 | 341 | fold_path = path |
|
344 | 342 | else: |
|
345 | 343 | fold_path = self._foldmap.setdefault(norm_path, |
|
346 | 344 | util.fspath(path, self._root)) |
|
347 | 345 | return fold_path |
|
348 | 346 | |
|
349 | 347 | def clear(self): |
|
350 | 348 | self._map = {} |
|
351 | 349 | if "_dirs" in self.__dict__: |
|
352 | 350 | delattr(self, "_dirs"); |
|
353 | 351 | self._copymap = {} |
|
354 | 352 | self._pl = [nullid, nullid] |
|
355 | 353 | self._dirty = True |
|
356 | 354 | |
|
357 | 355 | def rebuild(self, parent, files): |
|
358 | 356 | self.clear() |
|
359 | 357 | for f in files: |
|
360 | 358 | if 'x' in files.flags(f): |
|
361 | 359 | self._map[f] = ('n', 0777, -1, 0) |
|
362 | 360 | else: |
|
363 | 361 | self._map[f] = ('n', 0666, -1, 0) |
|
364 | 362 | self._pl = (parent, nullid) |
|
365 | 363 | self._dirty = True |
|
366 | 364 | |
|
367 | 365 | def write(self): |
|
368 | 366 | if not self._dirty: |
|
369 | 367 | return |
|
370 | 368 | st = self._opener("dirstate", "w", atomictemp=True) |
|
371 | 369 | |
|
372 | 370 | try: |
|
373 | 371 | gran = int(self._ui.config('dirstate', 'granularity', 1)) |
|
374 | 372 | except ValueError: |
|
375 | 373 | gran = 1 |
|
376 | 374 | limit = sys.maxint |
|
377 | 375 | if gran > 0: |
|
378 | 376 | limit = util.fstat(st).st_mtime - gran |
|
379 | 377 | |
|
380 | 378 | cs = cStringIO.StringIO() |
|
381 | 379 | copymap = self._copymap |
|
382 | 380 | pack = struct.pack |
|
383 | 381 | write = cs.write |
|
384 | 382 | write("".join(self._pl)) |
|
385 | 383 | for f, e in self._map.iteritems(): |
|
386 | 384 | if f in copymap: |
|
387 | 385 | f = "%s\0%s" % (f, copymap[f]) |
|
388 | 386 | if e[3] > limit and e[0] == 'n': |
|
389 | 387 | e = (e[0], 0, -1, -1) |
|
390 | 388 | e = pack(_format, e[0], e[1], e[2], e[3], len(f)) |
|
391 | 389 | write(e) |
|
392 | 390 | write(f) |
|
393 | 391 | st.write(cs.getvalue()) |
|
394 | 392 | st.rename() |
|
395 | 393 | self._dirty = self._dirtypl = False |
|
396 | 394 | |
|
397 | 395 | def _dirignore(self, f): |
|
398 | 396 | if f == '.': |
|
399 | 397 | return False |
|
400 | 398 | if self._ignore(f): |
|
401 | 399 | return True |
|
402 | 400 | for p in _finddirs(f): |
|
403 | 401 | if self._ignore(p): |
|
404 | 402 | return True |
|
405 | 403 | return False |
|
406 | 404 | |
|
407 | 405 | def walk(self, match, unknown, ignored): |
|
408 | 406 | ''' |
|
409 | 407 | walk recursively through the directory tree, finding all files |
|
410 | 408 | matched by the match function |
|
411 | 409 | |
|
412 | 410 | results are yielded in a tuple (filename, stat), where stat |
|
413 | 411 | and st is the stat result if the file was found in the directory. |
|
414 | 412 | ''' |
|
415 | 413 | |
|
416 | 414 | def fwarn(f, msg): |
|
417 | 415 | self._ui.warn('%s: %s\n' % (self.pathto(f), msg)) |
|
418 | 416 | return False |
|
419 | 417 | badfn = fwarn |
|
420 | 418 | if hasattr(match, 'bad'): |
|
421 | 419 | badfn = match.bad |
|
422 | 420 | |
|
423 | 421 | def badtype(f, mode): |
|
424 | 422 | kind = 'unknown' |
|
425 | 423 | if stat.S_ISCHR(mode): kind = _('character device') |
|
426 | 424 | elif stat.S_ISBLK(mode): kind = _('block device') |
|
427 | 425 | elif stat.S_ISFIFO(mode): kind = _('fifo') |
|
428 | 426 | elif stat.S_ISSOCK(mode): kind = _('socket') |
|
429 | 427 | elif stat.S_ISDIR(mode): kind = _('directory') |
|
430 | 428 | self._ui.warn(_('%s: unsupported file type (type is %s)\n') |
|
431 | 429 | % (self.pathto(f), kind)) |
|
432 | 430 | |
|
433 | 431 | ignore = self._ignore |
|
434 | 432 | dirignore = self._dirignore |
|
435 | 433 | if ignored: |
|
436 | 434 | ignore = util.never |
|
437 | 435 | dirignore = util.never |
|
438 | 436 | elif not unknown: |
|
439 | 437 | # if unknown and ignored are False, skip step 2 |
|
440 | 438 | ignore = util.always |
|
441 | 439 | dirignore = util.always |
|
442 | 440 | |
|
443 | 441 | matchfn = match.matchfn |
|
444 | 442 | dmap = self._map |
|
445 | 443 | normpath = util.normpath |
|
446 | 444 | normalize = self.normalize |
|
447 | 445 | listdir = osutil.listdir |
|
448 | 446 | lstat = os.lstat |
|
449 | 447 | getkind = stat.S_IFMT |
|
450 | 448 | dirkind = stat.S_IFDIR |
|
451 | 449 | regkind = stat.S_IFREG |
|
452 | 450 | lnkkind = stat.S_IFLNK |
|
453 | 451 | join = self._join |
|
454 | 452 | work = [] |
|
455 | 453 | wadd = work.append |
|
456 | 454 | |
|
457 | 455 | files = set(match.files()) |
|
458 | 456 | if not files or '.' in files: |
|
459 | 457 | files = [''] |
|
460 | 458 | results = {'.hg': None} |
|
461 | 459 | |
|
462 | 460 | # step 1: find all explicit files |
|
463 | 461 | for ff in sorted(files): |
|
464 | 462 | nf = normalize(normpath(ff)) |
|
465 | 463 | if nf in results: |
|
466 | 464 | continue |
|
467 | 465 | |
|
468 | 466 | try: |
|
469 | 467 | st = lstat(join(nf)) |
|
470 | 468 | kind = getkind(st.st_mode) |
|
471 | 469 | if kind == dirkind: |
|
472 | 470 | if not dirignore(nf): |
|
473 | 471 | wadd(nf) |
|
474 | 472 | elif kind == regkind or kind == lnkkind: |
|
475 | 473 | results[nf] = st |
|
476 | 474 | else: |
|
477 | 475 | badtype(ff, kind) |
|
478 | 476 | if nf in dmap: |
|
479 | 477 | results[nf] = None |
|
480 | 478 | except OSError, inst: |
|
481 | 479 | keep = False |
|
482 | 480 | prefix = nf + "/" |
|
483 | 481 | for fn in dmap: |
|
484 | 482 | if nf == fn or fn.startswith(prefix): |
|
485 | 483 | keep = True |
|
486 | 484 | break |
|
487 | 485 | if not keep: |
|
488 | 486 | if inst.errno != errno.ENOENT: |
|
489 | 487 | fwarn(ff, inst.strerror) |
|
490 | 488 | elif badfn(ff, inst.strerror): |
|
491 | 489 | if (nf in dmap or not ignore(nf)) and matchfn(nf): |
|
492 | 490 | results[nf] = None |
|
493 | 491 | |
|
494 | 492 | # step 2: visit subdirectories |
|
495 | 493 | while work: |
|
496 | 494 | nd = work.pop() |
|
497 | 495 | if hasattr(match, 'dir'): |
|
498 | 496 | match.dir(nd) |
|
499 | 497 | skip = None |
|
500 | 498 | if nd == '.': |
|
501 | 499 | nd = '' |
|
502 | 500 | else: |
|
503 | 501 | skip = '.hg' |
|
504 | 502 | try: |
|
505 | 503 | entries = listdir(join(nd), stat=True, skip=skip) |
|
506 | 504 | except OSError, inst: |
|
507 | 505 | if inst.errno == errno.EACCES: |
|
508 | 506 | fwarn(nd, inst.strerror) |
|
509 | 507 | continue |
|
510 | 508 | raise |
|
511 | 509 | for f, kind, st in entries: |
|
512 | 510 | nf = normalize(nd and (nd + "/" + f) or f, True) |
|
513 | 511 | if nf not in results: |
|
514 | 512 | if kind == dirkind: |
|
515 | 513 | if not ignore(nf): |
|
516 | 514 | wadd(nf) |
|
517 | 515 | if nf in dmap and matchfn(nf): |
|
518 | 516 | results[nf] = None |
|
519 | 517 | elif kind == regkind or kind == lnkkind: |
|
520 | 518 | if nf in dmap: |
|
521 | 519 | if matchfn(nf): |
|
522 | 520 | results[nf] = st |
|
523 | 521 | elif matchfn(nf) and not ignore(nf): |
|
524 | 522 | results[nf] = st |
|
525 | 523 | elif nf in dmap and matchfn(nf): |
|
526 | 524 | results[nf] = None |
|
527 | 525 | |
|
528 | 526 | # step 3: report unseen items in the dmap hash |
|
529 | 527 | visit = sorted([f for f in dmap if f not in results and match(f)]) |
|
530 | 528 | for nf, st in zip(visit, util.statfiles([join(i) for i in visit])): |
|
531 | 529 | if not st is None and not getkind(st.st_mode) in (regkind, lnkkind): |
|
532 | 530 | st = None |
|
533 | 531 | results[nf] = st |
|
534 | 532 | |
|
535 | 533 | del results['.hg'] |
|
536 | 534 | return results |
|
537 | 535 | |
|
538 | 536 | def status(self, match, ignored, clean, unknown): |
|
539 | 537 | listignored, listclean, listunknown = ignored, clean, unknown |
|
540 | 538 | lookup, modified, added, unknown, ignored = [], [], [], [], [] |
|
541 | 539 | removed, deleted, clean = [], [], [] |
|
542 | 540 | |
|
543 | 541 | dmap = self._map |
|
544 | 542 | ladd = lookup.append |
|
545 | 543 | madd = modified.append |
|
546 | 544 | aadd = added.append |
|
547 | 545 | uadd = unknown.append |
|
548 | 546 | iadd = ignored.append |
|
549 | 547 | radd = removed.append |
|
550 | 548 | dadd = deleted.append |
|
551 | 549 | cadd = clean.append |
|
552 | 550 | |
|
553 | 551 | for fn, st in self.walk(match, listunknown, listignored).iteritems(): |
|
554 | 552 | if fn not in dmap: |
|
555 | 553 | if (listignored or match.exact(fn)) and self._dirignore(fn): |
|
556 | 554 | if listignored: |
|
557 | 555 | iadd(fn) |
|
558 | 556 | elif listunknown: |
|
559 | 557 | uadd(fn) |
|
560 | 558 | continue |
|
561 | 559 | |
|
562 | 560 | state, mode, size, time = dmap[fn] |
|
563 | 561 | |
|
564 | 562 | if not st and state in "nma": |
|
565 | 563 | dadd(fn) |
|
566 | 564 | elif state == 'n': |
|
567 | 565 | if (size >= 0 and |
|
568 | 566 | (size != st.st_size |
|
569 | 567 | or ((mode ^ st.st_mode) & 0100 and self._checkexec)) |
|
570 | 568 | or size == -2 |
|
571 | 569 | or fn in self._copymap): |
|
572 | 570 | madd(fn) |
|
573 | 571 | elif time != int(st.st_mtime): |
|
574 | 572 | ladd(fn) |
|
575 | 573 | elif listclean: |
|
576 | 574 | cadd(fn) |
|
577 | 575 | elif state == 'm': |
|
578 | 576 | madd(fn) |
|
579 | 577 | elif state == 'a': |
|
580 | 578 | aadd(fn) |
|
581 | 579 | elif state == 'r': |
|
582 | 580 | radd(fn) |
|
583 | 581 | |
|
584 | 582 | return (lookup, modified, added, removed, deleted, unknown, ignored, |
|
585 | 583 | clean) |
@@ -1,76 +1,74 b'' | |||
|
1 | """ | |
|
2 | encoding.py - character transcoding support for Mercurial | |
|
3 | ||
|
4 | Copyright 2005-2009 Matt Mackall <mpm@selenic.com> and others | |
|
5 | ||
|
6 | This software may be used and distributed according to the terms of the | |
|
7 | GNU General Public License version 2, incorporated herein by reference. | |
|
8 | """ | |
|
1 | # encoding.py - character transcoding support for Mercurial | |
|
2 | # | |
|
3 | # Copyright 2005-2009 Matt Mackall <mpm@selenic.com> and others | |
|
4 | # | |
|
5 | # This software may be used and distributed according to the terms of the | |
|
6 | # GNU General Public License version 2, incorporated herein by reference. | |
|
9 | 7 | |
|
10 | 8 | import sys, unicodedata, locale, os, error |
|
11 | 9 | |
|
12 | 10 | _encodingfixup = {'646': 'ascii', 'ANSI_X3.4-1968': 'ascii'} |
|
13 | 11 | |
|
14 | 12 | try: |
|
15 | 13 | encoding = os.environ.get("HGENCODING") |
|
16 | 14 | if sys.platform == 'darwin' and not encoding: |
|
17 | 15 | # On darwin, getpreferredencoding ignores the locale environment and |
|
18 | 16 | # always returns mac-roman. We override this if the environment is |
|
19 | 17 | # not C (has been customized by the user). |
|
20 | 18 | locale.setlocale(locale.LC_CTYPE, '') |
|
21 | 19 | encoding = locale.getlocale()[1] |
|
22 | 20 | if not encoding: |
|
23 | 21 | encoding = locale.getpreferredencoding() or 'ascii' |
|
24 | 22 | encoding = _encodingfixup.get(encoding, encoding) |
|
25 | 23 | except locale.Error: |
|
26 | 24 | encoding = 'ascii' |
|
27 | 25 | encodingmode = os.environ.get("HGENCODINGMODE", "strict") |
|
28 | 26 | fallbackencoding = 'ISO-8859-1' |
|
29 | 27 | |
|
30 | 28 | def tolocal(s): |
|
31 | 29 | """ |
|
32 | 30 | Convert a string from internal UTF-8 to local encoding |
|
33 | 31 | |
|
34 | 32 | All internal strings should be UTF-8 but some repos before the |
|
35 | 33 | implementation of locale support may contain latin1 or possibly |
|
36 | 34 | other character sets. We attempt to decode everything strictly |
|
37 | 35 | using UTF-8, then Latin-1, and failing that, we use UTF-8 and |
|
38 | 36 | replace unknown characters. |
|
39 | 37 | """ |
|
40 | 38 | for e in ('UTF-8', fallbackencoding): |
|
41 | 39 | try: |
|
42 | 40 | u = s.decode(e) # attempt strict decoding |
|
43 | 41 | return u.encode(encoding, "replace") |
|
44 | 42 | except LookupError, k: |
|
45 | 43 | raise error.Abort("%s, please check your locale settings" % k) |
|
46 | 44 | except UnicodeDecodeError: |
|
47 | 45 | pass |
|
48 | 46 | u = s.decode("utf-8", "replace") # last ditch |
|
49 | 47 | return u.encode(encoding, "replace") |
|
50 | 48 | |
|
51 | 49 | def fromlocal(s): |
|
52 | 50 | """ |
|
53 | 51 | Convert a string from the local character encoding to UTF-8 |
|
54 | 52 | |
|
55 | 53 | We attempt to decode strings using the encoding mode set by |
|
56 | 54 | HGENCODINGMODE, which defaults to 'strict'. In this mode, unknown |
|
57 | 55 | characters will cause an error message. Other modes include |
|
58 | 56 | 'replace', which replaces unknown characters with a special |
|
59 | 57 | Unicode character, and 'ignore', which drops the character. |
|
60 | 58 | """ |
|
61 | 59 | try: |
|
62 | 60 | return s.decode(encoding, encodingmode).encode("utf-8") |
|
63 | 61 | except UnicodeDecodeError, inst: |
|
64 | 62 | sub = s[max(0, inst.start-10):inst.start+10] |
|
65 | 63 | raise error.Abort("decoding near '%s': %s!" % (sub, inst)) |
|
66 | 64 | except LookupError, k: |
|
67 | 65 | raise error.Abort("%s, please check your locale settings" % k) |
|
68 | 66 | |
|
69 | 67 | def colwidth(s): |
|
70 | 68 | "Find the column width of a UTF-8 string for display" |
|
71 | 69 | d = s.decode(encoding, 'replace') |
|
72 | 70 | if hasattr(unicodedata, 'east_asian_width'): |
|
73 | 71 | w = unicodedata.east_asian_width |
|
74 | 72 | return sum([w(c) in 'WF' and 2 or 1 for c in d]) |
|
75 | 73 | return len(d) |
|
76 | 74 |
@@ -1,70 +1,68 b'' | |||
|
1 | """ | |
|
2 | error.py - Mercurial exceptions | |
|
3 | ||
|
4 | This allows us to catch exceptions at higher levels without forcing imports | |
|
5 | ||
|
6 | Copyright 2005-2008 Matt Mackall <mpm@selenic.com> | |
|
7 | ||
|
8 | This software may be used and distributed according to the terms of the | |
|
9 | GNU General Public License version 2, incorporated herein by reference. | |
|
10 | """ | |
|
1 | # error.py - Mercurial exceptions | |
|
2 | # | |
|
3 | # This allows us to catch exceptions at higher levels without forcing imports | |
|
4 | # | |
|
5 | # Copyright 2005-2008 Matt Mackall <mpm@selenic.com> | |
|
6 | # | |
|
7 | # This software may be used and distributed according to the terms of the | |
|
8 | # GNU General Public License version 2, incorporated herein by reference. | |
|
11 | 9 | |
|
12 | 10 | # Do not import anything here, please |
|
13 | 11 | |
|
14 | 12 | class RevlogError(Exception): |
|
15 | 13 | pass |
|
16 | 14 | |
|
17 | 15 | class LookupError(RevlogError, KeyError): |
|
18 | 16 | def __init__(self, name, index, message): |
|
19 | 17 | self.name = name |
|
20 | 18 | if isinstance(name, str) and len(name) == 20: |
|
21 | 19 | from node import short |
|
22 | 20 | name = short(name) |
|
23 | 21 | RevlogError.__init__(self, '%s@%s: %s' % (index, name, message)) |
|
24 | 22 | |
|
25 | 23 | def __str__(self): |
|
26 | 24 | return RevlogError.__str__(self) |
|
27 | 25 | |
|
28 | 26 | class ParseError(Exception): |
|
29 | 27 | """Exception raised on errors in parsing the command line.""" |
|
30 | 28 | |
|
31 | 29 | class ConfigError(Exception): |
|
32 | 30 | 'Exception raised when parsing config files' |
|
33 | 31 | |
|
34 | 32 | class RepoError(Exception): |
|
35 | 33 | pass |
|
36 | 34 | |
|
37 | 35 | class CapabilityError(RepoError): |
|
38 | 36 | pass |
|
39 | 37 | |
|
40 | 38 | class LockError(IOError): |
|
41 | 39 | def __init__(self, errno, strerror, filename, desc): |
|
42 | 40 | IOError.__init__(self, errno, strerror, filename) |
|
43 | 41 | self.desc = desc |
|
44 | 42 | |
|
45 | 43 | class LockHeld(LockError): |
|
46 | 44 | def __init__(self, errno, filename, desc, locker): |
|
47 | 45 | LockError.__init__(self, errno, 'Lock held', filename, desc) |
|
48 | 46 | self.locker = locker |
|
49 | 47 | |
|
50 | 48 | class LockUnavailable(LockError): |
|
51 | 49 | pass |
|
52 | 50 | |
|
53 | 51 | class ResponseError(Exception): |
|
54 | 52 | """Raised to print an error with part of output and exit.""" |
|
55 | 53 | |
|
56 | 54 | class UnknownCommand(Exception): |
|
57 | 55 | """Exception raised if command is not in the command table.""" |
|
58 | 56 | |
|
59 | 57 | class AmbiguousCommand(Exception): |
|
60 | 58 | """Exception raised if command shortcut matches more than one command.""" |
|
61 | 59 | |
|
62 | 60 | # derived from KeyboardInterrupt to simplify some breakout code |
|
63 | 61 | class SignalInterrupt(KeyboardInterrupt): |
|
64 | 62 | """Exception raised on SIGTERM and SIGHUP.""" |
|
65 | 63 | |
|
66 | 64 | class SignatureError(Exception): |
|
67 | 65 | pass |
|
68 | 66 | |
|
69 | 67 | class Abort(Exception): |
|
70 | 68 | """Raised if a command needs to print an error and exit.""" |
@@ -1,49 +1,47 b'' | |||
|
1 | """ | |
|
2 | i18n.py - internationalization support for mercurial | |
|
3 | ||
|
4 | Copyright 2005, 2006 Matt Mackall <mpm@selenic.com> | |
|
5 | ||
|
6 | This software may be used and distributed according to the terms of the | |
|
7 | GNU General Public License version 2, incorporated herein by reference. | |
|
8 | """ | |
|
1 | # i18n.py - internationalization support for mercurial | |
|
2 | # | |
|
3 | # Copyright 2005, 2006 Matt Mackall <mpm@selenic.com> | |
|
4 | # | |
|
5 | # This software may be used and distributed according to the terms of the | |
|
6 | # GNU General Public License version 2, incorporated herein by reference. | |
|
9 | 7 | |
|
10 | 8 | import gettext, sys, os, encoding |
|
11 | 9 | |
|
12 | 10 | # modelled after templater.templatepath: |
|
13 | 11 | if hasattr(sys, 'frozen'): |
|
14 | 12 | module = sys.executable |
|
15 | 13 | else: |
|
16 | 14 | module = __file__ |
|
17 | 15 | |
|
18 | 16 | base = os.path.dirname(module) |
|
19 | 17 | for dir in ('.', '..'): |
|
20 | 18 | localedir = os.path.normpath(os.path.join(base, dir, 'locale')) |
|
21 | 19 | if os.path.isdir(localedir): |
|
22 | 20 | break |
|
23 | 21 | |
|
24 | 22 | t = gettext.translation('hg', localedir, fallback=True) |
|
25 | 23 | |
|
26 | 24 | def gettext(message): |
|
27 | 25 | """Translate message. |
|
28 | 26 | |
|
29 | 27 | The message is looked up in the catalog to get a Unicode string, |
|
30 | 28 | which is encoded in the local encoding before being returned. |
|
31 | 29 | |
|
32 | 30 | Important: message is restricted to characters in the encoding |
|
33 | 31 | given by sys.getdefaultencoding() which is most likely 'ascii'. |
|
34 | 32 | """ |
|
35 | 33 | # If message is None, t.ugettext will return u'None' as the |
|
36 | 34 | # translation whereas our callers expect us to return None. |
|
37 | 35 | if message is None: |
|
38 | 36 | return message |
|
39 | 37 | |
|
40 | 38 | # We cannot just run the text through encoding.tolocal since that |
|
41 | 39 | # leads to infinite recursion when encoding._encoding is invalid. |
|
42 | 40 | try: |
|
43 | 41 | u = t.ugettext(message) |
|
44 | 42 | return u.encode(encoding.encoding, "replace") |
|
45 | 43 | except LookupError: |
|
46 | 44 | return message |
|
47 | 45 | |
|
48 | 46 | _ = gettext |
|
49 | 47 |
@@ -1,20 +1,18 b'' | |||
|
1 | """ | |
|
2 | node.py - basic nodeid manipulation for mercurial | |
|
3 | ||
|
4 | Copyright 2005, 2006 Matt Mackall <mpm@selenic.com> | |
|
5 | ||
|
6 | This software may be used and distributed according to the terms of the | |
|
7 | GNU General Public License version 2, incorporated herein by reference. | |
|
8 | """ | |
|
1 | # node.py - basic nodeid manipulation for mercurial | |
|
2 | # | |
|
3 | # Copyright 2005, 2006 Matt Mackall <mpm@selenic.com> | |
|
4 | # | |
|
5 | # This software may be used and distributed according to the terms of the | |
|
6 | # GNU General Public License version 2, incorporated herein by reference. | |
|
9 | 7 | |
|
10 | 8 | import binascii |
|
11 | 9 | |
|
12 | 10 | nullrev = -1 |
|
13 | 11 | nullid = "\0" * 20 |
|
14 | 12 | |
|
15 | 13 | # This ugly style has a noticeable effect in manifest parsing |
|
16 | 14 | hex = binascii.hexlify |
|
17 | 15 | bin = binascii.unhexlify |
|
18 | 16 | |
|
19 | 17 | def short(node): |
|
20 | 18 | return hex(node[:6]) |
@@ -1,219 +1,217 b'' | |||
|
1 | """ | |
|
2 | posix.py - Posix utility function implementations for Mercurial | |
|
3 | ||
|
4 | Copyright 2005-2009 Matt Mackall <mpm@selenic.com> and others | |
|
5 | ||
|
6 | This software may be used and distributed according to the terms of the | |
|
7 | GNU General Public License version 2, incorporated herein by reference. | |
|
8 | """ | |
|
1 | # posix.py - Posix utility function implementations for Mercurial | |
|
2 | # | |
|
3 | # Copyright 2005-2009 Matt Mackall <mpm@selenic.com> and others | |
|
4 | # | |
|
5 | # This software may be used and distributed according to the terms of the | |
|
6 | # GNU General Public License version 2, incorporated herein by reference. | |
|
9 | 7 | |
|
10 | 8 | from i18n import _ |
|
11 | 9 | import os, sys, osutil, errno, stat, getpass, pwd, grp |
|
12 | 10 | |
|
13 | 11 | posixfile = file |
|
14 | 12 | nulldev = '/dev/null' |
|
15 | 13 | normpath = os.path.normpath |
|
16 | 14 | samestat = os.path.samestat |
|
17 | 15 | |
|
18 | 16 | umask = os.umask(0) |
|
19 | 17 | os.umask(umask) |
|
20 | 18 | |
|
21 | 19 | def openhardlinks(): |
|
22 | 20 | '''return true if it is safe to hold open file handles to hardlinks''' |
|
23 | 21 | return True |
|
24 | 22 | |
|
25 | 23 | def rcfiles(path): |
|
26 | 24 | rcs = [os.path.join(path, 'hgrc')] |
|
27 | 25 | rcdir = os.path.join(path, 'hgrc.d') |
|
28 | 26 | try: |
|
29 | 27 | rcs.extend([os.path.join(rcdir, f) |
|
30 | 28 | for f, kind in osutil.listdir(rcdir) |
|
31 | 29 | if f.endswith(".rc")]) |
|
32 | 30 | except OSError: |
|
33 | 31 | pass |
|
34 | 32 | return rcs |
|
35 | 33 | |
|
36 | 34 | def system_rcpath(): |
|
37 | 35 | path = [] |
|
38 | 36 | # old mod_python does not set sys.argv |
|
39 | 37 | if len(getattr(sys, 'argv', [])) > 0: |
|
40 | 38 | path.extend(rcfiles(os.path.dirname(sys.argv[0]) + |
|
41 | 39 | '/../etc/mercurial')) |
|
42 | 40 | path.extend(rcfiles('/etc/mercurial')) |
|
43 | 41 | return path |
|
44 | 42 | |
|
45 | 43 | def user_rcpath(): |
|
46 | 44 | return [os.path.expanduser('~/.hgrc')] |
|
47 | 45 | |
|
48 | 46 | def parse_patch_output(output_line): |
|
49 | 47 | """parses the output produced by patch and returns the file name""" |
|
50 | 48 | pf = output_line[14:] |
|
51 | 49 | if os.sys.platform == 'OpenVMS': |
|
52 | 50 | if pf[0] == '`': |
|
53 | 51 | pf = pf[1:-1] # Remove the quotes |
|
54 | 52 | else: |
|
55 | 53 | if pf.startswith("'") and pf.endswith("'") and " " in pf: |
|
56 | 54 | pf = pf[1:-1] # Remove the quotes |
|
57 | 55 | return pf |
|
58 | 56 | |
|
59 | 57 | def sshargs(sshcmd, host, user, port): |
|
60 | 58 | '''Build argument list for ssh''' |
|
61 | 59 | args = user and ("%s@%s" % (user, host)) or host |
|
62 | 60 | return port and ("%s -p %s" % (args, port)) or args |
|
63 | 61 | |
|
64 | 62 | def is_exec(f): |
|
65 | 63 | """check whether a file is executable""" |
|
66 | 64 | return (os.lstat(f).st_mode & 0100 != 0) |
|
67 | 65 | |
|
68 | 66 | def set_flags(f, l, x): |
|
69 | 67 | s = os.lstat(f).st_mode |
|
70 | 68 | if l: |
|
71 | 69 | if not stat.S_ISLNK(s): |
|
72 | 70 | # switch file to link |
|
73 | 71 | data = file(f).read() |
|
74 | 72 | os.unlink(f) |
|
75 | 73 | try: |
|
76 | 74 | os.symlink(data, f) |
|
77 | 75 | except: |
|
78 | 76 | # failed to make a link, rewrite file |
|
79 | 77 | file(f, "w").write(data) |
|
80 | 78 | # no chmod needed at this point |
|
81 | 79 | return |
|
82 | 80 | if stat.S_ISLNK(s): |
|
83 | 81 | # switch link to file |
|
84 | 82 | data = os.readlink(f) |
|
85 | 83 | os.unlink(f) |
|
86 | 84 | file(f, "w").write(data) |
|
87 | 85 | s = 0666 & ~umask # avoid restatting for chmod |
|
88 | 86 | |
|
89 | 87 | sx = s & 0100 |
|
90 | 88 | if x and not sx: |
|
91 | 89 | # Turn on +x for every +r bit when making a file executable |
|
92 | 90 | # and obey umask. |
|
93 | 91 | os.chmod(f, s | (s & 0444) >> 2 & ~umask) |
|
94 | 92 | elif not x and sx: |
|
95 | 93 | # Turn off all +x bits |
|
96 | 94 | os.chmod(f, s & 0666) |
|
97 | 95 | |
|
98 | 96 | def set_binary(fd): |
|
99 | 97 | pass |
|
100 | 98 | |
|
101 | 99 | def pconvert(path): |
|
102 | 100 | return path |
|
103 | 101 | |
|
104 | 102 | def localpath(path): |
|
105 | 103 | return path |
|
106 | 104 | |
|
107 | 105 | def shellquote(s): |
|
108 | 106 | if os.sys.platform == 'OpenVMS': |
|
109 | 107 | return '"%s"' % s |
|
110 | 108 | else: |
|
111 | 109 | return "'%s'" % s.replace("'", "'\\''") |
|
112 | 110 | |
|
113 | 111 | def quotecommand(cmd): |
|
114 | 112 | return cmd |
|
115 | 113 | |
|
116 | 114 | def popen(command, mode='r'): |
|
117 | 115 | return os.popen(command, mode) |
|
118 | 116 | |
|
119 | 117 | def testpid(pid): |
|
120 | 118 | '''return False if pid dead, True if running or not sure''' |
|
121 | 119 | if os.sys.platform == 'OpenVMS': |
|
122 | 120 | return True |
|
123 | 121 | try: |
|
124 | 122 | os.kill(pid, 0) |
|
125 | 123 | return True |
|
126 | 124 | except OSError, inst: |
|
127 | 125 | return inst.errno != errno.ESRCH |
|
128 | 126 | |
|
129 | 127 | def explain_exit(code): |
|
130 | 128 | """return a 2-tuple (desc, code) describing a process's status""" |
|
131 | 129 | if os.WIFEXITED(code): |
|
132 | 130 | val = os.WEXITSTATUS(code) |
|
133 | 131 | return _("exited with status %d") % val, val |
|
134 | 132 | elif os.WIFSIGNALED(code): |
|
135 | 133 | val = os.WTERMSIG(code) |
|
136 | 134 | return _("killed by signal %d") % val, val |
|
137 | 135 | elif os.WIFSTOPPED(code): |
|
138 | 136 | val = os.WSTOPSIG(code) |
|
139 | 137 | return _("stopped by signal %d") % val, val |
|
140 | 138 | raise ValueError(_("invalid exit code")) |
|
141 | 139 | |
|
142 | 140 | def isowner(fp, st=None): |
|
143 | 141 | """Return True if the file object f belongs to the current user. |
|
144 | 142 | |
|
145 | 143 | The return value of a util.fstat(f) may be passed as the st argument. |
|
146 | 144 | """ |
|
147 | 145 | if st is None: |
|
148 | 146 | st = fstat(fp) |
|
149 | 147 | return st.st_uid == os.getuid() |
|
150 | 148 | |
|
151 | 149 | def find_exe(command): |
|
152 | 150 | '''Find executable for command searching like which does. |
|
153 | 151 | If command is a basename then PATH is searched for command. |
|
154 | 152 | PATH isn't searched if command is an absolute or relative path. |
|
155 | 153 | If command isn't found None is returned.''' |
|
156 | 154 | if sys.platform == 'OpenVMS': |
|
157 | 155 | return command |
|
158 | 156 | |
|
159 | 157 | def findexisting(executable): |
|
160 | 158 | 'Will return executable if existing file' |
|
161 | 159 | if os.path.exists(executable): |
|
162 | 160 | return executable |
|
163 | 161 | return None |
|
164 | 162 | |
|
165 | 163 | if os.sep in command: |
|
166 | 164 | return findexisting(command) |
|
167 | 165 | |
|
168 | 166 | for path in os.environ.get('PATH', '').split(os.pathsep): |
|
169 | 167 | executable = findexisting(os.path.join(path, command)) |
|
170 | 168 | if executable is not None: |
|
171 | 169 | return executable |
|
172 | 170 | return None |
|
173 | 171 | |
|
174 | 172 | def set_signal_handler(): |
|
175 | 173 | pass |
|
176 | 174 | |
|
177 | 175 | def statfiles(files): |
|
178 | 176 | 'Stat each file in files and yield stat or None if file does not exist.' |
|
179 | 177 | lstat = os.lstat |
|
180 | 178 | for nf in files: |
|
181 | 179 | try: |
|
182 | 180 | st = lstat(nf) |
|
183 | 181 | except OSError, err: |
|
184 | 182 | if err.errno not in (errno.ENOENT, errno.ENOTDIR): |
|
185 | 183 | raise |
|
186 | 184 | st = None |
|
187 | 185 | yield st |
|
188 | 186 | |
|
189 | 187 | def getuser(): |
|
190 | 188 | '''return name of current user''' |
|
191 | 189 | return getpass.getuser() |
|
192 | 190 | |
|
193 | 191 | def expand_glob(pats): |
|
194 | 192 | '''On Windows, expand the implicit globs in a list of patterns''' |
|
195 | 193 | return list(pats) |
|
196 | 194 | |
|
197 | 195 | def username(uid=None): |
|
198 | 196 | """Return the name of the user with the given uid. |
|
199 | 197 | |
|
200 | 198 | If uid is None, return the name of the current user.""" |
|
201 | 199 | |
|
202 | 200 | if uid is None: |
|
203 | 201 | uid = os.getuid() |
|
204 | 202 | try: |
|
205 | 203 | return pwd.getpwuid(uid)[0] |
|
206 | 204 | except KeyError: |
|
207 | 205 | return str(uid) |
|
208 | 206 | |
|
209 | 207 | def groupname(gid=None): |
|
210 | 208 | """Return the name of the group with the given gid. |
|
211 | 209 | |
|
212 | 210 | If gid is None, return the name of the current group.""" |
|
213 | 211 | |
|
214 | 212 | if gid is None: |
|
215 | 213 | gid = os.getgid() |
|
216 | 214 | try: |
|
217 | 215 | return grp.getgrgid(gid)[0] |
|
218 | 216 | except KeyError: |
|
219 | 217 | return str(gid) |
@@ -1,1369 +1,1367 b'' | |||
|
1 | """ | |
|
2 | revlog.py - storage back-end for mercurial | |
|
3 | ||
|
4 | This provides efficient delta storage with O(1) retrieve and append | |
|
5 | and O(changes) merge between branches | |
|
6 | ||
|
7 | Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
|
8 | ||
|
9 | This software may be used and distributed according to the terms of the | |
|
10 | GNU General Public License version 2, incorporated herein by reference. | |
|
11 | """ | |
|
1 | # revlog.py - storage back-end for mercurial | |
|
2 | # | |
|
3 | # This provides efficient delta storage with O(1) retrieve and append | |
|
4 | # and O(changes) merge between branches | |
|
5 | # | |
|
6 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
|
7 | # | |
|
8 | # This software may be used and distributed according to the terms of the | |
|
9 | # GNU General Public License version 2, incorporated herein by reference. | |
|
12 | 10 | |
|
13 | 11 | # import stuff from node for others to import from revlog |
|
14 | 12 | from node import bin, hex, nullid, nullrev, short #@UnusedImport |
|
15 | 13 | from i18n import _ |
|
16 | 14 | import changegroup, errno, ancestor, mdiff, parsers |
|
17 | 15 | import struct, util, zlib, error |
|
18 | 16 | |
|
19 | 17 | _pack = struct.pack |
|
20 | 18 | _unpack = struct.unpack |
|
21 | 19 | _compress = zlib.compress |
|
22 | 20 | _decompress = zlib.decompress |
|
23 | 21 | _sha = util.sha1 |
|
24 | 22 | |
|
25 | 23 | # revlog flags |
|
26 | 24 | REVLOGV0 = 0 |
|
27 | 25 | REVLOGNG = 1 |
|
28 | 26 | REVLOGNGINLINEDATA = (1 << 16) |
|
29 | 27 | REVLOG_DEFAULT_FLAGS = REVLOGNGINLINEDATA |
|
30 | 28 | REVLOG_DEFAULT_FORMAT = REVLOGNG |
|
31 | 29 | REVLOG_DEFAULT_VERSION = REVLOG_DEFAULT_FORMAT | REVLOG_DEFAULT_FLAGS |
|
32 | 30 | |
|
33 | 31 | RevlogError = error.RevlogError |
|
34 | 32 | LookupError = error.LookupError |
|
35 | 33 | |
|
36 | 34 | def getoffset(q): |
|
37 | 35 | return int(q >> 16) |
|
38 | 36 | |
|
39 | 37 | def gettype(q): |
|
40 | 38 | return int(q & 0xFFFF) |
|
41 | 39 | |
|
42 | 40 | def offset_type(offset, type): |
|
43 | 41 | return long(long(offset) << 16 | type) |
|
44 | 42 | |
|
45 | 43 | nullhash = _sha(nullid) |
|
46 | 44 | |
|
47 | 45 | def hash(text, p1, p2): |
|
48 | 46 | """generate a hash from the given text and its parent hashes |
|
49 | 47 | |
|
50 | 48 | This hash combines both the current file contents and its history |
|
51 | 49 | in a manner that makes it easy to distinguish nodes with the same |
|
52 | 50 | content in the revision graph. |
|
53 | 51 | """ |
|
54 | 52 | # As of now, if one of the parent node is null, p2 is null |
|
55 | 53 | if p2 == nullid: |
|
56 | 54 | # deep copy of a hash is faster than creating one |
|
57 | 55 | s = nullhash.copy() |
|
58 | 56 | s.update(p1) |
|
59 | 57 | else: |
|
60 | 58 | # none of the parent nodes are nullid |
|
61 | 59 | l = [p1, p2] |
|
62 | 60 | l.sort() |
|
63 | 61 | s = _sha(l[0]) |
|
64 | 62 | s.update(l[1]) |
|
65 | 63 | s.update(text) |
|
66 | 64 | return s.digest() |
|
67 | 65 | |
|
68 | 66 | def compress(text): |
|
69 | 67 | """ generate a possibly-compressed representation of text """ |
|
70 | 68 | if not text: |
|
71 | 69 | return ("", text) |
|
72 | 70 | l = len(text) |
|
73 | 71 | bin = None |
|
74 | 72 | if l < 44: |
|
75 | 73 | pass |
|
76 | 74 | elif l > 1000000: |
|
77 | 75 | # zlib makes an internal copy, thus doubling memory usage for |
|
78 | 76 | # large files, so lets do this in pieces |
|
79 | 77 | z = zlib.compressobj() |
|
80 | 78 | p = [] |
|
81 | 79 | pos = 0 |
|
82 | 80 | while pos < l: |
|
83 | 81 | pos2 = pos + 2**20 |
|
84 | 82 | p.append(z.compress(text[pos:pos2])) |
|
85 | 83 | pos = pos2 |
|
86 | 84 | p.append(z.flush()) |
|
87 | 85 | if sum(map(len, p)) < l: |
|
88 | 86 | bin = "".join(p) |
|
89 | 87 | else: |
|
90 | 88 | bin = _compress(text) |
|
91 | 89 | if bin is None or len(bin) > l: |
|
92 | 90 | if text[0] == '\0': |
|
93 | 91 | return ("", text) |
|
94 | 92 | return ('u', text) |
|
95 | 93 | return ("", bin) |
|
96 | 94 | |
|
97 | 95 | def decompress(bin): |
|
98 | 96 | """ decompress the given input """ |
|
99 | 97 | if not bin: |
|
100 | 98 | return bin |
|
101 | 99 | t = bin[0] |
|
102 | 100 | if t == '\0': |
|
103 | 101 | return bin |
|
104 | 102 | if t == 'x': |
|
105 | 103 | return _decompress(bin) |
|
106 | 104 | if t == 'u': |
|
107 | 105 | return bin[1:] |
|
108 | 106 | raise RevlogError(_("unknown compression type %r") % t) |
|
109 | 107 | |
|
110 | 108 | class lazyparser(object): |
|
111 | 109 | """ |
|
112 | 110 | this class avoids the need to parse the entirety of large indices |
|
113 | 111 | """ |
|
114 | 112 | |
|
115 | 113 | # lazyparser is not safe to use on windows if win32 extensions not |
|
116 | 114 | # available. it keeps file handle open, which make it not possible |
|
117 | 115 | # to break hardlinks on local cloned repos. |
|
118 | 116 | |
|
119 | 117 | def __init__(self, dataf, size): |
|
120 | 118 | self.dataf = dataf |
|
121 | 119 | self.s = struct.calcsize(indexformatng) |
|
122 | 120 | self.datasize = size |
|
123 | 121 | self.l = size/self.s |
|
124 | 122 | self.index = [None] * self.l |
|
125 | 123 | self.map = {nullid: nullrev} |
|
126 | 124 | self.allmap = 0 |
|
127 | 125 | self.all = 0 |
|
128 | 126 | self.mapfind_count = 0 |
|
129 | 127 | |
|
130 | 128 | def loadmap(self): |
|
131 | 129 | """ |
|
132 | 130 | during a commit, we need to make sure the rev being added is |
|
133 | 131 | not a duplicate. This requires loading the entire index, |
|
134 | 132 | which is fairly slow. loadmap can load up just the node map, |
|
135 | 133 | which takes much less time. |
|
136 | 134 | """ |
|
137 | 135 | if self.allmap: |
|
138 | 136 | return |
|
139 | 137 | end = self.datasize |
|
140 | 138 | self.allmap = 1 |
|
141 | 139 | cur = 0 |
|
142 | 140 | count = 0 |
|
143 | 141 | blocksize = self.s * 256 |
|
144 | 142 | self.dataf.seek(0) |
|
145 | 143 | while cur < end: |
|
146 | 144 | data = self.dataf.read(blocksize) |
|
147 | 145 | off = 0 |
|
148 | 146 | for x in xrange(256): |
|
149 | 147 | n = data[off + ngshaoffset:off + ngshaoffset + 20] |
|
150 | 148 | self.map[n] = count |
|
151 | 149 | count += 1 |
|
152 | 150 | if count >= self.l: |
|
153 | 151 | break |
|
154 | 152 | off += self.s |
|
155 | 153 | cur += blocksize |
|
156 | 154 | |
|
157 | 155 | def loadblock(self, blockstart, blocksize, data=None): |
|
158 | 156 | if self.all: |
|
159 | 157 | return |
|
160 | 158 | if data is None: |
|
161 | 159 | self.dataf.seek(blockstart) |
|
162 | 160 | if blockstart + blocksize > self.datasize: |
|
163 | 161 | # the revlog may have grown since we've started running, |
|
164 | 162 | # but we don't have space in self.index for more entries. |
|
165 | 163 | # limit blocksize so that we don't get too much data. |
|
166 | 164 | blocksize = max(self.datasize - blockstart, 0) |
|
167 | 165 | data = self.dataf.read(blocksize) |
|
168 | 166 | lend = len(data) / self.s |
|
169 | 167 | i = blockstart / self.s |
|
170 | 168 | off = 0 |
|
171 | 169 | # lazyindex supports __delitem__ |
|
172 | 170 | if lend > len(self.index) - i: |
|
173 | 171 | lend = len(self.index) - i |
|
174 | 172 | for x in xrange(lend): |
|
175 | 173 | if self.index[i + x] == None: |
|
176 | 174 | b = data[off : off + self.s] |
|
177 | 175 | self.index[i + x] = b |
|
178 | 176 | n = b[ngshaoffset:ngshaoffset + 20] |
|
179 | 177 | self.map[n] = i + x |
|
180 | 178 | off += self.s |
|
181 | 179 | |
|
182 | 180 | def findnode(self, node): |
|
183 | 181 | """search backwards through the index file for a specific node""" |
|
184 | 182 | if self.allmap: |
|
185 | 183 | return None |
|
186 | 184 | |
|
187 | 185 | # hg log will cause many many searches for the manifest |
|
188 | 186 | # nodes. After we get called a few times, just load the whole |
|
189 | 187 | # thing. |
|
190 | 188 | if self.mapfind_count > 8: |
|
191 | 189 | self.loadmap() |
|
192 | 190 | if node in self.map: |
|
193 | 191 | return node |
|
194 | 192 | return None |
|
195 | 193 | self.mapfind_count += 1 |
|
196 | 194 | last = self.l - 1 |
|
197 | 195 | while self.index[last] != None: |
|
198 | 196 | if last == 0: |
|
199 | 197 | self.all = 1 |
|
200 | 198 | self.allmap = 1 |
|
201 | 199 | return None |
|
202 | 200 | last -= 1 |
|
203 | 201 | end = (last + 1) * self.s |
|
204 | 202 | blocksize = self.s * 256 |
|
205 | 203 | while end >= 0: |
|
206 | 204 | start = max(end - blocksize, 0) |
|
207 | 205 | self.dataf.seek(start) |
|
208 | 206 | data = self.dataf.read(end - start) |
|
209 | 207 | findend = end - start |
|
210 | 208 | while True: |
|
211 | 209 | # we're searching backwards, so we have to make sure |
|
212 | 210 | # we don't find a changeset where this node is a parent |
|
213 | 211 | off = data.find(node, 0, findend) |
|
214 | 212 | findend = off |
|
215 | 213 | if off >= 0: |
|
216 | 214 | i = off / self.s |
|
217 | 215 | off = i * self.s |
|
218 | 216 | n = data[off + ngshaoffset:off + ngshaoffset + 20] |
|
219 | 217 | if n == node: |
|
220 | 218 | self.map[n] = i + start / self.s |
|
221 | 219 | return node |
|
222 | 220 | else: |
|
223 | 221 | break |
|
224 | 222 | end -= blocksize |
|
225 | 223 | return None |
|
226 | 224 | |
|
227 | 225 | def loadindex(self, i=None, end=None): |
|
228 | 226 | if self.all: |
|
229 | 227 | return |
|
230 | 228 | all = False |
|
231 | 229 | if i == None: |
|
232 | 230 | blockstart = 0 |
|
233 | 231 | blocksize = (65536 / self.s) * self.s |
|
234 | 232 | end = self.datasize |
|
235 | 233 | all = True |
|
236 | 234 | else: |
|
237 | 235 | if end: |
|
238 | 236 | blockstart = i * self.s |
|
239 | 237 | end = end * self.s |
|
240 | 238 | blocksize = end - blockstart |
|
241 | 239 | else: |
|
242 | 240 | blockstart = (i & ~1023) * self.s |
|
243 | 241 | blocksize = self.s * 1024 |
|
244 | 242 | end = blockstart + blocksize |
|
245 | 243 | while blockstart < end: |
|
246 | 244 | self.loadblock(blockstart, blocksize) |
|
247 | 245 | blockstart += blocksize |
|
248 | 246 | if all: |
|
249 | 247 | self.all = True |
|
250 | 248 | |
|
251 | 249 | class lazyindex(object): |
|
252 | 250 | """a lazy version of the index array""" |
|
253 | 251 | def __init__(self, parser): |
|
254 | 252 | self.p = parser |
|
255 | 253 | def __len__(self): |
|
256 | 254 | return len(self.p.index) |
|
257 | 255 | def load(self, pos): |
|
258 | 256 | if pos < 0: |
|
259 | 257 | pos += len(self.p.index) |
|
260 | 258 | self.p.loadindex(pos) |
|
261 | 259 | return self.p.index[pos] |
|
262 | 260 | def __getitem__(self, pos): |
|
263 | 261 | return _unpack(indexformatng, self.p.index[pos] or self.load(pos)) |
|
264 | 262 | def __setitem__(self, pos, item): |
|
265 | 263 | self.p.index[pos] = _pack(indexformatng, *item) |
|
266 | 264 | def __delitem__(self, pos): |
|
267 | 265 | del self.p.index[pos] |
|
268 | 266 | def insert(self, pos, e): |
|
269 | 267 | self.p.index.insert(pos, _pack(indexformatng, *e)) |
|
270 | 268 | def append(self, e): |
|
271 | 269 | self.p.index.append(_pack(indexformatng, *e)) |
|
272 | 270 | |
|
273 | 271 | class lazymap(object): |
|
274 | 272 | """a lazy version of the node map""" |
|
275 | 273 | def __init__(self, parser): |
|
276 | 274 | self.p = parser |
|
277 | 275 | def load(self, key): |
|
278 | 276 | n = self.p.findnode(key) |
|
279 | 277 | if n == None: |
|
280 | 278 | raise KeyError(key) |
|
281 | 279 | def __contains__(self, key): |
|
282 | 280 | if key in self.p.map: |
|
283 | 281 | return True |
|
284 | 282 | self.p.loadmap() |
|
285 | 283 | return key in self.p.map |
|
286 | 284 | def __iter__(self): |
|
287 | 285 | yield nullid |
|
288 | 286 | for i in xrange(self.p.l): |
|
289 | 287 | ret = self.p.index[i] |
|
290 | 288 | if not ret: |
|
291 | 289 | self.p.loadindex(i) |
|
292 | 290 | ret = self.p.index[i] |
|
293 | 291 | if isinstance(ret, str): |
|
294 | 292 | ret = _unpack(indexformatng, ret) |
|
295 | 293 | yield ret[7] |
|
296 | 294 | def __getitem__(self, key): |
|
297 | 295 | try: |
|
298 | 296 | return self.p.map[key] |
|
299 | 297 | except KeyError: |
|
300 | 298 | try: |
|
301 | 299 | self.load(key) |
|
302 | 300 | return self.p.map[key] |
|
303 | 301 | except KeyError: |
|
304 | 302 | raise KeyError("node " + hex(key)) |
|
305 | 303 | def __setitem__(self, key, val): |
|
306 | 304 | self.p.map[key] = val |
|
307 | 305 | def __delitem__(self, key): |
|
308 | 306 | del self.p.map[key] |
|
309 | 307 | |
|
310 | 308 | indexformatv0 = ">4l20s20s20s" |
|
311 | 309 | v0shaoffset = 56 |
|
312 | 310 | |
|
313 | 311 | class revlogoldio(object): |
|
314 | 312 | def __init__(self): |
|
315 | 313 | self.size = struct.calcsize(indexformatv0) |
|
316 | 314 | |
|
317 | 315 | def parseindex(self, fp, inline): |
|
318 | 316 | s = self.size |
|
319 | 317 | index = [] |
|
320 | 318 | nodemap = {nullid: nullrev} |
|
321 | 319 | n = off = 0 |
|
322 | 320 | data = fp.read() |
|
323 | 321 | l = len(data) |
|
324 | 322 | while off + s <= l: |
|
325 | 323 | cur = data[off:off + s] |
|
326 | 324 | off += s |
|
327 | 325 | e = _unpack(indexformatv0, cur) |
|
328 | 326 | # transform to revlogv1 format |
|
329 | 327 | e2 = (offset_type(e[0], 0), e[1], -1, e[2], e[3], |
|
330 | 328 | nodemap.get(e[4], nullrev), nodemap.get(e[5], nullrev), e[6]) |
|
331 | 329 | index.append(e2) |
|
332 | 330 | nodemap[e[6]] = n |
|
333 | 331 | n += 1 |
|
334 | 332 | |
|
335 | 333 | return index, nodemap, None |
|
336 | 334 | |
|
337 | 335 | def packentry(self, entry, node, version, rev): |
|
338 | 336 | e2 = (getoffset(entry[0]), entry[1], entry[3], entry[4], |
|
339 | 337 | node(entry[5]), node(entry[6]), entry[7]) |
|
340 | 338 | return _pack(indexformatv0, *e2) |
|
341 | 339 | |
|
342 | 340 | # index ng: |
|
343 | 341 | # 6 bytes offset |
|
344 | 342 | # 2 bytes flags |
|
345 | 343 | # 4 bytes compressed length |
|
346 | 344 | # 4 bytes uncompressed length |
|
347 | 345 | # 4 bytes: base rev |
|
348 | 346 | # 4 bytes link rev |
|
349 | 347 | # 4 bytes parent 1 rev |
|
350 | 348 | # 4 bytes parent 2 rev |
|
351 | 349 | # 32 bytes: nodeid |
|
352 | 350 | indexformatng = ">Qiiiiii20s12x" |
|
353 | 351 | ngshaoffset = 32 |
|
354 | 352 | versionformat = ">I" |
|
355 | 353 | |
|
356 | 354 | class revlogio(object): |
|
357 | 355 | def __init__(self): |
|
358 | 356 | self.size = struct.calcsize(indexformatng) |
|
359 | 357 | |
|
360 | 358 | def parseindex(self, fp, inline): |
|
361 | 359 | try: |
|
362 | 360 | size = util.fstat(fp).st_size |
|
363 | 361 | except AttributeError: |
|
364 | 362 | size = 0 |
|
365 | 363 | |
|
366 | 364 | if util.openhardlinks() and not inline and size > 1000000: |
|
367 | 365 | # big index, let's parse it on demand |
|
368 | 366 | parser = lazyparser(fp, size) |
|
369 | 367 | index = lazyindex(parser) |
|
370 | 368 | nodemap = lazymap(parser) |
|
371 | 369 | e = list(index[0]) |
|
372 | 370 | type = gettype(e[0]) |
|
373 | 371 | e[0] = offset_type(0, type) |
|
374 | 372 | index[0] = e |
|
375 | 373 | return index, nodemap, None |
|
376 | 374 | |
|
377 | 375 | data = fp.read() |
|
378 | 376 | # call the C implementation to parse the index data |
|
379 | 377 | index, nodemap, cache = parsers.parse_index(data, inline) |
|
380 | 378 | return index, nodemap, cache |
|
381 | 379 | |
|
382 | 380 | def packentry(self, entry, node, version, rev): |
|
383 | 381 | p = _pack(indexformatng, *entry) |
|
384 | 382 | if rev == 0: |
|
385 | 383 | p = _pack(versionformat, version) + p[4:] |
|
386 | 384 | return p |
|
387 | 385 | |
|
388 | 386 | class revlog(object): |
|
389 | 387 | """ |
|
390 | 388 | the underlying revision storage object |
|
391 | 389 | |
|
392 | 390 | A revlog consists of two parts, an index and the revision data. |
|
393 | 391 | |
|
394 | 392 | The index is a file with a fixed record size containing |
|
395 | 393 | information on each revision, including its nodeid (hash), the |
|
396 | 394 | nodeids of its parents, the position and offset of its data within |
|
397 | 395 | the data file, and the revision it's based on. Finally, each entry |
|
398 | 396 | contains a linkrev entry that can serve as a pointer to external |
|
399 | 397 | data. |
|
400 | 398 | |
|
401 | 399 | The revision data itself is a linear collection of data chunks. |
|
402 | 400 | Each chunk represents a revision and is usually represented as a |
|
403 | 401 | delta against the previous chunk. To bound lookup time, runs of |
|
404 | 402 | deltas are limited to about 2 times the length of the original |
|
405 | 403 | version data. This makes retrieval of a version proportional to |
|
406 | 404 | its size, or O(1) relative to the number of revisions. |
|
407 | 405 | |
|
408 | 406 | Both pieces of the revlog are written to in an append-only |
|
409 | 407 | fashion, which means we never need to rewrite a file to insert or |
|
410 | 408 | remove data, and can use some simple techniques to avoid the need |
|
411 | 409 | for locking while reading. |
|
412 | 410 | """ |
|
413 | 411 | def __init__(self, opener, indexfile): |
|
414 | 412 | """ |
|
415 | 413 | create a revlog object |
|
416 | 414 | |
|
417 | 415 | opener is a function that abstracts the file opening operation |
|
418 | 416 | and can be used to implement COW semantics or the like. |
|
419 | 417 | """ |
|
420 | 418 | self.indexfile = indexfile |
|
421 | 419 | self.datafile = indexfile[:-2] + ".d" |
|
422 | 420 | self.opener = opener |
|
423 | 421 | self._cache = None |
|
424 | 422 | self._chunkcache = None |
|
425 | 423 | self.nodemap = {nullid: nullrev} |
|
426 | 424 | self.index = [] |
|
427 | 425 | |
|
428 | 426 | v = REVLOG_DEFAULT_VERSION |
|
429 | 427 | if hasattr(opener, "defversion"): |
|
430 | 428 | v = opener.defversion |
|
431 | 429 | if v & REVLOGNG: |
|
432 | 430 | v |= REVLOGNGINLINEDATA |
|
433 | 431 | |
|
434 | 432 | i = "" |
|
435 | 433 | try: |
|
436 | 434 | f = self.opener(self.indexfile) |
|
437 | 435 | i = f.read(4) |
|
438 | 436 | f.seek(0) |
|
439 | 437 | if len(i) > 0: |
|
440 | 438 | v = struct.unpack(versionformat, i)[0] |
|
441 | 439 | except IOError, inst: |
|
442 | 440 | if inst.errno != errno.ENOENT: |
|
443 | 441 | raise |
|
444 | 442 | |
|
445 | 443 | self.version = v |
|
446 | 444 | self._inline = v & REVLOGNGINLINEDATA |
|
447 | 445 | flags = v & ~0xFFFF |
|
448 | 446 | fmt = v & 0xFFFF |
|
449 | 447 | if fmt == REVLOGV0 and flags: |
|
450 | 448 | raise RevlogError(_("index %s unknown flags %#04x for format v0") |
|
451 | 449 | % (self.indexfile, flags >> 16)) |
|
452 | 450 | elif fmt == REVLOGNG and flags & ~REVLOGNGINLINEDATA: |
|
453 | 451 | raise RevlogError(_("index %s unknown flags %#04x for revlogng") |
|
454 | 452 | % (self.indexfile, flags >> 16)) |
|
455 | 453 | elif fmt > REVLOGNG: |
|
456 | 454 | raise RevlogError(_("index %s unknown format %d") |
|
457 | 455 | % (self.indexfile, fmt)) |
|
458 | 456 | |
|
459 | 457 | self._io = revlogio() |
|
460 | 458 | if self.version == REVLOGV0: |
|
461 | 459 | self._io = revlogoldio() |
|
462 | 460 | if i: |
|
463 | 461 | try: |
|
464 | 462 | d = self._io.parseindex(f, self._inline) |
|
465 | 463 | except (ValueError, IndexError), e: |
|
466 | 464 | raise RevlogError(_("index %s is corrupted") % (self.indexfile)) |
|
467 | 465 | self.index, self.nodemap, self._chunkcache = d |
|
468 | 466 | |
|
469 | 467 | # add the magic null revision at -1 (if it hasn't been done already) |
|
470 | 468 | if (self.index == [] or isinstance(self.index, lazyindex) or |
|
471 | 469 | self.index[-1][7] != nullid) : |
|
472 | 470 | self.index.append((0, 0, 0, -1, -1, -1, -1, nullid)) |
|
473 | 471 | |
|
474 | 472 | def _loadindex(self, start, end): |
|
475 | 473 | """load a block of indexes all at once from the lazy parser""" |
|
476 | 474 | if isinstance(self.index, lazyindex): |
|
477 | 475 | self.index.p.loadindex(start, end) |
|
478 | 476 | |
|
479 | 477 | def _loadindexmap(self): |
|
480 | 478 | """loads both the map and the index from the lazy parser""" |
|
481 | 479 | if isinstance(self.index, lazyindex): |
|
482 | 480 | p = self.index.p |
|
483 | 481 | p.loadindex() |
|
484 | 482 | self.nodemap = p.map |
|
485 | 483 | |
|
486 | 484 | def _loadmap(self): |
|
487 | 485 | """loads the map from the lazy parser""" |
|
488 | 486 | if isinstance(self.nodemap, lazymap): |
|
489 | 487 | self.nodemap.p.loadmap() |
|
490 | 488 | self.nodemap = self.nodemap.p.map |
|
491 | 489 | |
|
492 | 490 | def tip(self): |
|
493 | 491 | return self.node(len(self.index) - 2) |
|
494 | 492 | def __len__(self): |
|
495 | 493 | return len(self.index) - 1 |
|
496 | 494 | def __iter__(self): |
|
497 | 495 | for i in xrange(len(self)): |
|
498 | 496 | yield i |
|
499 | 497 | def rev(self, node): |
|
500 | 498 | try: |
|
501 | 499 | return self.nodemap[node] |
|
502 | 500 | except KeyError: |
|
503 | 501 | raise LookupError(node, self.indexfile, _('no node')) |
|
504 | 502 | def node(self, rev): |
|
505 | 503 | return self.index[rev][7] |
|
506 | 504 | def linkrev(self, rev): |
|
507 | 505 | return self.index[rev][4] |
|
508 | 506 | def parents(self, node): |
|
509 | 507 | i = self.index |
|
510 | 508 | d = i[self.rev(node)] |
|
511 | 509 | return i[d[5]][7], i[d[6]][7] # map revisions to nodes inline |
|
512 | 510 | def parentrevs(self, rev): |
|
513 | 511 | return self.index[rev][5:7] |
|
514 | 512 | def start(self, rev): |
|
515 | 513 | return int(self.index[rev][0] >> 16) |
|
516 | 514 | def end(self, rev): |
|
517 | 515 | return self.start(rev) + self.length(rev) |
|
518 | 516 | def length(self, rev): |
|
519 | 517 | return self.index[rev][1] |
|
520 | 518 | def base(self, rev): |
|
521 | 519 | return self.index[rev][3] |
|
522 | 520 | |
|
523 | 521 | def size(self, rev): |
|
524 | 522 | """return the length of the uncompressed text for a given revision""" |
|
525 | 523 | l = self.index[rev][2] |
|
526 | 524 | if l >= 0: |
|
527 | 525 | return l |
|
528 | 526 | |
|
529 | 527 | t = self.revision(self.node(rev)) |
|
530 | 528 | return len(t) |
|
531 | 529 | |
|
532 | 530 | # alternate implementation, The advantage to this code is it |
|
533 | 531 | # will be faster for a single revision. But, the results are not |
|
534 | 532 | # cached, so finding the size of every revision will be slower. |
|
535 | 533 | """ |
|
536 | 534 | if self.cache and self.cache[1] == rev: |
|
537 | 535 | return len(self.cache[2]) |
|
538 | 536 | |
|
539 | 537 | base = self.base(rev) |
|
540 | 538 | if self.cache and self.cache[1] >= base and self.cache[1] < rev: |
|
541 | 539 | base = self.cache[1] |
|
542 | 540 | text = self.cache[2] |
|
543 | 541 | else: |
|
544 | 542 | text = self.revision(self.node(base)) |
|
545 | 543 | |
|
546 | 544 | l = len(text) |
|
547 | 545 | for x in xrange(base + 1, rev + 1): |
|
548 | 546 | l = mdiff.patchedsize(l, self.chunk(x)) |
|
549 | 547 | return l |
|
550 | 548 | """ |
|
551 | 549 | |
|
552 | 550 | def reachable(self, node, stop=None): |
|
553 | 551 | """return a hash of all nodes ancestral to a given node, including |
|
554 | 552 | the node itself, stopping when stop is matched""" |
|
555 | 553 | reachable = {} |
|
556 | 554 | visit = [node] |
|
557 | 555 | reachable[node] = 1 |
|
558 | 556 | if stop: |
|
559 | 557 | stopn = self.rev(stop) |
|
560 | 558 | else: |
|
561 | 559 | stopn = 0 |
|
562 | 560 | while visit: |
|
563 | 561 | n = visit.pop(0) |
|
564 | 562 | if n == stop: |
|
565 | 563 | continue |
|
566 | 564 | if n == nullid: |
|
567 | 565 | continue |
|
568 | 566 | for p in self.parents(n): |
|
569 | 567 | if self.rev(p) < stopn: |
|
570 | 568 | continue |
|
571 | 569 | if p not in reachable: |
|
572 | 570 | reachable[p] = 1 |
|
573 | 571 | visit.append(p) |
|
574 | 572 | return reachable |
|
575 | 573 | |
|
576 | 574 | def ancestors(self, *revs): |
|
577 | 575 | 'Generate the ancestors of revs using a breadth-first visit' |
|
578 | 576 | visit = list(revs) |
|
579 | 577 | seen = set([nullrev]) |
|
580 | 578 | while visit: |
|
581 | 579 | for parent in self.parentrevs(visit.pop(0)): |
|
582 | 580 | if parent not in seen: |
|
583 | 581 | visit.append(parent) |
|
584 | 582 | seen.add(parent) |
|
585 | 583 | yield parent |
|
586 | 584 | |
|
587 | 585 | def descendants(self, *revs): |
|
588 | 586 | 'Generate the descendants of revs in topological order' |
|
589 | 587 | seen = set(revs) |
|
590 | 588 | for i in xrange(min(revs) + 1, len(self)): |
|
591 | 589 | for x in self.parentrevs(i): |
|
592 | 590 | if x != nullrev and x in seen: |
|
593 | 591 | seen.add(i) |
|
594 | 592 | yield i |
|
595 | 593 | break |
|
596 | 594 | |
|
597 | 595 | def findmissing(self, common=None, heads=None): |
|
598 | 596 | ''' |
|
599 | 597 | returns the topologically sorted list of nodes from the set: |
|
600 | 598 | missing = (ancestors(heads) \ ancestors(common)) |
|
601 | 599 | |
|
602 | 600 | where ancestors() is the set of ancestors from heads, heads included |
|
603 | 601 | |
|
604 | 602 | if heads is None, the heads of the revlog are used |
|
605 | 603 | if common is None, nullid is assumed to be a common node |
|
606 | 604 | ''' |
|
607 | 605 | if common is None: |
|
608 | 606 | common = [nullid] |
|
609 | 607 | if heads is None: |
|
610 | 608 | heads = self.heads() |
|
611 | 609 | |
|
612 | 610 | common = [self.rev(n) for n in common] |
|
613 | 611 | heads = [self.rev(n) for n in heads] |
|
614 | 612 | |
|
615 | 613 | # we want the ancestors, but inclusive |
|
616 | 614 | has = set(self.ancestors(*common)) |
|
617 | 615 | has.add(nullrev) |
|
618 | 616 | has.update(common) |
|
619 | 617 | |
|
620 | 618 | # take all ancestors from heads that aren't in has |
|
621 | 619 | missing = {} |
|
622 | 620 | visit = [r for r in heads if r not in has] |
|
623 | 621 | while visit: |
|
624 | 622 | r = visit.pop(0) |
|
625 | 623 | if r in missing: |
|
626 | 624 | continue |
|
627 | 625 | else: |
|
628 | 626 | missing[r] = None |
|
629 | 627 | for p in self.parentrevs(r): |
|
630 | 628 | if p not in has: |
|
631 | 629 | visit.append(p) |
|
632 | 630 | missing = missing.keys() |
|
633 | 631 | missing.sort() |
|
634 | 632 | return [self.node(r) for r in missing] |
|
635 | 633 | |
|
636 | 634 | def nodesbetween(self, roots=None, heads=None): |
|
637 | 635 | """Return a tuple containing three elements. Elements 1 and 2 contain |
|
638 | 636 | a final list bases and heads after all the unreachable ones have been |
|
639 | 637 | pruned. Element 0 contains a topologically sorted list of all |
|
640 | 638 | |
|
641 | 639 | nodes that satisfy these constraints: |
|
642 | 640 | 1. All nodes must be descended from a node in roots (the nodes on |
|
643 | 641 | roots are considered descended from themselves). |
|
644 | 642 | 2. All nodes must also be ancestors of a node in heads (the nodes in |
|
645 | 643 | heads are considered to be their own ancestors). |
|
646 | 644 | |
|
647 | 645 | If roots is unspecified, nullid is assumed as the only root. |
|
648 | 646 | If heads is unspecified, it is taken to be the output of the |
|
649 | 647 | heads method (i.e. a list of all nodes in the repository that |
|
650 | 648 | have no children).""" |
|
651 | 649 | nonodes = ([], [], []) |
|
652 | 650 | if roots is not None: |
|
653 | 651 | roots = list(roots) |
|
654 | 652 | if not roots: |
|
655 | 653 | return nonodes |
|
656 | 654 | lowestrev = min([self.rev(n) for n in roots]) |
|
657 | 655 | else: |
|
658 | 656 | roots = [nullid] # Everybody's a descendent of nullid |
|
659 | 657 | lowestrev = nullrev |
|
660 | 658 | if (lowestrev == nullrev) and (heads is None): |
|
661 | 659 | # We want _all_ the nodes! |
|
662 | 660 | return ([self.node(r) for r in self], [nullid], list(self.heads())) |
|
663 | 661 | if heads is None: |
|
664 | 662 | # All nodes are ancestors, so the latest ancestor is the last |
|
665 | 663 | # node. |
|
666 | 664 | highestrev = len(self) - 1 |
|
667 | 665 | # Set ancestors to None to signal that every node is an ancestor. |
|
668 | 666 | ancestors = None |
|
669 | 667 | # Set heads to an empty dictionary for later discovery of heads |
|
670 | 668 | heads = {} |
|
671 | 669 | else: |
|
672 | 670 | heads = list(heads) |
|
673 | 671 | if not heads: |
|
674 | 672 | return nonodes |
|
675 | 673 | ancestors = {} |
|
676 | 674 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
677 | 675 | # Also, later we will be using it to filter out the heads we can't |
|
678 | 676 | # find from roots. |
|
679 | 677 | heads = dict.fromkeys(heads, 0) |
|
680 | 678 | # Start at the top and keep marking parents until we're done. |
|
681 | 679 | nodestotag = set(heads) |
|
682 | 680 | # Remember where the top was so we can use it as a limit later. |
|
683 | 681 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
684 | 682 | while nodestotag: |
|
685 | 683 | # grab a node to tag |
|
686 | 684 | n = nodestotag.pop() |
|
687 | 685 | # Never tag nullid |
|
688 | 686 | if n == nullid: |
|
689 | 687 | continue |
|
690 | 688 | # A node's revision number represents its place in a |
|
691 | 689 | # topologically sorted list of nodes. |
|
692 | 690 | r = self.rev(n) |
|
693 | 691 | if r >= lowestrev: |
|
694 | 692 | if n not in ancestors: |
|
695 | 693 | # If we are possibly a descendent of one of the roots |
|
696 | 694 | # and we haven't already been marked as an ancestor |
|
697 | 695 | ancestors[n] = 1 # Mark as ancestor |
|
698 | 696 | # Add non-nullid parents to list of nodes to tag. |
|
699 | 697 | nodestotag.update([p for p in self.parents(n) if |
|
700 | 698 | p != nullid]) |
|
701 | 699 | elif n in heads: # We've seen it before, is it a fake head? |
|
702 | 700 | # So it is, real heads should not be the ancestors of |
|
703 | 701 | # any other heads. |
|
704 | 702 | heads.pop(n) |
|
705 | 703 | if not ancestors: |
|
706 | 704 | return nonodes |
|
707 | 705 | # Now that we have our set of ancestors, we want to remove any |
|
708 | 706 | # roots that are not ancestors. |
|
709 | 707 | |
|
710 | 708 | # If one of the roots was nullid, everything is included anyway. |
|
711 | 709 | if lowestrev > nullrev: |
|
712 | 710 | # But, since we weren't, let's recompute the lowest rev to not |
|
713 | 711 | # include roots that aren't ancestors. |
|
714 | 712 | |
|
715 | 713 | # Filter out roots that aren't ancestors of heads |
|
716 | 714 | roots = [n for n in roots if n in ancestors] |
|
717 | 715 | # Recompute the lowest revision |
|
718 | 716 | if roots: |
|
719 | 717 | lowestrev = min([self.rev(n) for n in roots]) |
|
720 | 718 | else: |
|
721 | 719 | # No more roots? Return empty list |
|
722 | 720 | return nonodes |
|
723 | 721 | else: |
|
724 | 722 | # We are descending from nullid, and don't need to care about |
|
725 | 723 | # any other roots. |
|
726 | 724 | lowestrev = nullrev |
|
727 | 725 | roots = [nullid] |
|
728 | 726 | # Transform our roots list into a set. |
|
729 | 727 | descendents = set(roots) |
|
730 | 728 | # Also, keep the original roots so we can filter out roots that aren't |
|
731 | 729 | # 'real' roots (i.e. are descended from other roots). |
|
732 | 730 | roots = descendents.copy() |
|
733 | 731 | # Our topologically sorted list of output nodes. |
|
734 | 732 | orderedout = [] |
|
735 | 733 | # Don't start at nullid since we don't want nullid in our output list, |
|
736 | 734 | # and if nullid shows up in descedents, empty parents will look like |
|
737 | 735 | # they're descendents. |
|
738 | 736 | for r in xrange(max(lowestrev, 0), highestrev + 1): |
|
739 | 737 | n = self.node(r) |
|
740 | 738 | isdescendent = False |
|
741 | 739 | if lowestrev == nullrev: # Everybody is a descendent of nullid |
|
742 | 740 | isdescendent = True |
|
743 | 741 | elif n in descendents: |
|
744 | 742 | # n is already a descendent |
|
745 | 743 | isdescendent = True |
|
746 | 744 | # This check only needs to be done here because all the roots |
|
747 | 745 | # will start being marked is descendents before the loop. |
|
748 | 746 | if n in roots: |
|
749 | 747 | # If n was a root, check if it's a 'real' root. |
|
750 | 748 | p = tuple(self.parents(n)) |
|
751 | 749 | # If any of its parents are descendents, it's not a root. |
|
752 | 750 | if (p[0] in descendents) or (p[1] in descendents): |
|
753 | 751 | roots.remove(n) |
|
754 | 752 | else: |
|
755 | 753 | p = tuple(self.parents(n)) |
|
756 | 754 | # A node is a descendent if either of its parents are |
|
757 | 755 | # descendents. (We seeded the dependents list with the roots |
|
758 | 756 | # up there, remember?) |
|
759 | 757 | if (p[0] in descendents) or (p[1] in descendents): |
|
760 | 758 | descendents.add(n) |
|
761 | 759 | isdescendent = True |
|
762 | 760 | if isdescendent and ((ancestors is None) or (n in ancestors)): |
|
763 | 761 | # Only include nodes that are both descendents and ancestors. |
|
764 | 762 | orderedout.append(n) |
|
765 | 763 | if (ancestors is not None) and (n in heads): |
|
766 | 764 | # We're trying to figure out which heads are reachable |
|
767 | 765 | # from roots. |
|
768 | 766 | # Mark this head as having been reached |
|
769 | 767 | heads[n] = 1 |
|
770 | 768 | elif ancestors is None: |
|
771 | 769 | # Otherwise, we're trying to discover the heads. |
|
772 | 770 | # Assume this is a head because if it isn't, the next step |
|
773 | 771 | # will eventually remove it. |
|
774 | 772 | heads[n] = 1 |
|
775 | 773 | # But, obviously its parents aren't. |
|
776 | 774 | for p in self.parents(n): |
|
777 | 775 | heads.pop(p, None) |
|
778 | 776 | heads = [n for n in heads.iterkeys() if heads[n] != 0] |
|
779 | 777 | roots = list(roots) |
|
780 | 778 | assert orderedout |
|
781 | 779 | assert roots |
|
782 | 780 | assert heads |
|
783 | 781 | return (orderedout, roots, heads) |
|
784 | 782 | |
|
785 | 783 | def heads(self, start=None, stop=None): |
|
786 | 784 | """return the list of all nodes that have no children |
|
787 | 785 | |
|
788 | 786 | if start is specified, only heads that are descendants of |
|
789 | 787 | start will be returned |
|
790 | 788 | if stop is specified, it will consider all the revs from stop |
|
791 | 789 | as if they had no children |
|
792 | 790 | """ |
|
793 | 791 | if start is None and stop is None: |
|
794 | 792 | count = len(self) |
|
795 | 793 | if not count: |
|
796 | 794 | return [nullid] |
|
797 | 795 | ishead = [1] * (count + 1) |
|
798 | 796 | index = self.index |
|
799 | 797 | for r in xrange(count): |
|
800 | 798 | e = index[r] |
|
801 | 799 | ishead[e[5]] = ishead[e[6]] = 0 |
|
802 | 800 | return [self.node(r) for r in xrange(count) if ishead[r]] |
|
803 | 801 | |
|
804 | 802 | if start is None: |
|
805 | 803 | start = nullid |
|
806 | 804 | if stop is None: |
|
807 | 805 | stop = [] |
|
808 | 806 | stoprevs = set([self.rev(n) for n in stop]) |
|
809 | 807 | startrev = self.rev(start) |
|
810 | 808 | reachable = {startrev: 1} |
|
811 | 809 | heads = {startrev: 1} |
|
812 | 810 | |
|
813 | 811 | parentrevs = self.parentrevs |
|
814 | 812 | for r in xrange(startrev + 1, len(self)): |
|
815 | 813 | for p in parentrevs(r): |
|
816 | 814 | if p in reachable: |
|
817 | 815 | if r not in stoprevs: |
|
818 | 816 | reachable[r] = 1 |
|
819 | 817 | heads[r] = 1 |
|
820 | 818 | if p in heads and p not in stoprevs: |
|
821 | 819 | del heads[p] |
|
822 | 820 | |
|
823 | 821 | return [self.node(r) for r in heads] |
|
824 | 822 | |
|
825 | 823 | def children(self, node): |
|
826 | 824 | """find the children of a given node""" |
|
827 | 825 | c = [] |
|
828 | 826 | p = self.rev(node) |
|
829 | 827 | for r in range(p + 1, len(self)): |
|
830 | 828 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
831 | 829 | if prevs: |
|
832 | 830 | for pr in prevs: |
|
833 | 831 | if pr == p: |
|
834 | 832 | c.append(self.node(r)) |
|
835 | 833 | elif p == nullrev: |
|
836 | 834 | c.append(self.node(r)) |
|
837 | 835 | return c |
|
838 | 836 | |
|
839 | 837 | def _match(self, id): |
|
840 | 838 | if isinstance(id, (long, int)): |
|
841 | 839 | # rev |
|
842 | 840 | return self.node(id) |
|
843 | 841 | if len(id) == 20: |
|
844 | 842 | # possibly a binary node |
|
845 | 843 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
846 | 844 | try: |
|
847 | 845 | node = id |
|
848 | 846 | self.rev(node) # quick search the index |
|
849 | 847 | return node |
|
850 | 848 | except LookupError: |
|
851 | 849 | pass # may be partial hex id |
|
852 | 850 | try: |
|
853 | 851 | # str(rev) |
|
854 | 852 | rev = int(id) |
|
855 | 853 | if str(rev) != id: |
|
856 | 854 | raise ValueError |
|
857 | 855 | if rev < 0: |
|
858 | 856 | rev = len(self) + rev |
|
859 | 857 | if rev < 0 or rev >= len(self): |
|
860 | 858 | raise ValueError |
|
861 | 859 | return self.node(rev) |
|
862 | 860 | except (ValueError, OverflowError): |
|
863 | 861 | pass |
|
864 | 862 | if len(id) == 40: |
|
865 | 863 | try: |
|
866 | 864 | # a full hex nodeid? |
|
867 | 865 | node = bin(id) |
|
868 | 866 | self.rev(node) |
|
869 | 867 | return node |
|
870 | 868 | except (TypeError, LookupError): |
|
871 | 869 | pass |
|
872 | 870 | |
|
873 | 871 | def _partialmatch(self, id): |
|
874 | 872 | if len(id) < 40: |
|
875 | 873 | try: |
|
876 | 874 | # hex(node)[:...] |
|
877 | 875 | l = len(id) / 2 # grab an even number of digits |
|
878 | 876 | bin_id = bin(id[:l*2]) |
|
879 | 877 | nl = [n for n in self.nodemap if n[:l] == bin_id] |
|
880 | 878 | nl = [n for n in nl if hex(n).startswith(id)] |
|
881 | 879 | if len(nl) > 0: |
|
882 | 880 | if len(nl) == 1: |
|
883 | 881 | return nl[0] |
|
884 | 882 | raise LookupError(id, self.indexfile, |
|
885 | 883 | _('ambiguous identifier')) |
|
886 | 884 | return None |
|
887 | 885 | except TypeError: |
|
888 | 886 | pass |
|
889 | 887 | |
|
890 | 888 | def lookup(self, id): |
|
891 | 889 | """locate a node based on: |
|
892 | 890 | - revision number or str(revision number) |
|
893 | 891 | - nodeid or subset of hex nodeid |
|
894 | 892 | """ |
|
895 | 893 | n = self._match(id) |
|
896 | 894 | if n is not None: |
|
897 | 895 | return n |
|
898 | 896 | n = self._partialmatch(id) |
|
899 | 897 | if n: |
|
900 | 898 | return n |
|
901 | 899 | |
|
902 | 900 | raise LookupError(id, self.indexfile, _('no match found')) |
|
903 | 901 | |
|
904 | 902 | def cmp(self, node, text): |
|
905 | 903 | """compare text with a given file revision""" |
|
906 | 904 | p1, p2 = self.parents(node) |
|
907 | 905 | return hash(text, p1, p2) != node |
|
908 | 906 | |
|
909 | 907 | def chunk(self, rev, df=None): |
|
910 | 908 | def loadcache(df): |
|
911 | 909 | if not df: |
|
912 | 910 | if self._inline: |
|
913 | 911 | df = self.opener(self.indexfile) |
|
914 | 912 | else: |
|
915 | 913 | df = self.opener(self.datafile) |
|
916 | 914 | df.seek(start) |
|
917 | 915 | self._chunkcache = (start, df.read(cache_length)) |
|
918 | 916 | |
|
919 | 917 | start, length = self.start(rev), self.length(rev) |
|
920 | 918 | if self._inline: |
|
921 | 919 | start += (rev + 1) * self._io.size |
|
922 | 920 | end = start + length |
|
923 | 921 | |
|
924 | 922 | offset = 0 |
|
925 | 923 | if not self._chunkcache: |
|
926 | 924 | cache_length = max(65536, length) |
|
927 | 925 | loadcache(df) |
|
928 | 926 | else: |
|
929 | 927 | cache_start = self._chunkcache[0] |
|
930 | 928 | cache_length = len(self._chunkcache[1]) |
|
931 | 929 | cache_end = cache_start + cache_length |
|
932 | 930 | if start >= cache_start and end <= cache_end: |
|
933 | 931 | # it is cached |
|
934 | 932 | offset = start - cache_start |
|
935 | 933 | else: |
|
936 | 934 | cache_length = max(65536, length) |
|
937 | 935 | loadcache(df) |
|
938 | 936 | |
|
939 | 937 | # avoid copying large chunks |
|
940 | 938 | c = self._chunkcache[1] |
|
941 | 939 | if cache_length != length: |
|
942 | 940 | c = c[offset:offset + length] |
|
943 | 941 | |
|
944 | 942 | return decompress(c) |
|
945 | 943 | |
|
946 | 944 | def revdiff(self, rev1, rev2): |
|
947 | 945 | """return or calculate a delta between two revisions""" |
|
948 | 946 | if rev1 + 1 == rev2 and self.base(rev1) == self.base(rev2): |
|
949 | 947 | return self.chunk(rev2) |
|
950 | 948 | |
|
951 | 949 | return mdiff.textdiff(self.revision(self.node(rev1)), |
|
952 | 950 | self.revision(self.node(rev2))) |
|
953 | 951 | |
|
954 | 952 | def revision(self, node): |
|
955 | 953 | """return an uncompressed revision of a given node""" |
|
956 | 954 | if node == nullid: |
|
957 | 955 | return "" |
|
958 | 956 | if self._cache and self._cache[0] == node: |
|
959 | 957 | return str(self._cache[2]) |
|
960 | 958 | |
|
961 | 959 | # look up what we need to read |
|
962 | 960 | text = None |
|
963 | 961 | rev = self.rev(node) |
|
964 | 962 | base = self.base(rev) |
|
965 | 963 | |
|
966 | 964 | # check rev flags |
|
967 | 965 | if self.index[rev][0] & 0xFFFF: |
|
968 | 966 | raise RevlogError(_('incompatible revision flag %x') % |
|
969 | 967 | (self.index[rev][0] & 0xFFFF)) |
|
970 | 968 | |
|
971 | 969 | df = None |
|
972 | 970 | |
|
973 | 971 | # do we have useful data cached? |
|
974 | 972 | if self._cache and self._cache[1] >= base and self._cache[1] < rev: |
|
975 | 973 | base = self._cache[1] |
|
976 | 974 | text = str(self._cache[2]) |
|
977 | 975 | self._loadindex(base, rev + 1) |
|
978 | 976 | if not self._inline and rev > base + 1: |
|
979 | 977 | df = self.opener(self.datafile) |
|
980 | 978 | else: |
|
981 | 979 | self._loadindex(base, rev + 1) |
|
982 | 980 | if not self._inline and rev > base: |
|
983 | 981 | df = self.opener(self.datafile) |
|
984 | 982 | text = self.chunk(base, df=df) |
|
985 | 983 | |
|
986 | 984 | bins = [self.chunk(r, df) for r in xrange(base + 1, rev + 1)] |
|
987 | 985 | text = mdiff.patches(text, bins) |
|
988 | 986 | p1, p2 = self.parents(node) |
|
989 | 987 | if node != hash(text, p1, p2): |
|
990 | 988 | raise RevlogError(_("integrity check failed on %s:%d") |
|
991 | 989 | % (self.datafile, rev)) |
|
992 | 990 | |
|
993 | 991 | self._cache = (node, rev, text) |
|
994 | 992 | return text |
|
995 | 993 | |
|
996 | 994 | def checkinlinesize(self, tr, fp=None): |
|
997 | 995 | if not self._inline: |
|
998 | 996 | return |
|
999 | 997 | if not fp: |
|
1000 | 998 | fp = self.opener(self.indexfile, 'r') |
|
1001 | 999 | fp.seek(0, 2) |
|
1002 | 1000 | size = fp.tell() |
|
1003 | 1001 | if size < 131072: |
|
1004 | 1002 | return |
|
1005 | 1003 | trinfo = tr.find(self.indexfile) |
|
1006 | 1004 | if trinfo == None: |
|
1007 | 1005 | raise RevlogError(_("%s not found in the transaction") |
|
1008 | 1006 | % self.indexfile) |
|
1009 | 1007 | |
|
1010 | 1008 | trindex = trinfo[2] |
|
1011 | 1009 | dataoff = self.start(trindex) |
|
1012 | 1010 | |
|
1013 | 1011 | tr.add(self.datafile, dataoff) |
|
1014 | 1012 | df = self.opener(self.datafile, 'w') |
|
1015 | 1013 | try: |
|
1016 | 1014 | calc = self._io.size |
|
1017 | 1015 | for r in self: |
|
1018 | 1016 | start = self.start(r) + (r + 1) * calc |
|
1019 | 1017 | length = self.length(r) |
|
1020 | 1018 | fp.seek(start) |
|
1021 | 1019 | d = fp.read(length) |
|
1022 | 1020 | df.write(d) |
|
1023 | 1021 | finally: |
|
1024 | 1022 | df.close() |
|
1025 | 1023 | |
|
1026 | 1024 | fp.close() |
|
1027 | 1025 | fp = self.opener(self.indexfile, 'w', atomictemp=True) |
|
1028 | 1026 | self.version &= ~(REVLOGNGINLINEDATA) |
|
1029 | 1027 | self._inline = False |
|
1030 | 1028 | for i in self: |
|
1031 | 1029 | e = self._io.packentry(self.index[i], self.node, self.version, i) |
|
1032 | 1030 | fp.write(e) |
|
1033 | 1031 | |
|
1034 | 1032 | # if we don't call rename, the temp file will never replace the |
|
1035 | 1033 | # real index |
|
1036 | 1034 | fp.rename() |
|
1037 | 1035 | |
|
1038 | 1036 | tr.replace(self.indexfile, trindex * calc) |
|
1039 | 1037 | self._chunkcache = None |
|
1040 | 1038 | |
|
1041 | 1039 | def addrevision(self, text, transaction, link, p1, p2, d=None): |
|
1042 | 1040 | """add a revision to the log |
|
1043 | 1041 | |
|
1044 | 1042 | text - the revision data to add |
|
1045 | 1043 | transaction - the transaction object used for rollback |
|
1046 | 1044 | link - the linkrev data to add |
|
1047 | 1045 | p1, p2 - the parent nodeids of the revision |
|
1048 | 1046 | d - an optional precomputed delta |
|
1049 | 1047 | """ |
|
1050 | 1048 | dfh = None |
|
1051 | 1049 | if not self._inline: |
|
1052 | 1050 | dfh = self.opener(self.datafile, "a") |
|
1053 | 1051 | ifh = self.opener(self.indexfile, "a+") |
|
1054 | 1052 | try: |
|
1055 | 1053 | return self._addrevision(text, transaction, link, p1, p2, d, ifh, dfh) |
|
1056 | 1054 | finally: |
|
1057 | 1055 | if dfh: |
|
1058 | 1056 | dfh.close() |
|
1059 | 1057 | ifh.close() |
|
1060 | 1058 | |
|
1061 | 1059 | def _addrevision(self, text, transaction, link, p1, p2, d, ifh, dfh): |
|
1062 | 1060 | node = hash(text, p1, p2) |
|
1063 | 1061 | if node in self.nodemap: |
|
1064 | 1062 | return node |
|
1065 | 1063 | |
|
1066 | 1064 | curr = len(self) |
|
1067 | 1065 | prev = curr - 1 |
|
1068 | 1066 | base = self.base(prev) |
|
1069 | 1067 | offset = self.end(prev) |
|
1070 | 1068 | |
|
1071 | 1069 | if curr: |
|
1072 | 1070 | if not d: |
|
1073 | 1071 | ptext = self.revision(self.node(prev)) |
|
1074 | 1072 | d = mdiff.textdiff(ptext, text) |
|
1075 | 1073 | data = compress(d) |
|
1076 | 1074 | l = len(data[1]) + len(data[0]) |
|
1077 | 1075 | dist = l + offset - self.start(base) |
|
1078 | 1076 | |
|
1079 | 1077 | # full versions are inserted when the needed deltas |
|
1080 | 1078 | # become comparable to the uncompressed text |
|
1081 | 1079 | if not curr or dist > len(text) * 2: |
|
1082 | 1080 | data = compress(text) |
|
1083 | 1081 | l = len(data[1]) + len(data[0]) |
|
1084 | 1082 | base = curr |
|
1085 | 1083 | |
|
1086 | 1084 | e = (offset_type(offset, 0), l, len(text), |
|
1087 | 1085 | base, link, self.rev(p1), self.rev(p2), node) |
|
1088 | 1086 | self.index.insert(-1, e) |
|
1089 | 1087 | self.nodemap[node] = curr |
|
1090 | 1088 | |
|
1091 | 1089 | entry = self._io.packentry(e, self.node, self.version, curr) |
|
1092 | 1090 | if not self._inline: |
|
1093 | 1091 | transaction.add(self.datafile, offset) |
|
1094 | 1092 | transaction.add(self.indexfile, curr * len(entry)) |
|
1095 | 1093 | if data[0]: |
|
1096 | 1094 | dfh.write(data[0]) |
|
1097 | 1095 | dfh.write(data[1]) |
|
1098 | 1096 | dfh.flush() |
|
1099 | 1097 | ifh.write(entry) |
|
1100 | 1098 | else: |
|
1101 | 1099 | offset += curr * self._io.size |
|
1102 | 1100 | transaction.add(self.indexfile, offset, curr) |
|
1103 | 1101 | ifh.write(entry) |
|
1104 | 1102 | ifh.write(data[0]) |
|
1105 | 1103 | ifh.write(data[1]) |
|
1106 | 1104 | self.checkinlinesize(transaction, ifh) |
|
1107 | 1105 | |
|
1108 | 1106 | self._cache = (node, curr, text) |
|
1109 | 1107 | return node |
|
1110 | 1108 | |
|
1111 | 1109 | def ancestor(self, a, b): |
|
1112 | 1110 | """calculate the least common ancestor of nodes a and b""" |
|
1113 | 1111 | |
|
1114 | 1112 | def parents(rev): |
|
1115 | 1113 | return [p for p in self.parentrevs(rev) if p != nullrev] |
|
1116 | 1114 | |
|
1117 | 1115 | c = ancestor.ancestor(self.rev(a), self.rev(b), parents) |
|
1118 | 1116 | if c is None: |
|
1119 | 1117 | return nullid |
|
1120 | 1118 | |
|
1121 | 1119 | return self.node(c) |
|
1122 | 1120 | |
|
1123 | 1121 | def group(self, nodelist, lookup, infocollect=None): |
|
1124 | 1122 | """calculate a delta group |
|
1125 | 1123 | |
|
1126 | 1124 | Given a list of changeset revs, return a set of deltas and |
|
1127 | 1125 | metadata corresponding to nodes. the first delta is |
|
1128 | 1126 | parent(nodes[0]) -> nodes[0] the receiver is guaranteed to |
|
1129 | 1127 | have this parent as it has all history before these |
|
1130 | 1128 | changesets. parent is parent[0] |
|
1131 | 1129 | """ |
|
1132 | 1130 | revs = [self.rev(n) for n in nodelist] |
|
1133 | 1131 | |
|
1134 | 1132 | # if we don't have any revisions touched by these changesets, bail |
|
1135 | 1133 | if not revs: |
|
1136 | 1134 | yield changegroup.closechunk() |
|
1137 | 1135 | return |
|
1138 | 1136 | |
|
1139 | 1137 | # add the parent of the first rev |
|
1140 | 1138 | p = self.parents(self.node(revs[0]))[0] |
|
1141 | 1139 | revs.insert(0, self.rev(p)) |
|
1142 | 1140 | |
|
1143 | 1141 | # build deltas |
|
1144 | 1142 | for d in xrange(0, len(revs) - 1): |
|
1145 | 1143 | a, b = revs[d], revs[d + 1] |
|
1146 | 1144 | nb = self.node(b) |
|
1147 | 1145 | |
|
1148 | 1146 | if infocollect is not None: |
|
1149 | 1147 | infocollect(nb) |
|
1150 | 1148 | |
|
1151 | 1149 | p = self.parents(nb) |
|
1152 | 1150 | meta = nb + p[0] + p[1] + lookup(nb) |
|
1153 | 1151 | if a == -1: |
|
1154 | 1152 | d = self.revision(nb) |
|
1155 | 1153 | meta += mdiff.trivialdiffheader(len(d)) |
|
1156 | 1154 | else: |
|
1157 | 1155 | d = self.revdiff(a, b) |
|
1158 | 1156 | yield changegroup.chunkheader(len(meta) + len(d)) |
|
1159 | 1157 | yield meta |
|
1160 | 1158 | if len(d) > 2**20: |
|
1161 | 1159 | pos = 0 |
|
1162 | 1160 | while pos < len(d): |
|
1163 | 1161 | pos2 = pos + 2 ** 18 |
|
1164 | 1162 | yield d[pos:pos2] |
|
1165 | 1163 | pos = pos2 |
|
1166 | 1164 | else: |
|
1167 | 1165 | yield d |
|
1168 | 1166 | |
|
1169 | 1167 | yield changegroup.closechunk() |
|
1170 | 1168 | |
|
1171 | 1169 | def addgroup(self, revs, linkmapper, transaction): |
|
1172 | 1170 | """ |
|
1173 | 1171 | add a delta group |
|
1174 | 1172 | |
|
1175 | 1173 | given a set of deltas, add them to the revision log. the |
|
1176 | 1174 | first delta is against its parent, which should be in our |
|
1177 | 1175 | log, the rest are against the previous delta. |
|
1178 | 1176 | """ |
|
1179 | 1177 | |
|
1180 | 1178 | #track the base of the current delta log |
|
1181 | 1179 | r = len(self) |
|
1182 | 1180 | t = r - 1 |
|
1183 | 1181 | node = None |
|
1184 | 1182 | |
|
1185 | 1183 | base = prev = nullrev |
|
1186 | 1184 | start = end = textlen = 0 |
|
1187 | 1185 | if r: |
|
1188 | 1186 | end = self.end(t) |
|
1189 | 1187 | |
|
1190 | 1188 | ifh = self.opener(self.indexfile, "a+") |
|
1191 | 1189 | isize = r * self._io.size |
|
1192 | 1190 | if self._inline: |
|
1193 | 1191 | transaction.add(self.indexfile, end + isize, r) |
|
1194 | 1192 | dfh = None |
|
1195 | 1193 | else: |
|
1196 | 1194 | transaction.add(self.indexfile, isize, r) |
|
1197 | 1195 | transaction.add(self.datafile, end) |
|
1198 | 1196 | dfh = self.opener(self.datafile, "a") |
|
1199 | 1197 | |
|
1200 | 1198 | try: |
|
1201 | 1199 | # loop through our set of deltas |
|
1202 | 1200 | chain = None |
|
1203 | 1201 | for chunk in revs: |
|
1204 | 1202 | node, p1, p2, cs = struct.unpack("20s20s20s20s", chunk[:80]) |
|
1205 | 1203 | link = linkmapper(cs) |
|
1206 | 1204 | if node in self.nodemap: |
|
1207 | 1205 | # this can happen if two branches make the same change |
|
1208 | 1206 | chain = node |
|
1209 | 1207 | continue |
|
1210 | 1208 | delta = buffer(chunk, 80) |
|
1211 | 1209 | del chunk |
|
1212 | 1210 | |
|
1213 | 1211 | for p in (p1, p2): |
|
1214 | 1212 | if not p in self.nodemap: |
|
1215 | 1213 | raise LookupError(p, self.indexfile, _('unknown parent')) |
|
1216 | 1214 | |
|
1217 | 1215 | if not chain: |
|
1218 | 1216 | # retrieve the parent revision of the delta chain |
|
1219 | 1217 | chain = p1 |
|
1220 | 1218 | if not chain in self.nodemap: |
|
1221 | 1219 | raise LookupError(chain, self.indexfile, _('unknown base')) |
|
1222 | 1220 | |
|
1223 | 1221 | # full versions are inserted when the needed deltas become |
|
1224 | 1222 | # comparable to the uncompressed text or when the previous |
|
1225 | 1223 | # version is not the one we have a delta against. We use |
|
1226 | 1224 | # the size of the previous full rev as a proxy for the |
|
1227 | 1225 | # current size. |
|
1228 | 1226 | |
|
1229 | 1227 | if chain == prev: |
|
1230 | 1228 | cdelta = compress(delta) |
|
1231 | 1229 | cdeltalen = len(cdelta[0]) + len(cdelta[1]) |
|
1232 | 1230 | textlen = mdiff.patchedsize(textlen, delta) |
|
1233 | 1231 | |
|
1234 | 1232 | if chain != prev or (end - start + cdeltalen) > textlen * 2: |
|
1235 | 1233 | # flush our writes here so we can read it in revision |
|
1236 | 1234 | if dfh: |
|
1237 | 1235 | dfh.flush() |
|
1238 | 1236 | ifh.flush() |
|
1239 | 1237 | text = self.revision(chain) |
|
1240 | 1238 | if len(text) == 0: |
|
1241 | 1239 | # skip over trivial delta header |
|
1242 | 1240 | text = buffer(delta, 12) |
|
1243 | 1241 | else: |
|
1244 | 1242 | text = mdiff.patches(text, [delta]) |
|
1245 | 1243 | del delta |
|
1246 | 1244 | chk = self._addrevision(text, transaction, link, p1, p2, None, |
|
1247 | 1245 | ifh, dfh) |
|
1248 | 1246 | if not dfh and not self._inline: |
|
1249 | 1247 | # addrevision switched from inline to conventional |
|
1250 | 1248 | # reopen the index |
|
1251 | 1249 | dfh = self.opener(self.datafile, "a") |
|
1252 | 1250 | ifh = self.opener(self.indexfile, "a") |
|
1253 | 1251 | if chk != node: |
|
1254 | 1252 | raise RevlogError(_("consistency error adding group")) |
|
1255 | 1253 | textlen = len(text) |
|
1256 | 1254 | else: |
|
1257 | 1255 | e = (offset_type(end, 0), cdeltalen, textlen, base, |
|
1258 | 1256 | link, self.rev(p1), self.rev(p2), node) |
|
1259 | 1257 | self.index.insert(-1, e) |
|
1260 | 1258 | self.nodemap[node] = r |
|
1261 | 1259 | entry = self._io.packentry(e, self.node, self.version, r) |
|
1262 | 1260 | if self._inline: |
|
1263 | 1261 | ifh.write(entry) |
|
1264 | 1262 | ifh.write(cdelta[0]) |
|
1265 | 1263 | ifh.write(cdelta[1]) |
|
1266 | 1264 | self.checkinlinesize(transaction, ifh) |
|
1267 | 1265 | if not self._inline: |
|
1268 | 1266 | dfh = self.opener(self.datafile, "a") |
|
1269 | 1267 | ifh = self.opener(self.indexfile, "a") |
|
1270 | 1268 | else: |
|
1271 | 1269 | dfh.write(cdelta[0]) |
|
1272 | 1270 | dfh.write(cdelta[1]) |
|
1273 | 1271 | ifh.write(entry) |
|
1274 | 1272 | |
|
1275 | 1273 | t, r, chain, prev = r, r + 1, node, node |
|
1276 | 1274 | base = self.base(t) |
|
1277 | 1275 | start = self.start(base) |
|
1278 | 1276 | end = self.end(t) |
|
1279 | 1277 | finally: |
|
1280 | 1278 | if dfh: |
|
1281 | 1279 | dfh.close() |
|
1282 | 1280 | ifh.close() |
|
1283 | 1281 | |
|
1284 | 1282 | return node |
|
1285 | 1283 | |
|
1286 | 1284 | def strip(self, minlink, transaction): |
|
1287 | 1285 | """truncate the revlog on the first revision with a linkrev >= minlink |
|
1288 | 1286 | |
|
1289 | 1287 | This function is called when we're stripping revision minlink and |
|
1290 | 1288 | its descendants from the repository. |
|
1291 | 1289 | |
|
1292 | 1290 | We have to remove all revisions with linkrev >= minlink, because |
|
1293 | 1291 | the equivalent changelog revisions will be renumbered after the |
|
1294 | 1292 | strip. |
|
1295 | 1293 | |
|
1296 | 1294 | So we truncate the revlog on the first of these revisions, and |
|
1297 | 1295 | trust that the caller has saved the revisions that shouldn't be |
|
1298 | 1296 | removed and that it'll readd them after this truncation. |
|
1299 | 1297 | """ |
|
1300 | 1298 | if len(self) == 0: |
|
1301 | 1299 | return |
|
1302 | 1300 | |
|
1303 | 1301 | if isinstance(self.index, lazyindex): |
|
1304 | 1302 | self._loadindexmap() |
|
1305 | 1303 | |
|
1306 | 1304 | for rev in self: |
|
1307 | 1305 | if self.index[rev][4] >= minlink: |
|
1308 | 1306 | break |
|
1309 | 1307 | else: |
|
1310 | 1308 | return |
|
1311 | 1309 | |
|
1312 | 1310 | # first truncate the files on disk |
|
1313 | 1311 | end = self.start(rev) |
|
1314 | 1312 | if not self._inline: |
|
1315 | 1313 | transaction.add(self.datafile, end) |
|
1316 | 1314 | end = rev * self._io.size |
|
1317 | 1315 | else: |
|
1318 | 1316 | end += rev * self._io.size |
|
1319 | 1317 | |
|
1320 | 1318 | transaction.add(self.indexfile, end) |
|
1321 | 1319 | |
|
1322 | 1320 | # then reset internal state in memory to forget those revisions |
|
1323 | 1321 | self._cache = None |
|
1324 | 1322 | self._chunkcache = None |
|
1325 | 1323 | for x in xrange(rev, len(self)): |
|
1326 | 1324 | del self.nodemap[self.node(x)] |
|
1327 | 1325 | |
|
1328 | 1326 | del self.index[rev:-1] |
|
1329 | 1327 | |
|
1330 | 1328 | def checksize(self): |
|
1331 | 1329 | expected = 0 |
|
1332 | 1330 | if len(self): |
|
1333 | 1331 | expected = max(0, self.end(len(self) - 1)) |
|
1334 | 1332 | |
|
1335 | 1333 | try: |
|
1336 | 1334 | f = self.opener(self.datafile) |
|
1337 | 1335 | f.seek(0, 2) |
|
1338 | 1336 | actual = f.tell() |
|
1339 | 1337 | dd = actual - expected |
|
1340 | 1338 | except IOError, inst: |
|
1341 | 1339 | if inst.errno != errno.ENOENT: |
|
1342 | 1340 | raise |
|
1343 | 1341 | dd = 0 |
|
1344 | 1342 | |
|
1345 | 1343 | try: |
|
1346 | 1344 | f = self.opener(self.indexfile) |
|
1347 | 1345 | f.seek(0, 2) |
|
1348 | 1346 | actual = f.tell() |
|
1349 | 1347 | s = self._io.size |
|
1350 | 1348 | i = max(0, actual / s) |
|
1351 | 1349 | di = actual - (i * s) |
|
1352 | 1350 | if self._inline: |
|
1353 | 1351 | databytes = 0 |
|
1354 | 1352 | for r in self: |
|
1355 | 1353 | databytes += max(0, self.length(r)) |
|
1356 | 1354 | dd = 0 |
|
1357 | 1355 | di = actual - len(self) * s - databytes |
|
1358 | 1356 | except IOError, inst: |
|
1359 | 1357 | if inst.errno != errno.ENOENT: |
|
1360 | 1358 | raise |
|
1361 | 1359 | di = 0 |
|
1362 | 1360 | |
|
1363 | 1361 | return (dd, di) |
|
1364 | 1362 | |
|
1365 | 1363 | def files(self): |
|
1366 | 1364 | res = [ self.indexfile ] |
|
1367 | 1365 | if not self._inline: |
|
1368 | 1366 | res.append(self.datafile) |
|
1369 | 1367 | return res |
@@ -1,1477 +1,1475 b'' | |||
|
1 | """ | |
|
2 | util.py - Mercurial utility functions and platform specfic implementations | |
|
3 | ||
|
4 | Copyright 2005 K. Thananchayan <thananck@yahoo.com> | |
|
5 | Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
|
6 | Copyright 2006 Vadim Gelfer <vadim.gelfer@gmail.com> | |
|
7 | ||
|
8 | This software may be used and distributed according to the terms of the | |
|
9 | GNU General Public License version 2, incorporated herein by reference. | |
|
10 | ||
|
11 | This contains helper routines that are independent of the SCM core and hide | |
|
12 | platform-specific details from the core. | |
|
13 | """ | |
|
1 | # util.py - Mercurial utility functions and platform specfic implementations | |
|
2 | # | |
|
3 | # Copyright 2005 K. Thananchayan <thananck@yahoo.com> | |
|
4 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
|
5 | # Copyright 2006 Vadim Gelfer <vadim.gelfer@gmail.com> | |
|
6 | # | |
|
7 | # This software may be used and distributed according to the terms of the | |
|
8 | # GNU General Public License version 2, incorporated herein by reference. | |
|
9 | # | |
|
10 | # This contains helper routines that are independent of the SCM core and hide | |
|
11 | # platform-specific details from the core. | |
|
14 | 12 | |
|
15 | 13 | from i18n import _ |
|
16 | 14 | import cStringIO, errno, re, shutil, sys, tempfile, traceback, error |
|
17 | 15 | import os, stat, threading, time, calendar, glob, osutil |
|
18 | 16 | import imp |
|
19 | 17 | |
|
20 | 18 | # Python compatibility |
|
21 | 19 | |
|
22 | 20 | _md5 = None |
|
23 | 21 | def md5(s): |
|
24 | 22 | global _md5 |
|
25 | 23 | if _md5 is None: |
|
26 | 24 | try: |
|
27 | 25 | import hashlib |
|
28 | 26 | _md5 = hashlib.md5 |
|
29 | 27 | except ImportError: |
|
30 | 28 | import md5 |
|
31 | 29 | _md5 = md5.md5 |
|
32 | 30 | return _md5(s) |
|
33 | 31 | |
|
34 | 32 | _sha1 = None |
|
35 | 33 | def sha1(s): |
|
36 | 34 | global _sha1 |
|
37 | 35 | if _sha1 is None: |
|
38 | 36 | try: |
|
39 | 37 | import hashlib |
|
40 | 38 | _sha1 = hashlib.sha1 |
|
41 | 39 | except ImportError: |
|
42 | 40 | import sha |
|
43 | 41 | _sha1 = sha.sha |
|
44 | 42 | return _sha1(s) |
|
45 | 43 | |
|
46 | 44 | try: |
|
47 | 45 | import subprocess |
|
48 | 46 | subprocess.Popen # trigger ImportError early |
|
49 | 47 | closefds = os.name == 'posix' |
|
50 | 48 | def popen2(cmd, mode='t', bufsize=-1): |
|
51 | 49 | p = subprocess.Popen(cmd, shell=True, bufsize=bufsize, |
|
52 | 50 | close_fds=closefds, |
|
53 | 51 | stdin=subprocess.PIPE, stdout=subprocess.PIPE) |
|
54 | 52 | return p.stdin, p.stdout |
|
55 | 53 | def popen3(cmd, mode='t', bufsize=-1): |
|
56 | 54 | p = subprocess.Popen(cmd, shell=True, bufsize=bufsize, |
|
57 | 55 | close_fds=closefds, |
|
58 | 56 | stdin=subprocess.PIPE, stdout=subprocess.PIPE, |
|
59 | 57 | stderr=subprocess.PIPE) |
|
60 | 58 | return p.stdin, p.stdout, p.stderr |
|
61 | 59 | def Popen3(cmd, capturestderr=False, bufsize=-1): |
|
62 | 60 | stderr = capturestderr and subprocess.PIPE or None |
|
63 | 61 | p = subprocess.Popen(cmd, shell=True, bufsize=bufsize, |
|
64 | 62 | close_fds=closefds, |
|
65 | 63 | stdin=subprocess.PIPE, stdout=subprocess.PIPE, |
|
66 | 64 | stderr=stderr) |
|
67 | 65 | p.fromchild = p.stdout |
|
68 | 66 | p.tochild = p.stdin |
|
69 | 67 | p.childerr = p.stderr |
|
70 | 68 | return p |
|
71 | 69 | except ImportError: |
|
72 | 70 | subprocess = None |
|
73 | 71 | from popen2 import Popen3 |
|
74 | 72 | popen2 = os.popen2 |
|
75 | 73 | popen3 = os.popen3 |
|
76 | 74 | |
|
77 | 75 | |
|
78 | 76 | def version(): |
|
79 | 77 | """Return version information if available.""" |
|
80 | 78 | try: |
|
81 | 79 | import __version__ |
|
82 | 80 | return __version__.version |
|
83 | 81 | except ImportError: |
|
84 | 82 | return 'unknown' |
|
85 | 83 | |
|
86 | 84 | # used by parsedate |
|
87 | 85 | defaultdateformats = ( |
|
88 | 86 | '%Y-%m-%d %H:%M:%S', |
|
89 | 87 | '%Y-%m-%d %I:%M:%S%p', |
|
90 | 88 | '%Y-%m-%d %H:%M', |
|
91 | 89 | '%Y-%m-%d %I:%M%p', |
|
92 | 90 | '%Y-%m-%d', |
|
93 | 91 | '%m-%d', |
|
94 | 92 | '%m/%d', |
|
95 | 93 | '%m/%d/%y', |
|
96 | 94 | '%m/%d/%Y', |
|
97 | 95 | '%a %b %d %H:%M:%S %Y', |
|
98 | 96 | '%a %b %d %I:%M:%S%p %Y', |
|
99 | 97 | '%a, %d %b %Y %H:%M:%S', # GNU coreutils "/bin/date --rfc-2822" |
|
100 | 98 | '%b %d %H:%M:%S %Y', |
|
101 | 99 | '%b %d %I:%M:%S%p %Y', |
|
102 | 100 | '%b %d %H:%M:%S', |
|
103 | 101 | '%b %d %I:%M:%S%p', |
|
104 | 102 | '%b %d %H:%M', |
|
105 | 103 | '%b %d %I:%M%p', |
|
106 | 104 | '%b %d %Y', |
|
107 | 105 | '%b %d', |
|
108 | 106 | '%H:%M:%S', |
|
109 | 107 | '%I:%M:%SP', |
|
110 | 108 | '%H:%M', |
|
111 | 109 | '%I:%M%p', |
|
112 | 110 | ) |
|
113 | 111 | |
|
114 | 112 | extendeddateformats = defaultdateformats + ( |
|
115 | 113 | "%Y", |
|
116 | 114 | "%Y-%m", |
|
117 | 115 | "%b", |
|
118 | 116 | "%b %Y", |
|
119 | 117 | ) |
|
120 | 118 | |
|
121 | 119 | def cachefunc(func): |
|
122 | 120 | '''cache the result of function calls''' |
|
123 | 121 | # XXX doesn't handle keywords args |
|
124 | 122 | cache = {} |
|
125 | 123 | if func.func_code.co_argcount == 1: |
|
126 | 124 | # we gain a small amount of time because |
|
127 | 125 | # we don't need to pack/unpack the list |
|
128 | 126 | def f(arg): |
|
129 | 127 | if arg not in cache: |
|
130 | 128 | cache[arg] = func(arg) |
|
131 | 129 | return cache[arg] |
|
132 | 130 | else: |
|
133 | 131 | def f(*args): |
|
134 | 132 | if args not in cache: |
|
135 | 133 | cache[args] = func(*args) |
|
136 | 134 | return cache[args] |
|
137 | 135 | |
|
138 | 136 | return f |
|
139 | 137 | |
|
140 | 138 | class propertycache(object): |
|
141 | 139 | def __init__(self, func): |
|
142 | 140 | self.func = func |
|
143 | 141 | self.name = func.__name__ |
|
144 | 142 | def __get__(self, obj, type=None): |
|
145 | 143 | result = self.func(obj) |
|
146 | 144 | setattr(obj, self.name, result) |
|
147 | 145 | return result |
|
148 | 146 | |
|
149 | 147 | def pipefilter(s, cmd): |
|
150 | 148 | '''filter string S through command CMD, returning its output''' |
|
151 | 149 | (pin, pout) = popen2(cmd, 'b') |
|
152 | 150 | def writer(): |
|
153 | 151 | try: |
|
154 | 152 | pin.write(s) |
|
155 | 153 | pin.close() |
|
156 | 154 | except IOError, inst: |
|
157 | 155 | if inst.errno != errno.EPIPE: |
|
158 | 156 | raise |
|
159 | 157 | |
|
160 | 158 | # we should use select instead on UNIX, but this will work on most |
|
161 | 159 | # systems, including Windows |
|
162 | 160 | w = threading.Thread(target=writer) |
|
163 | 161 | w.start() |
|
164 | 162 | f = pout.read() |
|
165 | 163 | pout.close() |
|
166 | 164 | w.join() |
|
167 | 165 | return f |
|
168 | 166 | |
|
169 | 167 | def tempfilter(s, cmd): |
|
170 | 168 | '''filter string S through a pair of temporary files with CMD. |
|
171 | 169 | CMD is used as a template to create the real command to be run, |
|
172 | 170 | with the strings INFILE and OUTFILE replaced by the real names of |
|
173 | 171 | the temporary files generated.''' |
|
174 | 172 | inname, outname = None, None |
|
175 | 173 | try: |
|
176 | 174 | infd, inname = tempfile.mkstemp(prefix='hg-filter-in-') |
|
177 | 175 | fp = os.fdopen(infd, 'wb') |
|
178 | 176 | fp.write(s) |
|
179 | 177 | fp.close() |
|
180 | 178 | outfd, outname = tempfile.mkstemp(prefix='hg-filter-out-') |
|
181 | 179 | os.close(outfd) |
|
182 | 180 | cmd = cmd.replace('INFILE', inname) |
|
183 | 181 | cmd = cmd.replace('OUTFILE', outname) |
|
184 | 182 | code = os.system(cmd) |
|
185 | 183 | if sys.platform == 'OpenVMS' and code & 1: |
|
186 | 184 | code = 0 |
|
187 | 185 | if code: raise Abort(_("command '%s' failed: %s") % |
|
188 | 186 | (cmd, explain_exit(code))) |
|
189 | 187 | return open(outname, 'rb').read() |
|
190 | 188 | finally: |
|
191 | 189 | try: |
|
192 | 190 | if inname: os.unlink(inname) |
|
193 | 191 | except: pass |
|
194 | 192 | try: |
|
195 | 193 | if outname: os.unlink(outname) |
|
196 | 194 | except: pass |
|
197 | 195 | |
|
198 | 196 | filtertable = { |
|
199 | 197 | 'tempfile:': tempfilter, |
|
200 | 198 | 'pipe:': pipefilter, |
|
201 | 199 | } |
|
202 | 200 | |
|
203 | 201 | def filter(s, cmd): |
|
204 | 202 | "filter a string through a command that transforms its input to its output" |
|
205 | 203 | for name, fn in filtertable.iteritems(): |
|
206 | 204 | if cmd.startswith(name): |
|
207 | 205 | return fn(s, cmd[len(name):].lstrip()) |
|
208 | 206 | return pipefilter(s, cmd) |
|
209 | 207 | |
|
210 | 208 | def binary(s): |
|
211 | 209 | """return true if a string is binary data""" |
|
212 | 210 | return bool(s and '\0' in s) |
|
213 | 211 | |
|
214 | 212 | def increasingchunks(source, min=1024, max=65536): |
|
215 | 213 | '''return no less than min bytes per chunk while data remains, |
|
216 | 214 | doubling min after each chunk until it reaches max''' |
|
217 | 215 | def log2(x): |
|
218 | 216 | if not x: |
|
219 | 217 | return 0 |
|
220 | 218 | i = 0 |
|
221 | 219 | while x: |
|
222 | 220 | x >>= 1 |
|
223 | 221 | i += 1 |
|
224 | 222 | return i - 1 |
|
225 | 223 | |
|
226 | 224 | buf = [] |
|
227 | 225 | blen = 0 |
|
228 | 226 | for chunk in source: |
|
229 | 227 | buf.append(chunk) |
|
230 | 228 | blen += len(chunk) |
|
231 | 229 | if blen >= min: |
|
232 | 230 | if min < max: |
|
233 | 231 | min = min << 1 |
|
234 | 232 | nmin = 1 << log2(blen) |
|
235 | 233 | if nmin > min: |
|
236 | 234 | min = nmin |
|
237 | 235 | if min > max: |
|
238 | 236 | min = max |
|
239 | 237 | yield ''.join(buf) |
|
240 | 238 | blen = 0 |
|
241 | 239 | buf = [] |
|
242 | 240 | if buf: |
|
243 | 241 | yield ''.join(buf) |
|
244 | 242 | |
|
245 | 243 | Abort = error.Abort |
|
246 | 244 | |
|
247 | 245 | def always(fn): return True |
|
248 | 246 | def never(fn): return False |
|
249 | 247 | |
|
250 | 248 | def patkind(name, default): |
|
251 | 249 | """Split a string into an optional pattern kind prefix and the |
|
252 | 250 | actual pattern.""" |
|
253 | 251 | for prefix in 're', 'glob', 'path', 'relglob', 'relpath', 'relre': |
|
254 | 252 | if name.startswith(prefix + ':'): return name.split(':', 1) |
|
255 | 253 | return default, name |
|
256 | 254 | |
|
257 | 255 | def globre(pat, head='^', tail='$'): |
|
258 | 256 | "convert a glob pattern into a regexp" |
|
259 | 257 | i, n = 0, len(pat) |
|
260 | 258 | res = '' |
|
261 | 259 | group = 0 |
|
262 | 260 | def peek(): return i < n and pat[i] |
|
263 | 261 | while i < n: |
|
264 | 262 | c = pat[i] |
|
265 | 263 | i = i+1 |
|
266 | 264 | if c == '*': |
|
267 | 265 | if peek() == '*': |
|
268 | 266 | i += 1 |
|
269 | 267 | res += '.*' |
|
270 | 268 | else: |
|
271 | 269 | res += '[^/]*' |
|
272 | 270 | elif c == '?': |
|
273 | 271 | res += '.' |
|
274 | 272 | elif c == '[': |
|
275 | 273 | j = i |
|
276 | 274 | if j < n and pat[j] in '!]': |
|
277 | 275 | j += 1 |
|
278 | 276 | while j < n and pat[j] != ']': |
|
279 | 277 | j += 1 |
|
280 | 278 | if j >= n: |
|
281 | 279 | res += '\\[' |
|
282 | 280 | else: |
|
283 | 281 | stuff = pat[i:j].replace('\\','\\\\') |
|
284 | 282 | i = j + 1 |
|
285 | 283 | if stuff[0] == '!': |
|
286 | 284 | stuff = '^' + stuff[1:] |
|
287 | 285 | elif stuff[0] == '^': |
|
288 | 286 | stuff = '\\' + stuff |
|
289 | 287 | res = '%s[%s]' % (res, stuff) |
|
290 | 288 | elif c == '{': |
|
291 | 289 | group += 1 |
|
292 | 290 | res += '(?:' |
|
293 | 291 | elif c == '}' and group: |
|
294 | 292 | res += ')' |
|
295 | 293 | group -= 1 |
|
296 | 294 | elif c == ',' and group: |
|
297 | 295 | res += '|' |
|
298 | 296 | elif c == '\\': |
|
299 | 297 | p = peek() |
|
300 | 298 | if p: |
|
301 | 299 | i += 1 |
|
302 | 300 | res += re.escape(p) |
|
303 | 301 | else: |
|
304 | 302 | res += re.escape(c) |
|
305 | 303 | else: |
|
306 | 304 | res += re.escape(c) |
|
307 | 305 | return head + res + tail |
|
308 | 306 | |
|
309 | 307 | _globchars = {'[': 1, '{': 1, '*': 1, '?': 1} |
|
310 | 308 | |
|
311 | 309 | def pathto(root, n1, n2): |
|
312 | 310 | '''return the relative path from one place to another. |
|
313 | 311 | root should use os.sep to separate directories |
|
314 | 312 | n1 should use os.sep to separate directories |
|
315 | 313 | n2 should use "/" to separate directories |
|
316 | 314 | returns an os.sep-separated path. |
|
317 | 315 | |
|
318 | 316 | If n1 is a relative path, it's assumed it's |
|
319 | 317 | relative to root. |
|
320 | 318 | n2 should always be relative to root. |
|
321 | 319 | ''' |
|
322 | 320 | if not n1: return localpath(n2) |
|
323 | 321 | if os.path.isabs(n1): |
|
324 | 322 | if os.path.splitdrive(root)[0] != os.path.splitdrive(n1)[0]: |
|
325 | 323 | return os.path.join(root, localpath(n2)) |
|
326 | 324 | n2 = '/'.join((pconvert(root), n2)) |
|
327 | 325 | a, b = splitpath(n1), n2.split('/') |
|
328 | 326 | a.reverse() |
|
329 | 327 | b.reverse() |
|
330 | 328 | while a and b and a[-1] == b[-1]: |
|
331 | 329 | a.pop() |
|
332 | 330 | b.pop() |
|
333 | 331 | b.reverse() |
|
334 | 332 | return os.sep.join((['..'] * len(a)) + b) or '.' |
|
335 | 333 | |
|
336 | 334 | def canonpath(root, cwd, myname): |
|
337 | 335 | """return the canonical path of myname, given cwd and root""" |
|
338 | 336 | if root == os.sep: |
|
339 | 337 | rootsep = os.sep |
|
340 | 338 | elif endswithsep(root): |
|
341 | 339 | rootsep = root |
|
342 | 340 | else: |
|
343 | 341 | rootsep = root + os.sep |
|
344 | 342 | name = myname |
|
345 | 343 | if not os.path.isabs(name): |
|
346 | 344 | name = os.path.join(root, cwd, name) |
|
347 | 345 | name = os.path.normpath(name) |
|
348 | 346 | audit_path = path_auditor(root) |
|
349 | 347 | if name != rootsep and name.startswith(rootsep): |
|
350 | 348 | name = name[len(rootsep):] |
|
351 | 349 | audit_path(name) |
|
352 | 350 | return pconvert(name) |
|
353 | 351 | elif name == root: |
|
354 | 352 | return '' |
|
355 | 353 | else: |
|
356 | 354 | # Determine whether `name' is in the hierarchy at or beneath `root', |
|
357 | 355 | # by iterating name=dirname(name) until that causes no change (can't |
|
358 | 356 | # check name == '/', because that doesn't work on windows). For each |
|
359 | 357 | # `name', compare dev/inode numbers. If they match, the list `rel' |
|
360 | 358 | # holds the reversed list of components making up the relative file |
|
361 | 359 | # name we want. |
|
362 | 360 | root_st = os.stat(root) |
|
363 | 361 | rel = [] |
|
364 | 362 | while True: |
|
365 | 363 | try: |
|
366 | 364 | name_st = os.stat(name) |
|
367 | 365 | except OSError: |
|
368 | 366 | break |
|
369 | 367 | if samestat(name_st, root_st): |
|
370 | 368 | if not rel: |
|
371 | 369 | # name was actually the same as root (maybe a symlink) |
|
372 | 370 | return '' |
|
373 | 371 | rel.reverse() |
|
374 | 372 | name = os.path.join(*rel) |
|
375 | 373 | audit_path(name) |
|
376 | 374 | return pconvert(name) |
|
377 | 375 | dirname, basename = os.path.split(name) |
|
378 | 376 | rel.append(basename) |
|
379 | 377 | if dirname == name: |
|
380 | 378 | break |
|
381 | 379 | name = dirname |
|
382 | 380 | |
|
383 | 381 | raise Abort('%s not under root' % myname) |
|
384 | 382 | |
|
385 | 383 | def matcher(canonroot, cwd='', names=[], inc=[], exc=[], src=None, dflt_pat='glob'): |
|
386 | 384 | """build a function to match a set of file patterns |
|
387 | 385 | |
|
388 | 386 | arguments: |
|
389 | 387 | canonroot - the canonical root of the tree you're matching against |
|
390 | 388 | cwd - the current working directory, if relevant |
|
391 | 389 | names - patterns to find |
|
392 | 390 | inc - patterns to include |
|
393 | 391 | exc - patterns to exclude |
|
394 | 392 | dflt_pat - if a pattern in names has no explicit type, assume this one |
|
395 | 393 | src - where these patterns came from (e.g. .hgignore) |
|
396 | 394 | |
|
397 | 395 | a pattern is one of: |
|
398 | 396 | 'glob:<glob>' - a glob relative to cwd |
|
399 | 397 | 're:<regexp>' - a regular expression |
|
400 | 398 | 'path:<path>' - a path relative to canonroot |
|
401 | 399 | 'relglob:<glob>' - an unrooted glob (*.c matches C files in all dirs) |
|
402 | 400 | 'relpath:<path>' - a path relative to cwd |
|
403 | 401 | 'relre:<regexp>' - a regexp that doesn't have to match the start of a name |
|
404 | 402 | '<something>' - one of the cases above, selected by the dflt_pat argument |
|
405 | 403 | |
|
406 | 404 | returns: |
|
407 | 405 | a 3-tuple containing |
|
408 | 406 | - list of roots (places where one should start a recursive walk of the fs); |
|
409 | 407 | this often matches the explicit non-pattern names passed in, but also |
|
410 | 408 | includes the initial part of glob: patterns that has no glob characters |
|
411 | 409 | - a bool match(filename) function |
|
412 | 410 | - a bool indicating if any patterns were passed in |
|
413 | 411 | """ |
|
414 | 412 | |
|
415 | 413 | # a common case: no patterns at all |
|
416 | 414 | if not names and not inc and not exc: |
|
417 | 415 | return [], always, False |
|
418 | 416 | |
|
419 | 417 | def contains_glob(name): |
|
420 | 418 | for c in name: |
|
421 | 419 | if c in _globchars: return True |
|
422 | 420 | return False |
|
423 | 421 | |
|
424 | 422 | def regex(kind, name, tail): |
|
425 | 423 | '''convert a pattern into a regular expression''' |
|
426 | 424 | if not name: |
|
427 | 425 | return '' |
|
428 | 426 | if kind == 're': |
|
429 | 427 | return name |
|
430 | 428 | elif kind == 'path': |
|
431 | 429 | return '^' + re.escape(name) + '(?:/|$)' |
|
432 | 430 | elif kind == 'relglob': |
|
433 | 431 | return globre(name, '(?:|.*/)', tail) |
|
434 | 432 | elif kind == 'relpath': |
|
435 | 433 | return re.escape(name) + '(?:/|$)' |
|
436 | 434 | elif kind == 'relre': |
|
437 | 435 | if name.startswith('^'): |
|
438 | 436 | return name |
|
439 | 437 | return '.*' + name |
|
440 | 438 | return globre(name, '', tail) |
|
441 | 439 | |
|
442 | 440 | def matchfn(pats, tail): |
|
443 | 441 | """build a matching function from a set of patterns""" |
|
444 | 442 | if not pats: |
|
445 | 443 | return |
|
446 | 444 | try: |
|
447 | 445 | pat = '(?:%s)' % '|'.join([regex(k, p, tail) for (k, p) in pats]) |
|
448 | 446 | if len(pat) > 20000: |
|
449 | 447 | raise OverflowError() |
|
450 | 448 | return re.compile(pat).match |
|
451 | 449 | except OverflowError: |
|
452 | 450 | # We're using a Python with a tiny regex engine and we |
|
453 | 451 | # made it explode, so we'll divide the pattern list in two |
|
454 | 452 | # until it works |
|
455 | 453 | l = len(pats) |
|
456 | 454 | if l < 2: |
|
457 | 455 | raise |
|
458 | 456 | a, b = matchfn(pats[:l//2], tail), matchfn(pats[l//2:], tail) |
|
459 | 457 | return lambda s: a(s) or b(s) |
|
460 | 458 | except re.error: |
|
461 | 459 | for k, p in pats: |
|
462 | 460 | try: |
|
463 | 461 | re.compile('(?:%s)' % regex(k, p, tail)) |
|
464 | 462 | except re.error: |
|
465 | 463 | if src: |
|
466 | 464 | raise Abort("%s: invalid pattern (%s): %s" % |
|
467 | 465 | (src, k, p)) |
|
468 | 466 | else: |
|
469 | 467 | raise Abort("invalid pattern (%s): %s" % (k, p)) |
|
470 | 468 | raise Abort("invalid pattern") |
|
471 | 469 | |
|
472 | 470 | def globprefix(pat): |
|
473 | 471 | '''return the non-glob prefix of a path, e.g. foo/* -> foo''' |
|
474 | 472 | root = [] |
|
475 | 473 | for p in pat.split('/'): |
|
476 | 474 | if contains_glob(p): break |
|
477 | 475 | root.append(p) |
|
478 | 476 | return '/'.join(root) or '.' |
|
479 | 477 | |
|
480 | 478 | def normalizepats(names, default): |
|
481 | 479 | pats = [] |
|
482 | 480 | roots = [] |
|
483 | 481 | anypats = False |
|
484 | 482 | for kind, name in [patkind(p, default) for p in names]: |
|
485 | 483 | if kind in ('glob', 'relpath'): |
|
486 | 484 | name = canonpath(canonroot, cwd, name) |
|
487 | 485 | elif kind in ('relglob', 'path'): |
|
488 | 486 | name = normpath(name) |
|
489 | 487 | |
|
490 | 488 | pats.append((kind, name)) |
|
491 | 489 | |
|
492 | 490 | if kind in ('glob', 're', 'relglob', 'relre'): |
|
493 | 491 | anypats = True |
|
494 | 492 | |
|
495 | 493 | if kind == 'glob': |
|
496 | 494 | root = globprefix(name) |
|
497 | 495 | roots.append(root) |
|
498 | 496 | elif kind in ('relpath', 'path'): |
|
499 | 497 | roots.append(name or '.') |
|
500 | 498 | elif kind == 'relglob': |
|
501 | 499 | roots.append('.') |
|
502 | 500 | return roots, pats, anypats |
|
503 | 501 | |
|
504 | 502 | roots, pats, anypats = normalizepats(names, dflt_pat) |
|
505 | 503 | |
|
506 | 504 | patmatch = matchfn(pats, '$') or always |
|
507 | 505 | incmatch = always |
|
508 | 506 | if inc: |
|
509 | 507 | dummy, inckinds, dummy = normalizepats(inc, 'glob') |
|
510 | 508 | incmatch = matchfn(inckinds, '(?:/|$)') |
|
511 | 509 | excmatch = never |
|
512 | 510 | if exc: |
|
513 | 511 | dummy, exckinds, dummy = normalizepats(exc, 'glob') |
|
514 | 512 | excmatch = matchfn(exckinds, '(?:/|$)') |
|
515 | 513 | |
|
516 | 514 | if not names and inc and not exc: |
|
517 | 515 | # common case: hgignore patterns |
|
518 | 516 | match = incmatch |
|
519 | 517 | else: |
|
520 | 518 | match = lambda fn: incmatch(fn) and not excmatch(fn) and patmatch(fn) |
|
521 | 519 | |
|
522 | 520 | return (roots, match, (inc or exc or anypats) and True) |
|
523 | 521 | |
|
524 | 522 | _hgexecutable = None |
|
525 | 523 | |
|
526 | 524 | def main_is_frozen(): |
|
527 | 525 | """return True if we are a frozen executable. |
|
528 | 526 | |
|
529 | 527 | The code supports py2exe (most common, Windows only) and tools/freeze |
|
530 | 528 | (portable, not much used). |
|
531 | 529 | """ |
|
532 | 530 | return (hasattr(sys, "frozen") or # new py2exe |
|
533 | 531 | hasattr(sys, "importers") or # old py2exe |
|
534 | 532 | imp.is_frozen("__main__")) # tools/freeze |
|
535 | 533 | |
|
536 | 534 | def hgexecutable(): |
|
537 | 535 | """return location of the 'hg' executable. |
|
538 | 536 | |
|
539 | 537 | Defaults to $HG or 'hg' in the search path. |
|
540 | 538 | """ |
|
541 | 539 | if _hgexecutable is None: |
|
542 | 540 | hg = os.environ.get('HG') |
|
543 | 541 | if hg: |
|
544 | 542 | set_hgexecutable(hg) |
|
545 | 543 | elif main_is_frozen(): |
|
546 | 544 | set_hgexecutable(sys.executable) |
|
547 | 545 | else: |
|
548 | 546 | set_hgexecutable(find_exe('hg') or 'hg') |
|
549 | 547 | return _hgexecutable |
|
550 | 548 | |
|
551 | 549 | def set_hgexecutable(path): |
|
552 | 550 | """set location of the 'hg' executable""" |
|
553 | 551 | global _hgexecutable |
|
554 | 552 | _hgexecutable = path |
|
555 | 553 | |
|
556 | 554 | def system(cmd, environ={}, cwd=None, onerr=None, errprefix=None): |
|
557 | 555 | '''enhanced shell command execution. |
|
558 | 556 | run with environment maybe modified, maybe in different dir. |
|
559 | 557 | |
|
560 | 558 | if command fails and onerr is None, return status. if ui object, |
|
561 | 559 | print error message and return status, else raise onerr object as |
|
562 | 560 | exception.''' |
|
563 | 561 | def py2shell(val): |
|
564 | 562 | 'convert python object into string that is useful to shell' |
|
565 | 563 | if val in (None, False): |
|
566 | 564 | return '0' |
|
567 | 565 | if val == True: |
|
568 | 566 | return '1' |
|
569 | 567 | return str(val) |
|
570 | 568 | oldenv = {} |
|
571 | 569 | for k in environ: |
|
572 | 570 | oldenv[k] = os.environ.get(k) |
|
573 | 571 | if cwd is not None: |
|
574 | 572 | oldcwd = os.getcwd() |
|
575 | 573 | origcmd = cmd |
|
576 | 574 | if os.name == 'nt': |
|
577 | 575 | cmd = '"%s"' % cmd |
|
578 | 576 | try: |
|
579 | 577 | for k, v in environ.iteritems(): |
|
580 | 578 | os.environ[k] = py2shell(v) |
|
581 | 579 | os.environ['HG'] = hgexecutable() |
|
582 | 580 | if cwd is not None and oldcwd != cwd: |
|
583 | 581 | os.chdir(cwd) |
|
584 | 582 | rc = os.system(cmd) |
|
585 | 583 | if sys.platform == 'OpenVMS' and rc & 1: |
|
586 | 584 | rc = 0 |
|
587 | 585 | if rc and onerr: |
|
588 | 586 | errmsg = '%s %s' % (os.path.basename(origcmd.split(None, 1)[0]), |
|
589 | 587 | explain_exit(rc)[0]) |
|
590 | 588 | if errprefix: |
|
591 | 589 | errmsg = '%s: %s' % (errprefix, errmsg) |
|
592 | 590 | try: |
|
593 | 591 | onerr.warn(errmsg + '\n') |
|
594 | 592 | except AttributeError: |
|
595 | 593 | raise onerr(errmsg) |
|
596 | 594 | return rc |
|
597 | 595 | finally: |
|
598 | 596 | for k, v in oldenv.iteritems(): |
|
599 | 597 | if v is None: |
|
600 | 598 | del os.environ[k] |
|
601 | 599 | else: |
|
602 | 600 | os.environ[k] = v |
|
603 | 601 | if cwd is not None and oldcwd != cwd: |
|
604 | 602 | os.chdir(oldcwd) |
|
605 | 603 | |
|
606 | 604 | def checksignature(func): |
|
607 | 605 | '''wrap a function with code to check for calling errors''' |
|
608 | 606 | def check(*args, **kwargs): |
|
609 | 607 | try: |
|
610 | 608 | return func(*args, **kwargs) |
|
611 | 609 | except TypeError: |
|
612 | 610 | if len(traceback.extract_tb(sys.exc_info()[2])) == 1: |
|
613 | 611 | raise error.SignatureError |
|
614 | 612 | raise |
|
615 | 613 | |
|
616 | 614 | return check |
|
617 | 615 | |
|
618 | 616 | # os.path.lexists is not available on python2.3 |
|
619 | 617 | def lexists(filename): |
|
620 | 618 | "test whether a file with this name exists. does not follow symlinks" |
|
621 | 619 | try: |
|
622 | 620 | os.lstat(filename) |
|
623 | 621 | except: |
|
624 | 622 | return False |
|
625 | 623 | return True |
|
626 | 624 | |
|
627 | 625 | def rename(src, dst): |
|
628 | 626 | """forcibly rename a file""" |
|
629 | 627 | try: |
|
630 | 628 | os.rename(src, dst) |
|
631 | 629 | except OSError, err: # FIXME: check err (EEXIST ?) |
|
632 | 630 | # on windows, rename to existing file is not allowed, so we |
|
633 | 631 | # must delete destination first. but if file is open, unlink |
|
634 | 632 | # schedules it for delete but does not delete it. rename |
|
635 | 633 | # happens immediately even for open files, so we rename |
|
636 | 634 | # destination to a temporary name, then delete that. then |
|
637 | 635 | # rename is safe to do. |
|
638 | 636 | temp = dst + "-force-rename" |
|
639 | 637 | os.rename(dst, temp) |
|
640 | 638 | os.unlink(temp) |
|
641 | 639 | os.rename(src, dst) |
|
642 | 640 | |
|
643 | 641 | def unlink(f): |
|
644 | 642 | """unlink and remove the directory if it is empty""" |
|
645 | 643 | os.unlink(f) |
|
646 | 644 | # try removing directories that might now be empty |
|
647 | 645 | try: |
|
648 | 646 | os.removedirs(os.path.dirname(f)) |
|
649 | 647 | except OSError: |
|
650 | 648 | pass |
|
651 | 649 | |
|
652 | 650 | def copyfile(src, dest): |
|
653 | 651 | "copy a file, preserving mode and atime/mtime" |
|
654 | 652 | if os.path.islink(src): |
|
655 | 653 | try: |
|
656 | 654 | os.unlink(dest) |
|
657 | 655 | except: |
|
658 | 656 | pass |
|
659 | 657 | os.symlink(os.readlink(src), dest) |
|
660 | 658 | else: |
|
661 | 659 | try: |
|
662 | 660 | shutil.copyfile(src, dest) |
|
663 | 661 | shutil.copystat(src, dest) |
|
664 | 662 | except shutil.Error, inst: |
|
665 | 663 | raise Abort(str(inst)) |
|
666 | 664 | |
|
667 | 665 | def copyfiles(src, dst, hardlink=None): |
|
668 | 666 | """Copy a directory tree using hardlinks if possible""" |
|
669 | 667 | |
|
670 | 668 | if hardlink is None: |
|
671 | 669 | hardlink = (os.stat(src).st_dev == |
|
672 | 670 | os.stat(os.path.dirname(dst)).st_dev) |
|
673 | 671 | |
|
674 | 672 | if os.path.isdir(src): |
|
675 | 673 | os.mkdir(dst) |
|
676 | 674 | for name, kind in osutil.listdir(src): |
|
677 | 675 | srcname = os.path.join(src, name) |
|
678 | 676 | dstname = os.path.join(dst, name) |
|
679 | 677 | copyfiles(srcname, dstname, hardlink) |
|
680 | 678 | else: |
|
681 | 679 | if hardlink: |
|
682 | 680 | try: |
|
683 | 681 | os_link(src, dst) |
|
684 | 682 | except (IOError, OSError): |
|
685 | 683 | hardlink = False |
|
686 | 684 | shutil.copy(src, dst) |
|
687 | 685 | else: |
|
688 | 686 | shutil.copy(src, dst) |
|
689 | 687 | |
|
690 | 688 | class path_auditor(object): |
|
691 | 689 | '''ensure that a filesystem path contains no banned components. |
|
692 | 690 | the following properties of a path are checked: |
|
693 | 691 | |
|
694 | 692 | - under top-level .hg |
|
695 | 693 | - starts at the root of a windows drive |
|
696 | 694 | - contains ".." |
|
697 | 695 | - traverses a symlink (e.g. a/symlink_here/b) |
|
698 | 696 | - inside a nested repository''' |
|
699 | 697 | |
|
700 | 698 | def __init__(self, root): |
|
701 | 699 | self.audited = set() |
|
702 | 700 | self.auditeddir = set() |
|
703 | 701 | self.root = root |
|
704 | 702 | |
|
705 | 703 | def __call__(self, path): |
|
706 | 704 | if path in self.audited: |
|
707 | 705 | return |
|
708 | 706 | normpath = os.path.normcase(path) |
|
709 | 707 | parts = splitpath(normpath) |
|
710 | 708 | if (os.path.splitdrive(path)[0] |
|
711 | 709 | or parts[0].lower() in ('.hg', '.hg.', '') |
|
712 | 710 | or os.pardir in parts): |
|
713 | 711 | raise Abort(_("path contains illegal component: %s") % path) |
|
714 | 712 | if '.hg' in path.lower(): |
|
715 | 713 | lparts = [p.lower() for p in parts] |
|
716 | 714 | for p in '.hg', '.hg.': |
|
717 | 715 | if p in lparts[1:]: |
|
718 | 716 | pos = lparts.index(p) |
|
719 | 717 | base = os.path.join(*parts[:pos]) |
|
720 | 718 | raise Abort(_('path %r is inside repo %r') % (path, base)) |
|
721 | 719 | def check(prefix): |
|
722 | 720 | curpath = os.path.join(self.root, prefix) |
|
723 | 721 | try: |
|
724 | 722 | st = os.lstat(curpath) |
|
725 | 723 | except OSError, err: |
|
726 | 724 | # EINVAL can be raised as invalid path syntax under win32. |
|
727 | 725 | # They must be ignored for patterns can be checked too. |
|
728 | 726 | if err.errno not in (errno.ENOENT, errno.ENOTDIR, errno.EINVAL): |
|
729 | 727 | raise |
|
730 | 728 | else: |
|
731 | 729 | if stat.S_ISLNK(st.st_mode): |
|
732 | 730 | raise Abort(_('path %r traverses symbolic link %r') % |
|
733 | 731 | (path, prefix)) |
|
734 | 732 | elif (stat.S_ISDIR(st.st_mode) and |
|
735 | 733 | os.path.isdir(os.path.join(curpath, '.hg'))): |
|
736 | 734 | raise Abort(_('path %r is inside repo %r') % |
|
737 | 735 | (path, prefix)) |
|
738 | 736 | parts.pop() |
|
739 | 737 | prefixes = [] |
|
740 | 738 | for n in range(len(parts)): |
|
741 | 739 | prefix = os.sep.join(parts) |
|
742 | 740 | if prefix in self.auditeddir: |
|
743 | 741 | break |
|
744 | 742 | check(prefix) |
|
745 | 743 | prefixes.append(prefix) |
|
746 | 744 | parts.pop() |
|
747 | 745 | |
|
748 | 746 | self.audited.add(path) |
|
749 | 747 | # only add prefixes to the cache after checking everything: we don't |
|
750 | 748 | # want to add "foo/bar/baz" before checking if there's a "foo/.hg" |
|
751 | 749 | self.auditeddir.update(prefixes) |
|
752 | 750 | |
|
753 | 751 | def nlinks(pathname): |
|
754 | 752 | """Return number of hardlinks for the given file.""" |
|
755 | 753 | return os.lstat(pathname).st_nlink |
|
756 | 754 | |
|
757 | 755 | if hasattr(os, 'link'): |
|
758 | 756 | os_link = os.link |
|
759 | 757 | else: |
|
760 | 758 | def os_link(src, dst): |
|
761 | 759 | raise OSError(0, _("Hardlinks not supported")) |
|
762 | 760 | |
|
763 | 761 | def lookup_reg(key, name=None, scope=None): |
|
764 | 762 | return None |
|
765 | 763 | |
|
766 | 764 | if os.name == 'nt': |
|
767 | 765 | from windows import * |
|
768 | 766 | def expand_glob(pats): |
|
769 | 767 | '''On Windows, expand the implicit globs in a list of patterns''' |
|
770 | 768 | ret = [] |
|
771 | 769 | for p in pats: |
|
772 | 770 | kind, name = patkind(p, None) |
|
773 | 771 | if kind is None: |
|
774 | 772 | globbed = glob.glob(name) |
|
775 | 773 | if globbed: |
|
776 | 774 | ret.extend(globbed) |
|
777 | 775 | continue |
|
778 | 776 | # if we couldn't expand the glob, just keep it around |
|
779 | 777 | ret.append(p) |
|
780 | 778 | return ret |
|
781 | 779 | else: |
|
782 | 780 | from posix import * |
|
783 | 781 | |
|
784 | 782 | def makelock(info, pathname): |
|
785 | 783 | try: |
|
786 | 784 | return os.symlink(info, pathname) |
|
787 | 785 | except OSError, why: |
|
788 | 786 | if why.errno == errno.EEXIST: |
|
789 | 787 | raise |
|
790 | 788 | except AttributeError: # no symlink in os |
|
791 | 789 | pass |
|
792 | 790 | |
|
793 | 791 | ld = os.open(pathname, os.O_CREAT | os.O_WRONLY | os.O_EXCL) |
|
794 | 792 | os.write(ld, info) |
|
795 | 793 | os.close(ld) |
|
796 | 794 | |
|
797 | 795 | def readlock(pathname): |
|
798 | 796 | try: |
|
799 | 797 | return os.readlink(pathname) |
|
800 | 798 | except OSError, why: |
|
801 | 799 | if why.errno not in (errno.EINVAL, errno.ENOSYS): |
|
802 | 800 | raise |
|
803 | 801 | except AttributeError: # no symlink in os |
|
804 | 802 | pass |
|
805 | 803 | return posixfile(pathname).read() |
|
806 | 804 | |
|
807 | 805 | def fstat(fp): |
|
808 | 806 | '''stat file object that may not have fileno method.''' |
|
809 | 807 | try: |
|
810 | 808 | return os.fstat(fp.fileno()) |
|
811 | 809 | except AttributeError: |
|
812 | 810 | return os.stat(fp.name) |
|
813 | 811 | |
|
814 | 812 | # File system features |
|
815 | 813 | |
|
816 | 814 | def checkcase(path): |
|
817 | 815 | """ |
|
818 | 816 | Check whether the given path is on a case-sensitive filesystem |
|
819 | 817 | |
|
820 | 818 | Requires a path (like /foo/.hg) ending with a foldable final |
|
821 | 819 | directory component. |
|
822 | 820 | """ |
|
823 | 821 | s1 = os.stat(path) |
|
824 | 822 | d, b = os.path.split(path) |
|
825 | 823 | p2 = os.path.join(d, b.upper()) |
|
826 | 824 | if path == p2: |
|
827 | 825 | p2 = os.path.join(d, b.lower()) |
|
828 | 826 | try: |
|
829 | 827 | s2 = os.stat(p2) |
|
830 | 828 | if s2 == s1: |
|
831 | 829 | return False |
|
832 | 830 | return True |
|
833 | 831 | except: |
|
834 | 832 | return True |
|
835 | 833 | |
|
836 | 834 | _fspathcache = {} |
|
837 | 835 | def fspath(name, root): |
|
838 | 836 | '''Get name in the case stored in the filesystem |
|
839 | 837 | |
|
840 | 838 | The name is either relative to root, or it is an absolute path starting |
|
841 | 839 | with root. Note that this function is unnecessary, and should not be |
|
842 | 840 | called, for case-sensitive filesystems (simply because it's expensive). |
|
843 | 841 | ''' |
|
844 | 842 | # If name is absolute, make it relative |
|
845 | 843 | if name.lower().startswith(root.lower()): |
|
846 | 844 | l = len(root) |
|
847 | 845 | if name[l] == os.sep or name[l] == os.altsep: |
|
848 | 846 | l = l + 1 |
|
849 | 847 | name = name[l:] |
|
850 | 848 | |
|
851 | 849 | if not os.path.exists(os.path.join(root, name)): |
|
852 | 850 | return None |
|
853 | 851 | |
|
854 | 852 | seps = os.sep |
|
855 | 853 | if os.altsep: |
|
856 | 854 | seps = seps + os.altsep |
|
857 | 855 | # Protect backslashes. This gets silly very quickly. |
|
858 | 856 | seps.replace('\\','\\\\') |
|
859 | 857 | pattern = re.compile(r'([^%s]+)|([%s]+)' % (seps, seps)) |
|
860 | 858 | dir = os.path.normcase(os.path.normpath(root)) |
|
861 | 859 | result = [] |
|
862 | 860 | for part, sep in pattern.findall(name): |
|
863 | 861 | if sep: |
|
864 | 862 | result.append(sep) |
|
865 | 863 | continue |
|
866 | 864 | |
|
867 | 865 | if dir not in _fspathcache: |
|
868 | 866 | _fspathcache[dir] = os.listdir(dir) |
|
869 | 867 | contents = _fspathcache[dir] |
|
870 | 868 | |
|
871 | 869 | lpart = part.lower() |
|
872 | 870 | for n in contents: |
|
873 | 871 | if n.lower() == lpart: |
|
874 | 872 | result.append(n) |
|
875 | 873 | break |
|
876 | 874 | else: |
|
877 | 875 | # Cannot happen, as the file exists! |
|
878 | 876 | result.append(part) |
|
879 | 877 | dir = os.path.join(dir, lpart) |
|
880 | 878 | |
|
881 | 879 | return ''.join(result) |
|
882 | 880 | |
|
883 | 881 | def checkexec(path): |
|
884 | 882 | """ |
|
885 | 883 | Check whether the given path is on a filesystem with UNIX-like exec flags |
|
886 | 884 | |
|
887 | 885 | Requires a directory (like /foo/.hg) |
|
888 | 886 | """ |
|
889 | 887 | |
|
890 | 888 | # VFAT on some Linux versions can flip mode but it doesn't persist |
|
891 | 889 | # a FS remount. Frequently we can detect it if files are created |
|
892 | 890 | # with exec bit on. |
|
893 | 891 | |
|
894 | 892 | try: |
|
895 | 893 | EXECFLAGS = stat.S_IXUSR | stat.S_IXGRP | stat.S_IXOTH |
|
896 | 894 | fh, fn = tempfile.mkstemp("", "", path) |
|
897 | 895 | try: |
|
898 | 896 | os.close(fh) |
|
899 | 897 | m = os.stat(fn).st_mode & 0777 |
|
900 | 898 | new_file_has_exec = m & EXECFLAGS |
|
901 | 899 | os.chmod(fn, m ^ EXECFLAGS) |
|
902 | 900 | exec_flags_cannot_flip = ((os.stat(fn).st_mode & 0777) == m) |
|
903 | 901 | finally: |
|
904 | 902 | os.unlink(fn) |
|
905 | 903 | except (IOError, OSError): |
|
906 | 904 | # we don't care, the user probably won't be able to commit anyway |
|
907 | 905 | return False |
|
908 | 906 | return not (new_file_has_exec or exec_flags_cannot_flip) |
|
909 | 907 | |
|
910 | 908 | def checklink(path): |
|
911 | 909 | """check whether the given path is on a symlink-capable filesystem""" |
|
912 | 910 | # mktemp is not racy because symlink creation will fail if the |
|
913 | 911 | # file already exists |
|
914 | 912 | name = tempfile.mktemp(dir=path) |
|
915 | 913 | try: |
|
916 | 914 | os.symlink(".", name) |
|
917 | 915 | os.unlink(name) |
|
918 | 916 | return True |
|
919 | 917 | except (OSError, AttributeError): |
|
920 | 918 | return False |
|
921 | 919 | |
|
922 | 920 | def needbinarypatch(): |
|
923 | 921 | """return True if patches should be applied in binary mode by default.""" |
|
924 | 922 | return os.name == 'nt' |
|
925 | 923 | |
|
926 | 924 | def endswithsep(path): |
|
927 | 925 | '''Check path ends with os.sep or os.altsep.''' |
|
928 | 926 | return path.endswith(os.sep) or os.altsep and path.endswith(os.altsep) |
|
929 | 927 | |
|
930 | 928 | def splitpath(path): |
|
931 | 929 | '''Split path by os.sep. |
|
932 | 930 | Note that this function does not use os.altsep because this is |
|
933 | 931 | an alternative of simple "xxx.split(os.sep)". |
|
934 | 932 | It is recommended to use os.path.normpath() before using this |
|
935 | 933 | function if need.''' |
|
936 | 934 | return path.split(os.sep) |
|
937 | 935 | |
|
938 | 936 | def gui(): |
|
939 | 937 | '''Are we running in a GUI?''' |
|
940 | 938 | return os.name == "nt" or os.name == "mac" or os.environ.get("DISPLAY") |
|
941 | 939 | |
|
942 | 940 | def mktempcopy(name, emptyok=False, createmode=None): |
|
943 | 941 | """Create a temporary file with the same contents from name |
|
944 | 942 | |
|
945 | 943 | The permission bits are copied from the original file. |
|
946 | 944 | |
|
947 | 945 | If the temporary file is going to be truncated immediately, you |
|
948 | 946 | can use emptyok=True as an optimization. |
|
949 | 947 | |
|
950 | 948 | Returns the name of the temporary file. |
|
951 | 949 | """ |
|
952 | 950 | d, fn = os.path.split(name) |
|
953 | 951 | fd, temp = tempfile.mkstemp(prefix='.%s-' % fn, dir=d) |
|
954 | 952 | os.close(fd) |
|
955 | 953 | # Temporary files are created with mode 0600, which is usually not |
|
956 | 954 | # what we want. If the original file already exists, just copy |
|
957 | 955 | # its mode. Otherwise, manually obey umask. |
|
958 | 956 | try: |
|
959 | 957 | st_mode = os.lstat(name).st_mode & 0777 |
|
960 | 958 | except OSError, inst: |
|
961 | 959 | if inst.errno != errno.ENOENT: |
|
962 | 960 | raise |
|
963 | 961 | st_mode = createmode |
|
964 | 962 | if st_mode is None: |
|
965 | 963 | st_mode = ~umask |
|
966 | 964 | st_mode &= 0666 |
|
967 | 965 | os.chmod(temp, st_mode) |
|
968 | 966 | if emptyok: |
|
969 | 967 | return temp |
|
970 | 968 | try: |
|
971 | 969 | try: |
|
972 | 970 | ifp = posixfile(name, "rb") |
|
973 | 971 | except IOError, inst: |
|
974 | 972 | if inst.errno == errno.ENOENT: |
|
975 | 973 | return temp |
|
976 | 974 | if not getattr(inst, 'filename', None): |
|
977 | 975 | inst.filename = name |
|
978 | 976 | raise |
|
979 | 977 | ofp = posixfile(temp, "wb") |
|
980 | 978 | for chunk in filechunkiter(ifp): |
|
981 | 979 | ofp.write(chunk) |
|
982 | 980 | ifp.close() |
|
983 | 981 | ofp.close() |
|
984 | 982 | except: |
|
985 | 983 | try: os.unlink(temp) |
|
986 | 984 | except: pass |
|
987 | 985 | raise |
|
988 | 986 | return temp |
|
989 | 987 | |
|
990 | 988 | class atomictempfile(posixfile): |
|
991 | 989 | """file-like object that atomically updates a file |
|
992 | 990 | |
|
993 | 991 | All writes will be redirected to a temporary copy of the original |
|
994 | 992 | file. When rename is called, the copy is renamed to the original |
|
995 | 993 | name, making the changes visible. |
|
996 | 994 | """ |
|
997 | 995 | def __init__(self, name, mode, createmode): |
|
998 | 996 | self.__name = name |
|
999 | 997 | self.temp = mktempcopy(name, emptyok=('w' in mode), |
|
1000 | 998 | createmode=createmode) |
|
1001 | 999 | posixfile.__init__(self, self.temp, mode) |
|
1002 | 1000 | |
|
1003 | 1001 | def rename(self): |
|
1004 | 1002 | if not self.closed: |
|
1005 | 1003 | posixfile.close(self) |
|
1006 | 1004 | rename(self.temp, localpath(self.__name)) |
|
1007 | 1005 | |
|
1008 | 1006 | def __del__(self): |
|
1009 | 1007 | if not self.closed: |
|
1010 | 1008 | try: |
|
1011 | 1009 | os.unlink(self.temp) |
|
1012 | 1010 | except: pass |
|
1013 | 1011 | posixfile.close(self) |
|
1014 | 1012 | |
|
1015 | 1013 | def makedirs(name, mode=None): |
|
1016 | 1014 | """recursive directory creation with parent mode inheritance""" |
|
1017 | 1015 | try: |
|
1018 | 1016 | os.mkdir(name) |
|
1019 | 1017 | if mode is not None: |
|
1020 | 1018 | os.chmod(name, mode) |
|
1021 | 1019 | return |
|
1022 | 1020 | except OSError, err: |
|
1023 | 1021 | if err.errno == errno.EEXIST: |
|
1024 | 1022 | return |
|
1025 | 1023 | if err.errno != errno.ENOENT: |
|
1026 | 1024 | raise |
|
1027 | 1025 | parent = os.path.abspath(os.path.dirname(name)) |
|
1028 | 1026 | makedirs(parent, mode) |
|
1029 | 1027 | makedirs(name, mode) |
|
1030 | 1028 | |
|
1031 | 1029 | class opener(object): |
|
1032 | 1030 | """Open files relative to a base directory |
|
1033 | 1031 | |
|
1034 | 1032 | This class is used to hide the details of COW semantics and |
|
1035 | 1033 | remote file access from higher level code. |
|
1036 | 1034 | """ |
|
1037 | 1035 | def __init__(self, base, audit=True): |
|
1038 | 1036 | self.base = base |
|
1039 | 1037 | if audit: |
|
1040 | 1038 | self.audit_path = path_auditor(base) |
|
1041 | 1039 | else: |
|
1042 | 1040 | self.audit_path = always |
|
1043 | 1041 | self.createmode = None |
|
1044 | 1042 | |
|
1045 | 1043 | def __getattr__(self, name): |
|
1046 | 1044 | if name == '_can_symlink': |
|
1047 | 1045 | self._can_symlink = checklink(self.base) |
|
1048 | 1046 | return self._can_symlink |
|
1049 | 1047 | raise AttributeError(name) |
|
1050 | 1048 | |
|
1051 | 1049 | def _fixfilemode(self, name): |
|
1052 | 1050 | if self.createmode is None: |
|
1053 | 1051 | return |
|
1054 | 1052 | os.chmod(name, self.createmode & 0666) |
|
1055 | 1053 | |
|
1056 | 1054 | def __call__(self, path, mode="r", text=False, atomictemp=False): |
|
1057 | 1055 | self.audit_path(path) |
|
1058 | 1056 | f = os.path.join(self.base, path) |
|
1059 | 1057 | |
|
1060 | 1058 | if not text and "b" not in mode: |
|
1061 | 1059 | mode += "b" # for that other OS |
|
1062 | 1060 | |
|
1063 | 1061 | nlink = -1 |
|
1064 | 1062 | if mode not in ("r", "rb"): |
|
1065 | 1063 | try: |
|
1066 | 1064 | nlink = nlinks(f) |
|
1067 | 1065 | except OSError: |
|
1068 | 1066 | nlink = 0 |
|
1069 | 1067 | d = os.path.dirname(f) |
|
1070 | 1068 | if not os.path.isdir(d): |
|
1071 | 1069 | makedirs(d, self.createmode) |
|
1072 | 1070 | if atomictemp: |
|
1073 | 1071 | return atomictempfile(f, mode, self.createmode) |
|
1074 | 1072 | if nlink > 1: |
|
1075 | 1073 | rename(mktempcopy(f), f) |
|
1076 | 1074 | fp = posixfile(f, mode) |
|
1077 | 1075 | if nlink == 0: |
|
1078 | 1076 | self._fixfilemode(f) |
|
1079 | 1077 | return fp |
|
1080 | 1078 | |
|
1081 | 1079 | def symlink(self, src, dst): |
|
1082 | 1080 | self.audit_path(dst) |
|
1083 | 1081 | linkname = os.path.join(self.base, dst) |
|
1084 | 1082 | try: |
|
1085 | 1083 | os.unlink(linkname) |
|
1086 | 1084 | except OSError: |
|
1087 | 1085 | pass |
|
1088 | 1086 | |
|
1089 | 1087 | dirname = os.path.dirname(linkname) |
|
1090 | 1088 | if not os.path.exists(dirname): |
|
1091 | 1089 | makedirs(dirname, self.createmode) |
|
1092 | 1090 | |
|
1093 | 1091 | if self._can_symlink: |
|
1094 | 1092 | try: |
|
1095 | 1093 | os.symlink(src, linkname) |
|
1096 | 1094 | except OSError, err: |
|
1097 | 1095 | raise OSError(err.errno, _('could not symlink to %r: %s') % |
|
1098 | 1096 | (src, err.strerror), linkname) |
|
1099 | 1097 | else: |
|
1100 | 1098 | f = self(dst, "w") |
|
1101 | 1099 | f.write(src) |
|
1102 | 1100 | f.close() |
|
1103 | 1101 | self._fixfilemode(dst) |
|
1104 | 1102 | |
|
1105 | 1103 | class chunkbuffer(object): |
|
1106 | 1104 | """Allow arbitrary sized chunks of data to be efficiently read from an |
|
1107 | 1105 | iterator over chunks of arbitrary size.""" |
|
1108 | 1106 | |
|
1109 | 1107 | def __init__(self, in_iter): |
|
1110 | 1108 | """in_iter is the iterator that's iterating over the input chunks. |
|
1111 | 1109 | targetsize is how big a buffer to try to maintain.""" |
|
1112 | 1110 | self.iter = iter(in_iter) |
|
1113 | 1111 | self.buf = '' |
|
1114 | 1112 | self.targetsize = 2**16 |
|
1115 | 1113 | |
|
1116 | 1114 | def read(self, l): |
|
1117 | 1115 | """Read L bytes of data from the iterator of chunks of data. |
|
1118 | 1116 | Returns less than L bytes if the iterator runs dry.""" |
|
1119 | 1117 | if l > len(self.buf) and self.iter: |
|
1120 | 1118 | # Clamp to a multiple of self.targetsize |
|
1121 | 1119 | targetsize = max(l, self.targetsize) |
|
1122 | 1120 | collector = cStringIO.StringIO() |
|
1123 | 1121 | collector.write(self.buf) |
|
1124 | 1122 | collected = len(self.buf) |
|
1125 | 1123 | for chunk in self.iter: |
|
1126 | 1124 | collector.write(chunk) |
|
1127 | 1125 | collected += len(chunk) |
|
1128 | 1126 | if collected >= targetsize: |
|
1129 | 1127 | break |
|
1130 | 1128 | if collected < targetsize: |
|
1131 | 1129 | self.iter = False |
|
1132 | 1130 | self.buf = collector.getvalue() |
|
1133 | 1131 | if len(self.buf) == l: |
|
1134 | 1132 | s, self.buf = str(self.buf), '' |
|
1135 | 1133 | else: |
|
1136 | 1134 | s, self.buf = self.buf[:l], buffer(self.buf, l) |
|
1137 | 1135 | return s |
|
1138 | 1136 | |
|
1139 | 1137 | def filechunkiter(f, size=65536, limit=None): |
|
1140 | 1138 | """Create a generator that produces the data in the file size |
|
1141 | 1139 | (default 65536) bytes at a time, up to optional limit (default is |
|
1142 | 1140 | to read all data). Chunks may be less than size bytes if the |
|
1143 | 1141 | chunk is the last chunk in the file, or the file is a socket or |
|
1144 | 1142 | some other type of file that sometimes reads less data than is |
|
1145 | 1143 | requested.""" |
|
1146 | 1144 | assert size >= 0 |
|
1147 | 1145 | assert limit is None or limit >= 0 |
|
1148 | 1146 | while True: |
|
1149 | 1147 | if limit is None: nbytes = size |
|
1150 | 1148 | else: nbytes = min(limit, size) |
|
1151 | 1149 | s = nbytes and f.read(nbytes) |
|
1152 | 1150 | if not s: break |
|
1153 | 1151 | if limit: limit -= len(s) |
|
1154 | 1152 | yield s |
|
1155 | 1153 | |
|
1156 | 1154 | def makedate(): |
|
1157 | 1155 | lt = time.localtime() |
|
1158 | 1156 | if lt[8] == 1 and time.daylight: |
|
1159 | 1157 | tz = time.altzone |
|
1160 | 1158 | else: |
|
1161 | 1159 | tz = time.timezone |
|
1162 | 1160 | return time.mktime(lt), tz |
|
1163 | 1161 | |
|
1164 | 1162 | def datestr(date=None, format='%a %b %d %H:%M:%S %Y %1%2'): |
|
1165 | 1163 | """represent a (unixtime, offset) tuple as a localized time. |
|
1166 | 1164 | unixtime is seconds since the epoch, and offset is the time zone's |
|
1167 | 1165 | number of seconds away from UTC. if timezone is false, do not |
|
1168 | 1166 | append time zone to string.""" |
|
1169 | 1167 | t, tz = date or makedate() |
|
1170 | 1168 | if "%1" in format or "%2" in format: |
|
1171 | 1169 | sign = (tz > 0) and "-" or "+" |
|
1172 | 1170 | minutes = abs(tz) / 60 |
|
1173 | 1171 | format = format.replace("%1", "%c%02d" % (sign, minutes / 60)) |
|
1174 | 1172 | format = format.replace("%2", "%02d" % (minutes % 60)) |
|
1175 | 1173 | s = time.strftime(format, time.gmtime(float(t) - tz)) |
|
1176 | 1174 | return s |
|
1177 | 1175 | |
|
1178 | 1176 | def shortdate(date=None): |
|
1179 | 1177 | """turn (timestamp, tzoff) tuple into iso 8631 date.""" |
|
1180 | 1178 | return datestr(date, format='%Y-%m-%d') |
|
1181 | 1179 | |
|
1182 | 1180 | def strdate(string, format, defaults=[]): |
|
1183 | 1181 | """parse a localized time string and return a (unixtime, offset) tuple. |
|
1184 | 1182 | if the string cannot be parsed, ValueError is raised.""" |
|
1185 | 1183 | def timezone(string): |
|
1186 | 1184 | tz = string.split()[-1] |
|
1187 | 1185 | if tz[0] in "+-" and len(tz) == 5 and tz[1:].isdigit(): |
|
1188 | 1186 | sign = (tz[0] == "+") and 1 or -1 |
|
1189 | 1187 | hours = int(tz[1:3]) |
|
1190 | 1188 | minutes = int(tz[3:5]) |
|
1191 | 1189 | return -sign * (hours * 60 + minutes) * 60 |
|
1192 | 1190 | if tz == "GMT" or tz == "UTC": |
|
1193 | 1191 | return 0 |
|
1194 | 1192 | return None |
|
1195 | 1193 | |
|
1196 | 1194 | # NOTE: unixtime = localunixtime + offset |
|
1197 | 1195 | offset, date = timezone(string), string |
|
1198 | 1196 | if offset != None: |
|
1199 | 1197 | date = " ".join(string.split()[:-1]) |
|
1200 | 1198 | |
|
1201 | 1199 | # add missing elements from defaults |
|
1202 | 1200 | for part in defaults: |
|
1203 | 1201 | found = [True for p in part if ("%"+p) in format] |
|
1204 | 1202 | if not found: |
|
1205 | 1203 | date += "@" + defaults[part] |
|
1206 | 1204 | format += "@%" + part[0] |
|
1207 | 1205 | |
|
1208 | 1206 | timetuple = time.strptime(date, format) |
|
1209 | 1207 | localunixtime = int(calendar.timegm(timetuple)) |
|
1210 | 1208 | if offset is None: |
|
1211 | 1209 | # local timezone |
|
1212 | 1210 | unixtime = int(time.mktime(timetuple)) |
|
1213 | 1211 | offset = unixtime - localunixtime |
|
1214 | 1212 | else: |
|
1215 | 1213 | unixtime = localunixtime + offset |
|
1216 | 1214 | return unixtime, offset |
|
1217 | 1215 | |
|
1218 | 1216 | def parsedate(date, formats=None, defaults=None): |
|
1219 | 1217 | """parse a localized date/time string and return a (unixtime, offset) tuple. |
|
1220 | 1218 | |
|
1221 | 1219 | The date may be a "unixtime offset" string or in one of the specified |
|
1222 | 1220 | formats. If the date already is a (unixtime, offset) tuple, it is returned. |
|
1223 | 1221 | """ |
|
1224 | 1222 | if not date: |
|
1225 | 1223 | return 0, 0 |
|
1226 | 1224 | if isinstance(date, tuple) and len(date) == 2: |
|
1227 | 1225 | return date |
|
1228 | 1226 | if not formats: |
|
1229 | 1227 | formats = defaultdateformats |
|
1230 | 1228 | date = date.strip() |
|
1231 | 1229 | try: |
|
1232 | 1230 | when, offset = map(int, date.split(' ')) |
|
1233 | 1231 | except ValueError: |
|
1234 | 1232 | # fill out defaults |
|
1235 | 1233 | if not defaults: |
|
1236 | 1234 | defaults = {} |
|
1237 | 1235 | now = makedate() |
|
1238 | 1236 | for part in "d mb yY HI M S".split(): |
|
1239 | 1237 | if part not in defaults: |
|
1240 | 1238 | if part[0] in "HMS": |
|
1241 | 1239 | defaults[part] = "00" |
|
1242 | 1240 | else: |
|
1243 | 1241 | defaults[part] = datestr(now, "%" + part[0]) |
|
1244 | 1242 | |
|
1245 | 1243 | for format in formats: |
|
1246 | 1244 | try: |
|
1247 | 1245 | when, offset = strdate(date, format, defaults) |
|
1248 | 1246 | except (ValueError, OverflowError): |
|
1249 | 1247 | pass |
|
1250 | 1248 | else: |
|
1251 | 1249 | break |
|
1252 | 1250 | else: |
|
1253 | 1251 | raise Abort(_('invalid date: %r ') % date) |
|
1254 | 1252 | # validate explicit (probably user-specified) date and |
|
1255 | 1253 | # time zone offset. values must fit in signed 32 bits for |
|
1256 | 1254 | # current 32-bit linux runtimes. timezones go from UTC-12 |
|
1257 | 1255 | # to UTC+14 |
|
1258 | 1256 | if abs(when) > 0x7fffffff: |
|
1259 | 1257 | raise Abort(_('date exceeds 32 bits: %d') % when) |
|
1260 | 1258 | if offset < -50400 or offset > 43200: |
|
1261 | 1259 | raise Abort(_('impossible time zone offset: %d') % offset) |
|
1262 | 1260 | return when, offset |
|
1263 | 1261 | |
|
1264 | 1262 | def matchdate(date): |
|
1265 | 1263 | """Return a function that matches a given date match specifier |
|
1266 | 1264 | |
|
1267 | 1265 | Formats include: |
|
1268 | 1266 | |
|
1269 | 1267 | '{date}' match a given date to the accuracy provided |
|
1270 | 1268 | |
|
1271 | 1269 | '<{date}' on or before a given date |
|
1272 | 1270 | |
|
1273 | 1271 | '>{date}' on or after a given date |
|
1274 | 1272 | |
|
1275 | 1273 | """ |
|
1276 | 1274 | |
|
1277 | 1275 | def lower(date): |
|
1278 | 1276 | d = dict(mb="1", d="1") |
|
1279 | 1277 | return parsedate(date, extendeddateformats, d)[0] |
|
1280 | 1278 | |
|
1281 | 1279 | def upper(date): |
|
1282 | 1280 | d = dict(mb="12", HI="23", M="59", S="59") |
|
1283 | 1281 | for days in "31 30 29".split(): |
|
1284 | 1282 | try: |
|
1285 | 1283 | d["d"] = days |
|
1286 | 1284 | return parsedate(date, extendeddateformats, d)[0] |
|
1287 | 1285 | except: |
|
1288 | 1286 | pass |
|
1289 | 1287 | d["d"] = "28" |
|
1290 | 1288 | return parsedate(date, extendeddateformats, d)[0] |
|
1291 | 1289 | |
|
1292 | 1290 | date = date.strip() |
|
1293 | 1291 | if date[0] == "<": |
|
1294 | 1292 | when = upper(date[1:]) |
|
1295 | 1293 | return lambda x: x <= when |
|
1296 | 1294 | elif date[0] == ">": |
|
1297 | 1295 | when = lower(date[1:]) |
|
1298 | 1296 | return lambda x: x >= when |
|
1299 | 1297 | elif date[0] == "-": |
|
1300 | 1298 | try: |
|
1301 | 1299 | days = int(date[1:]) |
|
1302 | 1300 | except ValueError: |
|
1303 | 1301 | raise Abort(_("invalid day spec: %s") % date[1:]) |
|
1304 | 1302 | when = makedate()[0] - days * 3600 * 24 |
|
1305 | 1303 | return lambda x: x >= when |
|
1306 | 1304 | elif " to " in date: |
|
1307 | 1305 | a, b = date.split(" to ") |
|
1308 | 1306 | start, stop = lower(a), upper(b) |
|
1309 | 1307 | return lambda x: x >= start and x <= stop |
|
1310 | 1308 | else: |
|
1311 | 1309 | start, stop = lower(date), upper(date) |
|
1312 | 1310 | return lambda x: x >= start and x <= stop |
|
1313 | 1311 | |
|
1314 | 1312 | def shortuser(user): |
|
1315 | 1313 | """Return a short representation of a user name or email address.""" |
|
1316 | 1314 | f = user.find('@') |
|
1317 | 1315 | if f >= 0: |
|
1318 | 1316 | user = user[:f] |
|
1319 | 1317 | f = user.find('<') |
|
1320 | 1318 | if f >= 0: |
|
1321 | 1319 | user = user[f+1:] |
|
1322 | 1320 | f = user.find(' ') |
|
1323 | 1321 | if f >= 0: |
|
1324 | 1322 | user = user[:f] |
|
1325 | 1323 | f = user.find('.') |
|
1326 | 1324 | if f >= 0: |
|
1327 | 1325 | user = user[:f] |
|
1328 | 1326 | return user |
|
1329 | 1327 | |
|
1330 | 1328 | def email(author): |
|
1331 | 1329 | '''get email of author.''' |
|
1332 | 1330 | r = author.find('>') |
|
1333 | 1331 | if r == -1: r = None |
|
1334 | 1332 | return author[author.find('<')+1:r] |
|
1335 | 1333 | |
|
1336 | 1334 | def ellipsis(text, maxlength=400): |
|
1337 | 1335 | """Trim string to at most maxlength (default: 400) characters.""" |
|
1338 | 1336 | if len(text) <= maxlength: |
|
1339 | 1337 | return text |
|
1340 | 1338 | else: |
|
1341 | 1339 | return "%s..." % (text[:maxlength-3]) |
|
1342 | 1340 | |
|
1343 | 1341 | def walkrepos(path, followsym=False, seen_dirs=None, recurse=False): |
|
1344 | 1342 | '''yield every hg repository under path, recursively.''' |
|
1345 | 1343 | def errhandler(err): |
|
1346 | 1344 | if err.filename == path: |
|
1347 | 1345 | raise err |
|
1348 | 1346 | if followsym and hasattr(os.path, 'samestat'): |
|
1349 | 1347 | def _add_dir_if_not_there(dirlst, dirname): |
|
1350 | 1348 | match = False |
|
1351 | 1349 | samestat = os.path.samestat |
|
1352 | 1350 | dirstat = os.stat(dirname) |
|
1353 | 1351 | for lstdirstat in dirlst: |
|
1354 | 1352 | if samestat(dirstat, lstdirstat): |
|
1355 | 1353 | match = True |
|
1356 | 1354 | break |
|
1357 | 1355 | if not match: |
|
1358 | 1356 | dirlst.append(dirstat) |
|
1359 | 1357 | return not match |
|
1360 | 1358 | else: |
|
1361 | 1359 | followsym = False |
|
1362 | 1360 | |
|
1363 | 1361 | if (seen_dirs is None) and followsym: |
|
1364 | 1362 | seen_dirs = [] |
|
1365 | 1363 | _add_dir_if_not_there(seen_dirs, path) |
|
1366 | 1364 | for root, dirs, files in os.walk(path, topdown=True, onerror=errhandler): |
|
1367 | 1365 | if '.hg' in dirs: |
|
1368 | 1366 | yield root # found a repository |
|
1369 | 1367 | qroot = os.path.join(root, '.hg', 'patches') |
|
1370 | 1368 | if os.path.isdir(os.path.join(qroot, '.hg')): |
|
1371 | 1369 | yield qroot # we have a patch queue repo here |
|
1372 | 1370 | if recurse: |
|
1373 | 1371 | # avoid recursing inside the .hg directory |
|
1374 | 1372 | dirs.remove('.hg') |
|
1375 | 1373 | else: |
|
1376 | 1374 | dirs[:] = [] # don't descend further |
|
1377 | 1375 | elif followsym: |
|
1378 | 1376 | newdirs = [] |
|
1379 | 1377 | for d in dirs: |
|
1380 | 1378 | fname = os.path.join(root, d) |
|
1381 | 1379 | if _add_dir_if_not_there(seen_dirs, fname): |
|
1382 | 1380 | if os.path.islink(fname): |
|
1383 | 1381 | for hgname in walkrepos(fname, True, seen_dirs): |
|
1384 | 1382 | yield hgname |
|
1385 | 1383 | else: |
|
1386 | 1384 | newdirs.append(d) |
|
1387 | 1385 | dirs[:] = newdirs |
|
1388 | 1386 | |
|
1389 | 1387 | _rcpath = None |
|
1390 | 1388 | |
|
1391 | 1389 | def os_rcpath(): |
|
1392 | 1390 | '''return default os-specific hgrc search path''' |
|
1393 | 1391 | path = system_rcpath() |
|
1394 | 1392 | path.extend(user_rcpath()) |
|
1395 | 1393 | path = [os.path.normpath(f) for f in path] |
|
1396 | 1394 | return path |
|
1397 | 1395 | |
|
1398 | 1396 | def rcpath(): |
|
1399 | 1397 | '''return hgrc search path. if env var HGRCPATH is set, use it. |
|
1400 | 1398 | for each item in path, if directory, use files ending in .rc, |
|
1401 | 1399 | else use item. |
|
1402 | 1400 | make HGRCPATH empty to only look in .hg/hgrc of current repo. |
|
1403 | 1401 | if no HGRCPATH, use default os-specific path.''' |
|
1404 | 1402 | global _rcpath |
|
1405 | 1403 | if _rcpath is None: |
|
1406 | 1404 | if 'HGRCPATH' in os.environ: |
|
1407 | 1405 | _rcpath = [] |
|
1408 | 1406 | for p in os.environ['HGRCPATH'].split(os.pathsep): |
|
1409 | 1407 | if not p: continue |
|
1410 | 1408 | if os.path.isdir(p): |
|
1411 | 1409 | for f, kind in osutil.listdir(p): |
|
1412 | 1410 | if f.endswith('.rc'): |
|
1413 | 1411 | _rcpath.append(os.path.join(p, f)) |
|
1414 | 1412 | else: |
|
1415 | 1413 | _rcpath.append(p) |
|
1416 | 1414 | else: |
|
1417 | 1415 | _rcpath = os_rcpath() |
|
1418 | 1416 | return _rcpath |
|
1419 | 1417 | |
|
1420 | 1418 | def bytecount(nbytes): |
|
1421 | 1419 | '''return byte count formatted as readable string, with units''' |
|
1422 | 1420 | |
|
1423 | 1421 | units = ( |
|
1424 | 1422 | (100, 1<<30, _('%.0f GB')), |
|
1425 | 1423 | (10, 1<<30, _('%.1f GB')), |
|
1426 | 1424 | (1, 1<<30, _('%.2f GB')), |
|
1427 | 1425 | (100, 1<<20, _('%.0f MB')), |
|
1428 | 1426 | (10, 1<<20, _('%.1f MB')), |
|
1429 | 1427 | (1, 1<<20, _('%.2f MB')), |
|
1430 | 1428 | (100, 1<<10, _('%.0f KB')), |
|
1431 | 1429 | (10, 1<<10, _('%.1f KB')), |
|
1432 | 1430 | (1, 1<<10, _('%.2f KB')), |
|
1433 | 1431 | (1, 1, _('%.0f bytes')), |
|
1434 | 1432 | ) |
|
1435 | 1433 | |
|
1436 | 1434 | for multiplier, divisor, format in units: |
|
1437 | 1435 | if nbytes >= divisor * multiplier: |
|
1438 | 1436 | return format % (nbytes / float(divisor)) |
|
1439 | 1437 | return units[-1][2] % nbytes |
|
1440 | 1438 | |
|
1441 | 1439 | def drop_scheme(scheme, path): |
|
1442 | 1440 | sc = scheme + ':' |
|
1443 | 1441 | if path.startswith(sc): |
|
1444 | 1442 | path = path[len(sc):] |
|
1445 | 1443 | if path.startswith('//'): |
|
1446 | 1444 | path = path[2:] |
|
1447 | 1445 | return path |
|
1448 | 1446 | |
|
1449 | 1447 | def uirepr(s): |
|
1450 | 1448 | # Avoid double backslash in Windows path repr() |
|
1451 | 1449 | return repr(s).replace('\\\\', '\\') |
|
1452 | 1450 | |
|
1453 | 1451 | def termwidth(): |
|
1454 | 1452 | if 'COLUMNS' in os.environ: |
|
1455 | 1453 | try: |
|
1456 | 1454 | return int(os.environ['COLUMNS']) |
|
1457 | 1455 | except ValueError: |
|
1458 | 1456 | pass |
|
1459 | 1457 | try: |
|
1460 | 1458 | import termios, array, fcntl |
|
1461 | 1459 | for dev in (sys.stdout, sys.stdin): |
|
1462 | 1460 | try: |
|
1463 | 1461 | fd = dev.fileno() |
|
1464 | 1462 | if not os.isatty(fd): |
|
1465 | 1463 | continue |
|
1466 | 1464 | arri = fcntl.ioctl(fd, termios.TIOCGWINSZ, '\0' * 8) |
|
1467 | 1465 | return array.array('h', arri)[1] |
|
1468 | 1466 | except ValueError: |
|
1469 | 1467 | pass |
|
1470 | 1468 | except ImportError: |
|
1471 | 1469 | pass |
|
1472 | 1470 | return 80 |
|
1473 | 1471 | |
|
1474 | 1472 | def iterlines(iterator): |
|
1475 | 1473 | for chunk in iterator: |
|
1476 | 1474 | for line in chunk.splitlines(): |
|
1477 | 1475 | yield line |
@@ -1,379 +1,377 b'' | |||
|
1 | ''' | |
|
2 | win32.py - utility functions that use win32 API | |
|
3 | ||
|
4 | Copyright 2005-2009 Matt Mackall <mpm@selenic.com> and others | |
|
5 | ||
|
6 | This software may be used and distributed according to the terms of the | |
|
7 | GNU General Public License version 2, incorporated herein by reference. | |
|
8 | ||
|
9 | Mark Hammond's win32all package allows better functionality on | |
|
10 | Windows. this module overrides definitions in util.py. if not | |
|
11 | available, import of this module will fail, and generic code will be | |
|
12 | used. | |
|
13 | ''' | |
|
1 | # win32.py - utility functions that use win32 API | |
|
2 | # | |
|
3 | # Copyright 2005-2009 Matt Mackall <mpm@selenic.com> and others | |
|
4 | # | |
|
5 | # This software may be used and distributed according to the terms of the | |
|
6 | # GNU General Public License version 2, incorporated herein by reference. | |
|
7 | # | |
|
8 | # Mark Hammond's win32all package allows better functionality on | |
|
9 | # Windows. this module overrides definitions in util.py. if not | |
|
10 | # available, import of this module will fail, and generic code will be | |
|
11 | # used. | |
|
14 | 12 | |
|
15 | 13 | import win32api |
|
16 | 14 | |
|
17 | 15 | import errno, os, sys, pywintypes, win32con, win32file, win32process |
|
18 | 16 | import cStringIO, winerror |
|
19 | 17 | import osutil, encoding |
|
20 | 18 | import util |
|
21 | 19 | from win32com.shell import shell,shellcon |
|
22 | 20 | |
|
23 | 21 | class WinError(Exception): |
|
24 | 22 | winerror_map = { |
|
25 | 23 | winerror.ERROR_ACCESS_DENIED: errno.EACCES, |
|
26 | 24 | winerror.ERROR_ACCOUNT_DISABLED: errno.EACCES, |
|
27 | 25 | winerror.ERROR_ACCOUNT_RESTRICTION: errno.EACCES, |
|
28 | 26 | winerror.ERROR_ALREADY_ASSIGNED: errno.EBUSY, |
|
29 | 27 | winerror.ERROR_ALREADY_EXISTS: errno.EEXIST, |
|
30 | 28 | winerror.ERROR_ARITHMETIC_OVERFLOW: errno.ERANGE, |
|
31 | 29 | winerror.ERROR_BAD_COMMAND: errno.EIO, |
|
32 | 30 | winerror.ERROR_BAD_DEVICE: errno.ENODEV, |
|
33 | 31 | winerror.ERROR_BAD_DRIVER_LEVEL: errno.ENXIO, |
|
34 | 32 | winerror.ERROR_BAD_EXE_FORMAT: errno.ENOEXEC, |
|
35 | 33 | winerror.ERROR_BAD_FORMAT: errno.ENOEXEC, |
|
36 | 34 | winerror.ERROR_BAD_LENGTH: errno.EINVAL, |
|
37 | 35 | winerror.ERROR_BAD_PATHNAME: errno.ENOENT, |
|
38 | 36 | winerror.ERROR_BAD_PIPE: errno.EPIPE, |
|
39 | 37 | winerror.ERROR_BAD_UNIT: errno.ENODEV, |
|
40 | 38 | winerror.ERROR_BAD_USERNAME: errno.EINVAL, |
|
41 | 39 | winerror.ERROR_BROKEN_PIPE: errno.EPIPE, |
|
42 | 40 | winerror.ERROR_BUFFER_OVERFLOW: errno.ENAMETOOLONG, |
|
43 | 41 | winerror.ERROR_BUSY: errno.EBUSY, |
|
44 | 42 | winerror.ERROR_BUSY_DRIVE: errno.EBUSY, |
|
45 | 43 | winerror.ERROR_CALL_NOT_IMPLEMENTED: errno.ENOSYS, |
|
46 | 44 | winerror.ERROR_CANNOT_MAKE: errno.EACCES, |
|
47 | 45 | winerror.ERROR_CANTOPEN: errno.EIO, |
|
48 | 46 | winerror.ERROR_CANTREAD: errno.EIO, |
|
49 | 47 | winerror.ERROR_CANTWRITE: errno.EIO, |
|
50 | 48 | winerror.ERROR_CRC: errno.EIO, |
|
51 | 49 | winerror.ERROR_CURRENT_DIRECTORY: errno.EACCES, |
|
52 | 50 | winerror.ERROR_DEVICE_IN_USE: errno.EBUSY, |
|
53 | 51 | winerror.ERROR_DEV_NOT_EXIST: errno.ENODEV, |
|
54 | 52 | winerror.ERROR_DIRECTORY: errno.EINVAL, |
|
55 | 53 | winerror.ERROR_DIR_NOT_EMPTY: errno.ENOTEMPTY, |
|
56 | 54 | winerror.ERROR_DISK_CHANGE: errno.EIO, |
|
57 | 55 | winerror.ERROR_DISK_FULL: errno.ENOSPC, |
|
58 | 56 | winerror.ERROR_DRIVE_LOCKED: errno.EBUSY, |
|
59 | 57 | winerror.ERROR_ENVVAR_NOT_FOUND: errno.EINVAL, |
|
60 | 58 | winerror.ERROR_EXE_MARKED_INVALID: errno.ENOEXEC, |
|
61 | 59 | winerror.ERROR_FILENAME_EXCED_RANGE: errno.ENAMETOOLONG, |
|
62 | 60 | winerror.ERROR_FILE_EXISTS: errno.EEXIST, |
|
63 | 61 | winerror.ERROR_FILE_INVALID: errno.ENODEV, |
|
64 | 62 | winerror.ERROR_FILE_NOT_FOUND: errno.ENOENT, |
|
65 | 63 | winerror.ERROR_GEN_FAILURE: errno.EIO, |
|
66 | 64 | winerror.ERROR_HANDLE_DISK_FULL: errno.ENOSPC, |
|
67 | 65 | winerror.ERROR_INSUFFICIENT_BUFFER: errno.ENOMEM, |
|
68 | 66 | winerror.ERROR_INVALID_ACCESS: errno.EACCES, |
|
69 | 67 | winerror.ERROR_INVALID_ADDRESS: errno.EFAULT, |
|
70 | 68 | winerror.ERROR_INVALID_BLOCK: errno.EFAULT, |
|
71 | 69 | winerror.ERROR_INVALID_DATA: errno.EINVAL, |
|
72 | 70 | winerror.ERROR_INVALID_DRIVE: errno.ENODEV, |
|
73 | 71 | winerror.ERROR_INVALID_EXE_SIGNATURE: errno.ENOEXEC, |
|
74 | 72 | winerror.ERROR_INVALID_FLAGS: errno.EINVAL, |
|
75 | 73 | winerror.ERROR_INVALID_FUNCTION: errno.ENOSYS, |
|
76 | 74 | winerror.ERROR_INVALID_HANDLE: errno.EBADF, |
|
77 | 75 | winerror.ERROR_INVALID_LOGON_HOURS: errno.EACCES, |
|
78 | 76 | winerror.ERROR_INVALID_NAME: errno.EINVAL, |
|
79 | 77 | winerror.ERROR_INVALID_OWNER: errno.EINVAL, |
|
80 | 78 | winerror.ERROR_INVALID_PARAMETER: errno.EINVAL, |
|
81 | 79 | winerror.ERROR_INVALID_PASSWORD: errno.EPERM, |
|
82 | 80 | winerror.ERROR_INVALID_PRIMARY_GROUP: errno.EINVAL, |
|
83 | 81 | winerror.ERROR_INVALID_SIGNAL_NUMBER: errno.EINVAL, |
|
84 | 82 | winerror.ERROR_INVALID_TARGET_HANDLE: errno.EIO, |
|
85 | 83 | winerror.ERROR_INVALID_WORKSTATION: errno.EACCES, |
|
86 | 84 | winerror.ERROR_IO_DEVICE: errno.EIO, |
|
87 | 85 | winerror.ERROR_IO_INCOMPLETE: errno.EINTR, |
|
88 | 86 | winerror.ERROR_LOCKED: errno.EBUSY, |
|
89 | 87 | winerror.ERROR_LOCK_VIOLATION: errno.EACCES, |
|
90 | 88 | winerror.ERROR_LOGON_FAILURE: errno.EACCES, |
|
91 | 89 | winerror.ERROR_MAPPED_ALIGNMENT: errno.EINVAL, |
|
92 | 90 | winerror.ERROR_META_EXPANSION_TOO_LONG: errno.E2BIG, |
|
93 | 91 | winerror.ERROR_MORE_DATA: errno.EPIPE, |
|
94 | 92 | winerror.ERROR_NEGATIVE_SEEK: errno.ESPIPE, |
|
95 | 93 | winerror.ERROR_NOACCESS: errno.EFAULT, |
|
96 | 94 | winerror.ERROR_NONE_MAPPED: errno.EINVAL, |
|
97 | 95 | winerror.ERROR_NOT_ENOUGH_MEMORY: errno.ENOMEM, |
|
98 | 96 | winerror.ERROR_NOT_READY: errno.EAGAIN, |
|
99 | 97 | winerror.ERROR_NOT_SAME_DEVICE: errno.EXDEV, |
|
100 | 98 | winerror.ERROR_NO_DATA: errno.EPIPE, |
|
101 | 99 | winerror.ERROR_NO_MORE_SEARCH_HANDLES: errno.EIO, |
|
102 | 100 | winerror.ERROR_NO_PROC_SLOTS: errno.EAGAIN, |
|
103 | 101 | winerror.ERROR_NO_SUCH_PRIVILEGE: errno.EACCES, |
|
104 | 102 | winerror.ERROR_OPEN_FAILED: errno.EIO, |
|
105 | 103 | winerror.ERROR_OPEN_FILES: errno.EBUSY, |
|
106 | 104 | winerror.ERROR_OPERATION_ABORTED: errno.EINTR, |
|
107 | 105 | winerror.ERROR_OUTOFMEMORY: errno.ENOMEM, |
|
108 | 106 | winerror.ERROR_PASSWORD_EXPIRED: errno.EACCES, |
|
109 | 107 | winerror.ERROR_PATH_BUSY: errno.EBUSY, |
|
110 | 108 | winerror.ERROR_PATH_NOT_FOUND: errno.ENOENT, |
|
111 | 109 | winerror.ERROR_PIPE_BUSY: errno.EBUSY, |
|
112 | 110 | winerror.ERROR_PIPE_CONNECTED: errno.EPIPE, |
|
113 | 111 | winerror.ERROR_PIPE_LISTENING: errno.EPIPE, |
|
114 | 112 | winerror.ERROR_PIPE_NOT_CONNECTED: errno.EPIPE, |
|
115 | 113 | winerror.ERROR_PRIVILEGE_NOT_HELD: errno.EACCES, |
|
116 | 114 | winerror.ERROR_READ_FAULT: errno.EIO, |
|
117 | 115 | winerror.ERROR_SEEK: errno.EIO, |
|
118 | 116 | winerror.ERROR_SEEK_ON_DEVICE: errno.ESPIPE, |
|
119 | 117 | winerror.ERROR_SHARING_BUFFER_EXCEEDED: errno.ENFILE, |
|
120 | 118 | winerror.ERROR_SHARING_VIOLATION: errno.EACCES, |
|
121 | 119 | winerror.ERROR_STACK_OVERFLOW: errno.ENOMEM, |
|
122 | 120 | winerror.ERROR_SWAPERROR: errno.ENOENT, |
|
123 | 121 | winerror.ERROR_TOO_MANY_MODULES: errno.EMFILE, |
|
124 | 122 | winerror.ERROR_TOO_MANY_OPEN_FILES: errno.EMFILE, |
|
125 | 123 | winerror.ERROR_UNRECOGNIZED_MEDIA: errno.ENXIO, |
|
126 | 124 | winerror.ERROR_UNRECOGNIZED_VOLUME: errno.ENODEV, |
|
127 | 125 | winerror.ERROR_WAIT_NO_CHILDREN: errno.ECHILD, |
|
128 | 126 | winerror.ERROR_WRITE_FAULT: errno.EIO, |
|
129 | 127 | winerror.ERROR_WRITE_PROTECT: errno.EROFS, |
|
130 | 128 | } |
|
131 | 129 | |
|
132 | 130 | def __init__(self, err): |
|
133 | 131 | self.win_errno, self.win_function, self.win_strerror = err |
|
134 | 132 | if self.win_strerror.endswith('.'): |
|
135 | 133 | self.win_strerror = self.win_strerror[:-1] |
|
136 | 134 | |
|
137 | 135 | class WinIOError(WinError, IOError): |
|
138 | 136 | def __init__(self, err, filename=None): |
|
139 | 137 | WinError.__init__(self, err) |
|
140 | 138 | IOError.__init__(self, self.winerror_map.get(self.win_errno, 0), |
|
141 | 139 | self.win_strerror) |
|
142 | 140 | self.filename = filename |
|
143 | 141 | |
|
144 | 142 | class WinOSError(WinError, OSError): |
|
145 | 143 | def __init__(self, err): |
|
146 | 144 | WinError.__init__(self, err) |
|
147 | 145 | OSError.__init__(self, self.winerror_map.get(self.win_errno, 0), |
|
148 | 146 | self.win_strerror) |
|
149 | 147 | |
|
150 | 148 | def os_link(src, dst): |
|
151 | 149 | try: |
|
152 | 150 | win32file.CreateHardLink(dst, src) |
|
153 | 151 | # CreateHardLink sometimes succeeds on mapped drives but |
|
154 | 152 | # following nlinks() returns 1. Check it now and bail out. |
|
155 | 153 | if nlinks(src) < 2: |
|
156 | 154 | try: |
|
157 | 155 | win32file.DeleteFile(dst) |
|
158 | 156 | except: |
|
159 | 157 | pass |
|
160 | 158 | # Fake hardlinking error |
|
161 | 159 | raise WinOSError((18, 'CreateHardLink', 'The system cannot ' |
|
162 | 160 | 'move the file to a different disk drive')) |
|
163 | 161 | except pywintypes.error, details: |
|
164 | 162 | raise WinOSError(details) |
|
165 | 163 | except NotImplementedError: # Another fake error win Win98 |
|
166 | 164 | raise WinOSError((18, 'CreateHardLink', 'Hardlinking not supported')) |
|
167 | 165 | |
|
168 | 166 | def nlinks(pathname): |
|
169 | 167 | """Return number of hardlinks for the given file.""" |
|
170 | 168 | try: |
|
171 | 169 | fh = win32file.CreateFile(pathname, |
|
172 | 170 | win32file.GENERIC_READ, win32file.FILE_SHARE_READ, |
|
173 | 171 | None, win32file.OPEN_EXISTING, 0, None) |
|
174 | 172 | res = win32file.GetFileInformationByHandle(fh) |
|
175 | 173 | fh.Close() |
|
176 | 174 | return res[7] |
|
177 | 175 | except pywintypes.error: |
|
178 | 176 | return os.lstat(pathname).st_nlink |
|
179 | 177 | |
|
180 | 178 | def testpid(pid): |
|
181 | 179 | '''return True if pid is still running or unable to |
|
182 | 180 | determine, False otherwise''' |
|
183 | 181 | try: |
|
184 | 182 | handle = win32api.OpenProcess( |
|
185 | 183 | win32con.PROCESS_QUERY_INFORMATION, False, pid) |
|
186 | 184 | if handle: |
|
187 | 185 | status = win32process.GetExitCodeProcess(handle) |
|
188 | 186 | return status == win32con.STILL_ACTIVE |
|
189 | 187 | except pywintypes.error, details: |
|
190 | 188 | return details[0] != winerror.ERROR_INVALID_PARAMETER |
|
191 | 189 | return True |
|
192 | 190 | |
|
193 | 191 | def lookup_reg(key, valname=None, scope=None): |
|
194 | 192 | ''' Look up a key/value name in the Windows registry. |
|
195 | 193 | |
|
196 | 194 | valname: value name. If unspecified, the default value for the key |
|
197 | 195 | is used. |
|
198 | 196 | scope: optionally specify scope for registry lookup, this can be |
|
199 | 197 | a sequence of scopes to look up in order. Default (CURRENT_USER, |
|
200 | 198 | LOCAL_MACHINE). |
|
201 | 199 | ''' |
|
202 | 200 | try: |
|
203 | 201 | from _winreg import HKEY_CURRENT_USER, HKEY_LOCAL_MACHINE, \ |
|
204 | 202 | QueryValueEx, OpenKey |
|
205 | 203 | except ImportError: |
|
206 | 204 | return None |
|
207 | 205 | |
|
208 | 206 | if scope is None: |
|
209 | 207 | scope = (HKEY_CURRENT_USER, HKEY_LOCAL_MACHINE) |
|
210 | 208 | elif not isinstance(scope, (list, tuple)): |
|
211 | 209 | scope = (scope,) |
|
212 | 210 | for s in scope: |
|
213 | 211 | try: |
|
214 | 212 | val = QueryValueEx(OpenKey(s, key), valname)[0] |
|
215 | 213 | # never let a Unicode string escape into the wild |
|
216 | 214 | return encoding.tolocal(val.encode('UTF-8')) |
|
217 | 215 | except EnvironmentError: |
|
218 | 216 | pass |
|
219 | 217 | |
|
220 | 218 | def system_rcpath_win32(): |
|
221 | 219 | '''return default os-specific hgrc search path''' |
|
222 | 220 | proc = win32api.GetCurrentProcess() |
|
223 | 221 | try: |
|
224 | 222 | # This will fail on windows < NT |
|
225 | 223 | filename = win32process.GetModuleFileNameEx(proc, 0) |
|
226 | 224 | except: |
|
227 | 225 | filename = win32api.GetModuleFileName(0) |
|
228 | 226 | # Use mercurial.ini found in directory with hg.exe |
|
229 | 227 | progrc = os.path.join(os.path.dirname(filename), 'mercurial.ini') |
|
230 | 228 | if os.path.isfile(progrc): |
|
231 | 229 | return [progrc] |
|
232 | 230 | # else look for a system rcpath in the registry |
|
233 | 231 | try: |
|
234 | 232 | value = win32api.RegQueryValue( |
|
235 | 233 | win32con.HKEY_LOCAL_MACHINE, 'SOFTWARE\\Mercurial') |
|
236 | 234 | rcpath = [] |
|
237 | 235 | for p in value.split(os.pathsep): |
|
238 | 236 | if p.lower().endswith('mercurial.ini'): |
|
239 | 237 | rcpath.append(p) |
|
240 | 238 | elif os.path.isdir(p): |
|
241 | 239 | for f, kind in osutil.listdir(p): |
|
242 | 240 | if f.endswith('.rc'): |
|
243 | 241 | rcpath.append(os.path.join(p, f)) |
|
244 | 242 | return rcpath |
|
245 | 243 | except pywintypes.error: |
|
246 | 244 | return [] |
|
247 | 245 | |
|
248 | 246 | def user_rcpath_win32(): |
|
249 | 247 | '''return os-specific hgrc search path to the user dir''' |
|
250 | 248 | userdir = os.path.expanduser('~') |
|
251 | 249 | if sys.getwindowsversion()[3] != 2 and userdir == '~': |
|
252 | 250 | # We are on win < nt: fetch the APPDATA directory location and use |
|
253 | 251 | # the parent directory as the user home dir. |
|
254 | 252 | appdir = shell.SHGetPathFromIDList( |
|
255 | 253 | shell.SHGetSpecialFolderLocation(0, shellcon.CSIDL_APPDATA)) |
|
256 | 254 | userdir = os.path.dirname(appdir) |
|
257 | 255 | return [os.path.join(userdir, 'mercurial.ini'), |
|
258 | 256 | os.path.join(userdir, '.hgrc')] |
|
259 | 257 | |
|
260 | 258 | class posixfile_nt(object): |
|
261 | 259 | '''file object with posix-like semantics. on windows, normal |
|
262 | 260 | files can not be deleted or renamed if they are open. must open |
|
263 | 261 | with win32file.FILE_SHARE_DELETE. this flag does not exist on |
|
264 | 262 | windows < nt, so do not use this class there.''' |
|
265 | 263 | |
|
266 | 264 | # tried to use win32file._open_osfhandle to pass fd to os.fdopen, |
|
267 | 265 | # but does not work at all. wrap win32 file api instead. |
|
268 | 266 | |
|
269 | 267 | def __init__(self, name, mode='rb'): |
|
270 | 268 | self.closed = False |
|
271 | 269 | self.name = name |
|
272 | 270 | self.mode = mode |
|
273 | 271 | access = 0 |
|
274 | 272 | if 'r' in mode or '+' in mode: |
|
275 | 273 | access |= win32file.GENERIC_READ |
|
276 | 274 | if 'w' in mode or 'a' in mode or '+' in mode: |
|
277 | 275 | access |= win32file.GENERIC_WRITE |
|
278 | 276 | if 'r' in mode: |
|
279 | 277 | creation = win32file.OPEN_EXISTING |
|
280 | 278 | elif 'a' in mode: |
|
281 | 279 | creation = win32file.OPEN_ALWAYS |
|
282 | 280 | else: |
|
283 | 281 | creation = win32file.CREATE_ALWAYS |
|
284 | 282 | try: |
|
285 | 283 | self.handle = win32file.CreateFile(name, |
|
286 | 284 | access, |
|
287 | 285 | win32file.FILE_SHARE_READ | |
|
288 | 286 | win32file.FILE_SHARE_WRITE | |
|
289 | 287 | win32file.FILE_SHARE_DELETE, |
|
290 | 288 | None, |
|
291 | 289 | creation, |
|
292 | 290 | win32file.FILE_ATTRIBUTE_NORMAL, |
|
293 | 291 | 0) |
|
294 | 292 | except pywintypes.error, err: |
|
295 | 293 | raise WinIOError(err, name) |
|
296 | 294 | |
|
297 | 295 | def __iter__(self): |
|
298 | 296 | for line in self.readlines(): |
|
299 | 297 | yield line |
|
300 | 298 | |
|
301 | 299 | def read(self, count=-1): |
|
302 | 300 | try: |
|
303 | 301 | cs = cStringIO.StringIO() |
|
304 | 302 | while count: |
|
305 | 303 | wincount = int(count) |
|
306 | 304 | if wincount == -1: |
|
307 | 305 | wincount = 1048576 |
|
308 | 306 | val, data = win32file.ReadFile(self.handle, wincount) |
|
309 | 307 | if not data: break |
|
310 | 308 | cs.write(data) |
|
311 | 309 | if count != -1: |
|
312 | 310 | count -= len(data) |
|
313 | 311 | return cs.getvalue() |
|
314 | 312 | except pywintypes.error, err: |
|
315 | 313 | raise WinIOError(err) |
|
316 | 314 | |
|
317 | 315 | def readlines(self, sizehint=None): |
|
318 | 316 | # splitlines() splits on single '\r' while readlines() |
|
319 | 317 | # does not. cStringIO has a well behaving readlines() and is fast. |
|
320 | 318 | return cStringIO.StringIO(self.read()).readlines() |
|
321 | 319 | |
|
322 | 320 | def write(self, data): |
|
323 | 321 | try: |
|
324 | 322 | if 'a' in self.mode: |
|
325 | 323 | win32file.SetFilePointer(self.handle, 0, win32file.FILE_END) |
|
326 | 324 | nwrit = 0 |
|
327 | 325 | while nwrit < len(data): |
|
328 | 326 | val, nwrit = win32file.WriteFile(self.handle, data) |
|
329 | 327 | data = data[nwrit:] |
|
330 | 328 | except pywintypes.error, err: |
|
331 | 329 | raise WinIOError(err) |
|
332 | 330 | |
|
333 | 331 | def writelines(self, sequence): |
|
334 | 332 | for s in sequence: |
|
335 | 333 | self.write(s) |
|
336 | 334 | |
|
337 | 335 | def seek(self, pos, whence=0): |
|
338 | 336 | try: |
|
339 | 337 | win32file.SetFilePointer(self.handle, int(pos), whence) |
|
340 | 338 | except pywintypes.error, err: |
|
341 | 339 | raise WinIOError(err) |
|
342 | 340 | |
|
343 | 341 | def tell(self): |
|
344 | 342 | try: |
|
345 | 343 | return win32file.SetFilePointer(self.handle, 0, |
|
346 | 344 | win32file.FILE_CURRENT) |
|
347 | 345 | except pywintypes.error, err: |
|
348 | 346 | raise WinIOError(err) |
|
349 | 347 | |
|
350 | 348 | def close(self): |
|
351 | 349 | if not self.closed: |
|
352 | 350 | self.handle = None |
|
353 | 351 | self.closed = True |
|
354 | 352 | |
|
355 | 353 | def flush(self): |
|
356 | 354 | # we have no application-level buffering |
|
357 | 355 | pass |
|
358 | 356 | |
|
359 | 357 | def truncate(self, pos=0): |
|
360 | 358 | try: |
|
361 | 359 | win32file.SetFilePointer(self.handle, int(pos), |
|
362 | 360 | win32file.FILE_BEGIN) |
|
363 | 361 | win32file.SetEndOfFile(self.handle) |
|
364 | 362 | except pywintypes.error, err: |
|
365 | 363 | raise WinIOError(err) |
|
366 | 364 | |
|
367 | 365 | def getuser(): |
|
368 | 366 | '''return name of current user''' |
|
369 | 367 | return win32api.GetUserName() |
|
370 | 368 | |
|
371 | 369 | def set_signal_handler_win32(): |
|
372 | 370 | """Register a termination handler for console events including |
|
373 | 371 | CTRL+C. python signal handlers do not work well with socket |
|
374 | 372 | operations. |
|
375 | 373 | """ |
|
376 | 374 | def handler(event): |
|
377 | 375 | win32process.ExitProcess(1) |
|
378 | 376 | win32api.SetConsoleCtrlHandler(handler) |
|
379 | 377 |
@@ -1,252 +1,250 b'' | |||
|
1 | """ | |
|
2 | windows.py - Windows utility function implementations for Mercurial | |
|
3 | ||
|
4 | Copyright 2005-2009 Matt Mackall <mpm@selenic.com> and others | |
|
5 | ||
|
6 | This software may be used and distributed according to the terms of the | |
|
7 | GNU General Public License version 2, incorporated herein by reference. | |
|
8 | """ | |
|
1 | # windows.py - Windows utility function implementations for Mercurial | |
|
2 | # | |
|
3 | # Copyright 2005-2009 Matt Mackall <mpm@selenic.com> and others | |
|
4 | # | |
|
5 | # This software may be used and distributed according to the terms of the | |
|
6 | # GNU General Public License version 2, incorporated herein by reference. | |
|
9 | 7 | |
|
10 | 8 | from i18n import _ |
|
11 | 9 | import errno, msvcrt, os, osutil, re, sys, error |
|
12 | 10 | nulldev = 'NUL:' |
|
13 | 11 | |
|
14 | 12 | umask = 002 |
|
15 | 13 | |
|
16 | 14 | class winstdout: |
|
17 | 15 | '''stdout on windows misbehaves if sent through a pipe''' |
|
18 | 16 | |
|
19 | 17 | def __init__(self, fp): |
|
20 | 18 | self.fp = fp |
|
21 | 19 | |
|
22 | 20 | def __getattr__(self, key): |
|
23 | 21 | return getattr(self.fp, key) |
|
24 | 22 | |
|
25 | 23 | def close(self): |
|
26 | 24 | try: |
|
27 | 25 | self.fp.close() |
|
28 | 26 | except: pass |
|
29 | 27 | |
|
30 | 28 | def write(self, s): |
|
31 | 29 | try: |
|
32 | 30 | # This is workaround for "Not enough space" error on |
|
33 | 31 | # writing large size of data to console. |
|
34 | 32 | limit = 16000 |
|
35 | 33 | l = len(s) |
|
36 | 34 | start = 0 |
|
37 | 35 | while start < l: |
|
38 | 36 | end = start + limit |
|
39 | 37 | self.fp.write(s[start:end]) |
|
40 | 38 | start = end |
|
41 | 39 | except IOError, inst: |
|
42 | 40 | if inst.errno != 0: raise |
|
43 | 41 | self.close() |
|
44 | 42 | raise IOError(errno.EPIPE, 'Broken pipe') |
|
45 | 43 | |
|
46 | 44 | def flush(self): |
|
47 | 45 | try: |
|
48 | 46 | return self.fp.flush() |
|
49 | 47 | except IOError, inst: |
|
50 | 48 | if inst.errno != errno.EINVAL: raise |
|
51 | 49 | self.close() |
|
52 | 50 | raise IOError(errno.EPIPE, 'Broken pipe') |
|
53 | 51 | |
|
54 | 52 | sys.stdout = winstdout(sys.stdout) |
|
55 | 53 | |
|
56 | 54 | def _is_win_9x(): |
|
57 | 55 | '''return true if run on windows 95, 98 or me.''' |
|
58 | 56 | try: |
|
59 | 57 | return sys.getwindowsversion()[3] == 1 |
|
60 | 58 | except AttributeError: |
|
61 | 59 | return 'command' in os.environ.get('comspec', '') |
|
62 | 60 | |
|
63 | 61 | def openhardlinks(): |
|
64 | 62 | return not _is_win_9x and "win32api" in locals() |
|
65 | 63 | |
|
66 | 64 | def system_rcpath(): |
|
67 | 65 | try: |
|
68 | 66 | return system_rcpath_win32() |
|
69 | 67 | except: |
|
70 | 68 | return [r'c:\mercurial\mercurial.ini'] |
|
71 | 69 | |
|
72 | 70 | def user_rcpath(): |
|
73 | 71 | '''return os-specific hgrc search path to the user dir''' |
|
74 | 72 | try: |
|
75 | 73 | path = user_rcpath_win32() |
|
76 | 74 | except: |
|
77 | 75 | home = os.path.expanduser('~') |
|
78 | 76 | path = [os.path.join(home, 'mercurial.ini'), |
|
79 | 77 | os.path.join(home, '.hgrc')] |
|
80 | 78 | userprofile = os.environ.get('USERPROFILE') |
|
81 | 79 | if userprofile: |
|
82 | 80 | path.append(os.path.join(userprofile, 'mercurial.ini')) |
|
83 | 81 | path.append(os.path.join(userprofile, '.hgrc')) |
|
84 | 82 | return path |
|
85 | 83 | |
|
86 | 84 | def parse_patch_output(output_line): |
|
87 | 85 | """parses the output produced by patch and returns the file name""" |
|
88 | 86 | pf = output_line[14:] |
|
89 | 87 | if pf[0] == '`': |
|
90 | 88 | pf = pf[1:-1] # Remove the quotes |
|
91 | 89 | return pf |
|
92 | 90 | |
|
93 | 91 | def sshargs(sshcmd, host, user, port): |
|
94 | 92 | '''Build argument list for ssh or Plink''' |
|
95 | 93 | pflag = 'plink' in sshcmd.lower() and '-P' or '-p' |
|
96 | 94 | args = user and ("%s@%s" % (user, host)) or host |
|
97 | 95 | return port and ("%s %s %s" % (args, pflag, port)) or args |
|
98 | 96 | |
|
99 | 97 | def testpid(pid): |
|
100 | 98 | '''return False if pid dead, True if running or not known''' |
|
101 | 99 | return True |
|
102 | 100 | |
|
103 | 101 | def set_flags(f, l, x): |
|
104 | 102 | pass |
|
105 | 103 | |
|
106 | 104 | def set_binary(fd): |
|
107 | 105 | # When run without console, pipes may expose invalid |
|
108 | 106 | # fileno(), usually set to -1. |
|
109 | 107 | if hasattr(fd, 'fileno') and fd.fileno() >= 0: |
|
110 | 108 | msvcrt.setmode(fd.fileno(), os.O_BINARY) |
|
111 | 109 | |
|
112 | 110 | def pconvert(path): |
|
113 | 111 | return '/'.join(path.split(os.sep)) |
|
114 | 112 | |
|
115 | 113 | def localpath(path): |
|
116 | 114 | return path.replace('/', '\\') |
|
117 | 115 | |
|
118 | 116 | def normpath(path): |
|
119 | 117 | return pconvert(os.path.normpath(path)) |
|
120 | 118 | |
|
121 | 119 | def samestat(s1, s2): |
|
122 | 120 | return False |
|
123 | 121 | |
|
124 | 122 | # A sequence of backslashes is special iff it precedes a double quote: |
|
125 | 123 | # - if there's an even number of backslashes, the double quote is not |
|
126 | 124 | # quoted (i.e. it ends the quoted region) |
|
127 | 125 | # - if there's an odd number of backslashes, the double quote is quoted |
|
128 | 126 | # - in both cases, every pair of backslashes is unquoted into a single |
|
129 | 127 | # backslash |
|
130 | 128 | # (See http://msdn2.microsoft.com/en-us/library/a1y7w461.aspx ) |
|
131 | 129 | # So, to quote a string, we must surround it in double quotes, double |
|
132 | 130 | # the number of backslashes that preceed double quotes and add another |
|
133 | 131 | # backslash before every double quote (being careful with the double |
|
134 | 132 | # quote we've appended to the end) |
|
135 | 133 | _quotere = None |
|
136 | 134 | def shellquote(s): |
|
137 | 135 | global _quotere |
|
138 | 136 | if _quotere is None: |
|
139 | 137 | _quotere = re.compile(r'(\\*)("|\\$)') |
|
140 | 138 | return '"%s"' % _quotere.sub(r'\1\1\\\2', s) |
|
141 | 139 | |
|
142 | 140 | def quotecommand(cmd): |
|
143 | 141 | """Build a command string suitable for os.popen* calls.""" |
|
144 | 142 | # The extra quotes are needed because popen* runs the command |
|
145 | 143 | # through the current COMSPEC. cmd.exe suppress enclosing quotes. |
|
146 | 144 | return '"' + cmd + '"' |
|
147 | 145 | |
|
148 | 146 | def popen(command, mode='r'): |
|
149 | 147 | # Work around "popen spawned process may not write to stdout |
|
150 | 148 | # under windows" |
|
151 | 149 | # http://bugs.python.org/issue1366 |
|
152 | 150 | command += " 2> %s" % nulldev |
|
153 | 151 | return os.popen(quotecommand(command), mode) |
|
154 | 152 | |
|
155 | 153 | def explain_exit(code): |
|
156 | 154 | return _("exited with status %d") % code, code |
|
157 | 155 | |
|
158 | 156 | # if you change this stub into a real check, please try to implement the |
|
159 | 157 | # username and groupname functions above, too. |
|
160 | 158 | def isowner(fp, st=None): |
|
161 | 159 | return True |
|
162 | 160 | |
|
163 | 161 | def find_exe(command): |
|
164 | 162 | '''Find executable for command searching like cmd.exe does. |
|
165 | 163 | If command is a basename then PATH is searched for command. |
|
166 | 164 | PATH isn't searched if command is an absolute or relative path. |
|
167 | 165 | An extension from PATHEXT is found and added if not present. |
|
168 | 166 | If command isn't found None is returned.''' |
|
169 | 167 | pathext = os.environ.get('PATHEXT', '.COM;.EXE;.BAT;.CMD') |
|
170 | 168 | pathexts = [ext for ext in pathext.lower().split(os.pathsep)] |
|
171 | 169 | if os.path.splitext(command)[1].lower() in pathexts: |
|
172 | 170 | pathexts = [''] |
|
173 | 171 | |
|
174 | 172 | def findexisting(pathcommand): |
|
175 | 173 | 'Will append extension (if needed) and return existing file' |
|
176 | 174 | for ext in pathexts: |
|
177 | 175 | executable = pathcommand + ext |
|
178 | 176 | if os.path.exists(executable): |
|
179 | 177 | return executable |
|
180 | 178 | return None |
|
181 | 179 | |
|
182 | 180 | if os.sep in command: |
|
183 | 181 | return findexisting(command) |
|
184 | 182 | |
|
185 | 183 | for path in os.environ.get('PATH', '').split(os.pathsep): |
|
186 | 184 | executable = findexisting(os.path.join(path, command)) |
|
187 | 185 | if executable is not None: |
|
188 | 186 | return executable |
|
189 | 187 | return None |
|
190 | 188 | |
|
191 | 189 | def set_signal_handler(): |
|
192 | 190 | try: |
|
193 | 191 | set_signal_handler_win32() |
|
194 | 192 | except NameError: |
|
195 | 193 | pass |
|
196 | 194 | |
|
197 | 195 | def statfiles(files): |
|
198 | 196 | '''Stat each file in files and yield stat or None if file does not exist. |
|
199 | 197 | Cluster and cache stat per directory to minimize number of OS stat calls.''' |
|
200 | 198 | ncase = os.path.normcase |
|
201 | 199 | sep = os.sep |
|
202 | 200 | dircache = {} # dirname -> filename -> status | None if file does not exist |
|
203 | 201 | for nf in files: |
|
204 | 202 | nf = ncase(nf) |
|
205 | 203 | pos = nf.rfind(sep) |
|
206 | 204 | if pos == -1: |
|
207 | 205 | dir, base = '.', nf |
|
208 | 206 | else: |
|
209 | 207 | dir, base = nf[:pos+1], nf[pos+1:] |
|
210 | 208 | cache = dircache.get(dir, None) |
|
211 | 209 | if cache is None: |
|
212 | 210 | try: |
|
213 | 211 | dmap = dict([(ncase(n), s) |
|
214 | 212 | for n, k, s in osutil.listdir(dir, True)]) |
|
215 | 213 | except OSError, err: |
|
216 | 214 | # handle directory not found in Python version prior to 2.5 |
|
217 | 215 | # Python <= 2.4 returns native Windows code 3 in errno |
|
218 | 216 | # Python >= 2.5 returns ENOENT and adds winerror field |
|
219 | 217 | # EINVAL is raised if dir is not a directory. |
|
220 | 218 | if err.errno not in (3, errno.ENOENT, errno.EINVAL, |
|
221 | 219 | errno.ENOTDIR): |
|
222 | 220 | raise |
|
223 | 221 | dmap = {} |
|
224 | 222 | cache = dircache.setdefault(dir, dmap) |
|
225 | 223 | yield cache.get(base, None) |
|
226 | 224 | |
|
227 | 225 | def getuser(): |
|
228 | 226 | '''return name of current user''' |
|
229 | 227 | raise error.Abort(_('user name not available - set USERNAME ' |
|
230 | 228 | 'environment variable')) |
|
231 | 229 | |
|
232 | 230 | def username(uid=None): |
|
233 | 231 | """Return the name of the user with the given uid. |
|
234 | 232 | |
|
235 | 233 | If uid is None, return the name of the current user.""" |
|
236 | 234 | return None |
|
237 | 235 | |
|
238 | 236 | def groupname(gid=None): |
|
239 | 237 | """Return the name of the group with the given gid. |
|
240 | 238 | |
|
241 | 239 | If gid is None, return the name of the current group.""" |
|
242 | 240 | return None |
|
243 | 241 | |
|
244 | 242 | posixfile = file |
|
245 | 243 | try: |
|
246 | 244 | # override functions with win32 versions if possible |
|
247 | 245 | from win32 import * |
|
248 | 246 | if not _is_win_9x(): |
|
249 | 247 | posixfile = posixfile_nt |
|
250 | 248 | except ImportError: |
|
251 | 249 | pass |
|
252 | 250 |
General Comments 0
You need to be logged in to leave comments.
Login now