Show More
@@ -1,353 +1,353 b'' | |||
|
1 | 1 | # convert.py Foreign SCM converter |
|
2 | 2 | # |
|
3 | 3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
4 | 4 | # |
|
5 | 5 | # This software may be used and distributed according to the terms |
|
6 | 6 | # of the GNU General Public License, incorporated herein by reference. |
|
7 | 7 | |
|
8 | 8 | from common import NoRepo, converter_source, converter_sink |
|
9 | 9 | from cvs import convert_cvs |
|
10 | 10 | from git import convert_git |
|
11 | 11 | from hg import convert_mercurial |
|
12 | 12 | from subversion import convert_svn |
|
13 | 13 | |
|
14 | 14 | import os, shutil |
|
15 | 15 | from mercurial import hg, ui, util, commands |
|
16 | 16 | |
|
17 | 17 | commands.norepo += " convert" |
|
18 | 18 | |
|
19 | 19 | converters = [convert_cvs, convert_git, convert_svn, convert_mercurial] |
|
20 | 20 | |
|
21 | 21 | def convertsource(ui, path, **opts): |
|
22 | 22 | for c in converters: |
|
23 | 23 | if not hasattr(c, 'getcommit'): |
|
24 | 24 | continue |
|
25 | 25 | try: |
|
26 | 26 | return c(ui, path, **opts) |
|
27 | 27 | except NoRepo: |
|
28 | 28 | pass |
|
29 | 29 | raise util.Abort('%s: unknown repository type' % path) |
|
30 | 30 | |
|
31 | 31 | def convertsink(ui, path): |
|
32 | 32 | if not os.path.isdir(path): |
|
33 | 33 | raise util.Abort("%s: not a directory" % path) |
|
34 | 34 | for c in converters: |
|
35 | 35 | if not hasattr(c, 'putcommit'): |
|
36 | 36 | continue |
|
37 | 37 | try: |
|
38 | 38 | return c(ui, path) |
|
39 | 39 | except NoRepo: |
|
40 | 40 | pass |
|
41 | 41 | raise util.Abort('%s: unknown repository type' % path) |
|
42 | 42 | |
|
43 | 43 | class convert(object): |
|
44 | 44 | def __init__(self, ui, source, dest, mapfile, opts): |
|
45 | 45 | |
|
46 | 46 | self.source = source |
|
47 | 47 | self.dest = dest |
|
48 | 48 | self.ui = ui |
|
49 | 49 | self.opts = opts |
|
50 | 50 | self.commitcache = {} |
|
51 | 51 | self.mapfile = mapfile |
|
52 | 52 | self.mapfilefd = None |
|
53 | 53 | self.authors = {} |
|
54 | 54 | self.authorfile = None |
|
55 | 55 | |
|
56 | 56 | self.map = {} |
|
57 | 57 | try: |
|
58 | 58 | origmapfile = open(self.mapfile, 'r') |
|
59 | 59 | for l in origmapfile: |
|
60 | 60 | sv, dv = l[:-1].split() |
|
61 | 61 | self.map[sv] = dv |
|
62 | 62 | origmapfile.close() |
|
63 | 63 | except IOError: |
|
64 | 64 | pass |
|
65 | 65 | |
|
66 | 66 | # Read first the dst author map if any |
|
67 | 67 | authorfile = self.dest.authorfile() |
|
68 | 68 | if authorfile and os.path.exists(authorfile): |
|
69 | 69 | self.readauthormap(authorfile) |
|
70 | 70 | # Extend/Override with new author map if necessary |
|
71 | 71 | if opts.get('authors'): |
|
72 | 72 | self.readauthormap(opts.get('authors')) |
|
73 | 73 | self.authorfile = self.dest.authorfile() |
|
74 | 74 | |
|
75 | 75 | def walktree(self, heads): |
|
76 | 76 | '''Return a mapping that identifies the uncommitted parents of every |
|
77 | 77 | uncommitted changeset.''' |
|
78 | 78 | visit = heads |
|
79 | 79 | known = {} |
|
80 | 80 | parents = {} |
|
81 | 81 | while visit: |
|
82 | 82 | n = visit.pop(0) |
|
83 | 83 | if n in known or n in self.map: continue |
|
84 | 84 | known[n] = 1 |
|
85 | 85 | self.commitcache[n] = self.source.getcommit(n) |
|
86 | 86 | cp = self.commitcache[n].parents |
|
87 | 87 | parents[n] = [] |
|
88 | 88 | for p in cp: |
|
89 | 89 | parents[n].append(p) |
|
90 | 90 | visit.append(p) |
|
91 | 91 | |
|
92 | 92 | return parents |
|
93 | 93 | |
|
94 | 94 | def toposort(self, parents): |
|
95 | 95 | '''Return an ordering such that every uncommitted changeset is |
|
96 | 96 | preceeded by all its uncommitted ancestors.''' |
|
97 | 97 | visit = parents.keys() |
|
98 | 98 | seen = {} |
|
99 | 99 | children = {} |
|
100 | 100 | |
|
101 | 101 | while visit: |
|
102 | 102 | n = visit.pop(0) |
|
103 | 103 | if n in seen: continue |
|
104 | 104 | seen[n] = 1 |
|
105 | 105 | # Ensure that nodes without parents are present in the 'children' |
|
106 | 106 | # mapping. |
|
107 | 107 | children.setdefault(n, []) |
|
108 | 108 | for p in parents[n]: |
|
109 | 109 | if not p in self.map: |
|
110 | 110 | visit.append(p) |
|
111 | 111 | children.setdefault(p, []).append(n) |
|
112 | 112 | |
|
113 | 113 | s = [] |
|
114 | 114 | removed = {} |
|
115 | 115 | visit = children.keys() |
|
116 | 116 | while visit: |
|
117 | 117 | n = visit.pop(0) |
|
118 | 118 | if n in removed: continue |
|
119 | 119 | dep = 0 |
|
120 | 120 | if n in parents: |
|
121 | 121 | for p in parents[n]: |
|
122 | 122 | if p in self.map: continue |
|
123 | 123 | if p not in removed: |
|
124 | 124 | # we're still dependent |
|
125 | 125 | visit.append(n) |
|
126 | 126 | dep = 1 |
|
127 | 127 | break |
|
128 | 128 | |
|
129 | 129 | if not dep: |
|
130 | 130 | # all n's parents are in the list |
|
131 | 131 | removed[n] = 1 |
|
132 | 132 | if n not in self.map: |
|
133 | 133 | s.append(n) |
|
134 | 134 | if n in children: |
|
135 | 135 | for c in children[n]: |
|
136 | 136 | visit.insert(0, c) |
|
137 | 137 | |
|
138 | 138 | if self.opts.get('datesort'): |
|
139 | 139 | depth = {} |
|
140 | 140 | for n in s: |
|
141 | 141 | depth[n] = 0 |
|
142 | 142 | pl = [p for p in self.commitcache[n].parents |
|
143 | 143 | if p not in self.map] |
|
144 | 144 | if pl: |
|
145 | 145 | depth[n] = max([depth[p] for p in pl]) + 1 |
|
146 | 146 | |
|
147 | 147 | s = [(depth[n], self.commitcache[n].date, n) for n in s] |
|
148 | 148 | s.sort() |
|
149 | 149 | s = [e[2] for e in s] |
|
150 | 150 | |
|
151 | 151 | return s |
|
152 | 152 | |
|
153 | 153 | def mapentry(self, src, dst): |
|
154 | 154 | if self.mapfilefd is None: |
|
155 | 155 | try: |
|
156 | 156 | self.mapfilefd = open(self.mapfile, "a") |
|
157 | 157 | except IOError, (errno, strerror): |
|
158 | 158 | raise util.Abort("Could not open map file %s: %s, %s\n" % (self.mapfile, errno, strerror)) |
|
159 | 159 | self.map[src] = dst |
|
160 | 160 | self.mapfilefd.write("%s %s\n" % (src, dst)) |
|
161 | 161 | self.mapfilefd.flush() |
|
162 | 162 | |
|
163 | 163 | def writeauthormap(self): |
|
164 | 164 | authorfile = self.authorfile |
|
165 | 165 | if authorfile: |
|
166 | 166 | self.ui.status('Writing author map file %s\n' % authorfile) |
|
167 | 167 | ofile = open(authorfile, 'w+') |
|
168 | 168 | for author in self.authors: |
|
169 | 169 | ofile.write("%s=%s\n" % (author, self.authors[author])) |
|
170 | 170 | ofile.close() |
|
171 | 171 | |
|
172 | 172 | def readauthormap(self, authorfile): |
|
173 | 173 | afile = open(authorfile, 'r') |
|
174 | 174 | for line in afile: |
|
175 | 175 | try: |
|
176 | 176 | srcauthor = line.split('=')[0].strip() |
|
177 | 177 | dstauthor = line.split('=')[1].strip() |
|
178 | 178 | if srcauthor in self.authors and dstauthor != self.authors[srcauthor]: |
|
179 | 179 | self.ui.status( |
|
180 | 180 | 'Overriding mapping for author %s, was %s, will be %s\n' |
|
181 | 181 | % (srcauthor, self.authors[srcauthor], dstauthor)) |
|
182 | 182 | else: |
|
183 | 183 | self.ui.debug('Mapping author %s to %s\n' |
|
184 | 184 | % (srcauthor, dstauthor)) |
|
185 | 185 | self.authors[srcauthor] = dstauthor |
|
186 | 186 | except IndexError: |
|
187 | 187 | self.ui.warn( |
|
188 | 188 | 'Ignoring bad line in author file map %s: %s\n' |
|
189 | 189 | % (authorfile, line)) |
|
190 | 190 | afile.close() |
|
191 | 191 | |
|
192 | 192 | def copy(self, rev): |
|
193 | 193 | c = self.commitcache[rev] |
|
194 | 194 | files = self.source.getchanges(rev) |
|
195 | ||
|
195 | ||
|
196 | 196 | do_copies = (hasattr(c, 'copies') and hasattr(self.dest, 'copyfile')) |
|
197 | 197 | |
|
198 | 198 | for f, v in files: |
|
199 | 199 | try: |
|
200 | 200 | data = self.source.getfile(f, v) |
|
201 | 201 | except IOError, inst: |
|
202 | 202 | self.dest.delfile(f) |
|
203 | 203 | else: |
|
204 | 204 | e = self.source.getmode(f, v) |
|
205 | 205 | self.dest.putfile(f, e, data) |
|
206 | 206 | if do_copies: |
|
207 | 207 | if f in c.copies: |
|
208 | 208 | # Merely marks that a copy happened. |
|
209 | 209 | self.dest.copyfile(c.copies[f], f) |
|
210 | 210 | |
|
211 | 211 | |
|
212 | 212 | r = [self.map[v] for v in c.parents] |
|
213 | 213 | f = [f for f, v in files] |
|
214 | 214 | newnode = self.dest.putcommit(f, r, c) |
|
215 | 215 | self.mapentry(rev, newnode) |
|
216 | 216 | |
|
217 | 217 | def convert(self): |
|
218 | 218 | try: |
|
219 | 219 | self.source.setrevmap(self.map) |
|
220 | 220 | self.ui.status("scanning source...\n") |
|
221 | 221 | heads = self.source.getheads() |
|
222 | 222 | parents = self.walktree(heads) |
|
223 | 223 | self.ui.status("sorting...\n") |
|
224 | 224 | t = self.toposort(parents) |
|
225 | 225 | num = len(t) |
|
226 | 226 | c = None |
|
227 | 227 | |
|
228 | 228 | self.ui.status("converting...\n") |
|
229 | 229 | for c in t: |
|
230 | 230 | num -= 1 |
|
231 | 231 | desc = self.commitcache[c].desc |
|
232 | 232 | if "\n" in desc: |
|
233 | 233 | desc = desc.splitlines()[0] |
|
234 | 234 | author = self.commitcache[c].author |
|
235 | 235 | author = self.authors.get(author, author) |
|
236 | 236 | self.commitcache[c].author = author |
|
237 | 237 | self.ui.status("%d %s\n" % (num, desc)) |
|
238 | 238 | self.copy(c) |
|
239 | 239 | |
|
240 | 240 | tags = self.source.gettags() |
|
241 | 241 | ctags = {} |
|
242 | 242 | for k in tags: |
|
243 | 243 | v = tags[k] |
|
244 | 244 | if v in self.map: |
|
245 | 245 | ctags[k] = self.map[v] |
|
246 | 246 | |
|
247 | 247 | if c and ctags: |
|
248 | 248 | nrev = self.dest.puttags(ctags) |
|
249 | 249 | # write another hash correspondence to override the previous |
|
250 | 250 | # one so we don't end up with extra tag heads |
|
251 | 251 | if nrev: |
|
252 | 252 | self.mapentry(c, nrev) |
|
253 | 253 | |
|
254 | 254 | self.writeauthormap() |
|
255 | 255 | finally: |
|
256 | 256 | self.cleanup() |
|
257 | 257 | |
|
258 | 258 | def cleanup(self): |
|
259 | 259 | if self.mapfilefd: |
|
260 | 260 | self.mapfilefd.close() |
|
261 | 261 | |
|
262 | 262 | def _convert(ui, src, dest=None, mapfile=None, **opts): |
|
263 | 263 | '''Convert a foreign SCM repository to a Mercurial one. |
|
264 | 264 | |
|
265 | 265 | Accepted source formats: |
|
266 | 266 | - GIT |
|
267 | 267 | - CVS |
|
268 | 268 | - SVN |
|
269 | 269 | |
|
270 | 270 | Accepted destination formats: |
|
271 | 271 | - Mercurial |
|
272 | 272 | |
|
273 | 273 | If no revision is given, all revisions will be converted. Otherwise, |
|
274 | 274 | convert will only import up to the named revision (given in a format |
|
275 | 275 | understood by the source). |
|
276 | 276 | |
|
277 | 277 | If no destination directory name is specified, it defaults to the |
|
278 | 278 | basename of the source with \'-hg\' appended. If the destination |
|
279 | 279 | repository doesn\'t exist, it will be created. |
|
280 | 280 | |
|
281 | 281 | If <mapfile> isn\'t given, it will be put in a default location |
|
282 | 282 | (<dest>/.hg/shamap by default). The <mapfile> is a simple text |
|
283 | 283 | file that maps each source commit ID to the destination ID for |
|
284 | 284 | that revision, like so: |
|
285 | 285 | <source ID> <destination ID> |
|
286 | 286 | |
|
287 | 287 | If the file doesn\'t exist, it\'s automatically created. It\'s updated |
|
288 | 288 | on each commit copied, so convert-repo can be interrupted and can |
|
289 | 289 | be run repeatedly to copy new commits. |
|
290 | 290 | |
|
291 | 291 | The [username mapping] file is a simple text file that maps each source |
|
292 | 292 | commit author to a destination commit author. It is handy for source SCMs |
|
293 | 293 | that use unix logins to identify authors (eg: CVS). One line per author |
|
294 | 294 | mapping and the line format is: |
|
295 | 295 | srcauthor=whatever string you want |
|
296 | 296 | ''' |
|
297 | 297 | |
|
298 | 298 | util._encoding = 'UTF-8' |
|
299 | 299 | |
|
300 | 300 | if not dest: |
|
301 | 301 | dest = hg.defaultdest(src) + "-hg" |
|
302 | 302 | ui.status("assuming destination %s\n" % dest) |
|
303 | 303 | |
|
304 | 304 | # Try to be smart and initalize things when required |
|
305 | 305 | created = False |
|
306 | 306 | if os.path.isdir(dest): |
|
307 | 307 | if len(os.listdir(dest)) > 0: |
|
308 | 308 | try: |
|
309 | 309 | hg.repository(ui, dest) |
|
310 | 310 | ui.status("destination %s is a Mercurial repository\n" % dest) |
|
311 | 311 | except hg.RepoError: |
|
312 | 312 | raise util.Abort( |
|
313 | 313 | "destination directory %s is not empty.\n" |
|
314 | 314 | "Please specify an empty directory to be initialized\n" |
|
315 | 315 | "or an already initialized mercurial repository" |
|
316 | 316 | % dest) |
|
317 | 317 | else: |
|
318 | 318 | ui.status("initializing destination %s repository\n" % dest) |
|
319 | 319 | hg.repository(ui, dest, create=True) |
|
320 | 320 | created = True |
|
321 | 321 | elif os.path.exists(dest): |
|
322 | 322 | raise util.Abort("destination %s exists and is not a directory" % dest) |
|
323 | 323 | else: |
|
324 | 324 | ui.status("initializing destination %s repository\n" % dest) |
|
325 | 325 | hg.repository(ui, dest, create=True) |
|
326 | 326 | created = True |
|
327 | 327 | |
|
328 | 328 | destc = convertsink(ui, dest) |
|
329 | 329 | |
|
330 | 330 | try: |
|
331 | 331 | srcc = convertsource(ui, src, rev=opts.get('rev')) |
|
332 | 332 | except Exception: |
|
333 | 333 | if created: |
|
334 | 334 | shutil.rmtree(dest, True) |
|
335 | 335 | raise |
|
336 | 336 | |
|
337 | 337 | if not mapfile: |
|
338 | 338 | try: |
|
339 | 339 | mapfile = destc.mapfile() |
|
340 | 340 | except: |
|
341 | 341 | mapfile = os.path.join(destc, "map") |
|
342 | 342 | |
|
343 | 343 | c = convert(ui, srcc, destc, mapfile, opts) |
|
344 | 344 | c.convert() |
|
345 | 345 | |
|
346 | 346 | cmdtable = { |
|
347 | 347 | "convert": |
|
348 | 348 | (_convert, |
|
349 | 349 | [('A', 'authors', '', 'username mapping filename'), |
|
350 | 350 | ('r', 'rev', '', 'import up to target revision REV'), |
|
351 | 351 | ('', 'datesort', None, 'try to sort changesets by date')], |
|
352 | 352 | 'hg convert [OPTION]... SOURCE [DEST [MAPFILE]]'), |
|
353 | 353 | } |
@@ -1,121 +1,121 b'' | |||
|
1 | 1 | # common code for the convert extension |
|
2 | 2 | |
|
3 | 3 | class NoRepo(Exception): pass |
|
4 | 4 | |
|
5 | 5 | class commit(object): |
|
6 | 6 | def __init__(self, **parts): |
|
7 | 7 | self.rev = None |
|
8 | 8 | self.branch = None |
|
9 | 9 | |
|
10 | 10 | for x in "author date desc parents".split(): |
|
11 | 11 | if not x in parts: |
|
12 | 12 | raise util.Abort("commit missing field %s" % x) |
|
13 | 13 | self.__dict__.update(parts) |
|
14 | 14 | if not self.desc or self.desc.isspace(): |
|
15 | 15 | self.desc = '*** empty log message ***' |
|
16 | 16 | |
|
17 | 17 | class converter_source(object): |
|
18 | 18 | """Conversion source interface""" |
|
19 | 19 | |
|
20 | 20 | def __init__(self, ui, path, rev=None): |
|
21 | 21 | """Initialize conversion source (or raise NoRepo("message") |
|
22 | 22 | exception if path is not a valid repository)""" |
|
23 | 23 | self.ui = ui |
|
24 | 24 | self.path = path |
|
25 | 25 | self.rev = rev |
|
26 | 26 | |
|
27 | 27 | self.encoding = 'utf-8' |
|
28 | 28 | |
|
29 | 29 | def setrevmap(self, revmap): |
|
30 | 30 | """set the map of already-converted revisions""" |
|
31 | 31 | pass |
|
32 | 32 | |
|
33 | 33 | def getheads(self): |
|
34 | 34 | """Return a list of this repository's heads""" |
|
35 | 35 | raise NotImplementedError() |
|
36 | 36 | |
|
37 | 37 | def getfile(self, name, rev): |
|
38 | 38 | """Return file contents as a string""" |
|
39 | 39 | raise NotImplementedError() |
|
40 | 40 | |
|
41 | 41 | def getmode(self, name, rev): |
|
42 | 42 | """Return file mode, eg. '', 'x', or 'l'""" |
|
43 | 43 | raise NotImplementedError() |
|
44 | 44 | |
|
45 | 45 | def getchanges(self, version): |
|
46 | 46 | """Return sorted list of (filename, id) tuples for all files changed in rev. |
|
47 | 47 | |
|
48 | 48 | id just tells us which revision to return in getfile(), e.g. in |
|
49 | 49 | git it's an object hash.""" |
|
50 | 50 | raise NotImplementedError() |
|
51 | 51 | |
|
52 | 52 | def getcommit(self, version): |
|
53 | 53 | """Return the commit object for version""" |
|
54 | 54 | raise NotImplementedError() |
|
55 | 55 | |
|
56 | 56 | def gettags(self): |
|
57 | 57 | """Return the tags as a dictionary of name: revision""" |
|
58 | 58 | raise NotImplementedError() |
|
59 | 59 | |
|
60 | 60 | def recode(self, s, encoding=None): |
|
61 | 61 | if not encoding: |
|
62 | 62 | encoding = self.encoding or 'utf-8' |
|
63 | ||
|
63 | ||
|
64 | 64 | try: |
|
65 | 65 | return s.decode(encoding).encode("utf-8") |
|
66 | 66 | except: |
|
67 | 67 | try: |
|
68 | 68 | return s.decode("latin-1").encode("utf-8") |
|
69 | 69 | except: |
|
70 | 70 | return s.decode(encoding, "replace").encode("utf-8") |
|
71 | 71 | |
|
72 | 72 | class converter_sink(object): |
|
73 | 73 | """Conversion sink (target) interface""" |
|
74 | 74 | |
|
75 | 75 | def __init__(self, ui, path): |
|
76 | 76 | """Initialize conversion sink (or raise NoRepo("message") |
|
77 | 77 | exception if path is not a valid repository)""" |
|
78 | 78 | raise NotImplementedError() |
|
79 | 79 | |
|
80 | 80 | def getheads(self): |
|
81 | 81 | """Return a list of this repository's heads""" |
|
82 | 82 | raise NotImplementedError() |
|
83 | 83 | |
|
84 | 84 | def mapfile(self): |
|
85 | 85 | """Path to a file that will contain lines |
|
86 | 86 | source_rev_id sink_rev_id |
|
87 | 87 | mapping equivalent revision identifiers for each system.""" |
|
88 | 88 | raise NotImplementedError() |
|
89 | 89 | |
|
90 | 90 | def authorfile(self): |
|
91 | 91 | """Path to a file that will contain lines |
|
92 | 92 | srcauthor=dstauthor |
|
93 | 93 | mapping equivalent authors identifiers for each system.""" |
|
94 | 94 | return None |
|
95 | 95 | |
|
96 | 96 | def putfile(self, f, e, data): |
|
97 | 97 | """Put file for next putcommit(). |
|
98 | 98 | f: path to file |
|
99 | 99 | e: '', 'x', or 'l' (regular file, executable, or symlink) |
|
100 | 100 | data: file contents""" |
|
101 | 101 | raise NotImplementedError() |
|
102 | 102 | |
|
103 | 103 | def delfile(self, f): |
|
104 | 104 | """Delete file for next putcommit(). |
|
105 | 105 | f: path to file""" |
|
106 | 106 | raise NotImplementedError() |
|
107 | 107 | |
|
108 | 108 | def putcommit(self, files, parents, commit): |
|
109 | 109 | """Create a revision with all changed files listed in 'files' |
|
110 | 110 | and having listed parents. 'commit' is a commit object containing |
|
111 | 111 | at a minimum the author, date, and message for this changeset. |
|
112 | 112 | Called after putfile() and delfile() calls. Note that the sink |
|
113 | 113 | repository is not told to update itself to a particular revision |
|
114 | 114 | (or even what that revision would be) before it receives the |
|
115 | 115 | file data.""" |
|
116 | 116 | raise NotImplementedError() |
|
117 | 117 | |
|
118 | 118 | def puttags(self, tags): |
|
119 | 119 | """Put tags into sink. |
|
120 | 120 | tags: {tagname: sink_rev_id, ...}""" |
|
121 | 121 | raise NotImplementedError() |
@@ -1,97 +1,97 b'' | |||
|
1 | 1 | # hg backend for convert extension |
|
2 | 2 | |
|
3 | 3 | import os, time |
|
4 | 4 | from mercurial import hg |
|
5 | 5 | |
|
6 | 6 | from common import NoRepo, converter_sink |
|
7 | 7 | |
|
8 | 8 | class convert_mercurial(converter_sink): |
|
9 | 9 | def __init__(self, ui, path): |
|
10 | 10 | self.path = path |
|
11 | 11 | self.ui = ui |
|
12 | 12 | try: |
|
13 | 13 | self.repo = hg.repository(self.ui, path) |
|
14 | 14 | except: |
|
15 | 15 | raise NoRepo("could open hg repo %s" % path) |
|
16 | 16 | |
|
17 | 17 | def mapfile(self): |
|
18 | 18 | return os.path.join(self.path, ".hg", "shamap") |
|
19 | 19 | |
|
20 | 20 | def authorfile(self): |
|
21 | 21 | return os.path.join(self.path, ".hg", "authormap") |
|
22 | 22 | |
|
23 | 23 | def getheads(self): |
|
24 | 24 | h = self.repo.changelog.heads() |
|
25 | 25 | return [ hg.hex(x) for x in h ] |
|
26 | 26 | |
|
27 | 27 | def putfile(self, f, e, data): |
|
28 | 28 | self.repo.wwrite(f, data, e) |
|
29 | 29 | if self.repo.dirstate.state(f) == '?': |
|
30 | 30 | self.repo.dirstate.update([f], "a") |
|
31 | 31 | |
|
32 | 32 | def copyfile(self, source, dest): |
|
33 | 33 | self.repo.copy(source, dest) |
|
34 | 34 | |
|
35 | 35 | def delfile(self, f): |
|
36 | 36 | try: |
|
37 | 37 | os.unlink(self.repo.wjoin(f)) |
|
38 | 38 | #self.repo.remove([f]) |
|
39 | 39 | except: |
|
40 | 40 | pass |
|
41 | 41 | |
|
42 | 42 | def putcommit(self, files, parents, commit): |
|
43 | 43 | seen = {} |
|
44 | 44 | pl = [] |
|
45 | 45 | for p in parents: |
|
46 | 46 | if p not in seen: |
|
47 | 47 | pl.append(p) |
|
48 | 48 | seen[p] = 1 |
|
49 | 49 | parents = pl |
|
50 | 50 | |
|
51 | 51 | if len(parents) < 2: parents.append("0" * 40) |
|
52 | 52 | if len(parents) < 2: parents.append("0" * 40) |
|
53 | 53 | p2 = parents.pop(0) |
|
54 | 54 | |
|
55 | 55 | text = commit.desc |
|
56 | 56 | extra = {} |
|
57 | 57 | if commit.branch: |
|
58 | 58 | extra['branch'] = commit.branch |
|
59 | 59 | if commit.rev: |
|
60 | 60 | extra['convert_revision'] = commit.rev |
|
61 | ||
|
61 | ||
|
62 | 62 | while parents: |
|
63 | 63 | p1 = p2 |
|
64 | 64 | p2 = parents.pop(0) |
|
65 | 65 | a = self.repo.rawcommit(files, text, commit.author, commit.date, |
|
66 | 66 | hg.bin(p1), hg.bin(p2), extra=extra) |
|
67 | 67 | text = "(octopus merge fixup)\n" |
|
68 | 68 | p2 = hg.hex(self.repo.changelog.tip()) |
|
69 | 69 | |
|
70 | 70 | return p2 |
|
71 | 71 | |
|
72 | 72 | def puttags(self, tags): |
|
73 | 73 | try: |
|
74 | 74 | old = self.repo.wfile(".hgtags").read() |
|
75 | 75 | oldlines = old.splitlines(1) |
|
76 | 76 | oldlines.sort() |
|
77 | 77 | except: |
|
78 | 78 | oldlines = [] |
|
79 | 79 | |
|
80 | 80 | k = tags.keys() |
|
81 | 81 | k.sort() |
|
82 | 82 | newlines = [] |
|
83 | 83 | for tag in k: |
|
84 | 84 | newlines.append("%s %s\n" % (tags[tag], tag)) |
|
85 | 85 | |
|
86 | 86 | newlines.sort() |
|
87 | 87 | |
|
88 | 88 | if newlines != oldlines: |
|
89 | 89 | self.ui.status("updating tags\n") |
|
90 | 90 | f = self.repo.wfile(".hgtags", "w") |
|
91 | 91 | f.write("".join(newlines)) |
|
92 | 92 | f.close() |
|
93 | 93 | if not oldlines: self.repo.add([".hgtags"]) |
|
94 | 94 | date = "%s 0" % int(time.mktime(time.gmtime())) |
|
95 | 95 | self.repo.rawcommit([".hgtags"], "update tags", "convert-repo", |
|
96 | 96 | date, self.repo.changelog.tip(), hg.nullid) |
|
97 | 97 | return hg.hex(self.repo.changelog.tip()) |
@@ -1,669 +1,669 b'' | |||
|
1 | 1 | # Subversion 1.4/1.5 Python API backend |
|
2 | 2 | # |
|
3 | 3 | # Copyright(C) 2007 Daniel Holth et al |
|
4 | 4 | # |
|
5 | 5 | # Configuration options: |
|
6 | 6 | # |
|
7 | 7 | # convert.svn.trunk |
|
8 | 8 | # Relative path to the trunk (default: "trunk") |
|
9 | 9 | # convert.svn.branches |
|
10 | 10 | # Relative path to tree of branches (default: "branches") |
|
11 | 11 | # |
|
12 | 12 | # Set these in a hgrc, or on the command line as follows: |
|
13 | 13 | # |
|
14 | 14 | # hg convert --config convert.svn.trunk=wackoname [...] |
|
15 | 15 | |
|
16 | 16 | import pprint |
|
17 | 17 | import locale |
|
18 | 18 | import os |
|
19 | 19 | import cPickle as pickle |
|
20 | 20 | from mercurial import util |
|
21 | 21 | |
|
22 | 22 | # Subversion stuff. Works best with very recent Python SVN bindings |
|
23 | 23 | # e.g. SVN 1.5 or backports. Thanks to the bzr folks for enhancing |
|
24 | 24 | # these bindings. |
|
25 | 25 | |
|
26 | 26 | from cStringIO import StringIO |
|
27 | 27 | |
|
28 | 28 | from common import NoRepo, commit, converter_source |
|
29 | 29 | |
|
30 | 30 | try: |
|
31 | 31 | from svn.core import SubversionException, Pool |
|
32 | 32 | import svn.core |
|
33 | 33 | import svn.ra |
|
34 | 34 | import svn.delta |
|
35 | 35 | import svn |
|
36 | 36 | import transport |
|
37 | 37 | except ImportError: |
|
38 | 38 | pass |
|
39 | 39 | |
|
40 | 40 | class CompatibilityException(Exception): pass |
|
41 | 41 | |
|
42 | 42 | class changedpath(object): |
|
43 | 43 | def __init__(self, p): |
|
44 | 44 | self.copyfrom_path = p.copyfrom_path |
|
45 | 45 | self.copyfrom_rev = p.copyfrom_rev |
|
46 | 46 | self.action = p.action |
|
47 | 47 | |
|
48 | 48 | # SVN conversion code stolen from bzr-svn and tailor |
|
49 | 49 | class convert_svn(converter_source): |
|
50 | 50 | def __init__(self, ui, url, rev=None): |
|
51 | 51 | super(convert_svn, self).__init__(ui, url, rev=rev) |
|
52 | 52 | |
|
53 | 53 | try: |
|
54 | 54 | SubversionException |
|
55 | 55 | except NameError: |
|
56 | 56 | msg = 'subversion python bindings could not be loaded\n' |
|
57 | 57 | ui.warn(msg) |
|
58 | 58 | raise NoRepo(msg) |
|
59 | 59 | |
|
60 | 60 | self.encoding = locale.getpreferredencoding() |
|
61 | 61 | self.lastrevs = {} |
|
62 | 62 | |
|
63 | 63 | latest = None |
|
64 | 64 | if rev: |
|
65 | 65 | try: |
|
66 | 66 | latest = int(rev) |
|
67 | 67 | except ValueError: |
|
68 | 68 | raise util.Abort('svn: revision %s is not an integer' % rev) |
|
69 | 69 | try: |
|
70 | 70 | # Support file://path@rev syntax. Useful e.g. to convert |
|
71 | 71 | # deleted branches. |
|
72 | 72 | at = url.rfind('@') |
|
73 | 73 | if at >= 0: |
|
74 | 74 | latest = int(url[at+1:]) |
|
75 | 75 | url = url[:at] |
|
76 | 76 | except ValueError, e: |
|
77 | 77 | pass |
|
78 | 78 | self.url = url |
|
79 | 79 | self.encoding = 'UTF-8' # Subversion is always nominal UTF-8 |
|
80 | 80 | try: |
|
81 | 81 | self.transport = transport.SvnRaTransport(url=url) |
|
82 | 82 | self.ra = self.transport.ra |
|
83 | 83 | self.ctx = self.transport.client |
|
84 | 84 | self.base = svn.ra.get_repos_root(self.ra) |
|
85 | 85 | self.module = self.url[len(self.base):] |
|
86 | 86 | self.modulemap = {} # revision, module |
|
87 | 87 | self.commits = {} |
|
88 | 88 | self.files = {} |
|
89 | 89 | self.uuid = svn.ra.get_uuid(self.ra).decode(self.encoding) |
|
90 | 90 | except SubversionException, e: |
|
91 | 91 | raise NoRepo("couldn't open SVN repo %s" % url) |
|
92 | 92 | |
|
93 | 93 | try: |
|
94 | 94 | self.get_blacklist() |
|
95 | 95 | except IOError, e: |
|
96 | 96 | pass |
|
97 | 97 | |
|
98 | 98 | self.last_changed = self.latest(self.module, latest) |
|
99 | 99 | |
|
100 | 100 | self.head = self.revid(self.last_changed) |
|
101 | 101 | |
|
102 | 102 | def setrevmap(self, revmap): |
|
103 | 103 | lastrevs = {} |
|
104 | 104 | for revid in revmap.keys(): |
|
105 | 105 | uuid, module, revnum = self.revsplit(revid) |
|
106 | 106 | lastrevnum = lastrevs.setdefault(module, revnum) |
|
107 | 107 | if revnum > lastrevnum: |
|
108 | 108 | lastrevs[module] = revnum |
|
109 | 109 | self.lastrevs = lastrevs |
|
110 | 110 | |
|
111 | 111 | def exists(self, path, optrev): |
|
112 | 112 | try: |
|
113 | 113 | return svn.client.ls(self.url.rstrip('/') + '/' + path, |
|
114 | 114 | optrev, False, self.ctx) |
|
115 | 115 | except SubversionException, err: |
|
116 | 116 | return [] |
|
117 | 117 | |
|
118 | 118 | def getheads(self): |
|
119 | 119 | # detect standard /branches, /tags, /trunk layout |
|
120 | 120 | optrev = svn.core.svn_opt_revision_t() |
|
121 | 121 | optrev.kind = svn.core.svn_opt_revision_number |
|
122 | 122 | optrev.value.number = self.last_changed |
|
123 | 123 | rpath = self.url.strip('/') |
|
124 | 124 | cfgtrunk = self.ui.config('convert', 'svn.trunk') |
|
125 | 125 | cfgbranches = self.ui.config('convert', 'svn.branches') |
|
126 | 126 | trunk = (cfgtrunk or 'trunk').strip('/') |
|
127 | 127 | branches = (cfgbranches or 'branches').strip('/') |
|
128 | 128 | if self.exists(trunk, optrev) and self.exists(branches, optrev): |
|
129 | 129 | self.ui.note('found trunk at %r and branches at %r\n' % |
|
130 | 130 | (trunk, branches)) |
|
131 | 131 | oldmodule = self.module |
|
132 | 132 | self.module += '/' + trunk |
|
133 | 133 | lt = self.latest(self.module, self.last_changed) |
|
134 | 134 | self.head = self.revid(lt) |
|
135 | 135 | self.heads = [self.head] |
|
136 | 136 | branchnames = svn.client.ls(rpath + '/' + branches, optrev, False, |
|
137 | 137 | self.ctx) |
|
138 | 138 | for branch in branchnames.keys(): |
|
139 | 139 | if oldmodule: |
|
140 | 140 | module = '/' + oldmodule + '/' + branches + '/' + branch |
|
141 | 141 | else: |
|
142 | 142 | module = '/' + branches + '/' + branch |
|
143 | 143 | brevnum = self.latest(module, self.last_changed) |
|
144 | 144 | brev = self.revid(brevnum, module) |
|
145 | 145 | self.ui.note('found branch %s at %d\n' % (branch, brevnum)) |
|
146 | 146 | self.heads.append(brev) |
|
147 | 147 | elif cfgtrunk or cfgbranches: |
|
148 | 148 | raise util.Abort(_('trunk/branch layout expected, ' |
|
149 | 149 | 'but not found')) |
|
150 | 150 | else: |
|
151 | 151 | self.ui.note('working with one branch\n') |
|
152 | 152 | self.heads = [self.head] |
|
153 | 153 | return self.heads |
|
154 | 154 | |
|
155 | 155 | def getfile(self, file, rev): |
|
156 | 156 | data, mode = self._getfile(file, rev) |
|
157 | 157 | self.modecache[(file, rev)] = mode |
|
158 | 158 | return data |
|
159 | 159 | |
|
160 |
def getmode(self, file, rev): |
|
|
160 | def getmode(self, file, rev): | |
|
161 | 161 | return self.modecache[(file, rev)] |
|
162 | 162 | |
|
163 | 163 | def getchanges(self, rev): |
|
164 | 164 | self.modecache = {} |
|
165 | 165 | files = self.files[rev] |
|
166 | 166 | cl = files |
|
167 | 167 | cl.sort() |
|
168 | 168 | # caller caches the result, so free it here to release memory |
|
169 | 169 | del self.files[rev] |
|
170 | 170 | return cl |
|
171 | 171 | |
|
172 | 172 | def getcommit(self, rev): |
|
173 | 173 | if rev not in self.commits: |
|
174 | 174 | uuid, module, revnum = self.revsplit(rev) |
|
175 | 175 | self.module = module |
|
176 | 176 | self.reparent(module) |
|
177 | 177 | stop = self.lastrevs.get(module, 0) |
|
178 | 178 | self._fetch_revisions(from_revnum=revnum, to_revnum=stop) |
|
179 | 179 | commit = self.commits[rev] |
|
180 | 180 | # caller caches the result, so free it here to release memory |
|
181 | 181 | del self.commits[rev] |
|
182 | 182 | return commit |
|
183 | 183 | |
|
184 | 184 | def get_log(self, paths, start, end, limit=0, discover_changed_paths=True, |
|
185 | 185 | strict_node_history=False): |
|
186 | 186 | '''wrapper for svn.ra.get_log. |
|
187 | 187 | on a large repository, svn.ra.get_log pins huge amounts of |
|
188 | 188 | memory that cannot be recovered. work around it by forking |
|
189 | 189 | and writing results over a pipe.''' |
|
190 | 190 | |
|
191 | 191 | def child(fp): |
|
192 | 192 | protocol = -1 |
|
193 | 193 | def receiver(orig_paths, revnum, author, date, message, pool): |
|
194 | 194 | if orig_paths is not None: |
|
195 | 195 | for k, v in orig_paths.iteritems(): |
|
196 | 196 | orig_paths[k] = changedpath(v) |
|
197 | 197 | pickle.dump((orig_paths, revnum, author, date, message), |
|
198 | 198 | fp, protocol) |
|
199 | 199 | |
|
200 | 200 | try: |
|
201 | 201 | # Use an ra of our own so that our parent can consume |
|
202 | 202 | # our results without confusing the server. |
|
203 | 203 | t = transport.SvnRaTransport(url=self.url) |
|
204 | 204 | svn.ra.get_log(t.ra, paths, start, end, limit, |
|
205 | 205 | discover_changed_paths, |
|
206 | 206 | strict_node_history, |
|
207 | 207 | receiver) |
|
208 | 208 | except SubversionException, (_, num): |
|
209 | 209 | self.ui.print_exc() |
|
210 | 210 | pickle.dump(num, fp, protocol) |
|
211 | 211 | else: |
|
212 | 212 | pickle.dump(None, fp, protocol) |
|
213 | 213 | fp.close() |
|
214 | 214 | |
|
215 | 215 | def parent(fp): |
|
216 | 216 | while True: |
|
217 | 217 | entry = pickle.load(fp) |
|
218 | 218 | try: |
|
219 | 219 | orig_paths, revnum, author, date, message = entry |
|
220 | 220 | except: |
|
221 | 221 | if entry is None: |
|
222 | 222 | break |
|
223 | 223 | raise SubversionException("child raised exception", entry) |
|
224 | 224 | yield entry |
|
225 | 225 | |
|
226 | 226 | rfd, wfd = os.pipe() |
|
227 | 227 | pid = os.fork() |
|
228 | 228 | if pid: |
|
229 | 229 | os.close(wfd) |
|
230 | 230 | for p in parent(os.fdopen(rfd, 'rb')): |
|
231 | 231 | yield p |
|
232 | 232 | ret = os.waitpid(pid, 0)[1] |
|
233 | 233 | if ret: |
|
234 | 234 | raise util.Abort(_('get_log %s') % util.explain_exit(ret)) |
|
235 | 235 | else: |
|
236 | 236 | os.close(rfd) |
|
237 | 237 | child(os.fdopen(wfd, 'wb')) |
|
238 | 238 | os._exit(0) |
|
239 | 239 | |
|
240 | 240 | def gettags(self): |
|
241 | 241 | tags = {} |
|
242 | 242 | start = self.revnum(self.head) |
|
243 | 243 | try: |
|
244 | 244 | for entry in self.get_log(['/tags'], 0, start): |
|
245 | 245 | orig_paths, revnum, author, date, message = entry |
|
246 | 246 | for path in orig_paths: |
|
247 | 247 | if not path.startswith('/tags/'): |
|
248 | 248 | continue |
|
249 | 249 | ent = orig_paths[path] |
|
250 | 250 | source = ent.copyfrom_path |
|
251 | 251 | rev = ent.copyfrom_rev |
|
252 | 252 | tag = path.split('/', 2)[2] |
|
253 | 253 | tags[tag] = self.revid(rev, module=source) |
|
254 | 254 | except SubversionException, (_, num): |
|
255 | 255 | self.ui.note('no tags found at revision %d\n' % start) |
|
256 | 256 | return tags |
|
257 | 257 | |
|
258 | 258 | # -- helper functions -- |
|
259 | 259 | |
|
260 | 260 | def revid(self, revnum, module=None): |
|
261 | 261 | if not module: |
|
262 | 262 | module = self.module |
|
263 | 263 | return (u"svn:%s%s@%s" % (self.uuid, module, revnum)).decode(self.encoding) |
|
264 | 264 | |
|
265 | 265 | def revnum(self, rev): |
|
266 | 266 | return int(rev.split('@')[-1]) |
|
267 | 267 | |
|
268 | 268 | def revsplit(self, rev): |
|
269 | 269 | url, revnum = rev.encode(self.encoding).split('@', 1) |
|
270 | 270 | revnum = int(revnum) |
|
271 | 271 | parts = url.split('/', 1) |
|
272 | 272 | uuid = parts.pop(0)[4:] |
|
273 | 273 | mod = '' |
|
274 | 274 | if parts: |
|
275 | 275 | mod = '/' + parts[0] |
|
276 | 276 | return uuid, mod, revnum |
|
277 | 277 | |
|
278 | 278 | def latest(self, path, stop=0): |
|
279 | 279 | 'find the latest revision affecting path, up to stop' |
|
280 | 280 | if not stop: |
|
281 | 281 | stop = svn.ra.get_latest_revnum(self.ra) |
|
282 | 282 | try: |
|
283 | 283 | self.reparent('') |
|
284 | 284 | dirent = svn.ra.stat(self.ra, path.strip('/'), stop) |
|
285 | 285 | self.reparent(self.module) |
|
286 | 286 | except SubversionException: |
|
287 | 287 | dirent = None |
|
288 | 288 | if not dirent: |
|
289 | 289 | print self.base, path |
|
290 | 290 | raise util.Abort('%s not found up to revision %d' % (path, stop)) |
|
291 | 291 | |
|
292 | 292 | return dirent.created_rev |
|
293 | 293 | |
|
294 | 294 | def get_blacklist(self): |
|
295 | 295 | """Avoid certain revision numbers. |
|
296 | 296 | It is not uncommon for two nearby revisions to cancel each other |
|
297 | 297 | out, e.g. 'I copied trunk into a subdirectory of itself instead |
|
298 | 298 | of making a branch'. The converted repository is significantly |
|
299 | 299 | smaller if we ignore such revisions.""" |
|
300 | 300 | self.blacklist = set() |
|
301 | 301 | blacklist = self.blacklist |
|
302 | 302 | for line in file("blacklist.txt", "r"): |
|
303 | 303 | if not line.startswith("#"): |
|
304 | 304 | try: |
|
305 | 305 | svn_rev = int(line.strip()) |
|
306 | 306 | blacklist.add(svn_rev) |
|
307 | 307 | except ValueError, e: |
|
308 | 308 | pass # not an integer or a comment |
|
309 | 309 | |
|
310 | 310 | def is_blacklisted(self, svn_rev): |
|
311 | 311 | return svn_rev in self.blacklist |
|
312 | 312 | |
|
313 | 313 | def reparent(self, module): |
|
314 | 314 | svn_url = self.base + module |
|
315 | 315 | self.ui.debug("reparent to %s\n" % svn_url.encode(self.encoding)) |
|
316 | 316 | svn.ra.reparent(self.ra, svn_url.encode(self.encoding)) |
|
317 | 317 | |
|
318 | 318 | def _fetch_revisions(self, from_revnum = 0, to_revnum = 347): |
|
319 | 319 | def get_entry_from_path(path, module=self.module): |
|
320 | 320 | # Given the repository url of this wc, say |
|
321 | 321 | # "http://server/plone/CMFPlone/branches/Plone-2_0-branch" |
|
322 | 322 | # extract the "entry" portion (a relative path) from what |
|
323 | 323 | # svn log --xml says, ie |
|
324 | 324 | # "/CMFPlone/branches/Plone-2_0-branch/tests/PloneTestCase.py" |
|
325 | 325 | # that is to say "tests/PloneTestCase.py" |
|
326 | 326 | |
|
327 | 327 | if path.startswith(module): |
|
328 | 328 | relative = path[len(module):] |
|
329 | 329 | if relative.startswith('/'): |
|
330 | 330 | return relative[1:] |
|
331 | 331 | else: |
|
332 | 332 | return relative |
|
333 | 333 | |
|
334 | 334 | # The path is outside our tracked tree... |
|
335 | 335 | self.ui.debug('Ignoring %r since it is not under %r\n' % (path, module)) |
|
336 | 336 | return None |
|
337 | 337 | |
|
338 | 338 | self.child_cset = None |
|
339 | 339 | def parselogentry(orig_paths, revnum, author, date, message): |
|
340 | 340 | self.ui.debug("parsing revision %d (%d changes)\n" % |
|
341 | 341 | (revnum, len(orig_paths))) |
|
342 | 342 | |
|
343 | 343 | if revnum in self.modulemap: |
|
344 | 344 | new_module = self.modulemap[revnum] |
|
345 | 345 | if new_module != self.module: |
|
346 | 346 | self.module = new_module |
|
347 | 347 | self.reparent(self.module) |
|
348 | 348 | |
|
349 | 349 | copyfrom = {} # Map of entrypath, revision for finding source of deleted revisions. |
|
350 | 350 | copies = {} |
|
351 | 351 | entries = [] |
|
352 | 352 | rev = self.revid(revnum) |
|
353 | 353 | parents = [] |
|
354 | 354 | |
|
355 | 355 | # branch log might return entries for a parent we already have |
|
356 | 356 | if (rev in self.commits or |
|
357 | 357 | (revnum < self.lastrevs.get(self.module, 0))): |
|
358 | 358 | return |
|
359 | 359 | |
|
360 | 360 | try: |
|
361 | 361 | branch = self.module.split("/")[-1] |
|
362 | 362 | if branch == 'trunk': |
|
363 | 363 | branch = '' |
|
364 | 364 | except IndexError: |
|
365 | 365 | branch = None |
|
366 | 366 | |
|
367 | 367 | orig_paths = orig_paths.items() |
|
368 | 368 | orig_paths.sort() |
|
369 | 369 | for path, ent in orig_paths: |
|
370 | 370 | # self.ui.write("path %s\n" % path) |
|
371 | 371 | if path == self.module: # Follow branching back in history |
|
372 | 372 | if ent: |
|
373 | 373 | if ent.copyfrom_path: |
|
374 | 374 | # ent.copyfrom_rev may not be the actual last revision |
|
375 | 375 | prev = self.latest(ent.copyfrom_path, ent.copyfrom_rev) |
|
376 | 376 | self.modulemap[prev] = ent.copyfrom_path |
|
377 | 377 | parents = [self.revid(prev, ent.copyfrom_path)] |
|
378 | 378 | self.ui.note('found parent of branch %s at %d: %s\n' % \ |
|
379 | 379 | (self.module, prev, ent.copyfrom_path)) |
|
380 | 380 | else: |
|
381 | 381 | self.ui.debug("No copyfrom path, don't know what to do.\n") |
|
382 | 382 | # Maybe it was added and there is no more history. |
|
383 | 383 | entrypath = get_entry_from_path(path, module=self.module) |
|
384 | 384 | # self.ui.write("entrypath %s\n" % entrypath) |
|
385 | 385 | if entrypath is None: |
|
386 | 386 | # Outside our area of interest |
|
387 | 387 | self.ui.debug("boring@%s: %s\n" % (revnum, path)) |
|
388 | 388 | continue |
|
389 | 389 | entry = entrypath.decode(self.encoding) |
|
390 | 390 | |
|
391 | 391 | kind = svn.ra.check_path(self.ra, entrypath, revnum) |
|
392 | 392 | if kind == svn.core.svn_node_file: |
|
393 | 393 | if ent.copyfrom_path: |
|
394 | 394 | copyfrom_path = get_entry_from_path(ent.copyfrom_path) |
|
395 | 395 | if copyfrom_path: |
|
396 | 396 | self.ui.debug("Copied to %s from %s@%s\n" % (entry, copyfrom_path, ent.copyfrom_rev)) |
|
397 | 397 | # It's probably important for hg that the source |
|
398 | 398 | # exists in the revision's parent, not just the |
|
399 | 399 | # ent.copyfrom_rev |
|
400 | 400 | fromkind = svn.ra.check_path(self.ra, copyfrom_path, ent.copyfrom_rev) |
|
401 | 401 | if fromkind != 0: |
|
402 | 402 | copies[self.recode(entry)] = self.recode(copyfrom_path) |
|
403 | 403 | entries.append(self.recode(entry)) |
|
404 | 404 | elif kind == 0: # gone, but had better be a deleted *file* |
|
405 | 405 | self.ui.debug("gone from %s\n" % ent.copyfrom_rev) |
|
406 | 406 | |
|
407 | 407 | # if a branch is created but entries are removed in the same |
|
408 | 408 | # changeset, get the right fromrev |
|
409 | 409 | if parents: |
|
410 | 410 | uuid, old_module, fromrev = self.revsplit(parents[0]) |
|
411 | 411 | else: |
|
412 | 412 | fromrev = revnum - 1 |
|
413 | 413 | # might always need to be revnum - 1 in these 3 lines? |
|
414 | 414 | old_module = self.modulemap.get(fromrev, self.module) |
|
415 | 415 | |
|
416 | 416 | basepath = old_module + "/" + get_entry_from_path(path, module=self.module) |
|
417 | 417 | entrypath = old_module + "/" + get_entry_from_path(path, module=self.module) |
|
418 | 418 | |
|
419 | 419 | def lookup_parts(p): |
|
420 | 420 | rc = None |
|
421 | 421 | parts = p.split("/") |
|
422 | 422 | for i in range(len(parts)): |
|
423 | 423 | part = "/".join(parts[:i]) |
|
424 | 424 | info = part, copyfrom.get(part, None) |
|
425 | 425 | if info[1] is not None: |
|
426 | 426 | self.ui.debug("Found parent directory %s\n" % info[1]) |
|
427 | 427 | rc = info |
|
428 | 428 | return rc |
|
429 | 429 | |
|
430 | 430 | self.ui.debug("base, entry %s %s\n" % (basepath, entrypath)) |
|
431 | 431 | |
|
432 | 432 | frompath, froment = lookup_parts(entrypath) or (None, revnum - 1) |
|
433 | 433 | |
|
434 | 434 | # need to remove fragment from lookup_parts and replace with copyfrom_path |
|
435 | 435 | if frompath is not None: |
|
436 | 436 | self.ui.debug("munge-o-matic\n") |
|
437 | 437 | self.ui.debug(entrypath + '\n') |
|
438 | 438 | self.ui.debug(entrypath[len(frompath):] + '\n') |
|
439 | 439 | entrypath = froment.copyfrom_path + entrypath[len(frompath):] |
|
440 | 440 | fromrev = froment.copyfrom_rev |
|
441 | 441 | self.ui.debug("Info: %s %s %s %s\n" % (frompath, froment, ent, entrypath)) |
|
442 | 442 | |
|
443 | 443 | fromkind = svn.ra.check_path(self.ra, entrypath, fromrev) |
|
444 | 444 | if fromkind == svn.core.svn_node_file: # a deleted file |
|
445 | 445 | entries.append(self.recode(entry)) |
|
446 | 446 | elif fromkind == svn.core.svn_node_dir: |
|
447 | 447 | # print "Deleted/moved non-file:", revnum, path, ent |
|
448 | 448 | # children = self._find_children(path, revnum - 1) |
|
449 | 449 | # print "find children %s@%d from %d action %s" % (path, revnum, ent.copyfrom_rev, ent.action) |
|
450 | 450 | # Sometimes this is tricky. For example: in |
|
451 | 451 | # The Subversion Repository revision 6940 a dir |
|
452 |
# was copied and one of its files was deleted |
|
|
452 | # was copied and one of its files was deleted | |
|
453 | 453 | # from the new location in the same commit. This |
|
454 | 454 | # code can't deal with that yet. |
|
455 | 455 | if ent.action == 'C': |
|
456 | 456 | children = self._find_children(path, fromrev) |
|
457 | 457 | else: |
|
458 | 458 | oroot = entrypath.strip('/') |
|
459 | 459 | nroot = path.strip('/') |
|
460 | 460 | children = self._find_children(oroot, fromrev) |
|
461 | 461 | children = [s.replace(oroot,nroot) for s in children] |
|
462 | 462 | # Mark all [files, not directories] as deleted. |
|
463 | 463 | for child in children: |
|
464 | 464 | # Can we move a child directory and its |
|
465 | 465 | # parent in the same commit? (probably can). Could |
|
466 |
# cause problems if instead of revnum -1, |
|
|
466 | # cause problems if instead of revnum -1, | |
|
467 | 467 | # we have to look in (copyfrom_path, revnum - 1) |
|
468 | 468 | entrypath = get_entry_from_path("/" + child, module=old_module) |
|
469 | 469 | if entrypath: |
|
470 | 470 | entry = self.recode(entrypath.decode(self.encoding)) |
|
471 | 471 | if entry in copies: |
|
472 | 472 | # deleted file within a copy |
|
473 | 473 | del copies[entry] |
|
474 | 474 | else: |
|
475 | 475 | entries.append(entry) |
|
476 | 476 | else: |
|
477 | 477 | self.ui.debug('unknown path in revision %d: %s\n' % \ |
|
478 | 478 | (revnum, path)) |
|
479 | 479 | elif kind == svn.core.svn_node_dir: |
|
480 | 480 | # Should probably synthesize normal file entries |
|
481 | 481 | # and handle as above to clean up copy/rename handling. |
|
482 | 482 | |
|
483 | 483 | # If the directory just had a prop change, |
|
484 | 484 | # then we shouldn't need to look for its children. |
|
485 | 485 | # Also this could create duplicate entries. Not sure |
|
486 | 486 | # whether this will matter. Maybe should make entries a set. |
|
487 | 487 | # print "Changed directory", revnum, path, ent.action, ent.copyfrom_path, ent.copyfrom_rev |
|
488 | 488 | # This will fail if a directory was copied |
|
489 | 489 | # from another branch and then some of its files |
|
490 | 490 | # were deleted in the same transaction. |
|
491 | 491 | children = self._find_children(path, revnum) |
|
492 | 492 | children.sort() |
|
493 | 493 | for child in children: |
|
494 | 494 | # Can we move a child directory and its |
|
495 | 495 | # parent in the same commit? (probably can). Could |
|
496 |
# cause problems if instead of revnum -1, |
|
|
496 | # cause problems if instead of revnum -1, | |
|
497 | 497 | # we have to look in (copyfrom_path, revnum - 1) |
|
498 | 498 | entrypath = get_entry_from_path("/" + child, module=self.module) |
|
499 | 499 | # print child, self.module, entrypath |
|
500 | 500 | if entrypath: |
|
501 | 501 | # Need to filter out directories here... |
|
502 | 502 | kind = svn.ra.check_path(self.ra, entrypath, revnum) |
|
503 | 503 | if kind != svn.core.svn_node_dir: |
|
504 | 504 | entries.append(self.recode(entrypath)) |
|
505 | 505 | |
|
506 | 506 | # Copies here (must copy all from source) |
|
507 | 507 | # Probably not a real problem for us if |
|
508 | 508 | # source does not exist |
|
509 | 509 | |
|
510 | 510 | # Can do this with the copy command "hg copy" |
|
511 | 511 | # if ent.copyfrom_path: |
|
512 | 512 | # copyfrom_entry = get_entry_from_path(ent.copyfrom_path.decode(self.encoding), |
|
513 | 513 | # module=self.module) |
|
514 | 514 | # copyto_entry = entrypath |
|
515 | 515 | # |
|
516 | 516 | # print "copy directory", copyfrom_entry, 'to', copyto_entry |
|
517 | 517 | # |
|
518 | 518 | # copies.append((copyfrom_entry, copyto_entry)) |
|
519 | 519 | |
|
520 | 520 | if ent.copyfrom_path: |
|
521 | 521 | copyfrom_path = ent.copyfrom_path.decode(self.encoding) |
|
522 | 522 | copyfrom_entry = get_entry_from_path(copyfrom_path, module=self.module) |
|
523 | 523 | if copyfrom_entry: |
|
524 | 524 | copyfrom[path] = ent |
|
525 | 525 | self.ui.debug("mark %s came from %s\n" % (path, copyfrom[path])) |
|
526 | 526 | |
|
527 | 527 | # Good, /probably/ a regular copy. Really should check |
|
528 | 528 | # to see whether the parent revision actually contains |
|
529 | 529 | # the directory in question. |
|
530 | 530 | children = self._find_children(self.recode(copyfrom_path), ent.copyfrom_rev) |
|
531 | 531 | children.sort() |
|
532 | 532 | for child in children: |
|
533 | 533 | entrypath = get_entry_from_path("/" + child, module=self.module) |
|
534 | 534 | if entrypath: |
|
535 | 535 | entry = entrypath.decode(self.encoding) |
|
536 | 536 | # print "COPY COPY From", copyfrom_entry, entry |
|
537 | 537 | copyto_path = path + entry[len(copyfrom_entry):] |
|
538 | 538 | copyto_entry = get_entry_from_path(copyto_path, module=self.module) |
|
539 | 539 | # print "COPY", entry, "COPY To", copyto_entry |
|
540 | 540 | copies[self.recode(copyto_entry)] = self.recode(entry) |
|
541 | 541 | # copy from quux splort/quuxfile |
|
542 | 542 | |
|
543 | 543 | self.modulemap[revnum] = self.module # track backwards in time |
|
544 | 544 | # a list of (filename, id) where id lets us retrieve the file. |
|
545 |
# eg in git, id is the object hash. for svn it'll be the |
|
|
545 | # eg in git, id is the object hash. for svn it'll be the | |
|
546 | 546 | self.files[rev] = zip(entries, [rev] * len(entries)) |
|
547 | 547 | if not entries: |
|
548 | 548 | return |
|
549 | 549 | |
|
550 | 550 | # Example SVN datetime. Includes microseconds. |
|
551 | 551 | # ISO-8601 conformant |
|
552 | 552 | # '2007-01-04T17:35:00.902377Z' |
|
553 | 553 | date = util.parsedate(date[:18] + " UTC", ["%Y-%m-%dT%H:%M:%S"]) |
|
554 | 554 | |
|
555 | 555 | log = message and self.recode(message) |
|
556 | 556 | author = author and self.recode(author) or '' |
|
557 | 557 | |
|
558 | 558 | cset = commit(author=author, |
|
559 |
date=util.datestr(date), |
|
|
560 |
desc=log, |
|
|
559 | date=util.datestr(date), | |
|
560 | desc=log, | |
|
561 | 561 | parents=parents, |
|
562 | 562 | copies=copies, |
|
563 | 563 | branch=branch, |
|
564 | 564 | rev=rev.encode('utf-8')) |
|
565 | 565 | |
|
566 | 566 | self.commits[rev] = cset |
|
567 | 567 | if self.child_cset and not self.child_cset.parents: |
|
568 | 568 | self.child_cset.parents = [rev] |
|
569 | 569 | self.child_cset = cset |
|
570 | 570 | |
|
571 | 571 | self.ui.note('fetching revision log for "%s" from %d to %d\n' % |
|
572 | 572 | (self.module, from_revnum, to_revnum)) |
|
573 | 573 | |
|
574 | 574 | try: |
|
575 | 575 | discover_changed_paths = True |
|
576 | 576 | strict_node_history = False |
|
577 | 577 | for entry in self.get_log([self.module], from_revnum, to_revnum): |
|
578 | 578 | orig_paths, revnum, author, date, message = entry |
|
579 | 579 | if self.is_blacklisted(revnum): |
|
580 | 580 | self.ui.note('skipping blacklisted revision %d\n' % revnum) |
|
581 | 581 | continue |
|
582 | 582 | if orig_paths is None: |
|
583 | 583 | self.ui.debug('revision %d has no entries\n' % revnum) |
|
584 | 584 | continue |
|
585 | 585 | parselogentry(orig_paths, revnum, author, date, message) |
|
586 | 586 | except SubversionException, (_, num): |
|
587 | 587 | if num == svn.core.SVN_ERR_FS_NO_SUCH_REVISION: |
|
588 |
raise NoSuchRevision(branch=self, |
|
|
588 | raise NoSuchRevision(branch=self, | |
|
589 | 589 | revision="Revision number %d" % to_revnum) |
|
590 | 590 | raise |
|
591 | 591 | |
|
592 | 592 | def _getfile(self, file, rev): |
|
593 | 593 | io = StringIO() |
|
594 | 594 | # TODO: ra.get_file transmits the whole file instead of diffs. |
|
595 | 595 | mode = '' |
|
596 | 596 | try: |
|
597 | 597 | revnum = self.revnum(rev) |
|
598 | 598 | if self.module != self.modulemap[revnum]: |
|
599 | 599 | self.module = self.modulemap[revnum] |
|
600 | 600 | self.reparent(self.module) |
|
601 | 601 | info = svn.ra.get_file(self.ra, file, revnum, io) |
|
602 | 602 | if isinstance(info, list): |
|
603 | 603 | info = info[-1] |
|
604 | 604 | mode = ("svn:executable" in info) and 'x' or '' |
|
605 | 605 | mode = ("svn:special" in info) and 'l' or mode |
|
606 | 606 | except SubversionException, e: |
|
607 | 607 | notfound = (svn.core.SVN_ERR_FS_NOT_FOUND, |
|
608 | 608 | svn.core.SVN_ERR_RA_DAV_PATH_NOT_FOUND) |
|
609 | 609 | if e.apr_err in notfound: # File not found |
|
610 | 610 | raise IOError() |
|
611 | 611 | raise |
|
612 | 612 | data = io.getvalue() |
|
613 | 613 | if mode == 'l': |
|
614 | 614 | link_prefix = "link " |
|
615 | 615 | if data.startswith(link_prefix): |
|
616 | 616 | data = data[len(link_prefix):] |
|
617 | 617 | return data, mode |
|
618 | 618 | |
|
619 | 619 | def _find_children(self, path, revnum): |
|
620 | 620 | path = path.strip("/") |
|
621 | 621 | |
|
622 | 622 | def _find_children_fallback(path, revnum): |
|
623 | 623 | # SWIG python bindings for getdir are broken up to at least 1.4.3 |
|
624 | 624 | pool = Pool() |
|
625 | 625 | optrev = svn.core.svn_opt_revision_t() |
|
626 | 626 | optrev.kind = svn.core.svn_opt_revision_number |
|
627 | 627 | optrev.value.number = revnum |
|
628 | 628 | rpath = '/'.join([self.base, path]).strip('/') |
|
629 | 629 | return ['%s/%s' % (path, x) for x in svn.client.ls(rpath, optrev, True, self.ctx, pool).keys()] |
|
630 | 630 | |
|
631 | 631 | if hasattr(self, '_find_children_fallback'): |
|
632 | 632 | return _find_children_fallback(path, revnum) |
|
633 | 633 | |
|
634 | 634 | self.reparent("/" + path) |
|
635 | 635 | pool = Pool() |
|
636 | 636 | |
|
637 | 637 | children = [] |
|
638 | 638 | def find_children_inner(children, path, revnum = revnum): |
|
639 | 639 | if hasattr(svn.ra, 'get_dir2'): # Since SVN 1.4 |
|
640 | 640 | fields = 0xffffffff # Binding does not provide SVN_DIRENT_ALL |
|
641 | 641 | getdir = svn.ra.get_dir2(self.ra, path, revnum, fields, pool) |
|
642 | 642 | else: |
|
643 | 643 | getdir = svn.ra.get_dir(self.ra, path, revnum, pool) |
|
644 | 644 | if type(getdir) == dict: |
|
645 | 645 | # python binding for getdir is broken up to at least 1.4.3 |
|
646 | 646 | raise CompatibilityException() |
|
647 | 647 | dirents = getdir[0] |
|
648 | 648 | if type(dirents) == int: |
|
649 | 649 | # got here once due to infinite recursion bug |
|
650 | 650 | # pprint.pprint(getdir) |
|
651 | 651 | return |
|
652 | 652 | c = dirents.keys() |
|
653 | 653 | c.sort() |
|
654 | 654 | for child in c: |
|
655 | 655 | dirent = dirents[child] |
|
656 | 656 | if dirent.kind == svn.core.svn_node_dir: |
|
657 | 657 | find_children_inner(children, (path + "/" + child).strip("/")) |
|
658 | 658 | else: |
|
659 | 659 | children.append((path + "/" + child).strip("/")) |
|
660 | 660 | |
|
661 | 661 | try: |
|
662 | 662 | find_children_inner(children, "") |
|
663 | 663 | except CompatibilityException: |
|
664 | 664 | self._find_children_fallback = True |
|
665 | 665 | self.reparent(self.module) |
|
666 | 666 | return _find_children_fallback(path, revnum) |
|
667 | 667 | |
|
668 | 668 | self.reparent(self.module) |
|
669 | 669 | return [path + "/" + c for c in children] |
@@ -1,125 +1,125 b'' | |||
|
1 | 1 | # -*- coding: utf-8 -*- |
|
2 | 2 | |
|
3 | 3 | # Copyright (C) 2007 Daniel Holth <dholth@fastmail.fm> |
|
4 | 4 | # This is a stripped-down version of the original bzr-svn transport.py, |
|
5 | 5 | # Copyright (C) 2006 Jelmer Vernooij <jelmer@samba.org> |
|
6 | 6 | |
|
7 | 7 | # This program is free software; you can redistribute it and/or modify |
|
8 | 8 | # it under the terms of the GNU General Public License as published by |
|
9 | 9 | # the Free Software Foundation; either version 2 of the License, or |
|
10 | 10 | # (at your option) any later version. |
|
11 | 11 | |
|
12 | 12 | # This program is distributed in the hope that it will be useful, |
|
13 | 13 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
14 | 14 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
15 | 15 | # GNU General Public License for more details. |
|
16 | 16 | |
|
17 | 17 | # You should have received a copy of the GNU General Public License |
|
18 | 18 | # along with this program; if not, write to the Free Software |
|
19 | 19 | # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
|
20 | 20 | |
|
21 | 21 | from cStringIO import StringIO |
|
22 | 22 | import os |
|
23 | 23 | from tempfile import mktemp |
|
24 | 24 | |
|
25 | 25 | from svn.core import SubversionException, Pool |
|
26 | 26 | import svn.ra |
|
27 | 27 | import svn.client |
|
28 | 28 | import svn.core |
|
29 | 29 | |
|
30 |
# Some older versions of the Python bindings need to be |
|
|
30 | # Some older versions of the Python bindings need to be | |
|
31 | 31 | # explicitly initialized. But what we want to do probably |
|
32 | 32 | # won't work worth a darn against those libraries anyway! |
|
33 | 33 | svn.ra.initialize() |
|
34 | 34 | |
|
35 | 35 | svn_config = svn.core.svn_config_get_config(None) |
|
36 | 36 | |
|
37 | 37 | |
|
38 | 38 | def _create_auth_baton(pool): |
|
39 | 39 | """Create a Subversion authentication baton. """ |
|
40 | 40 | import svn.client |
|
41 | 41 | # Give the client context baton a suite of authentication |
|
42 | 42 | # providers.h |
|
43 | 43 | providers = [ |
|
44 | 44 | svn.client.get_simple_provider(pool), |
|
45 | 45 | svn.client.get_username_provider(pool), |
|
46 | 46 | svn.client.get_ssl_client_cert_file_provider(pool), |
|
47 | 47 | svn.client.get_ssl_client_cert_pw_file_provider(pool), |
|
48 | 48 | svn.client.get_ssl_server_trust_file_provider(pool), |
|
49 | 49 | ] |
|
50 | 50 | return svn.core.svn_auth_open(providers, pool) |
|
51 | 51 | |
|
52 | 52 | class NotBranchError(SubversionException): |
|
53 | 53 | pass |
|
54 | 54 | |
|
55 | 55 | class SvnRaTransport(object): |
|
56 | 56 | """ |
|
57 | 57 | Open an ra connection to a Subversion repository. |
|
58 | 58 | """ |
|
59 | 59 | def __init__(self, url="", ra=None): |
|
60 | 60 | self.pool = Pool() |
|
61 | 61 | self.svn_url = url |
|
62 | 62 | self.username = '' |
|
63 | 63 | self.password = '' |
|
64 | 64 | |
|
65 | 65 | # Only Subversion 1.4 has reparent() |
|
66 | 66 | if ra is None or not hasattr(svn.ra, 'reparent'): |
|
67 | 67 | self.client = svn.client.create_context(self.pool) |
|
68 | 68 | ab = _create_auth_baton(self.pool) |
|
69 | 69 | if False: |
|
70 | 70 | svn.core.svn_auth_set_parameter( |
|
71 | 71 | ab, svn.core.SVN_AUTH_PARAM_DEFAULT_USERNAME, self.username) |
|
72 | 72 | svn.core.svn_auth_set_parameter( |
|
73 | 73 | ab, svn.core.SVN_AUTH_PARAM_DEFAULT_PASSWORD, self.password) |
|
74 | 74 | self.client.auth_baton = ab |
|
75 | 75 | self.client.config = svn_config |
|
76 | 76 | try: |
|
77 | 77 | self.ra = svn.client.open_ra_session( |
|
78 |
self.svn_url.encode('utf8'), |
|
|
78 | self.svn_url.encode('utf8'), | |
|
79 | 79 | self.client, self.pool) |
|
80 | 80 | except SubversionException, (_, num): |
|
81 | 81 | if num in (svn.core.SVN_ERR_RA_ILLEGAL_URL, |
|
82 | 82 | svn.core.SVN_ERR_RA_LOCAL_REPOS_OPEN_FAILED, |
|
83 | 83 | svn.core.SVN_ERR_BAD_URL): |
|
84 | 84 | raise NotBranchError(url) |
|
85 | 85 | raise |
|
86 | 86 | else: |
|
87 | 87 | self.ra = ra |
|
88 | 88 | svn.ra.reparent(self.ra, self.svn_url.encode('utf8')) |
|
89 | 89 | |
|
90 | 90 | class Reporter: |
|
91 | 91 | def __init__(self, (reporter, report_baton)): |
|
92 | 92 | self._reporter = reporter |
|
93 | 93 | self._baton = report_baton |
|
94 | 94 | |
|
95 | 95 | def set_path(self, path, revnum, start_empty, lock_token, pool=None): |
|
96 | 96 | svn.ra.reporter2_invoke_set_path(self._reporter, self._baton, |
|
97 | 97 | path, revnum, start_empty, lock_token, pool) |
|
98 | 98 | |
|
99 | 99 | def delete_path(self, path, pool=None): |
|
100 | 100 | svn.ra.reporter2_invoke_delete_path(self._reporter, self._baton, |
|
101 | 101 | path, pool) |
|
102 | 102 | |
|
103 | 103 | def link_path(self, path, url, revision, start_empty, lock_token, |
|
104 | 104 | pool=None): |
|
105 | 105 | svn.ra.reporter2_invoke_link_path(self._reporter, self._baton, |
|
106 | 106 | path, url, revision, start_empty, lock_token, |
|
107 | 107 | pool) |
|
108 | 108 | |
|
109 | 109 | def finish_report(self, pool=None): |
|
110 | 110 | svn.ra.reporter2_invoke_finish_report(self._reporter, |
|
111 | 111 | self._baton, pool) |
|
112 | 112 | |
|
113 | 113 | def abort_report(self, pool=None): |
|
114 | 114 | svn.ra.reporter2_invoke_abort_report(self._reporter, |
|
115 | 115 | self._baton, pool) |
|
116 | 116 | |
|
117 | 117 | def do_update(self, revnum, path, *args, **kwargs): |
|
118 | 118 | return self.Reporter(svn.ra.do_update(self.ra, revnum, path, *args, **kwargs)) |
|
119 | 119 | |
|
120 | 120 | def clone(self, offset=None): |
|
121 | 121 | """See Transport.clone().""" |
|
122 | 122 | if offset is None: |
|
123 | 123 | return self.__class__(self.base) |
|
124 | 124 | |
|
125 | 125 | return SvnRaTransport(urlutils.join(self.base, offset), ra=self.ra) |
@@ -1,168 +1,168 b'' | |||
|
1 | 1 | # Copyright (C) 2006 - Marco Barisione <marco@barisione.org> |
|
2 | 2 | # |
|
3 | 3 | # This is a small extension for Mercurial (http://www.selenic.com/mercurial) |
|
4 | 4 | # that removes files not known to mercurial |
|
5 | 5 | # |
|
6 | 6 | # This program was inspired by the "cvspurge" script contained in CVS utilities |
|
7 | 7 | # (http://www.red-bean.com/cvsutils/). |
|
8 | 8 | # |
|
9 | 9 | # To enable the "purge" extension put these lines in your ~/.hgrc: |
|
10 | 10 | # [extensions] |
|
11 | 11 | # hgext.purge = |
|
12 | 12 | # |
|
13 | 13 | # For help on the usage of "hg purge" use: |
|
14 | 14 | # hg help purge |
|
15 | 15 | # |
|
16 | 16 | # This program is free software; you can redistribute it and/or modify |
|
17 | 17 | # it under the terms of the GNU General Public License as published by |
|
18 | 18 | # the Free Software Foundation; either version 2 of the License, or |
|
19 | 19 | # (at your option) any later version. |
|
20 | 20 | # |
|
21 | 21 | # This program is distributed in the hope that it will be useful, |
|
22 | 22 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
23 | 23 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
24 | 24 | # GNU General Public License for more details. |
|
25 | 25 | # |
|
26 | 26 | # You should have received a copy of the GNU General Public License |
|
27 | 27 | # along with this program; if not, write to the Free Software |
|
28 | 28 | # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. |
|
29 | 29 | |
|
30 | 30 | from mercurial import hg, util |
|
31 | 31 | from mercurial.i18n import _ |
|
32 | 32 | import os |
|
33 | 33 | |
|
34 |
def dopurge(ui, repo, dirs=None, act=True, ignored=False, |
|
|
34 | def dopurge(ui, repo, dirs=None, act=True, ignored=False, | |
|
35 | 35 | abort_on_err=False, eol='\n', |
|
36 | 36 | force=False, include=None, exclude=None): |
|
37 | 37 | def error(msg): |
|
38 | 38 | if abort_on_err: |
|
39 | 39 | raise util.Abort(msg) |
|
40 | 40 | else: |
|
41 | 41 | ui.warn(_('warning: %s\n') % msg) |
|
42 | 42 | |
|
43 | 43 | def remove(remove_func, name): |
|
44 | 44 | if act: |
|
45 | 45 | try: |
|
46 | 46 | remove_func(os.path.join(repo.root, name)) |
|
47 | 47 | except OSError, e: |
|
48 | 48 | error(_('%s cannot be removed') % name) |
|
49 | 49 | else: |
|
50 | 50 | ui.write('%s%s' % (name, eol)) |
|
51 | 51 | |
|
52 | 52 | directories = [] |
|
53 | 53 | files = [] |
|
54 | 54 | missing = [] |
|
55 | 55 | roots, match, anypats = util.cmdmatcher(repo.root, repo.getcwd(), dirs, |
|
56 | 56 | include, exclude) |
|
57 | 57 | for src, f, st in repo.dirstate.statwalk(files=roots, match=match, |
|
58 | 58 | ignored=ignored, directories=True): |
|
59 | 59 | if src == 'd': |
|
60 | 60 | directories.append(f) |
|
61 | 61 | elif src == 'm': |
|
62 | 62 | missing.append(f) |
|
63 | 63 | elif src == 'f' and f not in repo.dirstate: |
|
64 | 64 | files.append(f) |
|
65 | 65 | |
|
66 | 66 | _check_missing(ui, repo, missing, force) |
|
67 | 67 | |
|
68 | 68 | directories.sort() |
|
69 | 69 | |
|
70 | 70 | for f in files: |
|
71 | 71 | if f not in repo.dirstate: |
|
72 | 72 | ui.note(_('Removing file %s\n') % f) |
|
73 | 73 | remove(os.remove, f) |
|
74 | 74 | |
|
75 | 75 | for f in directories[::-1]: |
|
76 | 76 | if match(f) and not os.listdir(repo.wjoin(f)): |
|
77 | 77 | ui.note(_('Removing directory %s\n') % f) |
|
78 | 78 | remove(os.rmdir, f) |
|
79 | 79 | |
|
80 | 80 | def _check_missing(ui, repo, missing, force=False): |
|
81 | 81 | """Abort if there is the chance of having problems with name-mangling fs |
|
82 | 82 | |
|
83 | 83 | In a name mangling filesystem (e.g. a case insensitive one) |
|
84 | 84 | dirstate.walk() can yield filenames different from the ones |
|
85 | 85 | stored in the dirstate. This already confuses the status and |
|
86 | 86 | add commands, but with purge this may cause data loss. |
|
87 | 87 | |
|
88 | 88 | To prevent this, _check_missing will abort if there are missing |
|
89 | 89 | files. The force option will let the user skip the check if he |
|
90 | 90 | knows it is safe. |
|
91 | 91 | |
|
92 | 92 | Even with the force option this function will check if any of the |
|
93 | 93 | missing files is still available in the working dir: if so there |
|
94 | 94 | may be some problem with the underlying filesystem, so it |
|
95 | 95 | aborts unconditionally.""" |
|
96 | 96 | |
|
97 | 97 | found = [f for f in missing if util.lexists(repo.wjoin(f))] |
|
98 | 98 | |
|
99 | 99 | if found: |
|
100 | 100 | if not ui.quiet: |
|
101 | 101 | ui.warn(_("The following tracked files weren't listed by the " |
|
102 | 102 | "filesystem, but could still be found:\n")) |
|
103 | 103 | for f in found: |
|
104 | 104 | ui.warn("%s\n" % f) |
|
105 | 105 | if util.checkfolding(repo.path): |
|
106 | 106 | ui.warn(_("This is probably due to a case-insensitive " |
|
107 | 107 | "filesystem\n")) |
|
108 | 108 | raise util.Abort(_("purging on name mangling filesystems is not " |
|
109 | 109 | "yet fully supported")) |
|
110 | 110 | |
|
111 | 111 | if missing and not force: |
|
112 | 112 | raise util.Abort(_("there are missing files in the working dir and " |
|
113 | 113 | "purge still has problems with them due to name " |
|
114 | 114 | "mangling filesystems. " |
|
115 | 115 | "Use --force if you know what you are doing")) |
|
116 | 116 | |
|
117 | 117 | |
|
118 | 118 | def purge(ui, repo, *dirs, **opts): |
|
119 | 119 | '''removes files not tracked by mercurial |
|
120 | 120 | |
|
121 | 121 | Delete files not known to mercurial, this is useful to test local and |
|
122 | 122 | uncommitted changes in the otherwise clean source tree. |
|
123 | 123 | |
|
124 | 124 | This means that purge will delete: |
|
125 | 125 | - Unknown files: files marked with "?" by "hg status" |
|
126 | 126 | - Ignored files: files usually ignored by Mercurial because they match |
|
127 | 127 | a pattern in a ".hgignore" file |
|
128 | 128 | - Empty directories: in fact Mercurial ignores directories unless they |
|
129 | 129 | contain files under source control managment |
|
130 | 130 | But it will leave untouched: |
|
131 | 131 | - Unmodified tracked files |
|
132 | 132 | - Modified tracked files |
|
133 | 133 | - New files added to the repository (with "hg add") |
|
134 | 134 | |
|
135 | 135 | If directories are given on the command line, only files in these |
|
136 | 136 | directories are considered. |
|
137 | 137 | |
|
138 | 138 | Be careful with purge, you could irreversibly delete some files you |
|
139 | 139 | forgot to add to the repository. If you only want to print the list of |
|
140 | 140 | files that this program would delete use the --print option. |
|
141 | 141 | ''' |
|
142 | 142 | act = not opts['print'] |
|
143 | 143 | ignored = bool(opts['all']) |
|
144 | 144 | abort_on_err = bool(opts['abort_on_err']) |
|
145 | 145 | eol = opts['print0'] and '\0' or '\n' |
|
146 | 146 | if eol == '\0': |
|
147 | 147 | # --print0 implies --print |
|
148 | 148 | act = False |
|
149 | 149 | force = bool(opts['force']) |
|
150 | 150 | include = opts['include'] |
|
151 | 151 | exclude = opts['exclude'] |
|
152 | 152 | dopurge(ui, repo, dirs, act, ignored, abort_on_err, |
|
153 | 153 | eol, force, include, exclude) |
|
154 | 154 | |
|
155 | 155 | |
|
156 | 156 | cmdtable = { |
|
157 | 157 | 'purge|clean': |
|
158 | 158 | (purge, |
|
159 | 159 | [('a', 'abort-on-err', None, _('abort if an error occurs')), |
|
160 | 160 | ('', 'all', None, _('purge ignored files too')), |
|
161 | 161 | ('f', 'force', None, _('purge even when missing files are detected')), |
|
162 | 162 | ('p', 'print', None, _('print the file names instead of deleting them')), |
|
163 | 163 | ('0', 'print0', None, _('end filenames with NUL, for use with xargs' |
|
164 | 164 | ' (implies -p)')), |
|
165 | 165 | ('I', 'include', [], _('include names matching the given patterns')), |
|
166 | 166 | ('X', 'exclude', [], _('exclude names matching the given patterns'))], |
|
167 | 167 | _('hg purge [OPTION]... [DIR]...')) |
|
168 | 168 | } |
@@ -1,258 +1,258 b'' | |||
|
1 | 1 | # hgweb/hgwebdir_mod.py - Web interface for a directory of repositories. |
|
2 | 2 | # |
|
3 | 3 | # Copyright 21 May 2005 - (c) 2005 Jake Edge <jake@edge2.net> |
|
4 | 4 | # Copyright 2005, 2006 Matt Mackall <mpm@selenic.com> |
|
5 | 5 | # |
|
6 | 6 | # This software may be used and distributed according to the terms |
|
7 | 7 | # of the GNU General Public License, incorporated herein by reference. |
|
8 | 8 | |
|
9 | 9 | from mercurial import demandimport; demandimport.enable() |
|
10 | 10 | import os, mimetools, cStringIO |
|
11 | 11 | from mercurial.i18n import gettext as _ |
|
12 | 12 | from mercurial import ui, hg, util, templater |
|
13 | 13 | from common import get_mtime, staticfile, style_map, paritygen |
|
14 | 14 | from hgweb_mod import hgweb |
|
15 | 15 | |
|
16 | 16 | # This is a stopgap |
|
17 | 17 | class hgwebdir(object): |
|
18 | 18 | def __init__(self, config, parentui=None): |
|
19 | 19 | def cleannames(items): |
|
20 | 20 | return [(name.strip(os.sep), path) for name, path in items] |
|
21 | 21 | |
|
22 | 22 | self.parentui = parentui |
|
23 | 23 | self.motd = None |
|
24 | 24 | self.style = None |
|
25 | 25 | self.stripecount = None |
|
26 | 26 | self.repos_sorted = ('name', False) |
|
27 | 27 | if isinstance(config, (list, tuple)): |
|
28 | 28 | self.repos = cleannames(config) |
|
29 | 29 | self.repos_sorted = ('', False) |
|
30 | 30 | elif isinstance(config, dict): |
|
31 | 31 | self.repos = cleannames(config.items()) |
|
32 | 32 | self.repos.sort() |
|
33 | 33 | else: |
|
34 | 34 | if isinstance(config, util.configparser): |
|
35 | 35 | cp = config |
|
36 | 36 | else: |
|
37 | 37 | cp = util.configparser() |
|
38 | 38 | cp.read(config) |
|
39 | 39 | self.repos = [] |
|
40 | 40 | if cp.has_section('web'): |
|
41 | 41 | if cp.has_option('web', 'motd'): |
|
42 | 42 | self.motd = cp.get('web', 'motd') |
|
43 | 43 | if cp.has_option('web', 'style'): |
|
44 | 44 | self.style = cp.get('web', 'style') |
|
45 | 45 | if cp.has_option('web', 'stripes'): |
|
46 | 46 | self.stripecount = int(cp.get('web', 'stripes')) |
|
47 | 47 | if cp.has_section('paths'): |
|
48 | 48 | self.repos.extend(cleannames(cp.items('paths'))) |
|
49 | 49 | if cp.has_section('collections'): |
|
50 | 50 | for prefix, root in cp.items('collections'): |
|
51 | 51 | for path in util.walkrepos(root): |
|
52 | 52 | repo = os.path.normpath(path) |
|
53 | 53 | name = repo |
|
54 | 54 | if name.startswith(prefix): |
|
55 | 55 | name = name[len(prefix):] |
|
56 | 56 | self.repos.append((name.lstrip(os.sep), repo)) |
|
57 | 57 | self.repos.sort() |
|
58 | 58 | |
|
59 | 59 | def run(self): |
|
60 | 60 | if not os.environ.get('GATEWAY_INTERFACE', '').startswith("CGI/1."): |
|
61 | 61 | raise RuntimeError("This function is only intended to be called while running as a CGI script.") |
|
62 | 62 | import mercurial.hgweb.wsgicgi as wsgicgi |
|
63 | 63 | from request import wsgiapplication |
|
64 | 64 | def make_web_app(): |
|
65 | 65 | return self |
|
66 | 66 | wsgicgi.launch(wsgiapplication(make_web_app)) |
|
67 | 67 | |
|
68 | 68 | def run_wsgi(self, req): |
|
69 | 69 | def header(**map): |
|
70 | 70 | header_file = cStringIO.StringIO( |
|
71 | 71 | ''.join(tmpl("header", encoding=util._encoding, **map))) |
|
72 | 72 | msg = mimetools.Message(header_file, 0) |
|
73 | 73 | req.header(msg.items()) |
|
74 | 74 | yield header_file.read() |
|
75 | 75 | |
|
76 | 76 | def footer(**map): |
|
77 | 77 | yield tmpl("footer", **map) |
|
78 | 78 | |
|
79 | 79 | def motd(**map): |
|
80 | 80 | if self.motd is not None: |
|
81 | 81 | yield self.motd |
|
82 | 82 | else: |
|
83 | 83 | yield config('web', 'motd', '') |
|
84 | 84 | |
|
85 | 85 | parentui = self.parentui or ui.ui(report_untrusted=False) |
|
86 | 86 | |
|
87 | 87 | def config(section, name, default=None, untrusted=True): |
|
88 | 88 | return parentui.config(section, name, default, untrusted) |
|
89 | 89 | |
|
90 | 90 | url = req.env['REQUEST_URI'].split('?')[0] |
|
91 | 91 | if not url.endswith('/'): |
|
92 | 92 | url += '/' |
|
93 | 93 | pathinfo = req.env.get('PATH_INFO', '').strip('/') + '/' |
|
94 | 94 | base = url[:len(url) - len(pathinfo)] |
|
95 | 95 | if not base.endswith('/'): |
|
96 | 96 | base += '/' |
|
97 | 97 | |
|
98 | 98 | staticurl = config('web', 'staticurl') or base + 'static/' |
|
99 | 99 | if not staticurl.endswith('/'): |
|
100 | 100 | staticurl += '/' |
|
101 | 101 | |
|
102 | 102 | style = self.style |
|
103 | 103 | if style is None: |
|
104 | 104 | style = config('web', 'style', '') |
|
105 | 105 | if req.form.has_key('style'): |
|
106 | 106 | style = req.form['style'][0] |
|
107 | 107 | if self.stripecount is None: |
|
108 | 108 | self.stripecount = int(config('web', 'stripes', 1)) |
|
109 | 109 | mapfile = style_map(templater.templatepath(), style) |
|
110 | 110 | tmpl = templater.templater(mapfile, templater.common_filters, |
|
111 | 111 | defaults={"header": header, |
|
112 | 112 | "footer": footer, |
|
113 | 113 | "motd": motd, |
|
114 | 114 | "url": url, |
|
115 | 115 | "staticurl": staticurl}) |
|
116 | 116 | |
|
117 | 117 | def archivelist(ui, nodeid, url): |
|
118 | 118 | allowed = ui.configlist("web", "allow_archive", untrusted=True) |
|
119 | 119 | for i in [('zip', '.zip'), ('gz', '.tar.gz'), ('bz2', '.tar.bz2')]: |
|
120 | 120 | if i[0] in allowed or ui.configbool("web", "allow" + i[0], |
|
121 | 121 | untrusted=True): |
|
122 | 122 | yield {"type" : i[0], "extension": i[1], |
|
123 | 123 | "node": nodeid, "url": url} |
|
124 | 124 | |
|
125 | 125 | def entries(sortcolumn="", descending=False, subdir="", **map): |
|
126 | 126 | def sessionvars(**map): |
|
127 | 127 | fields = [] |
|
128 | 128 | if req.form.has_key('style'): |
|
129 | 129 | style = req.form['style'][0] |
|
130 | 130 | if style != get('web', 'style', ''): |
|
131 | 131 | fields.append(('style', style)) |
|
132 | 132 | |
|
133 | 133 | separator = url[-1] == '?' and ';' or '?' |
|
134 | 134 | for name, value in fields: |
|
135 | 135 | yield dict(name=name, value=value, separator=separator) |
|
136 | 136 | separator = ';' |
|
137 | 137 | |
|
138 | 138 | rows = [] |
|
139 | 139 | parity = paritygen(self.stripecount) |
|
140 | 140 | for name, path in self.repos: |
|
141 | 141 | if not name.startswith(subdir): |
|
142 | 142 | continue |
|
143 | 143 | name = name[len(subdir):] |
|
144 | 144 | |
|
145 | 145 | u = ui.ui(parentui=parentui) |
|
146 | 146 | try: |
|
147 | 147 | u.readconfig(os.path.join(path, '.hg', 'hgrc')) |
|
148 | 148 | except IOError: |
|
149 | 149 | pass |
|
150 | 150 | def get(section, name, default=None): |
|
151 | 151 | return u.config(section, name, default, untrusted=True) |
|
152 | 152 | |
|
153 | 153 | if u.configbool("web", "hidden", untrusted=True): |
|
154 | 154 | continue |
|
155 | 155 | |
|
156 | 156 | url = ('/'.join([req.env["REQUEST_URI"].split('?')[0], name]) |
|
157 | 157 | .replace("//", "/")) + '/' |
|
158 | 158 | |
|
159 | 159 | # update time with local timezone |
|
160 | 160 | try: |
|
161 | 161 | d = (get_mtime(path), util.makedate()[1]) |
|
162 | 162 | except OSError: |
|
163 | 163 | continue |
|
164 | 164 | |
|
165 | 165 | contact = (get("ui", "username") or # preferred |
|
166 | 166 | get("web", "contact") or # deprecated |
|
167 | 167 | get("web", "author", "")) # also |
|
168 | 168 | description = get("web", "description", "") |
|
169 | 169 | name = get("web", "name", name) |
|
170 | 170 | row = dict(contact=contact or "unknown", |
|
171 | 171 | contact_sort=contact.upper() or "unknown", |
|
172 | 172 | name=name, |
|
173 | 173 | name_sort=name, |
|
174 | 174 | url=url, |
|
175 | 175 | description=description or "unknown", |
|
176 | 176 | description_sort=description.upper() or "unknown", |
|
177 | 177 | lastchange=d, |
|
178 | 178 | lastchange_sort=d[1]-d[0], |
|
179 | 179 | sessionvars=sessionvars, |
|
180 | 180 | archives=archivelist(u, "tip", url)) |
|
181 | 181 | if (not sortcolumn |
|
182 | 182 | or (sortcolumn, descending) == self.repos_sorted): |
|
183 | 183 | # fast path for unsorted output |
|
184 | 184 | row['parity'] = parity.next() |
|
185 | 185 | yield row |
|
186 | 186 | else: |
|
187 | 187 | rows.append((row["%s_sort" % sortcolumn], row)) |
|
188 | 188 | if rows: |
|
189 | 189 | rows.sort() |
|
190 | 190 | if descending: |
|
191 | 191 | rows.reverse() |
|
192 | 192 | for key, row in rows: |
|
193 | 193 | row['parity'] = parity.next() |
|
194 | 194 | yield row |
|
195 | 195 | |
|
196 | 196 | def makeindex(req, subdir=""): |
|
197 | 197 | sortable = ["name", "description", "contact", "lastchange"] |
|
198 | 198 | sortcolumn, descending = self.repos_sorted |
|
199 | 199 | if req.form.has_key('sort'): |
|
200 | 200 | sortcolumn = req.form['sort'][0] |
|
201 | 201 | descending = sortcolumn.startswith('-') |
|
202 | 202 | if descending: |
|
203 | 203 | sortcolumn = sortcolumn[1:] |
|
204 | 204 | if sortcolumn not in sortable: |
|
205 | 205 | sortcolumn = "" |
|
206 | 206 | |
|
207 | 207 | sort = [("sort_%s" % column, |
|
208 | 208 | "%s%s" % ((not descending and column == sortcolumn) |
|
209 | 209 | and "-" or "", column)) |
|
210 | 210 | for column in sortable] |
|
211 | 211 | req.write(tmpl("index", entries=entries, subdir=subdir, |
|
212 | 212 | sortcolumn=sortcolumn, descending=descending, |
|
213 | 213 | **dict(sort))) |
|
214 | 214 | |
|
215 | 215 | try: |
|
216 | 216 | virtual = req.env.get("PATH_INFO", "").strip('/') |
|
217 | 217 | if virtual.startswith('static/'): |
|
218 | 218 | static = os.path.join(templater.templatepath(), 'static') |
|
219 | 219 | fname = virtual[7:] |
|
220 | 220 | req.write(staticfile(static, fname, req) or |
|
221 | 221 | tmpl('error', error='%r not found' % fname)) |
|
222 | 222 | elif virtual: |
|
223 | 223 | repos = dict(self.repos) |
|
224 | 224 | while virtual: |
|
225 | 225 | real = repos.get(virtual) |
|
226 | 226 | if real: |
|
227 | 227 | req.env['REPO_NAME'] = virtual |
|
228 | 228 | try: |
|
229 | 229 | repo = hg.repository(parentui, real) |
|
230 | 230 | hgweb(repo).run_wsgi(req) |
|
231 | 231 | except IOError, inst: |
|
232 | 232 | req.write(tmpl("error", error=inst.strerror)) |
|
233 | 233 | except hg.RepoError, inst: |
|
234 | 234 | req.write(tmpl("error", error=str(inst))) |
|
235 | 235 | return |
|
236 | 236 | |
|
237 | 237 | # browse subdirectories |
|
238 | 238 | subdir = virtual + '/' |
|
239 | 239 | if [r for r in repos if r.startswith(subdir)]: |
|
240 | 240 | makeindex(req, subdir) |
|
241 | 241 | return |
|
242 | 242 | |
|
243 | 243 | up = virtual.rfind('/') |
|
244 | 244 | if up < 0: |
|
245 | 245 | break |
|
246 | 246 | virtual = virtual[:up] |
|
247 | ||
|
247 | ||
|
248 | 248 | req.write(tmpl("notfound", repo=virtual)) |
|
249 | 249 | else: |
|
250 | 250 | if req.form.has_key('static'): |
|
251 | 251 | static = os.path.join(templater.templatepath(), "static") |
|
252 | 252 | fname = req.form['static'][0] |
|
253 | 253 | req.write(staticfile(static, fname, req) |
|
254 | 254 | or tmpl("error", error="%r not found" % fname)) |
|
255 | 255 | else: |
|
256 | 256 | makeindex(req) |
|
257 | 257 | finally: |
|
258 | 258 | tmpl = None |
@@ -1,289 +1,289 b'' | |||
|
1 | 1 | # hgweb/server.py - The standalone hg web server. |
|
2 | 2 | # |
|
3 | 3 | # Copyright 21 May 2005 - (c) 2005 Jake Edge <jake@edge2.net> |
|
4 | 4 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
5 | 5 | # |
|
6 | 6 | # This software may be used and distributed according to the terms |
|
7 | 7 | # of the GNU General Public License, incorporated herein by reference. |
|
8 | 8 | |
|
9 | 9 | import os, sys, errno, urllib, BaseHTTPServer, socket, SocketServer, traceback |
|
10 | 10 | from mercurial import ui, hg, util, templater |
|
11 | 11 | from hgweb_mod import hgweb |
|
12 | 12 | from hgwebdir_mod import hgwebdir |
|
13 | 13 | from request import wsgiapplication |
|
14 | 14 | from mercurial.i18n import gettext as _ |
|
15 | 15 | |
|
16 | 16 | def _splitURI(uri): |
|
17 | 17 | """ Return path and query splited from uri |
|
18 | 18 | |
|
19 | 19 | Just like CGI environment, the path is unquoted, the query is |
|
20 | 20 | not. |
|
21 | 21 | """ |
|
22 | 22 | if '?' in uri: |
|
23 | 23 | path, query = uri.split('?', 1) |
|
24 | 24 | else: |
|
25 | 25 | path, query = uri, '' |
|
26 | 26 | return urllib.unquote(path), query |
|
27 | 27 | |
|
28 | 28 | class _error_logger(object): |
|
29 | 29 | def __init__(self, handler): |
|
30 | 30 | self.handler = handler |
|
31 | 31 | def flush(self): |
|
32 | 32 | pass |
|
33 | 33 | def write(self, str): |
|
34 | 34 | self.writelines(str.split('\n')) |
|
35 | 35 | def writelines(self, seq): |
|
36 | 36 | for msg in seq: |
|
37 | 37 | self.handler.log_error("HG error: %s", msg) |
|
38 | 38 | |
|
39 | 39 | class _hgwebhandler(object, BaseHTTPServer.BaseHTTPRequestHandler): |
|
40 | 40 | |
|
41 | 41 | url_scheme = 'http' |
|
42 | ||
|
42 | ||
|
43 | 43 | def __init__(self, *args, **kargs): |
|
44 | 44 | self.protocol_version = 'HTTP/1.1' |
|
45 | 45 | BaseHTTPServer.BaseHTTPRequestHandler.__init__(self, *args, **kargs) |
|
46 | 46 | |
|
47 | 47 | def log_error(self, format, *args): |
|
48 | 48 | errorlog = self.server.errorlog |
|
49 | 49 | errorlog.write("%s - - [%s] %s\n" % (self.client_address[0], |
|
50 | 50 | self.log_date_time_string(), |
|
51 | 51 | format % args)) |
|
52 | 52 | |
|
53 | 53 | def log_message(self, format, *args): |
|
54 | 54 | accesslog = self.server.accesslog |
|
55 | 55 | accesslog.write("%s - - [%s] %s\n" % (self.client_address[0], |
|
56 | 56 | self.log_date_time_string(), |
|
57 | 57 | format % args)) |
|
58 | 58 | |
|
59 | 59 | def do_write(self): |
|
60 | 60 | try: |
|
61 | 61 | self.do_hgweb() |
|
62 | 62 | except socket.error, inst: |
|
63 | 63 | if inst[0] != errno.EPIPE: |
|
64 | 64 | raise |
|
65 | 65 | |
|
66 | 66 | def do_POST(self): |
|
67 | 67 | try: |
|
68 | 68 | self.do_write() |
|
69 | 69 | except StandardError, inst: |
|
70 | 70 | self._start_response("500 Internal Server Error", []) |
|
71 | 71 | self._write("Internal Server Error") |
|
72 | 72 | tb = "".join(traceback.format_exception(*sys.exc_info())) |
|
73 | 73 | self.log_error("Exception happened during processing request '%s':\n%s", |
|
74 | 74 | self.path, tb) |
|
75 | 75 | |
|
76 | 76 | def do_GET(self): |
|
77 | 77 | self.do_POST() |
|
78 | 78 | |
|
79 | 79 | def do_hgweb(self): |
|
80 | 80 | path_info, query = _splitURI(self.path) |
|
81 | 81 | |
|
82 | 82 | env = {} |
|
83 | 83 | env['GATEWAY_INTERFACE'] = 'CGI/1.1' |
|
84 | 84 | env['REQUEST_METHOD'] = self.command |
|
85 | 85 | env['SERVER_NAME'] = self.server.server_name |
|
86 | 86 | env['SERVER_PORT'] = str(self.server.server_port) |
|
87 | 87 | env['REQUEST_URI'] = self.path |
|
88 | 88 | env['PATH_INFO'] = path_info |
|
89 | 89 | env['REMOTE_HOST'] = self.client_address[0] |
|
90 | 90 | env['REMOTE_ADDR'] = self.client_address[0] |
|
91 | 91 | if query: |
|
92 | 92 | env['QUERY_STRING'] = query |
|
93 | 93 | |
|
94 | 94 | if self.headers.typeheader is None: |
|
95 | 95 | env['CONTENT_TYPE'] = self.headers.type |
|
96 | 96 | else: |
|
97 | 97 | env['CONTENT_TYPE'] = self.headers.typeheader |
|
98 | 98 | length = self.headers.getheader('content-length') |
|
99 | 99 | if length: |
|
100 | 100 | env['CONTENT_LENGTH'] = length |
|
101 | 101 | for header in [h for h in self.headers.keys() |
|
102 | 102 | if h not in ('content-type', 'content-length')]: |
|
103 | 103 | hkey = 'HTTP_' + header.replace('-', '_').upper() |
|
104 | 104 | hval = self.headers.getheader(header) |
|
105 | 105 | hval = hval.replace('\n', '').strip() |
|
106 | 106 | if hval: |
|
107 | 107 | env[hkey] = hval |
|
108 | 108 | env['SERVER_PROTOCOL'] = self.request_version |
|
109 | 109 | env['wsgi.version'] = (1, 0) |
|
110 | 110 | env['wsgi.url_scheme'] = self.url_scheme |
|
111 | 111 | env['wsgi.input'] = self.rfile |
|
112 | 112 | env['wsgi.errors'] = _error_logger(self) |
|
113 | 113 | env['wsgi.multithread'] = isinstance(self.server, |
|
114 | 114 | SocketServer.ThreadingMixIn) |
|
115 | 115 | env['wsgi.multiprocess'] = isinstance(self.server, |
|
116 | 116 | SocketServer.ForkingMixIn) |
|
117 | 117 | env['wsgi.run_once'] = 0 |
|
118 | 118 | |
|
119 | 119 | self.close_connection = True |
|
120 | 120 | self.saved_status = None |
|
121 | 121 | self.saved_headers = [] |
|
122 | 122 | self.sent_headers = False |
|
123 | 123 | self.length = None |
|
124 | 124 | req = self.server.reqmaker(env, self._start_response) |
|
125 | 125 | for data in req: |
|
126 | 126 | if data: |
|
127 | 127 | self._write(data) |
|
128 | 128 | |
|
129 | 129 | def send_headers(self): |
|
130 | 130 | if not self.saved_status: |
|
131 | 131 | raise AssertionError("Sending headers before start_response() called") |
|
132 | 132 | saved_status = self.saved_status.split(None, 1) |
|
133 | 133 | saved_status[0] = int(saved_status[0]) |
|
134 | 134 | self.send_response(*saved_status) |
|
135 | 135 | should_close = True |
|
136 | 136 | for h in self.saved_headers: |
|
137 | 137 | self.send_header(*h) |
|
138 | 138 | if h[0].lower() == 'content-length': |
|
139 | 139 | should_close = False |
|
140 | 140 | self.length = int(h[1]) |
|
141 | 141 | # The value of the Connection header is a list of case-insensitive |
|
142 | 142 | # tokens separated by commas and optional whitespace. |
|
143 | 143 | if 'close' in [token.strip().lower() for token in |
|
144 | 144 | self.headers.get('connection', '').split(',')]: |
|
145 | 145 | should_close = True |
|
146 | 146 | if should_close: |
|
147 | 147 | self.send_header('Connection', 'close') |
|
148 | 148 | self.close_connection = should_close |
|
149 | 149 | self.end_headers() |
|
150 | 150 | self.sent_headers = True |
|
151 | 151 | |
|
152 | 152 | def _start_response(self, http_status, headers, exc_info=None): |
|
153 | 153 | code, msg = http_status.split(None, 1) |
|
154 | 154 | code = int(code) |
|
155 | 155 | self.saved_status = http_status |
|
156 | 156 | bad_headers = ('connection', 'transfer-encoding') |
|
157 | 157 | self.saved_headers = [h for h in headers |
|
158 | 158 | if h[0].lower() not in bad_headers] |
|
159 | 159 | return self._write |
|
160 | 160 | |
|
161 | 161 | def _write(self, data): |
|
162 | 162 | if not self.saved_status: |
|
163 | 163 | raise AssertionError("data written before start_response() called") |
|
164 | 164 | elif not self.sent_headers: |
|
165 | 165 | self.send_headers() |
|
166 | 166 | if self.length is not None: |
|
167 | 167 | if len(data) > self.length: |
|
168 | 168 | raise AssertionError("Content-length header sent, but more bytes than specified are being written.") |
|
169 | 169 | self.length = self.length - len(data) |
|
170 | 170 | self.wfile.write(data) |
|
171 | 171 | self.wfile.flush() |
|
172 | 172 | |
|
173 | 173 | class _shgwebhandler(_hgwebhandler): |
|
174 | 174 | |
|
175 | 175 | url_scheme = 'https' |
|
176 | ||
|
176 | ||
|
177 | 177 | def setup(self): |
|
178 | 178 | self.connection = self.request |
|
179 | 179 | self.rfile = socket._fileobject(self.request, "rb", self.rbufsize) |
|
180 | 180 | self.wfile = socket._fileobject(self.request, "wb", self.wbufsize) |
|
181 | 181 | |
|
182 | 182 | def do_write(self): |
|
183 | 183 | from OpenSSL.SSL import SysCallError |
|
184 | 184 | try: |
|
185 | 185 | super(_shgwebhandler, self).do_write() |
|
186 | 186 | except SysCallError, inst: |
|
187 | 187 | if inst.args[0] != errno.EPIPE: |
|
188 | 188 | raise |
|
189 | 189 | |
|
190 | 190 | def handle_one_request(self): |
|
191 | 191 | from OpenSSL.SSL import SysCallError, ZeroReturnError |
|
192 | 192 | try: |
|
193 | 193 | super(_shgwebhandler, self).handle_one_request() |
|
194 | 194 | except (SysCallError, ZeroReturnError): |
|
195 | 195 | self.close_connection = True |
|
196 | 196 | pass |
|
197 | 197 | |
|
198 | 198 | def create_server(ui, repo): |
|
199 | 199 | use_threads = True |
|
200 | 200 | |
|
201 | 201 | def openlog(opt, default): |
|
202 | 202 | if opt and opt != '-': |
|
203 | 203 | return open(opt, 'w') |
|
204 | 204 | return default |
|
205 | 205 | |
|
206 | 206 | address = repo.ui.config("web", "address", "") |
|
207 | 207 | port = int(repo.ui.config("web", "port", 8000)) |
|
208 | 208 | use_ipv6 = repo.ui.configbool("web", "ipv6") |
|
209 | 209 | webdir_conf = repo.ui.config("web", "webdir_conf") |
|
210 | 210 | ssl_cert = repo.ui.config("web", "certificate") |
|
211 | 211 | accesslog = openlog(repo.ui.config("web", "accesslog", "-"), sys.stdout) |
|
212 | 212 | errorlog = openlog(repo.ui.config("web", "errorlog", "-"), sys.stderr) |
|
213 | 213 | |
|
214 | 214 | if use_threads: |
|
215 | 215 | try: |
|
216 | 216 | from threading import activeCount |
|
217 | 217 | except ImportError: |
|
218 | 218 | use_threads = False |
|
219 | 219 | |
|
220 | 220 | if use_threads: |
|
221 | 221 | _mixin = SocketServer.ThreadingMixIn |
|
222 | 222 | else: |
|
223 | 223 | if hasattr(os, "fork"): |
|
224 | 224 | _mixin = SocketServer.ForkingMixIn |
|
225 | 225 | else: |
|
226 | 226 | class _mixin: |
|
227 | 227 | pass |
|
228 | 228 | |
|
229 | 229 | class MercurialHTTPServer(object, _mixin, BaseHTTPServer.HTTPServer): |
|
230 | 230 | |
|
231 | 231 | # SO_REUSEADDR has broken semantics on windows |
|
232 | 232 | if os.name == 'nt': |
|
233 | 233 | allow_reuse_address = 0 |
|
234 | 234 | |
|
235 | 235 | def __init__(self, *args, **kargs): |
|
236 | 236 | BaseHTTPServer.HTTPServer.__init__(self, *args, **kargs) |
|
237 | 237 | self.accesslog = accesslog |
|
238 | 238 | self.errorlog = errorlog |
|
239 | 239 | self.daemon_threads = True |
|
240 | 240 | def make_handler(): |
|
241 | 241 | if webdir_conf: |
|
242 | 242 | hgwebobj = hgwebdir(webdir_conf, ui) |
|
243 | 243 | elif repo is not None: |
|
244 | 244 | hgwebobj = hgweb(hg.repository(repo.ui, repo.root)) |
|
245 | 245 | else: |
|
246 | 246 | raise hg.RepoError(_("There is no Mercurial repository here" |
|
247 | 247 | " (.hg not found)")) |
|
248 | 248 | return hgwebobj |
|
249 | 249 | self.reqmaker = wsgiapplication(make_handler) |
|
250 | 250 | |
|
251 | 251 | addr = address |
|
252 | 252 | if addr in ('', '::'): |
|
253 | 253 | addr = socket.gethostname() |
|
254 | 254 | |
|
255 | 255 | self.addr, self.port = addr, port |
|
256 | 256 | |
|
257 | 257 | if ssl_cert: |
|
258 | 258 | try: |
|
259 | 259 | from OpenSSL import SSL |
|
260 | 260 | ctx = SSL.Context(SSL.SSLv23_METHOD) |
|
261 | 261 | except ImportError: |
|
262 | 262 | raise util.Abort("SSL support is unavailable") |
|
263 | 263 | ctx.use_privatekey_file(ssl_cert) |
|
264 | 264 | ctx.use_certificate_file(ssl_cert) |
|
265 | 265 | sock = socket.socket(self.address_family, self.socket_type) |
|
266 | 266 | self.socket = SSL.Connection(ctx, sock) |
|
267 | 267 | self.server_bind() |
|
268 | 268 | self.server_activate() |
|
269 | 269 | |
|
270 | 270 | class IPv6HTTPServer(MercurialHTTPServer): |
|
271 | 271 | address_family = getattr(socket, 'AF_INET6', None) |
|
272 | 272 | |
|
273 | 273 | def __init__(self, *args, **kwargs): |
|
274 | 274 | if self.address_family is None: |
|
275 | 275 | raise hg.RepoError(_('IPv6 not available on this system')) |
|
276 | 276 | super(IPv6HTTPServer, self).__init__(*args, **kwargs) |
|
277 | 277 | |
|
278 | 278 | if ssl_cert: |
|
279 | 279 | handler = _shgwebhandler |
|
280 | 280 | else: |
|
281 | 281 | handler = _hgwebhandler |
|
282 | 282 | |
|
283 | 283 | try: |
|
284 | 284 | if use_ipv6: |
|
285 | 285 | return IPv6HTTPServer((address, port), handler) |
|
286 | 286 | else: |
|
287 | 287 | return MercurialHTTPServer((address, port), handler) |
|
288 | 288 | except socket.error, inst: |
|
289 | 289 | raise util.Abort(_('cannot start server: %s') % inst.args[1]) |
General Comments 0
You need to be logged in to leave comments.
Login now