Show More
@@ -1,313 +1,318 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 |
|
9 | 9 | from cvs import convert_cvs |
|
10 | 10 | from git import convert_git |
|
11 | 11 | from hg import convert_mercurial |
|
12 | 12 | |
|
13 | 13 | import os |
|
14 | 14 | from mercurial import hg, ui, util, commands |
|
15 | 15 | |
|
16 | 16 | commands.norepo += " convert" |
|
17 | 17 | |
|
18 | 18 | converters = [convert_cvs, convert_git, convert_mercurial] |
|
19 | 19 | |
|
20 | 20 | def converter(ui, path): |
|
21 | 21 | if not os.path.isdir(path): |
|
22 | 22 | raise util.Abort("%s: not a directory" % path) |
|
23 | 23 | for c in converters: |
|
24 | 24 | try: |
|
25 | 25 | return c(ui, path) |
|
26 | 26 | except NoRepo: |
|
27 | 27 | pass |
|
28 | 28 | raise util.Abort("%s: unknown repository type" % path) |
|
29 | 29 | |
|
30 | 30 | class convert(object): |
|
31 | 31 | def __init__(self, ui, source, dest, mapfile, opts): |
|
32 | 32 | |
|
33 | 33 | self.source = source |
|
34 | 34 | self.dest = dest |
|
35 | 35 | self.ui = ui |
|
36 | 36 | self.opts = opts |
|
37 | 37 | self.commitcache = {} |
|
38 | 38 | self.mapfile = mapfile |
|
39 | 39 | self.mapfilefd = None |
|
40 | 40 | self.authors = {} |
|
41 | 41 | self.authorfile = None |
|
42 | 42 | |
|
43 | 43 | self.map = {} |
|
44 | 44 | try: |
|
45 | 45 | origmapfile = open(self.mapfile, 'r') |
|
46 | 46 | for l in origmapfile: |
|
47 | 47 | sv, dv = l[:-1].split() |
|
48 | 48 | self.map[sv] = dv |
|
49 | 49 | origmapfile.close() |
|
50 | 50 | except IOError: |
|
51 | 51 | pass |
|
52 | 52 | |
|
53 | 53 | # Read first the dst author map if any |
|
54 | 54 | authorfile = self.dest.authorfile() |
|
55 | 55 | if authorfile and os.path.exists(authorfile): |
|
56 | 56 | self.readauthormap(authorfile) |
|
57 | 57 | # Extend/Override with new author map if necessary |
|
58 | 58 | if opts.get('authors'): |
|
59 | 59 | self.readauthormap(opts.get('authors')) |
|
60 | 60 | self.authorfile = self.dest.authorfile() |
|
61 | 61 | |
|
62 | 62 | def walktree(self, heads): |
|
63 | '''Return a mapping that identifies the uncommitted parents of every | |
|
64 | uncommitted changeset.''' | |
|
63 | 65 | visit = heads |
|
64 | 66 | known = {} |
|
65 | 67 | parents = {} |
|
66 | 68 | while visit: |
|
67 | 69 | n = visit.pop(0) |
|
68 | 70 | if n in known or n in self.map: continue |
|
69 | 71 | known[n] = 1 |
|
70 | 72 | self.commitcache[n] = self.source.getcommit(n) |
|
71 | 73 | cp = self.commitcache[n].parents |
|
74 | parents[n] = [] | |
|
72 | 75 | for p in cp: |
|
73 |
parents |
|
|
76 | parents[n].append(p) | |
|
74 | 77 | visit.append(p) |
|
75 | 78 | |
|
76 | 79 | return parents |
|
77 | 80 | |
|
78 | 81 | def toposort(self, parents): |
|
82 | '''Return an ordering such that every uncommitted changeset is | |
|
83 | preceeded by all its uncommitted ancestors.''' | |
|
79 | 84 | visit = parents.keys() |
|
80 | 85 | seen = {} |
|
81 | 86 | children = {} |
|
82 | 87 | |
|
83 | 88 | while visit: |
|
84 | 89 | n = visit.pop(0) |
|
85 | 90 | if n in seen: continue |
|
86 | 91 | seen[n] = 1 |
|
87 | pc = 0 | |
|
88 |
|
|
|
89 | for p in parents[n]: | |
|
90 | if p not in self.map: pc += 1 | |
|
92 | # Ensure that nodes without parents are present in the 'children' | |
|
93 | # mapping. | |
|
94 | children.setdefault(n, []) | |
|
95 | for p in parents[n]: | |
|
96 | if not p in self.map: | |
|
91 | 97 | visit.append(p) |
|
92 |
|
|
|
93 | if not pc: root = n | |
|
98 | children.setdefault(p, []).append(n) | |
|
94 | 99 | |
|
95 | 100 | s = [] |
|
96 | 101 | removed = {} |
|
97 | 102 | visit = children.keys() |
|
98 | 103 | while visit: |
|
99 | 104 | n = visit.pop(0) |
|
100 | 105 | if n in removed: continue |
|
101 | 106 | dep = 0 |
|
102 | 107 | if n in parents: |
|
103 | 108 | for p in parents[n]: |
|
104 | 109 | if p in self.map: continue |
|
105 | 110 | if p not in removed: |
|
106 | 111 | # we're still dependent |
|
107 | 112 | visit.append(n) |
|
108 | 113 | dep = 1 |
|
109 | 114 | break |
|
110 | 115 | |
|
111 | 116 | if not dep: |
|
112 | 117 | # all n's parents are in the list |
|
113 | 118 | removed[n] = 1 |
|
114 | 119 | if n not in self.map: |
|
115 | 120 | s.append(n) |
|
116 | 121 | if n in children: |
|
117 | 122 | for c in children[n]: |
|
118 | 123 | visit.insert(0, c) |
|
119 | 124 | |
|
120 | 125 | if self.opts.get('datesort'): |
|
121 | 126 | depth = {} |
|
122 | 127 | for n in s: |
|
123 | 128 | depth[n] = 0 |
|
124 | 129 | pl = [p for p in self.commitcache[n].parents |
|
125 | 130 | if p not in self.map] |
|
126 | 131 | if pl: |
|
127 | 132 | depth[n] = max([depth[p] for p in pl]) + 1 |
|
128 | 133 | |
|
129 | 134 | s = [(depth[n], self.commitcache[n].date, n) for n in s] |
|
130 | 135 | s.sort() |
|
131 | 136 | s = [e[2] for e in s] |
|
132 | 137 | |
|
133 | 138 | return s |
|
134 | 139 | |
|
135 | 140 | def mapentry(self, src, dst): |
|
136 | 141 | if self.mapfilefd is None: |
|
137 | 142 | try: |
|
138 | 143 | self.mapfilefd = open(self.mapfile, "a") |
|
139 | 144 | except IOError, (errno, strerror): |
|
140 | 145 | raise util.Abort("Could not open map file %s: %s, %s\n" % (self.mapfile, errno, strerror)) |
|
141 | 146 | self.map[src] = dst |
|
142 | 147 | self.mapfilefd.write("%s %s\n" % (src, dst)) |
|
143 | 148 | self.mapfilefd.flush() |
|
144 | 149 | |
|
145 | 150 | def writeauthormap(self): |
|
146 | 151 | authorfile = self.authorfile |
|
147 | 152 | if authorfile: |
|
148 | 153 | self.ui.status('Writing author map file %s\n' % authorfile) |
|
149 | 154 | ofile = open(authorfile, 'w+') |
|
150 | 155 | for author in self.authors: |
|
151 | 156 | ofile.write("%s=%s\n" % (author, self.authors[author])) |
|
152 | 157 | ofile.close() |
|
153 | 158 | |
|
154 | 159 | def readauthormap(self, authorfile): |
|
155 | 160 | afile = open(authorfile, 'r') |
|
156 | 161 | for line in afile: |
|
157 | 162 | try: |
|
158 | 163 | srcauthor = line.split('=')[0].strip() |
|
159 | 164 | dstauthor = line.split('=')[1].strip() |
|
160 | 165 | if srcauthor in self.authors and dstauthor != self.authors[srcauthor]: |
|
161 | 166 | self.ui.status( |
|
162 | 167 | 'Overriding mapping for author %s, was %s, will be %s\n' |
|
163 | 168 | % (srcauthor, self.authors[srcauthor], dstauthor)) |
|
164 | 169 | else: |
|
165 | 170 | self.ui.debug('Mapping author %s to %s\n' |
|
166 | 171 | % (srcauthor, dstauthor)) |
|
167 | 172 | self.authors[srcauthor] = dstauthor |
|
168 | 173 | except IndexError: |
|
169 | 174 | self.ui.warn( |
|
170 | 175 | 'Ignoring bad line in author file map %s: %s\n' |
|
171 | 176 | % (authorfile, line)) |
|
172 | 177 | afile.close() |
|
173 | 178 | |
|
174 | 179 | def copy(self, rev): |
|
175 | 180 | c = self.commitcache[rev] |
|
176 | 181 | files = self.source.getchanges(rev) |
|
177 | 182 | |
|
178 | 183 | for f, v in files: |
|
179 | 184 | try: |
|
180 | 185 | data = self.source.getfile(f, v) |
|
181 | 186 | except IOError, inst: |
|
182 | 187 | self.dest.delfile(f) |
|
183 | 188 | else: |
|
184 | 189 | e = self.source.getmode(f, v) |
|
185 | 190 | self.dest.putfile(f, e, data) |
|
186 | 191 | |
|
187 | 192 | r = [self.map[v] for v in c.parents] |
|
188 | 193 | f = [f for f, v in files] |
|
189 | 194 | newnode = self.dest.putcommit(f, r, c) |
|
190 | 195 | self.mapentry(rev, newnode) |
|
191 | 196 | |
|
192 | 197 | def convert(self): |
|
193 | 198 | try: |
|
194 | 199 | self.ui.status("scanning source...\n") |
|
195 | 200 | heads = self.source.getheads() |
|
196 | 201 | parents = self.walktree(heads) |
|
197 | 202 | self.ui.status("sorting...\n") |
|
198 | 203 | t = self.toposort(parents) |
|
199 | 204 | num = len(t) |
|
200 | 205 | c = None |
|
201 | 206 | |
|
202 | 207 | self.ui.status("converting...\n") |
|
203 | 208 | for c in t: |
|
204 | 209 | num -= 1 |
|
205 | 210 | desc = self.commitcache[c].desc |
|
206 | 211 | if "\n" in desc: |
|
207 | 212 | desc = desc.splitlines()[0] |
|
208 | 213 | author = self.commitcache[c].author |
|
209 | 214 | author = self.authors.get(author, author) |
|
210 | 215 | self.commitcache[c].author = author |
|
211 | 216 | self.ui.status("%d %s\n" % (num, desc)) |
|
212 | 217 | self.copy(c) |
|
213 | 218 | |
|
214 | 219 | tags = self.source.gettags() |
|
215 | 220 | ctags = {} |
|
216 | 221 | for k in tags: |
|
217 | 222 | v = tags[k] |
|
218 | 223 | if v in self.map: |
|
219 | 224 | ctags[k] = self.map[v] |
|
220 | 225 | |
|
221 | 226 | if c and ctags: |
|
222 | 227 | nrev = self.dest.puttags(ctags) |
|
223 | 228 | # write another hash correspondence to override the previous |
|
224 | 229 | # one so we don't end up with extra tag heads |
|
225 | 230 | if nrev: |
|
226 | 231 | self.mapentry(c, nrev) |
|
227 | 232 | |
|
228 | 233 | self.writeauthormap() |
|
229 | 234 | finally: |
|
230 | 235 | self.cleanup() |
|
231 | 236 | |
|
232 | 237 | def cleanup(self): |
|
233 | 238 | if self.mapfilefd: |
|
234 | 239 | self.mapfilefd.close() |
|
235 | 240 | |
|
236 | 241 | def _convert(ui, src, dest=None, mapfile=None, **opts): |
|
237 | 242 | '''Convert a foreign SCM repository to a Mercurial one. |
|
238 | 243 | |
|
239 | 244 | Accepted source formats: |
|
240 | 245 | - GIT |
|
241 | 246 | - CVS |
|
242 | 247 | |
|
243 | 248 | Accepted destination formats: |
|
244 | 249 | - Mercurial |
|
245 | 250 | |
|
246 | 251 | If destination isn't given, a new Mercurial repo named <src>-hg will |
|
247 | 252 | be created. If <mapfile> isn't given, it will be put in a default |
|
248 | 253 | location (<dest>/.hg/shamap by default) |
|
249 | 254 | |
|
250 | 255 | The <mapfile> is a simple text file that maps each source commit ID to |
|
251 | 256 | the destination ID for that revision, like so: |
|
252 | 257 | <source ID> <destination ID> |
|
253 | 258 | |
|
254 | 259 | If the file doesn't exist, it's automatically created. It's updated |
|
255 | 260 | on each commit copied, so convert-repo can be interrupted and can |
|
256 | 261 | be run repeatedly to copy new commits. |
|
257 | 262 | |
|
258 | 263 | The [username mapping] file is a simple text file that maps each source |
|
259 | 264 | commit author to a destination commit author. It is handy for source SCMs |
|
260 | 265 | that use unix logins to identify authors (eg: CVS). One line per author |
|
261 | 266 | mapping and the line format is: |
|
262 | 267 | srcauthor=whatever string you want |
|
263 | 268 | ''' |
|
264 | 269 | |
|
265 | 270 | srcc = converter(ui, src) |
|
266 | 271 | if not hasattr(srcc, "getcommit"): |
|
267 | 272 | raise util.Abort("%s: can't read from this repo type" % src) |
|
268 | 273 | |
|
269 | 274 | if not dest: |
|
270 | 275 | dest = src + "-hg" |
|
271 | 276 | ui.status("assuming destination %s\n" % dest) |
|
272 | 277 | |
|
273 | 278 | # Try to be smart and initalize things when required |
|
274 | 279 | if os.path.isdir(dest): |
|
275 | 280 | if len(os.listdir(dest)) > 0: |
|
276 | 281 | try: |
|
277 | 282 | hg.repository(ui, dest) |
|
278 | 283 | ui.status("destination %s is a Mercurial repository\n" % dest) |
|
279 | 284 | except hg.RepoError: |
|
280 | 285 | raise util.Abort( |
|
281 | 286 | "destination directory %s is not empty.\n" |
|
282 | 287 | "Please specify an empty directory to be initialized\n" |
|
283 | 288 | "or an already initialized mercurial repository" |
|
284 | 289 | % dest) |
|
285 | 290 | else: |
|
286 | 291 | ui.status("initializing destination %s repository\n" % dest) |
|
287 | 292 | hg.repository(ui, dest, create=True) |
|
288 | 293 | elif os.path.exists(dest): |
|
289 | 294 | raise util.Abort("destination %s exists and is not a directory" % dest) |
|
290 | 295 | else: |
|
291 | 296 | ui.status("initializing destination %s repository\n" % dest) |
|
292 | 297 | hg.repository(ui, dest, create=True) |
|
293 | 298 | |
|
294 | 299 | destc = converter(ui, dest) |
|
295 | 300 | if not hasattr(destc, "putcommit"): |
|
296 | 301 | raise util.Abort("%s: can't write to this repo type" % src) |
|
297 | 302 | |
|
298 | 303 | if not mapfile: |
|
299 | 304 | try: |
|
300 | 305 | mapfile = destc.mapfile() |
|
301 | 306 | except: |
|
302 | 307 | mapfile = os.path.join(destc, "map") |
|
303 | 308 | |
|
304 | 309 | c = convert(ui, srcc, destc, mapfile, opts) |
|
305 | 310 | c.convert() |
|
306 | 311 | |
|
307 | 312 | cmdtable = { |
|
308 | 313 | "convert": |
|
309 | 314 | (_convert, |
|
310 | 315 | [('A', 'authors', '', 'username mapping filename'), |
|
311 | 316 | ('', 'datesort', None, 'try to sort changesets by date')], |
|
312 | 317 | 'hg convert [OPTION]... SOURCE [DEST [MAPFILE]]'), |
|
313 | 318 | } |
General Comments 0
You need to be logged in to leave comments.
Login now