Show More
@@ -1,109 +1,121 | |||
|
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 | import convcmd |
|
9 | 9 | from mercurial import commands |
|
10 | 10 | |
|
11 | 11 | # Commands definition was moved elsewhere to ease demandload job. |
|
12 | 12 | |
|
13 | 13 | def convert(ui, src, dest=None, revmapfile=None, **opts): |
|
14 | 14 | """Convert a foreign SCM repository to a Mercurial one. |
|
15 | 15 | |
|
16 | 16 | Accepted source formats: |
|
17 | 17 | - Mercurial |
|
18 | 18 | - CVS |
|
19 | 19 | - Darcs |
|
20 | 20 | - git |
|
21 | 21 | - Subversion |
|
22 | 22 | - GNU Arch |
|
23 | 23 | |
|
24 | 24 | Accepted destination formats: |
|
25 | 25 | - Mercurial |
|
26 | 26 | - Subversion (history on branches is not preserved) |
|
27 | 27 | |
|
28 | 28 | If no revision is given, all revisions will be converted. Otherwise, |
|
29 | 29 | convert will only import up to the named revision (given in a format |
|
30 | 30 | understood by the source). |
|
31 | 31 | |
|
32 | 32 | If no destination directory name is specified, it defaults to the |
|
33 | 33 | basename of the source with '-hg' appended. If the destination |
|
34 | 34 | repository doesn't exist, it will be created. |
|
35 | 35 | |
|
36 | 36 | If <MAPFILE> isn't given, it will be put in a default location |
|
37 | 37 | (<dest>/.hg/shamap by default). The <MAPFILE> is a simple text |
|
38 | 38 | file that maps each source commit ID to the destination ID for |
|
39 | 39 | that revision, like so: |
|
40 | 40 | <source ID> <destination ID> |
|
41 | 41 | |
|
42 | 42 | If the file doesn't exist, it's automatically created. It's updated |
|
43 | 43 | on each commit copied, so convert-repo can be interrupted and can |
|
44 | 44 | be run repeatedly to copy new commits. |
|
45 | 45 | |
|
46 | 46 | The [username mapping] file is a simple text file that maps each source |
|
47 | 47 | commit author to a destination commit author. It is handy for source SCMs |
|
48 | 48 | that use unix logins to identify authors (eg: CVS). One line per author |
|
49 | 49 | mapping and the line format is: |
|
50 | 50 | srcauthor=whatever string you want |
|
51 | 51 | |
|
52 | 52 | The filemap is a file that allows filtering and remapping of files |
|
53 | 53 | and directories. Comment lines start with '#'. Each line can |
|
54 | 54 | contain one of the following directives: |
|
55 | 55 | |
|
56 | 56 | include path/to/file |
|
57 | 57 | |
|
58 | 58 | exclude path/to/file |
|
59 | 59 | |
|
60 | 60 | rename from/file to/file |
|
61 | 61 | |
|
62 | 62 | The 'include' directive causes a file, or all files under a |
|
63 | 63 | directory, to be included in the destination repository, and the |
|
64 | 64 | exclusion of all other files and dirs not explicitely included. |
|
65 | 65 | The 'exclude' directive causes files or directories to be omitted. |
|
66 | 66 | The 'rename' directive renames a file or directory. To rename from a |
|
67 | 67 | subdirectory into the root of the repository, use '.' as the path to |
|
68 | 68 | rename to. |
|
69 | 69 | |
|
70 | The splicemap is a file that allows insertion of synthetic | |
|
71 | history, letting you specify the parents of a revision. This is | |
|
72 | useful if you want to e.g. give a Subversion merge two parents, or | |
|
73 | graft two disconnected series of history together. Each entry | |
|
74 | contains a key, followed by a space, followed by one or two | |
|
75 | values, separated by spaces. The key is the revision ID in the | |
|
76 | source revision control system whose parents should be modified | |
|
77 | (same format as a key in .hg/shamap). The values are the revision | |
|
78 | IDs (in either the source or destination revision control system) | |
|
79 | that should be used as the new parents for that node. | |
|
80 | ||
|
70 | 81 | Back end options: |
|
71 | 82 | |
|
72 | 83 | --config convert.hg.clonebranches=False (boolean) |
|
73 | 84 | hg target: XXX not documented |
|
74 | 85 | --config convert.hg.saverev=True (boolean) |
|
75 | 86 | hg source: allow target to preserve source revision ID |
|
76 | 87 | --config convert.hg.tagsbranch=default (branch name) |
|
77 | 88 | hg target: XXX not documented |
|
78 | 89 | --config convert.hg.usebranchnames=True (boolean) |
|
79 | 90 | hg target: preserve branch names |
|
80 | 91 | |
|
81 | 92 | --config convert.svn.branches=branches (directory name) |
|
82 | 93 | svn source: specify the directory containing branches |
|
83 | 94 | --config convert.svn.tags=tags (directory name) |
|
84 | 95 | svn source: specify the directory containing tags |
|
85 | 96 | --config convert.svn.trunk=trunk (directory name) |
|
86 | 97 | svn source: specify the name of the trunk branch |
|
87 | 98 | """ |
|
88 | 99 | return convcmd.convert(ui, src, dest, revmapfile, **opts) |
|
89 | 100 | |
|
90 | 101 | def debugsvnlog(ui, **opts): |
|
91 | 102 | return convcmd.debugsvnlog(ui, **opts) |
|
92 | 103 | |
|
93 | 104 | commands.norepo += " convert debugsvnlog" |
|
94 | 105 | |
|
95 | 106 | cmdtable = { |
|
96 | 107 | "convert": |
|
97 | 108 | (convert, |
|
98 | 109 | [('A', 'authors', '', 'username mapping filename'), |
|
99 | 110 | ('d', 'dest-type', '', 'destination repository type'), |
|
100 | 111 | ('', 'filemap', '', 'remap file names using contents of file'), |
|
101 | 112 | ('r', 'rev', '', 'import up to target revision REV'), |
|
102 | 113 | ('s', 'source-type', '', 'source repository type'), |
|
114 | ('', 'splicemap', '', 'splice synthesized history into place'), | |
|
103 | 115 | ('', 'datesort', None, 'try to sort changesets by date')], |
|
104 | 116 | 'hg convert [OPTION]... SOURCE [DEST [MAPFILE]]'), |
|
105 | 117 | "debugsvnlog": |
|
106 | 118 | (debugsvnlog, |
|
107 | 119 | [], |
|
108 | 120 | 'hg debugsvnlog'), |
|
109 | 121 | } |
@@ -1,348 +1,349 | |||
|
1 | 1 | # convcmd - convert extension commands definition |
|
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, SKIPREV, converter_source, converter_sink, mapfile |
|
9 | 9 | from cvs import convert_cvs |
|
10 | 10 | from darcs import darcs_source |
|
11 | 11 | from git import convert_git |
|
12 | 12 | from hg import mercurial_source, mercurial_sink |
|
13 | 13 | from subversion import debugsvnlog, svn_source, svn_sink |
|
14 | 14 | from gnuarch import gnuarch_source |
|
15 | 15 | import filemap |
|
16 | 16 | |
|
17 | 17 | import os, shutil |
|
18 | 18 | from mercurial import hg, util |
|
19 | 19 | from mercurial.i18n import _ |
|
20 | 20 | |
|
21 | 21 | orig_encoding = 'ascii' |
|
22 | 22 | |
|
23 | 23 | def recode(s): |
|
24 | 24 | if isinstance(s, unicode): |
|
25 | 25 | return s.encode(orig_encoding, 'replace') |
|
26 | 26 | else: |
|
27 | 27 | return s.decode('utf-8').encode(orig_encoding, 'replace') |
|
28 | 28 | |
|
29 | 29 | source_converters = [ |
|
30 | 30 | ('cvs', convert_cvs), |
|
31 | 31 | ('git', convert_git), |
|
32 | 32 | ('svn', svn_source), |
|
33 | 33 | ('hg', mercurial_source), |
|
34 | 34 | ('darcs', darcs_source), |
|
35 | 35 | ('gnuarch', gnuarch_source), |
|
36 | 36 | ] |
|
37 | 37 | |
|
38 | 38 | sink_converters = [ |
|
39 | 39 | ('hg', mercurial_sink), |
|
40 | 40 | ('svn', svn_sink), |
|
41 | 41 | ] |
|
42 | 42 | |
|
43 | 43 | def convertsource(ui, path, type, rev): |
|
44 | 44 | exceptions = [] |
|
45 | 45 | for name, source in source_converters: |
|
46 | 46 | try: |
|
47 | 47 | if not type or name == type: |
|
48 | 48 | return source(ui, path, rev) |
|
49 | 49 | except NoRepo, inst: |
|
50 | 50 | exceptions.append(inst) |
|
51 | 51 | if not ui.quiet: |
|
52 | 52 | for inst in exceptions: |
|
53 | 53 | ui.write(_("%s\n") % inst) |
|
54 | 54 | raise util.Abort('%s: unknown repository type' % path) |
|
55 | 55 | |
|
56 | 56 | def convertsink(ui, path, type): |
|
57 | 57 | for name, sink in sink_converters: |
|
58 | 58 | try: |
|
59 | 59 | if not type or name == type: |
|
60 | 60 | return sink(ui, path) |
|
61 | 61 | except NoRepo, inst: |
|
62 | 62 | ui.note(_("convert: %s\n") % inst) |
|
63 | 63 | raise util.Abort('%s: unknown repository type' % path) |
|
64 | 64 | |
|
65 | 65 | class converter(object): |
|
66 | 66 | def __init__(self, ui, source, dest, revmapfile, opts): |
|
67 | 67 | |
|
68 | 68 | self.source = source |
|
69 | 69 | self.dest = dest |
|
70 | 70 | self.ui = ui |
|
71 | 71 | self.opts = opts |
|
72 | 72 | self.commitcache = {} |
|
73 | 73 | self.authors = {} |
|
74 | 74 | self.authorfile = None |
|
75 | 75 | |
|
76 | 76 | self.map = mapfile(ui, revmapfile) |
|
77 | 77 | |
|
78 | 78 | # Read first the dst author map if any |
|
79 | 79 | authorfile = self.dest.authorfile() |
|
80 | 80 | if authorfile and os.path.exists(authorfile): |
|
81 | 81 | self.readauthormap(authorfile) |
|
82 | 82 | # Extend/Override with new author map if necessary |
|
83 | 83 | if opts.get('authors'): |
|
84 | 84 | self.readauthormap(opts.get('authors')) |
|
85 | 85 | self.authorfile = self.dest.authorfile() |
|
86 | 86 | |
|
87 |
self.splicemap = mapfile(ui, |
|
|
87 | self.splicemap = mapfile(ui, opts.get('splicemap')) | |
|
88 | 88 | |
|
89 | 89 | def walktree(self, heads): |
|
90 | 90 | '''Return a mapping that identifies the uncommitted parents of every |
|
91 | 91 | uncommitted changeset.''' |
|
92 | 92 | visit = heads |
|
93 | 93 | known = {} |
|
94 | 94 | parents = {} |
|
95 | 95 | while visit: |
|
96 | 96 | n = visit.pop(0) |
|
97 | 97 | if n in known or n in self.map: continue |
|
98 | 98 | known[n] = 1 |
|
99 | 99 | commit = self.cachecommit(n) |
|
100 | 100 | parents[n] = [] |
|
101 | 101 | for p in commit.parents: |
|
102 | 102 | parents[n].append(p) |
|
103 | 103 | visit.append(p) |
|
104 | 104 | |
|
105 | 105 | return parents |
|
106 | 106 | |
|
107 | 107 | def toposort(self, parents): |
|
108 | 108 | '''Return an ordering such that every uncommitted changeset is |
|
109 | 109 | preceeded by all its uncommitted ancestors.''' |
|
110 | 110 | visit = parents.keys() |
|
111 | 111 | seen = {} |
|
112 | 112 | children = {} |
|
113 | 113 | actives = [] |
|
114 | 114 | |
|
115 | 115 | while visit: |
|
116 | 116 | n = visit.pop(0) |
|
117 | 117 | if n in seen: continue |
|
118 | 118 | seen[n] = 1 |
|
119 | 119 | # Ensure that nodes without parents are present in the 'children' |
|
120 | 120 | # mapping. |
|
121 | 121 | children.setdefault(n, []) |
|
122 | 122 | hasparent = False |
|
123 | 123 | for p in parents[n]: |
|
124 | 124 | if not p in self.map: |
|
125 | 125 | visit.append(p) |
|
126 | 126 | hasparent = True |
|
127 | 127 | children.setdefault(p, []).append(n) |
|
128 | 128 | if not hasparent: |
|
129 | 129 | actives.append(n) |
|
130 | 130 | |
|
131 | 131 | del seen |
|
132 | 132 | del visit |
|
133 | 133 | |
|
134 | 134 | if self.opts.get('datesort'): |
|
135 | 135 | dates = {} |
|
136 | 136 | def getdate(n): |
|
137 | 137 | if n not in dates: |
|
138 | 138 | dates[n] = util.parsedate(self.commitcache[n].date) |
|
139 | 139 | return dates[n] |
|
140 | 140 | |
|
141 | 141 | def picknext(nodes): |
|
142 | 142 | return min([(getdate(n), n) for n in nodes])[1] |
|
143 | 143 | else: |
|
144 | 144 | prev = [None] |
|
145 | 145 | def picknext(nodes): |
|
146 | 146 | # Return the first eligible child of the previously converted |
|
147 | 147 | # revision, or any of them. |
|
148 | 148 | next = nodes[0] |
|
149 | 149 | for n in nodes: |
|
150 | 150 | if prev[0] in parents[n]: |
|
151 | 151 | next = n |
|
152 | 152 | break |
|
153 | 153 | prev[0] = next |
|
154 | 154 | return next |
|
155 | 155 | |
|
156 | 156 | s = [] |
|
157 | 157 | pendings = {} |
|
158 | 158 | while actives: |
|
159 | 159 | n = picknext(actives) |
|
160 | 160 | actives.remove(n) |
|
161 | 161 | s.append(n) |
|
162 | 162 | |
|
163 | 163 | # Update dependents list |
|
164 | 164 | for c in children.get(n, []): |
|
165 | 165 | if c not in pendings: |
|
166 | 166 | pendings[c] = [p for p in parents[c] if p not in self.map] |
|
167 | 167 | try: |
|
168 | 168 | pendings[c].remove(n) |
|
169 | 169 | except ValueError: |
|
170 | 170 | raise util.Abort(_('cycle detected between %s and %s') |
|
171 | 171 | % (recode(c), recode(n))) |
|
172 | 172 | if not pendings[c]: |
|
173 | 173 | # Parents are converted, node is eligible |
|
174 | 174 | actives.insert(0, c) |
|
175 | 175 | pendings[c] = None |
|
176 | 176 | |
|
177 | 177 | if len(s) != len(parents): |
|
178 | 178 | raise util.Abort(_("not all revisions were sorted")) |
|
179 | 179 | |
|
180 | 180 | return s |
|
181 | 181 | |
|
182 | 182 | def writeauthormap(self): |
|
183 | 183 | authorfile = self.authorfile |
|
184 | 184 | if authorfile: |
|
185 | 185 | self.ui.status('Writing author map file %s\n' % authorfile) |
|
186 | 186 | ofile = open(authorfile, 'w+') |
|
187 | 187 | for author in self.authors: |
|
188 | 188 | ofile.write("%s=%s\n" % (author, self.authors[author])) |
|
189 | 189 | ofile.close() |
|
190 | 190 | |
|
191 | 191 | def readauthormap(self, authorfile): |
|
192 | 192 | afile = open(authorfile, 'r') |
|
193 | 193 | for line in afile: |
|
194 | 194 | try: |
|
195 | 195 | srcauthor = line.split('=')[0].strip() |
|
196 | 196 | dstauthor = line.split('=')[1].strip() |
|
197 | 197 | if srcauthor in self.authors and dstauthor != self.authors[srcauthor]: |
|
198 | 198 | self.ui.status( |
|
199 | 199 | 'Overriding mapping for author %s, was %s, will be %s\n' |
|
200 | 200 | % (srcauthor, self.authors[srcauthor], dstauthor)) |
|
201 | 201 | else: |
|
202 | 202 | self.ui.debug('Mapping author %s to %s\n' |
|
203 | 203 | % (srcauthor, dstauthor)) |
|
204 | 204 | self.authors[srcauthor] = dstauthor |
|
205 | 205 | except IndexError: |
|
206 | 206 | self.ui.warn( |
|
207 | 207 | 'Ignoring bad line in author file map %s: %s\n' |
|
208 | 208 | % (authorfile, line)) |
|
209 | 209 | afile.close() |
|
210 | 210 | |
|
211 | 211 | def cachecommit(self, rev): |
|
212 | 212 | commit = self.source.getcommit(rev) |
|
213 | 213 | commit.author = self.authors.get(commit.author, commit.author) |
|
214 | 214 | self.commitcache[rev] = commit |
|
215 | 215 | return commit |
|
216 | 216 | |
|
217 | 217 | def copy(self, rev): |
|
218 | 218 | commit = self.commitcache[rev] |
|
219 | 219 | do_copies = hasattr(self.dest, 'copyfile') |
|
220 | 220 | filenames = [] |
|
221 | 221 | |
|
222 | 222 | changes = self.source.getchanges(rev) |
|
223 | 223 | if isinstance(changes, basestring): |
|
224 | 224 | if changes == SKIPREV: |
|
225 | 225 | dest = SKIPREV |
|
226 | 226 | else: |
|
227 | 227 | dest = self.map[changes] |
|
228 | 228 | self.map[rev] = dest |
|
229 | 229 | return |
|
230 | 230 | files, copies = changes |
|
231 | 231 | pbranches = [] |
|
232 | 232 | if commit.parents: |
|
233 | 233 | for prev in commit.parents: |
|
234 | 234 | if prev not in self.commitcache: |
|
235 | 235 | self.cachecommit(prev) |
|
236 | 236 | pbranches.append((self.map[prev], |
|
237 | 237 | self.commitcache[prev].branch)) |
|
238 | 238 | self.dest.setbranch(commit.branch, pbranches) |
|
239 | 239 | for f, v in files: |
|
240 | 240 | filenames.append(f) |
|
241 | 241 | try: |
|
242 | 242 | data = self.source.getfile(f, v) |
|
243 | 243 | except IOError, inst: |
|
244 | 244 | self.dest.delfile(f) |
|
245 | 245 | else: |
|
246 | 246 | e = self.source.getmode(f, v) |
|
247 | 247 | self.dest.putfile(f, e, data) |
|
248 | 248 | if do_copies: |
|
249 | 249 | if f in copies: |
|
250 | 250 | copyf = copies[f] |
|
251 | 251 | # Merely marks that a copy happened. |
|
252 | 252 | self.dest.copyfile(copyf, f) |
|
253 | 253 | |
|
254 | 254 | try: |
|
255 |
parents = |
|
|
256 |
self.ui. |
|
|
257 | (parents, rev)) | |
|
255 | parents = self.splicemap[rev].replace(',', ' ').split() | |
|
256 | self.ui.status('spliced in %s as parents of %s\n' % | |
|
257 | (parents, rev)) | |
|
258 | parents = [self.map.get(p, p) for p in parents] | |
|
258 | 259 | except KeyError: |
|
259 | 260 | parents = [b[0] for b in pbranches] |
|
260 | 261 | newnode = self.dest.putcommit(filenames, parents, commit) |
|
261 | 262 | self.source.converted(rev, newnode) |
|
262 | 263 | self.map[rev] = newnode |
|
263 | 264 | |
|
264 | 265 | def convert(self): |
|
265 | 266 | |
|
266 | 267 | try: |
|
267 | 268 | self.source.before() |
|
268 | 269 | self.dest.before() |
|
269 | 270 | self.source.setrevmap(self.map) |
|
270 | 271 | self.ui.status("scanning source...\n") |
|
271 | 272 | heads = self.source.getheads() |
|
272 | 273 | parents = self.walktree(heads) |
|
273 | 274 | self.ui.status("sorting...\n") |
|
274 | 275 | t = self.toposort(parents) |
|
275 | 276 | num = len(t) |
|
276 | 277 | c = None |
|
277 | 278 | |
|
278 | 279 | self.ui.status("converting...\n") |
|
279 | 280 | for c in t: |
|
280 | 281 | num -= 1 |
|
281 | 282 | desc = self.commitcache[c].desc |
|
282 | 283 | if "\n" in desc: |
|
283 | 284 | desc = desc.splitlines()[0] |
|
284 | 285 | # convert log message to local encoding without using |
|
285 | 286 | # tolocal() because util._encoding conver() use it as |
|
286 | 287 | # 'utf-8' |
|
287 | 288 | self.ui.status("%d %s\n" % (num, recode(desc))) |
|
288 | 289 | self.ui.note(_("source: %s\n" % recode(c))) |
|
289 | 290 | self.copy(c) |
|
290 | 291 | |
|
291 | 292 | tags = self.source.gettags() |
|
292 | 293 | ctags = {} |
|
293 | 294 | for k in tags: |
|
294 | 295 | v = tags[k] |
|
295 | 296 | if self.map.get(v, SKIPREV) != SKIPREV: |
|
296 | 297 | ctags[k] = self.map[v] |
|
297 | 298 | |
|
298 | 299 | if c and ctags: |
|
299 | 300 | nrev = self.dest.puttags(ctags) |
|
300 | 301 | # write another hash correspondence to override the previous |
|
301 | 302 | # one so we don't end up with extra tag heads |
|
302 | 303 | if nrev: |
|
303 | 304 | self.map[c] = nrev |
|
304 | 305 | |
|
305 | 306 | self.writeauthormap() |
|
306 | 307 | finally: |
|
307 | 308 | self.cleanup() |
|
308 | 309 | |
|
309 | 310 | def cleanup(self): |
|
310 | 311 | try: |
|
311 | 312 | self.dest.after() |
|
312 | 313 | finally: |
|
313 | 314 | self.source.after() |
|
314 | 315 | self.map.close() |
|
315 | 316 | |
|
316 | 317 | def convert(ui, src, dest=None, revmapfile=None, **opts): |
|
317 | 318 | global orig_encoding |
|
318 | 319 | orig_encoding = util._encoding |
|
319 | 320 | util._encoding = 'UTF-8' |
|
320 | 321 | |
|
321 | 322 | if not dest: |
|
322 | 323 | dest = hg.defaultdest(src) + "-hg" |
|
323 | 324 | ui.status("assuming destination %s\n" % dest) |
|
324 | 325 | |
|
325 | 326 | destc = convertsink(ui, dest, opts.get('dest_type')) |
|
326 | 327 | |
|
327 | 328 | try: |
|
328 | 329 | srcc = convertsource(ui, src, opts.get('source_type'), |
|
329 | 330 | opts.get('rev')) |
|
330 | 331 | except Exception: |
|
331 | 332 | for path in destc.created: |
|
332 | 333 | shutil.rmtree(path, True) |
|
333 | 334 | raise |
|
334 | 335 | |
|
335 | 336 | fmap = opts.get('filemap') |
|
336 | 337 | if fmap: |
|
337 | 338 | srcc = filemap.filemap_source(ui, srcc, fmap) |
|
338 | 339 | destc.setfilemapmode(True) |
|
339 | 340 | |
|
340 | 341 | if not revmapfile: |
|
341 | 342 | try: |
|
342 | 343 | revmapfile = destc.revmapfile() |
|
343 | 344 | except: |
|
344 | 345 | revmapfile = os.path.join(destc, "map") |
|
345 | 346 | |
|
346 | 347 | c = converter(ui, srcc, destc, revmapfile, opts) |
|
347 | 348 | c.convert() |
|
348 | 349 |
General Comments 0
You need to be logged in to leave comments.
Login now