Show More
@@ -1,284 +1,284 b'' | |||||
1 | #!/usr/bin/env python |
|
1 | #!/usr/bin/env python | |
2 | # |
|
2 | # | |
3 | # This is a generalized framework for converting between SCM |
|
3 | # This is a generalized framework for converting between SCM | |
4 | # repository formats. |
|
4 | # repository formats. | |
5 | # |
|
5 | # | |
6 | # In its current form, it's hardcoded to convert incrementally between |
|
6 | # In its current form, it's hardcoded to convert incrementally between | |
7 | # git and Mercurial. |
|
7 | # git and Mercurial. | |
8 | # |
|
8 | # | |
9 | # To use, you must first import the first git version into Mercurial, |
|
9 | # To use, you must first import the first git version into Mercurial, | |
10 | # and establish a mapping between the git commit hash and the hash in |
|
10 | # and establish a mapping between the git commit hash and the hash in | |
11 | # Mercurial for that version. This mapping is kept in a simple text |
|
11 | # Mercurial for that version. This mapping is kept in a simple text | |
12 | # file with lines like so: |
|
12 | # file with lines like so: | |
13 | # |
|
13 | # | |
14 | # <git hash> <mercurial hash> |
|
14 | # <git hash> <mercurial hash> | |
15 | # |
|
15 | # | |
16 | # To convert the rest of the repo, run: |
|
16 | # To convert the rest of the repo, run: | |
17 | # |
|
17 | # | |
18 | # convert-repo <git-dir> <hg-dir> <mapfile> |
|
18 | # convert-repo <git-dir> <hg-dir> <mapfile> | |
19 | # |
|
19 | # | |
20 | # This updates the mapfile on each commit copied, so it can be |
|
20 | # This updates the mapfile on each commit copied, so it can be | |
21 | # interrupted and can be run repeatedly to copy new commits. |
|
21 | # interrupted and can be run repeatedly to copy new commits. | |
22 |
|
22 | |||
23 | import sys, os, zlib, sha, time |
|
23 | import sys, os, zlib, sha, time | |
24 | from mercurial import hg, ui, util |
|
24 | from mercurial import hg, ui, util | |
25 |
|
25 | |||
26 | class convert_git: |
|
26 | class convert_git: | |
27 | def __init__(self, path): |
|
27 | def __init__(self, path): | |
28 | self.path = path |
|
28 | self.path = path | |
29 |
|
29 | |||
30 | def getheads(self): |
|
30 | def getheads(self): | |
31 | return [file(self.path + "/HEAD").read()[:-1]] |
|
31 | return [file(self.path + "/HEAD").read()[:-1]] | |
32 |
|
32 | |||
33 | def catfile(self, rev, type): |
|
33 | def catfile(self, rev, type): | |
34 | if rev == "0" * 40: raise IOError() |
|
34 | if rev == "0" * 40: raise IOError() | |
35 | fh = os.popen("GIT_DIR=%s git-cat-file %s %s 2>/dev/null" % (self.path, type, rev)) |
|
35 | fh = os.popen("GIT_DIR=%s git-cat-file %s %s 2>/dev/null" % (self.path, type, rev)) | |
36 | return fh.read() |
|
36 | return fh.read() | |
37 |
|
37 | |||
38 | def getfile(self, name, rev): |
|
38 | def getfile(self, name, rev): | |
39 | return self.catfile(rev, "blob") |
|
39 | return self.catfile(rev, "blob") | |
40 |
|
40 | |||
41 | def getchanges(self, version): |
|
41 | def getchanges(self, version): | |
42 | fh = os.popen("GIT_DIR=%s git-diff-tree --root -m -r %s" % (self.path, version)) |
|
42 | fh = os.popen("GIT_DIR=%s git-diff-tree --root -m -r %s" % (self.path, version)) | |
43 | changes = [] |
|
43 | changes = [] | |
44 | for l in fh: |
|
44 | for l in fh: | |
45 | if "\t" not in l: continue |
|
45 | if "\t" not in l: continue | |
46 | m, f = l[:-1].split("\t") |
|
46 | m, f = l[:-1].split("\t") | |
47 | m = m.split() |
|
47 | m = m.split() | |
48 | h = m[3] |
|
48 | h = m[3] | |
49 | p = (m[1] == "100755") |
|
49 | p = (m[1] == "100755") | |
50 | changes.append((f, h, p)) |
|
50 | changes.append((f, h, p)) | |
51 | return changes |
|
51 | return changes | |
52 |
|
52 | |||
53 | def getcommit(self, version): |
|
53 | def getcommit(self, version): | |
54 | c = self.catfile(version, "commit") # read the commit hash |
|
54 | c = self.catfile(version, "commit") # read the commit hash | |
55 | end = c.find("\n\n") |
|
55 | end = c.find("\n\n") | |
56 | message = c[end+2:] |
|
56 | message = c[end+2:] | |
57 | l = c[:end].splitlines() |
|
57 | l = c[:end].splitlines() | |
58 | manifest = l[0].split()[1] |
|
58 | manifest = l[0].split()[1] | |
59 | parents = [] |
|
59 | parents = [] | |
60 | for e in l[1:]: |
|
60 | for e in l[1:]: | |
61 | n,v = e.split(" ", 1) |
|
61 | n,v = e.split(" ", 1) | |
62 | if n == "author": |
|
62 | if n == "author": | |
63 | p = v.split() |
|
63 | p = v.split() | |
64 | tm, tz = p[-2:] |
|
64 | tm, tz = p[-2:] | |
65 | author = " ".join(p[:-2]) |
|
65 | author = " ".join(p[:-2]) | |
66 | if author[0] == "<": author = author[1:-1] |
|
66 | if author[0] == "<": author = author[1:-1] | |
67 | if n == "committer": |
|
67 | if n == "committer": | |
68 | p = v.split() |
|
68 | p = v.split() | |
69 | tm, tz = p[-2:] |
|
69 | tm, tz = p[-2:] | |
70 | committer = " ".join(p[:-2]) |
|
70 | committer = " ".join(p[:-2]) | |
71 | if committer[0] == "<": committer = committer[1:-1] |
|
71 | if committer[0] == "<": committer = committer[1:-1] | |
72 | message += "\ncommitter: %s\n" % v |
|
72 | message += "\ncommitter: %s\n" % v | |
73 | if n == "parent": parents.append(v) |
|
73 | if n == "parent": parents.append(v) | |
74 |
|
74 | |||
75 | tzs, tzh, tzm = tz[-5:-4] + "1", tz[-4:-2], tz[-2:] |
|
75 | tzs, tzh, tzm = tz[-5:-4] + "1", tz[-4:-2], tz[-2:] | |
76 | tz = int(tzs) * (int(tzh) * 3600 + int(tzm)) |
|
76 | tz = int(tzs) * (int(tzh) * 3600 + int(tzm)) | |
77 | date = tm + " " + str(tz) |
|
77 | date = tm + " " + str(tz) | |
78 | return (parents, author, date, message) |
|
78 | return (parents, author, date, message) | |
79 |
|
79 | |||
80 | def gettags(self): |
|
80 | def gettags(self): | |
81 | tags = {} |
|
81 | tags = {} | |
82 | for f in os.listdir(self.path + "/refs/tags"): |
|
82 | for f in os.listdir(self.path + "/refs/tags"): | |
83 | try: |
|
83 | try: | |
84 | h = file(self.path + "/refs/tags/" + f).read().strip() |
|
84 | h = file(self.path + "/refs/tags/" + f).read().strip() | |
85 | c = self.catfile(h, "tag") # read the commit hash |
|
85 | c = self.catfile(h, "tag") # read the commit hash | |
86 | h = c.splitlines()[0].split()[1] |
|
86 | h = c.splitlines()[0].split()[1] | |
87 | tags[f] = h |
|
87 | tags[f] = h | |
88 | except: |
|
88 | except: | |
89 | pass |
|
89 | pass | |
90 | return tags |
|
90 | return tags | |
91 |
|
91 | |||
92 | class convert_mercurial: |
|
92 | class convert_mercurial: | |
93 | def __init__(self, path): |
|
93 | def __init__(self, path): | |
94 | self.path = path |
|
94 | self.path = path | |
95 | u = ui.ui() |
|
95 | u = ui.ui() | |
96 | self.repo = hg.repository(u, path) |
|
96 | self.repo = hg.repository(u, path) | |
97 |
|
97 | |||
98 | def getheads(self): |
|
98 | def getheads(self): | |
99 | h = self.repo.changelog.heads() |
|
99 | h = self.repo.changelog.heads() | |
100 | return [ hg.hex(x) for x in h ] |
|
100 | return [ hg.hex(x) for x in h ] | |
101 |
|
101 | |||
102 | def putfile(self, f, e, data): |
|
102 | def putfile(self, f, e, data): | |
103 | self.repo.wfile(f, "w").write(data) |
|
103 | self.repo.wfile(f, "w").write(data) | |
104 | if self.repo.dirstate.state(f) == '?': |
|
104 | if self.repo.dirstate.state(f) == '?': | |
105 | self.repo.dirstate.update([f], "a") |
|
105 | self.repo.dirstate.update([f], "a") | |
106 |
|
106 | |||
107 | util.set_exec(self.repo.wjoin(f), e) |
|
107 | util.set_exec(self.repo.wjoin(f), e) | |
108 |
|
108 | |||
109 | def delfile(self, f): |
|
109 | def delfile(self, f): | |
110 | try: |
|
110 | try: | |
111 | os.unlink(self.repo.wjoin(f)) |
|
111 | os.unlink(self.repo.wjoin(f)) | |
112 | #self.repo.remove([f]) |
|
112 | #self.repo.remove([f]) | |
113 | except: |
|
113 | except: | |
114 | pass |
|
114 | pass | |
115 |
|
115 | |||
116 | def putcommit(self, files, parents, author, dest, text): |
|
116 | def putcommit(self, files, parents, author, dest, text): | |
117 | seen = {} |
|
117 | seen = {} | |
118 | pl = [] |
|
118 | pl = [] | |
119 | for p in parents: |
|
119 | for p in parents: | |
120 | if p not in seen: |
|
120 | if p not in seen: | |
121 | pl.append(p) |
|
121 | pl.append(p) | |
122 | seen[p] = 1 |
|
122 | seen[p] = 1 | |
123 | parents = pl |
|
123 | parents = pl | |
124 |
|
124 | |||
125 | if len(parents) < 2: parents.append("0" * 40) |
|
125 | if len(parents) < 2: parents.append("0" * 40) | |
126 | if len(parents) < 2: parents.append("0" * 40) |
|
126 | if len(parents) < 2: parents.append("0" * 40) | |
127 | p2 = parents.pop(0) |
|
127 | p2 = parents.pop(0) | |
128 |
|
128 | |||
129 | while parents: |
|
129 | while parents: | |
130 | p1 = p2 |
|
130 | p1 = p2 | |
131 | p2 = parents.pop(0) |
|
131 | p2 = parents.pop(0) | |
132 | self.repo.rawcommit(files, text, author, dest, |
|
132 | self.repo.rawcommit(files, text, author, dest, | |
133 | hg.bin(p1), hg.bin(p2)) |
|
133 | hg.bin(p1), hg.bin(p2)) | |
134 | text = "(octopus merge fixup)\n" |
|
134 | text = "(octopus merge fixup)\n" | |
135 |
|
135 | |||
136 | return hg.hex(self.repo.changelog.tip()) |
|
136 | return hg.hex(self.repo.changelog.tip()) | |
137 |
|
137 | |||
138 | def puttags(self, tags): |
|
138 | def puttags(self, tags): | |
139 | try: |
|
139 | try: | |
140 | old = self.repo.wfile(".hgtags").read() |
|
140 | old = self.repo.wfile(".hgtags").read() | |
141 | oldlines = old.splitlines(1) |
|
141 | oldlines = old.splitlines(1) | |
142 | oldlines.sort() |
|
142 | oldlines.sort() | |
143 | except: |
|
143 | except: | |
144 | oldlines = [] |
|
144 | oldlines = [] | |
145 |
|
145 | |||
146 | k = tags.keys() |
|
146 | k = tags.keys() | |
147 | k.sort() |
|
147 | k.sort() | |
148 | newlines = [] |
|
148 | newlines = [] | |
149 | for tag in k: |
|
149 | for tag in k: | |
150 | newlines.append("%s %s\n" % (tags[tag], tag)) |
|
150 | newlines.append("%s %s\n" % (tags[tag], tag)) | |
151 |
|
151 | |||
152 | newlines.sort() |
|
152 | newlines.sort() | |
153 |
|
153 | |||
154 | if newlines != oldlines: |
|
154 | if newlines != oldlines: | |
155 | #print "updating tags" |
|
155 | #print "updating tags" | |
156 | f = self.repo.wfile(".hgtags", "w") |
|
156 | f = self.repo.wfile(".hgtags", "w") | |
157 | f.write("".join(newlines)) |
|
157 | f.write("".join(newlines)) | |
158 | f.close() |
|
158 | f.close() | |
159 | if not oldlines: self.repo.add([".hgtags"]) |
|
159 | if not oldlines: self.repo.add([".hgtags"]) | |
160 | date = "%s 0" % int(time.mktime(time.gmtime())) |
|
160 | date = "%s 0" % int(time.mktime(time.gmtime())) | |
161 | self.repo.rawcommit([".hgtags"], "update tags", "convert-repo", |
|
161 | self.repo.rawcommit([".hgtags"], "update tags", "convert-repo", | |
162 | date, self.repo.changelog.tip(), hg.nullid) |
|
162 | date, self.repo.changelog.tip(), hg.nullid) | |
163 | return hg.hex(self.repo.changelog.tip()) |
|
163 | return hg.hex(self.repo.changelog.tip()) | |
164 |
|
164 | |||
165 | class convert: |
|
165 | class convert: | |
166 | def __init__(self, source, dest, mapfile): |
|
166 | def __init__(self, source, dest, mapfile): | |
167 | self.source = source |
|
167 | self.source = source | |
168 | self.dest = dest |
|
168 | self.dest = dest | |
169 | self.mapfile = mapfile |
|
169 | self.mapfile = mapfile | |
170 | self.commitcache = {} |
|
170 | self.commitcache = {} | |
171 |
|
171 | |||
172 | self.map = {} |
|
172 | self.map = {} | |
173 | for l in file(self.mapfile): |
|
173 | for l in file(self.mapfile): | |
174 | sv, dv = l[:-1].split() |
|
174 | sv, dv = l[:-1].split() | |
175 | self.map[sv] = dv |
|
175 | self.map[sv] = dv | |
176 |
|
176 | |||
177 | def walktree(self, heads): |
|
177 | def walktree(self, heads): | |
178 | visit = heads |
|
178 | visit = heads | |
179 | known = {} |
|
179 | known = {} | |
180 | parents = {} |
|
180 | parents = {} | |
181 | while visit: |
|
181 | while visit: | |
182 | n = visit.pop(0) |
|
182 | n = visit.pop(0) | |
183 | if n in known or n in self.map: continue |
|
183 | if n in known or n in self.map: continue | |
184 | known[n] = 1 |
|
184 | known[n] = 1 | |
185 | self.commitcache[n] = self.source.getcommit(n) |
|
185 | self.commitcache[n] = self.source.getcommit(n) | |
186 | cp = self.commitcache[n][0] |
|
186 | cp = self.commitcache[n][0] | |
187 | for p in cp: |
|
187 | for p in cp: | |
188 | parents.setdefault(n, []).append(p) |
|
188 | parents.setdefault(n, []).append(p) | |
189 | visit.append(p) |
|
189 | visit.append(p) | |
190 |
|
190 | |||
191 | return parents |
|
191 | return parents | |
192 |
|
192 | |||
193 | def toposort(self, parents): |
|
193 | def toposort(self, parents): | |
194 | visit = parents.keys() |
|
194 | visit = parents.keys() | |
195 | seen = {} |
|
195 | seen = {} | |
196 | children = {} |
|
196 | children = {} | |
197 |
|
197 | |||
198 | while visit: |
|
198 | while visit: | |
199 | n = visit.pop(0) |
|
199 | n = visit.pop(0) | |
200 | if n in seen: continue |
|
200 | if n in seen: continue | |
201 | seen[n] = 1 |
|
201 | seen[n] = 1 | |
202 | pc = 0 |
|
202 | pc = 0 | |
203 | if n in parents: |
|
203 | if n in parents: | |
204 | for p in parents[n]: |
|
204 | for p in parents[n]: | |
205 | if p not in self.map: pc += 1 |
|
205 | if p not in self.map: pc += 1 | |
206 | visit.append(p) |
|
206 | visit.append(p) | |
207 | children.setdefault(p, []).append(n) |
|
207 | children.setdefault(p, []).append(n) | |
208 | if not pc: root = n |
|
208 | if not pc: root = n | |
209 |
|
209 | |||
210 | s = [] |
|
210 | s = [] | |
211 | removed = {} |
|
211 | removed = {} | |
212 | visit = children.keys() |
|
212 | visit = children.keys() | |
213 | while visit: |
|
213 | while visit: | |
214 | n = visit.pop(0) |
|
214 | n = visit.pop(0) | |
215 | if n in removed: continue |
|
215 | if n in removed: continue | |
216 | dep = 0 |
|
216 | dep = 0 | |
217 | if n in parents: |
|
217 | if n in parents: | |
218 | for p in parents[n]: |
|
218 | for p in parents[n]: | |
219 | if p in self.map: continue |
|
219 | if p in self.map: continue | |
220 | if p not in removed: |
|
220 | if p not in removed: | |
221 | # we're still dependent |
|
221 | # we're still dependent | |
222 | visit.append(n) |
|
222 | visit.append(n) | |
223 | dep = 1 |
|
223 | dep = 1 | |
224 | break |
|
224 | break | |
225 |
|
225 | |||
226 | if not dep: |
|
226 | if not dep: | |
227 | # all n's parents are in the list |
|
227 | # all n's parents are in the list | |
228 | removed[n] = 1 |
|
228 | removed[n] = 1 | |
229 | s.append(n) |
|
229 | s.append(n) | |
230 | if n in children: |
|
230 | if n in children: | |
231 | for c in children[n]: |
|
231 | for c in children[n]: | |
232 | visit.insert(0, c) |
|
232 | visit.insert(0, c) | |
233 |
|
233 | |||
234 | return s |
|
234 | return s | |
235 |
|
235 | |||
236 | def copy(self, rev): |
|
236 | def copy(self, rev): | |
237 | p, a, d, t = self.commitcache[rev] |
|
237 | p, a, d, t = self.commitcache[rev] | |
238 | files = self.source.getchanges(rev) |
|
238 | files = self.source.getchanges(rev) | |
239 |
|
239 | |||
240 | for f,v,e in files: |
|
240 | for f,v,e in files: | |
241 | try: |
|
241 | try: | |
242 | data = self.source.getfile(f, v) |
|
242 | data = self.source.getfile(f, v) | |
243 | except IOError, inst: |
|
243 | except IOError, inst: | |
244 | self.dest.delfile(f) |
|
244 | self.dest.delfile(f) | |
245 | else: |
|
245 | else: | |
246 | self.dest.putfile(f, e, data) |
|
246 | self.dest.putfile(f, e, data) | |
247 |
|
247 | |||
248 | r = [self.map[v] for v in p] |
|
248 | r = [self.map[v] for v in p] | |
249 | f = [f for f,v,e in files] |
|
249 | f = [f for f,v,e in files] | |
250 | self.map[rev] = self.dest.putcommit(f, r, a, d, t) |
|
250 | self.map[rev] = self.dest.putcommit(f, r, a, d, t) | |
251 | file(self.mapfile, "a").write("%s %s\n" % (rev, self.map[rev])) |
|
251 | file(self.mapfile, "a").write("%s %s\n" % (rev, self.map[rev])) | |
252 |
|
252 | |||
253 | def convert(self): |
|
253 | def convert(self): | |
254 | heads = self.source.getheads() |
|
254 | heads = self.source.getheads() | |
255 | parents = self.walktree(heads) |
|
255 | parents = self.walktree(heads) | |
256 | t = self.toposort(parents) |
|
256 | t = self.toposort(parents) | |
|
257 | t = [n for n in t if n not in self.map] | |||
257 | num = len(t) |
|
258 | num = len(t) | |
258 |
|
259 | |||
259 | for c in t: |
|
260 | for c in t: | |
260 | num -= 1 |
|
261 | num -= 1 | |
261 | if c in self.map: continue |
|
|||
262 | desc = self.commitcache[c][3].splitlines()[0] |
|
262 | desc = self.commitcache[c][3].splitlines()[0] | |
263 | #print num, desc |
|
263 | #print num, desc | |
264 | self.copy(c) |
|
264 | self.copy(c) | |
265 |
|
265 | |||
266 | tags = self.source.gettags() |
|
266 | tags = self.source.gettags() | |
267 | ctags = {} |
|
267 | ctags = {} | |
268 | for k in tags: |
|
268 | for k in tags: | |
269 | v = tags[k] |
|
269 | v = tags[k] | |
270 | if v in self.map: |
|
270 | if v in self.map: | |
271 | ctags[k] = self.map[v] |
|
271 | ctags[k] = self.map[v] | |
272 |
|
272 | |||
273 | if ctags: |
|
273 | if ctags: | |
274 | nrev = self.dest.puttags(ctags) |
|
274 | nrev = self.dest.puttags(ctags) | |
275 | # write another hash correspondence to override the previous |
|
275 | # write another hash correspondence to override the previous | |
276 | # one so we don't end up with extra tag heads |
|
276 | # one so we don't end up with extra tag heads | |
277 | file(self.mapfile, "a").write("%s %s\n" % (c, nrev)) |
|
277 | file(self.mapfile, "a").write("%s %s\n" % (c, nrev)) | |
278 |
|
278 | |||
279 | gitpath, hgpath, mapfile = sys.argv[1:] |
|
279 | gitpath, hgpath, mapfile = sys.argv[1:] | |
280 | if os.path.isdir(gitpath + "/.git"): |
|
280 | if os.path.isdir(gitpath + "/.git"): | |
281 | gitpath += "/.git" |
|
281 | gitpath += "/.git" | |
282 |
|
282 | |||
283 | c = convert(convert_git(gitpath), convert_mercurial(hgpath), mapfile) |
|
283 | c = convert(convert_git(gitpath), convert_mercurial(hgpath), mapfile) | |
284 | c.convert() |
|
284 | c.convert() |
General Comments 0
You need to be logged in to leave comments.
Login now