annotate contrib/convert-repo @ 3821:158fce02dc40

Teach convert-repo to deal with mixed charsets in git
author Matt Mackall <mpm@selenic.com>
date Thu, 07 Dec 2006 18:03:28 -0600
parents e6a7a6a33a62
children 4bc5a2405b12
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
1 #!/usr/bin/env python
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
2 #
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
3 # This is a generalized framework for converting between SCM
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
4 # repository formats.
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
5 #
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
6 # In its current form, it's hardcoded to convert incrementally between
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
7 # git and Mercurial.
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
8 #
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
9 # To use, you must first import the first git version into Mercurial,
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
10 # and establish a mapping between the git commit hash and the hash in
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
11 # Mercurial for that version. This mapping is kept in a simple text
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
12 # file with lines like so:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
13 #
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
14 # <git hash> <mercurial hash>
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
15 #
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
16 # To convert the rest of the repo, run:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
17 #
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
18 # convert-repo <git-dir> <hg-dir> <mapfile>
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
19 #
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
20 # This updates the mapfile on each commit copied, so it can be
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
21 # interrupted and can be run repeatedly to copy new commits.
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
22
694
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
23 import sys, os, zlib, sha, time
3821
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
24
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
25 os.environ["HGENCODING"] = "utf-8"
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
26
1715
40346aa66b0f Revert convert-repo changes
Matt Mackall <mpm@selenic.com>
parents: 1656
diff changeset
27 from mercurial import hg, ui, util
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
28
3821
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
29 def recode(s):
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
30 try:
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
31 return s.decode("utf-8").encode("utf-8")
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
32 except:
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
33 try:
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
34 return s.decode("latin-1").encode("utf-8")
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
35 except:
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
36 return s.decode("utf-8", "replace").encode("utf-8")
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
37
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
38 class convert_git:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
39 def __init__(self, path):
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
40 self.path = path
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
41
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
42 def getheads(self):
2657
e6a7a6a33a62 make convert-repo deal with git symbolic refs.
Alexis S. L. Carvalho <alexis@cecm.usp.br>
parents: 2093
diff changeset
43 fh = os.popen("GIT_DIR=%s git-rev-parse --verify HEAD" % self.path)
e6a7a6a33a62 make convert-repo deal with git symbolic refs.
Alexis S. L. Carvalho <alexis@cecm.usp.br>
parents: 2093
diff changeset
44 return [fh.read()[:-1]]
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
45
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
46 def catfile(self, rev, type):
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
47 if rev == "0" * 40: raise IOError()
1335
bea6356b8bca git -> hg conversion script
Florian La Roche <laroche@redhat.com>
parents: 1237
diff changeset
48 fh = os.popen("GIT_DIR=%s git-cat-file %s %s 2>/dev/null" % (self.path, type, rev))
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
49 return fh.read()
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
50
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
51 def getfile(self, name, rev):
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
52 return self.catfile(rev, "blob")
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
53
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
54 def getchanges(self, version):
1335
bea6356b8bca git -> hg conversion script
Florian La Roche <laroche@redhat.com>
parents: 1237
diff changeset
55 fh = os.popen("GIT_DIR=%s git-diff-tree --root -m -r %s" % (self.path, version))
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
56 changes = []
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
57 for l in fh:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
58 if "\t" not in l: continue
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
59 m, f = l[:-1].split("\t")
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
60 m = m.split()
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
61 h = m[3]
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
62 p = (m[1] == "100755")
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
63 changes.append((f, h, p))
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
64 return changes
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
65
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
66 def getcommit(self, version):
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
67 c = self.catfile(version, "commit") # read the commit hash
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
68 end = c.find("\n\n")
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
69 message = c[end+2:]
3821
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
70 message = recode(message)
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
71 l = c[:end].splitlines()
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
72 manifest = l[0].split()[1]
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
73 parents = []
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
74 for e in l[1:]:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
75 n,v = e.split(" ", 1)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
76 if n == "author":
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
77 p = v.split()
1385
adb3de56635b convert-repo: Fix timezone handling
Matt Mackall <mpm@selenic.com>
parents: 1335
diff changeset
78 tm, tz = p[-2:]
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
79 author = " ".join(p[:-2])
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
80 if author[0] == "<": author = author[1:-1]
3821
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
81 author = recode(author)
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
82 if n == "committer":
431
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
83 p = v.split()
1385
adb3de56635b convert-repo: Fix timezone handling
Matt Mackall <mpm@selenic.com>
parents: 1335
diff changeset
84 tm, tz = p[-2:]
431
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
85 committer = " ".join(p[:-2])
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
86 if committer[0] == "<": committer = committer[1:-1]
3821
158fce02dc40 Teach convert-repo to deal with mixed charsets in git
Matt Mackall <mpm@selenic.com>
parents: 2657
diff changeset
87 committer = recode(committer)
1385
adb3de56635b convert-repo: Fix timezone handling
Matt Mackall <mpm@selenic.com>
parents: 1335
diff changeset
88 message += "\ncommitter: %s\n" % v
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
89 if n == "parent": parents.append(v)
1385
adb3de56635b convert-repo: Fix timezone handling
Matt Mackall <mpm@selenic.com>
parents: 1335
diff changeset
90
adb3de56635b convert-repo: Fix timezone handling
Matt Mackall <mpm@selenic.com>
parents: 1335
diff changeset
91 tzs, tzh, tzm = tz[-5:-4] + "1", tz[-4:-2], tz[-2:]
2093
5cc414722587 convert-repo: fix reversed time zone offset
Vadim Gelfer <vadim.gelfer@gmail.com>
parents: 1715
diff changeset
92 tz = -int(tzs) * (int(tzh) * 3600 + int(tzm))
1385
adb3de56635b convert-repo: Fix timezone handling
Matt Mackall <mpm@selenic.com>
parents: 1335
diff changeset
93 date = tm + " " + str(tz)
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
94 return (parents, author, date, message)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
95
694
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
96 def gettags(self):
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
97 tags = {}
1335
bea6356b8bca git -> hg conversion script
Florian La Roche <laroche@redhat.com>
parents: 1237
diff changeset
98 for f in os.listdir(self.path + "/refs/tags"):
694
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
99 try:
1335
bea6356b8bca git -> hg conversion script
Florian La Roche <laroche@redhat.com>
parents: 1237
diff changeset
100 h = file(self.path + "/refs/tags/" + f).read().strip()
1386
a1040345fdda convert-repo: retrieve the commit hash from the tag object for tag import
Matt Mackall <mpm@selenic.com>
parents: 1385
diff changeset
101 c = self.catfile(h, "tag") # read the commit hash
a1040345fdda convert-repo: retrieve the commit hash from the tag object for tag import
Matt Mackall <mpm@selenic.com>
parents: 1385
diff changeset
102 h = c.splitlines()[0].split()[1]
1237
227cfbe34109 Fix off by one in convert-repo tags
mason@suse.com
parents: 705
diff changeset
103 tags[f] = h
694
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
104 except:
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
105 pass
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
106 return tags
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
107
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
108 class convert_mercurial:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
109 def __init__(self, path):
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
110 self.path = path
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
111 u = ui.ui()
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
112 self.repo = hg.repository(u, path)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
113
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
114 def getheads(self):
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
115 h = self.repo.changelog.heads()
1335
bea6356b8bca git -> hg conversion script
Florian La Roche <laroche@redhat.com>
parents: 1237
diff changeset
116 return [ hg.hex(x) for x in h ]
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
117
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
118 def putfile(self, f, e, data):
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
119 self.repo.wfile(f, "w").write(data)
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
120 if self.repo.dirstate.state(f) == '?':
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
121 self.repo.dirstate.update([f], "a")
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
122
450
9d785fd7deec Get set_exec from util in convert_repo
mpm@selenic.com
parents: 431
diff changeset
123 util.set_exec(self.repo.wjoin(f), e)
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
124
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
125 def delfile(self, f):
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
126 try:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
127 os.unlink(self.repo.wjoin(f))
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
128 #self.repo.remove([f])
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
129 except:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
130 pass
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
131
1715
40346aa66b0f Revert convert-repo changes
Matt Mackall <mpm@selenic.com>
parents: 1656
diff changeset
132 def putcommit(self, files, parents, author, dest, text):
431
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
133 seen = {}
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
134 pl = []
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
135 for p in parents:
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
136 if p not in seen:
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
137 pl.append(p)
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
138 seen[p] = 1
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
139 parents = pl
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
140
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
141 if len(parents) < 2: parents.append("0" * 40)
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
142 if len(parents) < 2: parents.append("0" * 40)
431
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
143 p2 = parents.pop(0)
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
144
431
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
145 while parents:
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
146 p1 = p2
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
147 p2 = parents.pop(0)
1715
40346aa66b0f Revert convert-repo changes
Matt Mackall <mpm@selenic.com>
parents: 1656
diff changeset
148 self.repo.rawcommit(files, text, author, dest,
40346aa66b0f Revert convert-repo changes
Matt Mackall <mpm@selenic.com>
parents: 1656
diff changeset
149 hg.bin(p1), hg.bin(p2))
431
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
150 text = "(octopus merge fixup)\n"
1389
9b3ef6f3cef5 convert-repo: fix up octopus merge conversion
Matt Mackall <mpm@selenic.com>
parents: 1388
diff changeset
151 p2 = hg.hex(self.repo.changelog.tip())
431
dfc44f3f587c convert-repo fixups
mpm@selenic.com
parents: 316
diff changeset
152
1389
9b3ef6f3cef5 convert-repo: fix up octopus merge conversion
Matt Mackall <mpm@selenic.com>
parents: 1388
diff changeset
153 return p2
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
154
694
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
155 def puttags(self, tags):
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
156 try:
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
157 old = self.repo.wfile(".hgtags").read()
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
158 oldlines = old.splitlines(1)
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
159 oldlines.sort()
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
160 except:
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
161 oldlines = []
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
162
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
163 k = tags.keys()
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
164 k.sort()
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
165 newlines = []
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
166 for tag in k:
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
167 newlines.append("%s %s\n" % (tags[tag], tag))
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
168
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
169 newlines.sort()
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
170
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
171 if newlines != oldlines:
1335
bea6356b8bca git -> hg conversion script
Florian La Roche <laroche@redhat.com>
parents: 1237
diff changeset
172 #print "updating tags"
694
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
173 f = self.repo.wfile(".hgtags", "w")
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
174 f.write("".join(newlines))
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
175 f.close()
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
176 if not oldlines: self.repo.add([".hgtags"])
1335
bea6356b8bca git -> hg conversion script
Florian La Roche <laroche@redhat.com>
parents: 1237
diff changeset
177 date = "%s 0" % int(time.mktime(time.gmtime()))
694
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
178 self.repo.rawcommit([".hgtags"], "update tags", "convert-repo",
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
179 date, self.repo.changelog.tip(), hg.nullid)
1387
0c7e8d345564 convert-repo: linearize the tag commit
Matt Mackall <mpm@selenic.com>
parents: 1386
diff changeset
180 return hg.hex(self.repo.changelog.tip())
694
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
181
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
182 class convert:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
183 def __init__(self, source, dest, mapfile):
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
184 self.source = source
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
185 self.dest = dest
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
186 self.mapfile = mapfile
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
187 self.commitcache = {}
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
188
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
189 self.map = {}
1655
7bfd4724932a convert-repo: automatically create empty map file
Matt Mackall <mpm@selenic.com>
parents: 1389
diff changeset
190 try:
7bfd4724932a convert-repo: automatically create empty map file
Matt Mackall <mpm@selenic.com>
parents: 1389
diff changeset
191 for l in file(self.mapfile):
7bfd4724932a convert-repo: automatically create empty map file
Matt Mackall <mpm@selenic.com>
parents: 1389
diff changeset
192 sv, dv = l[:-1].split()
7bfd4724932a convert-repo: automatically create empty map file
Matt Mackall <mpm@selenic.com>
parents: 1389
diff changeset
193 self.map[sv] = dv
7bfd4724932a convert-repo: automatically create empty map file
Matt Mackall <mpm@selenic.com>
parents: 1389
diff changeset
194 except IOError:
7bfd4724932a convert-repo: automatically create empty map file
Matt Mackall <mpm@selenic.com>
parents: 1389
diff changeset
195 pass
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
196
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
197 def walktree(self, heads):
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
198 visit = heads
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
199 known = {}
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
200 parents = {}
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
201 while visit:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
202 n = visit.pop(0)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
203 if n in known or n in self.map: continue
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
204 known[n] = 1
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
205 self.commitcache[n] = self.source.getcommit(n)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
206 cp = self.commitcache[n][0]
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
207 for p in cp:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
208 parents.setdefault(n, []).append(p)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
209 visit.append(p)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
210
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
211 return parents
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
212
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
213 def toposort(self, parents):
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
214 visit = parents.keys()
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
215 seen = {}
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
216 children = {}
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
217
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
218 while visit:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
219 n = visit.pop(0)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
220 if n in seen: continue
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
221 seen[n] = 1
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
222 pc = 0
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
223 if n in parents:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
224 for p in parents[n]:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
225 if p not in self.map: pc += 1
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
226 visit.append(p)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
227 children.setdefault(p, []).append(n)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
228 if not pc: root = n
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
229
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
230 s = []
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
231 removed = {}
692
695dd9a491da convert-repo: deal with packed git and other fixes
mpm@selenic.com
parents: 450
diff changeset
232 visit = children.keys()
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
233 while visit:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
234 n = visit.pop(0)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
235 if n in removed: continue
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
236 dep = 0
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
237 if n in parents:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
238 for p in parents[n]:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
239 if p in self.map: continue
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
240 if p not in removed:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
241 # we're still dependent
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
242 visit.append(n)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
243 dep = 1
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
244 break
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
245
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
246 if not dep:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
247 # all n's parents are in the list
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
248 removed[n] = 1
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
249 s.append(n)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
250 if n in children:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
251 for c in children[n]:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
252 visit.insert(0, c)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
253
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
254 return s
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
255
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
256 def copy(self, rev):
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
257 p, a, d, t = self.commitcache[rev]
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
258 files = self.source.getchanges(rev)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
259
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
260 for f,v,e in files:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
261 try:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
262 data = self.source.getfile(f, v)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
263 except IOError, inst:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
264 self.dest.delfile(f)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
265 else:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
266 self.dest.putfile(f, e, data)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
267
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
268 r = [self.map[v] for v in p]
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
269 f = [f for f,v,e in files]
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
270 self.map[rev] = self.dest.putcommit(f, r, a, d, t)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
271 file(self.mapfile, "a").write("%s %s\n" % (rev, self.map[rev]))
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
272
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
273 def convert(self):
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
274 heads = self.source.getheads()
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
275 parents = self.walktree(heads)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
276 t = self.toposort(parents)
1388
5eb2d3c54165 convert-repo: change duplicate elimination
Matt Mackall <mpm@selenic.com>
parents: 1387
diff changeset
277 t = [n for n in t if n not in self.map]
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
278 num = len(t)
1715
40346aa66b0f Revert convert-repo changes
Matt Mackall <mpm@selenic.com>
parents: 1656
diff changeset
279 c = None
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
280
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
281 for c in t:
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
282 num -= 1
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
283 desc = self.commitcache[c][3].splitlines()[0]
1335
bea6356b8bca git -> hg conversion script
Florian La Roche <laroche@redhat.com>
parents: 1237
diff changeset
284 #print num, desc
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
285 self.copy(c)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
286
694
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
287 tags = self.source.gettags()
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
288 ctags = {}
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
289 for k in tags:
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
290 v = tags[k]
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
291 if v in self.map:
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
292 ctags[k] = self.map[v]
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
293
1715
40346aa66b0f Revert convert-repo changes
Matt Mackall <mpm@selenic.com>
parents: 1656
diff changeset
294 if c and ctags:
1387
0c7e8d345564 convert-repo: linearize the tag commit
Matt Mackall <mpm@selenic.com>
parents: 1386
diff changeset
295 nrev = self.dest.puttags(ctags)
0c7e8d345564 convert-repo: linearize the tag commit
Matt Mackall <mpm@selenic.com>
parents: 1386
diff changeset
296 # write another hash correspondence to override the previous
0c7e8d345564 convert-repo: linearize the tag commit
Matt Mackall <mpm@selenic.com>
parents: 1386
diff changeset
297 # one so we don't end up with extra tag heads
0c7e8d345564 convert-repo: linearize the tag commit
Matt Mackall <mpm@selenic.com>
parents: 1386
diff changeset
298 file(self.mapfile, "a").write("%s %s\n" % (c, nrev))
694
51eb248d3348 Teach convert-repo about tags
mpm@selenic.com
parents: 692
diff changeset
299
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
300 gitpath, hgpath, mapfile = sys.argv[1:]
1335
bea6356b8bca git -> hg conversion script
Florian La Roche <laroche@redhat.com>
parents: 1237
diff changeset
301 if os.path.isdir(gitpath + "/.git"):
bea6356b8bca git -> hg conversion script
Florian La Roche <laroche@redhat.com>
parents: 1237
diff changeset
302 gitpath += "/.git"
316
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
303
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
304 c = convert(convert_git(gitpath), convert_mercurial(hgpath), mapfile)
c48d069163d6 Add new convert-repo script
mpm@selenic.com
parents:
diff changeset
305 c.convert()