Teach convert-repo to deal with mixed charsets in git

contrib/convert-repo

@@ -1,290 +1,305 b''
 #!/usr/bin/env python
 #
 # This is a generalized framework for converting between SCM
 # repository formats.
 #
 # In its current form, it's hardcoded to convert incrementally between
 # git and Mercurial.
 #
 # To use, you must first import the first git version into Mercurial,
 # and establish a mapping between the git commit hash and the hash in
 # Mercurial for that version. This mapping is kept in a simple text
 # file with lines like so:
 #
 # <git hash> <mercurial hash>
 #
 # To convert the rest of the repo, run:
 #
 # convert-repo <git-dir> <hg-dir> <mapfile>
 #
 # This updates the mapfile on each commit copied, so it can be
 # interrupted and can be run repeatedly to copy new commits.
 
 import sys, os, zlib, sha, time
+
+os.environ["HGENCODING"] = "utf-8"
+
 from mercurial import hg, ui, util
 
+def recode(s):
+    try:
+        return s.decode("utf-8").encode("utf-8")
+    except:
+        try:
+            return s.decode("latin-1").encode("utf-8")
+        except:
+            return s.decode("utf-8", "replace").encode("utf-8")
+
 class convert_git:
     def __init__(self, path):
         self.path = path
 
     def getheads(self):
         fh = os.popen("GIT_DIR=%s git-rev-parse --verify HEAD" % self.path)
         return [fh.read()[:-1]]
 
     def catfile(self, rev, type):
         if rev == "0" * 40: raise IOError()
         fh = os.popen("GIT_DIR=%s git-cat-file %s %s 2>/dev/null" % (self.path, type, rev))
         return fh.read()
 
     def getfile(self, name, rev):
         return self.catfile(rev, "blob")
 
     def getchanges(self, version):
         fh = os.popen("GIT_DIR=%s git-diff-tree --root -m -r %s" % (self.path, version))
         changes = []
         for l in fh:
             if "\t" not in l: continue
             m, f = l[:-1].split("\t")
             m = m.split()
             h = m[3]
             p = (m[1] == "100755")
             changes.append((f, h, p))
         return changes
 
     def getcommit(self, version):
         c = self.catfile(version, "commit") # read the commit hash
         end = c.find("\n\n")
         message = c[end+2:]
+        message = recode(message)
         l = c[:end].splitlines()
         manifest = l[0].split()[1]
         parents = []
         for e in l[1:]:
             n,v = e.split(" ", 1)
             if n == "author":
                 p = v.split()
                 tm, tz = p[-2:]
                 author = " ".join(p[:-2])
                 if author[0] == "<": author = author[1:-1]
+                author = recode(author)
             if n == "committer":
                 p = v.split()
                 tm, tz = p[-2:]
                 committer = " ".join(p[:-2])
                 if committer[0] == "<": committer = committer[1:-1]
+                committer = recode(committer)
                 message += "\ncommitter: %s\n" % v
             if n == "parent": parents.append(v)
 
         tzs, tzh, tzm = tz[-5:-4] + "1", tz[-4:-2], tz[-2:]
         tz = -int(tzs) * (int(tzh) * 3600 + int(tzm))
         date = tm + " " + str(tz)
         return (parents, author, date, message)
 
     def gettags(self):
         tags = {}
         for f in os.listdir(self.path + "/refs/tags"):
             try:
                 h = file(self.path + "/refs/tags/" + f).read().strip()
                 c = self.catfile(h, "tag") # read the commit hash
                 h = c.splitlines()[0].split()[1]
                 tags[f] = h
             except:
                 pass
         return tags
 
 class convert_mercurial:
     def __init__(self, path):
         self.path = path
         u = ui.ui()
         self.repo = hg.repository(u, path)
 
     def getheads(self):
         h = self.repo.changelog.heads()
         return [ hg.hex(x) for x in h ]
 
     def putfile(self, f, e, data):
         self.repo.wfile(f, "w").write(data)
         if self.repo.dirstate.state(f) == '?':
             self.repo.dirstate.update([f], "a")
 
         util.set_exec(self.repo.wjoin(f), e)
 
     def delfile(self, f):
         try:
             os.unlink(self.repo.wjoin(f))
             #self.repo.remove([f])
         except:
             pass
 
     def putcommit(self, files, parents, author, dest, text):
         seen = {}
         pl = []
         for p in parents:
             if p not in seen:
                 pl.append(p)
                 seen[p] = 1
         parents = pl
 
         if len(parents) < 2: parents.append("0" * 40)
         if len(parents) < 2: parents.append("0" * 40)
         p2 = parents.pop(0)
 
         while parents:
             p1 = p2
             p2 = parents.pop(0)
             self.repo.rawcommit(files, text, author, dest,
                                 hg.bin(p1), hg.bin(p2))
             text = "(octopus merge fixup)\n"
             p2 = hg.hex(self.repo.changelog.tip())
 
         return p2
 
     def puttags(self, tags):
         try:
             old = self.repo.wfile(".hgtags").read()
             oldlines = old.splitlines(1)
             oldlines.sort()
         except:
             oldlines = []
 
         k = tags.keys()
         k.sort()
         newlines = []
         for tag in k:
             newlines.append("%s %s\n" % (tags[tag], tag))
 
         newlines.sort()
 
         if newlines != oldlines:
             #print "updating tags"
             f = self.repo.wfile(".hgtags", "w")
             f.write("".join(newlines))
             f.close()
             if not oldlines: self.repo.add([".hgtags"])
             date = "%s 0" % int(time.mktime(time.gmtime()))
             self.repo.rawcommit([".hgtags"], "update tags", "convert-repo",
                                 date, self.repo.changelog.tip(), hg.nullid)
             return hg.hex(self.repo.changelog.tip())
 
 class convert:
     def __init__(self, source, dest, mapfile):
         self.source = source
         self.dest = dest
         self.mapfile = mapfile
         self.commitcache = {}
 
         self.map = {}
         try:
             for l in file(self.mapfile):
                 sv, dv = l[:-1].split()
                 self.map[sv] = dv
         except IOError:
             pass
 
     def walktree(self, heads):
         visit = heads
         known = {}
         parents = {}
         while visit:
             n = visit.pop(0)
             if n in known or n in self.map: continue
             known[n] = 1
             self.commitcache[n] = self.source.getcommit(n)
             cp = self.commitcache[n][0]
             for p in cp:
                 parents.setdefault(n, []).append(p)
                 visit.append(p)
 
         return parents
 
     def toposort(self, parents):
         visit = parents.keys()
         seen = {}
         children = {}
 
         while visit:
             n = visit.pop(0)
             if n in seen: continue
             seen[n] = 1
             pc = 0
             if n in parents:
                 for p in parents[n]:
                     if p not in self.map: pc += 1
                     visit.append(p)
                     children.setdefault(p, []).append(n)
             if not pc: root = n
 
         s = []
         removed = {}
         visit = children.keys()
         while visit:
             n = visit.pop(0)
             if n in removed: continue
             dep = 0
             if n in parents:
                 for p in parents[n]:
                     if p in self.map: continue
                     if p not in removed:
                         # we're still dependent
                         visit.append(n)
                         dep = 1
                         break
 
             if not dep:
                 # all n's parents are in the list
                 removed[n] = 1
                 s.append(n)
                 if n in children:
                     for c in children[n]:
                         visit.insert(0, c)
 
         return s
 
     def copy(self, rev):
         p, a, d, t = self.commitcache[rev]
         files = self.source.getchanges(rev)
 
         for f,v,e in files:
             try:
                 data = self.source.getfile(f, v)
             except IOError, inst:
                 self.dest.delfile(f)
             else:
                 self.dest.putfile(f, e, data)
 
         r = [self.map[v] for v in p]
         f = [f for f,v,e in files]
         self.map[rev] = self.dest.putcommit(f, r, a, d, t)
         file(self.mapfile, "a").write("%s %s\n" % (rev, self.map[rev]))
 
     def convert(self):
         heads = self.source.getheads()
         parents = self.walktree(heads)
         t = self.toposort(parents)
         t = [n for n in t if n not in self.map]
         num = len(t)
         c = None
 
         for c in t:
             num -= 1
             desc = self.commitcache[c][3].splitlines()[0]
             #print num, desc
             self.copy(c)
 
         tags = self.source.gettags()
         ctags = {}
         for k in tags:
             v = tags[k]
             if v in self.map:
                 ctags[k] = self.map[v]
 
         if c and ctags:
             nrev = self.dest.puttags(ctags)
             # write another hash correspondence to override the previous
             # one so we don't end up with extra tag heads
             file(self.mapfile, "a").write("%s %s\n" % (c, nrev))
 
 gitpath, hgpath, mapfile = sys.argv[1:]
 if os.path.isdir(gitpath + "/.git"):
     gitpath += "/.git"
 
 c = convert(convert_git(gitpath), convert_mercurial(hgpath), mapfile)
 c.convert()