convert-repo: add basic CVS import support
authorMatt Mackall <mpm@selenic.com>
Fri, 22 Dec 2006 17:59:34 -0600
changeset 3953 fad134931327
parent 3952 32c1653b7dad
child 3954 9af4b853ed4d
convert-repo: add basic CVS import support
contrib/convert-repo
--- a/contrib/convert-repo	Fri Dec 22 22:53:16 2006 +0100
+++ b/contrib/convert-repo	Fri Dec 22 17:59:34 2006 -0600
@@ -7,7 +7,7 @@
 #
 # convert-repo <source> [<dest> [<mapfile>]]
 #
-# Currently accepted source formats: git
+# Currently accepted source formats: git, cvs
 # Currently accepted destination formats: hg
 #
 # If destination isn't given, a new Mercurial repo named <src>-hg will
@@ -23,11 +23,12 @@
 # on each commit copied, so convert-repo can be interrupted and can
 # be run repeatedly to copy new commits.
 
-import sys, os, zlib, sha, time
+import sys, os, zlib, sha, time, re, locale
 os.environ["HGENCODING"] = "utf-8"
 from mercurial import hg, ui, util, fancyopts
 
 class Abort(Exception): pass
+class NoRepo(Exception): pass
 
 quiet = 0
 def status(msg):
@@ -48,13 +49,198 @@
         except:
             return s.decode("utf-8", "replace").encode("utf-8")
 
+# CVS conversion code inspired by hg-cvs-import and git-cvsimport
+class convert_cvs:
+    def __init__(self, path):
+        self.path = path
+        cvs = os.path.join(path, "CVS")
+        if not os.path.exists(cvs):
+            raise NoRepo("couldn't open CVS repo %s" % path)
+
+        self.changeset = {}
+        self.tags = {}
+        self.lastbranch = {}
+        self.parent = {}
+        self.socket = None
+        self.cvsroot = file(os.path.join(cvs, "Root")).read()[:-1]
+        self.cvsrepo = file(os.path.join(cvs, "Repository")).read()[:-1]
+        self.encoding = locale.getpreferredencoding()
+        self._parse()
+        self._connect()
+
+    def _parse(self):
+        if self.changeset:
+            return
+
+        d = os.getcwd()
+        try:
+            os.chdir(self.path)
+            id = None
+            state = 0
+            for l in os.popen("cvsps -A"):
+                if state == 0: # header
+                    if l.startswith("PatchSet"):
+                        id = l[9:-2]
+                    elif l.startswith("Date"):
+                        date = util.parsedate(l[6:-1], ["%Y/%m/%d %H:%M:%S"])
+                        date = util.datestr(date)
+                    elif l.startswith("Branch"):
+                        branch = l[8:-1]
+                        self.parent[id] = self.lastbranch.get(branch,'bad')
+                        self.lastbranch[branch] = id
+                    elif l.startswith("Ancestor branch"):
+                        ancestor = l[17:-1]
+                        self.parent[id] = self.lastbranch[ancestor]
+                    elif l.startswith("Author"):
+                        author = self.recode(l[8:-1])
+                    elif l.startswith("Tag: "):
+                        t = l[5:-1]
+                        if t != "(none) ":
+                            self.tags[t] = id
+                    elif l.startswith("Log:"):
+                        state = 1
+                        log = ""
+                elif state == 1: # log
+                    if l == "Members: \n":
+                        files = {}
+                        log = self.recode(log[:-1])
+                        if log.isspace():
+                            log = "*** empty log message ***\n"
+                        state = 2
+                    else:
+                        log += l
+                elif state == 2:
+                    if l == "\n": #
+                        state = 0
+                        self.changeset[id] = (date, author, log, files)
+                    else:
+                        file,rev = l[1:-2].rsplit(':',1)
+                        rev = rev.split("->")[1]
+                        files[file] = rev
+
+            self.heads = self.lastbranch.values()
+        finally:
+            os.chdir(d)
+
+    def _connect(self):
+        root = self.cvsroot
+        local = False
+        user, host = None, None
+        cmd = ['cvs', 'server']
+
+        status("connecting to %s\n" % root)
+
+        # only non-pserver for now
+        if root.startswith(":pserver"):
+            abort("can't handle pserver mode yet: %s\n" % root)
+
+        if root.startswith(":local:"):
+            local = True
+            root = root[7:]
+        else:
+            # :ext:user@host/home/user/path/to/cvsroot
+            if root.startswith(":ext:"):
+                root = root[5:]
+            m = re.match(r'(?:([^@:/]+)@)?([^:/]+):?(.*)', root)
+            if not m:
+                local = True
+            else:
+                local = False
+                user, host, root = m.group(1), m.group(2), m.group(3)
+
+        if not local:
+            rsh = os.environ.get("CVS_RSH" or "rsh")
+            if user:
+                cmd = [rsh, '-l', user, host] + cmd
+            else:
+                cmd = [rsh, host] + cmd
+
+        self.writep, self.readp = os.popen2(cmd)
+        self.realroot = root
+
+        self.writep.write("Root %s\n" % root)
+        self.writep.write("Valid-responses ok error Valid-requests Mode"
+                          " M Mbinary E Checked-in Created Updated"
+                          " Merged Removed\n")
+        self.writep.write("valid-requests\n")
+        self.writep.flush()
+        r = self.readp.readline()
+        if not r.startswith("Valid-requests"):
+            abort("server sucks\n")
+        if "UseUnchanged" in r:
+            self.writep.write("UseUnchanged\n")
+            self.writep.flush()
+            r = self.readp.readline()
+
+    def getheads(self):
+        return self.heads
+
+    def getfile(self, name, rev):
+        if rev.endswith("(DEAD)"):
+            raise IOError
+
+        args = ("-N -P -kk -r %s --" % rev).split()
+        args.append(os.path.join(self.cvsrepo, name))
+        for x in args:
+            self.writep.write("Argument %s\n" % x)
+        self.writep.write("Directory .\n%s\nco\n" % self.realroot)
+        self.writep.flush()
+
+        data = ""
+        while 1:
+            line = self.readp.readline()
+            if line.startswith("Created ") or line.startswith("Updated "):
+                self.readp.readline() # path
+                self.readp.readline() # entries
+                mode = self.readp.readline()[:-1]
+                count = int(self.readp.readline()[:-1])
+                data = self.readp.read(count)
+            elif line.startswith(" "):
+                data += line[1:]
+            elif line.startswith("M "):
+                pass
+            elif line.startswith("Mbinary "):
+                count = int(self.readp.readline()[:-1])
+                data = self.readp.read(count)
+            else:
+                if line == "ok\n":
+                    return data
+                elif line.startswith("E "):
+                    warn("cvs server: %s\n" % line[2:])
+                elif line.startswith("Remove"):
+                    l = self.readp.readline()
+                    l = self.readp.readline()
+                    if l != "ok\n":
+                        abort("unknown CVS response: %s\n" % l)
+                else:
+                    abort("unknown CVS response: %s\n" % line)
+
+    def getchanges(self, rev):
+        files = self.changeset[rev][3]
+        cl = [ (f, r, 0) for f,r in files.items() ]
+        cl.sort()
+        return cl
+
+    def recode(self, text):
+        return text.decode(self.encoding, "replace").encode("utf-8")
+
+    def getcommit(self, rev):
+        cs = self.changeset[rev]
+        parents = [self.parent[rev]]
+        if rev == "1":
+            parents = []
+        return (parents, cs[1], cs[0], cs[2])
+
+    def gettags(self):
+        return self.tags
+
 class convert_git:
     def __init__(self, path):
         if os.path.isdir(path + "/.git"):
             path += "/.git"
         self.path = path
         if not os.path.exists(path + "/HEAD"):
-            raise TypeError("couldn't open GIT repo %s" % path)
+            raise NoRepo("couldn't open GIT repo %s" % path)
 
     def getheads(self):
         fh = os.popen("GIT_DIR=%s git-rev-parse --verify HEAD" % self.path)
@@ -129,7 +315,7 @@
         try:
             self.repo = hg.repository(u, path)
         except:
-            raise TypeError("could open hg repo %s" % path)
+            raise NoRepo("could open hg repo %s" % path)
 
     def mapfile(self):
         return os.path.join(self.path, ".hg", "shamap")
@@ -168,8 +354,8 @@
         while parents:
             p1 = p2
             p2 = parents.pop(0)
-            self.repo.rawcommit(files, text, author, dest,
-                                hg.bin(p1), hg.bin(p2))
+            a = 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())
 
@@ -202,7 +388,7 @@
                                 date, self.repo.changelog.tip(), hg.nullid)
             return hg.hex(self.repo.changelog.tip())
 
-converters = [convert_git, convert_mercurial]
+converters = [convert_cvs, convert_git, convert_mercurial]
 
 def converter(path):
     if not os.path.isdir(path):
@@ -210,7 +396,7 @@
     for c in converters:
         try:
             return c(path)
-        except TypeError:
+        except NoRepo:
             pass
     abort("%s: unknown repository type\n" % path)
 
@@ -319,7 +505,9 @@
         status("converting...\n")
         for c in t:
             num -= 1
-            desc = self.commitcache[c][3].splitlines()[0]
+            desc = self.commitcache[c][3]
+            if "\n" in desc:
+                desc = desc.splitlines()[0]
             status("%d %s\n" % (num, desc))
             self.copy(c)