summaryrefslogtreecommitdiff
path: root/contrib/remote-helpers/git-remote-bzr
diff options
context:
space:
mode:
Diffstat (limited to 'contrib/remote-helpers/git-remote-bzr')
-rwxr-xr-xcontrib/remote-helpers/git-remote-bzr986
1 files changed, 8 insertions, 978 deletions
diff --git a/contrib/remote-helpers/git-remote-bzr b/contrib/remote-helpers/git-remote-bzr
index 9abb58e6ab..712a1377e2 100755
--- a/contrib/remote-helpers/git-remote-bzr
+++ b/contrib/remote-helpers/git-remote-bzr
@@ -1,983 +1,13 @@
#!/usr/bin/env python
-#
-# Copyright (c) 2012 Felipe Contreras
-#
-
-#
-# Just copy to your ~/bin, or anywhere in your $PATH.
-# Then you can clone with:
-# % git clone bzr::/path/to/bzr/repo/or/url
-#
-# For example:
-# % git clone bzr::$HOME/myrepo
-# or
-# % git clone bzr::lp:myrepo
-#
-# If you want to specify which branches you want to track (per repo):
-# % git config remote.origin.bzr-branches 'trunk, devel, test'
-#
-# Where 'origin' is the name of the repository you want to specify the
-# branches.
-#
-
-import sys
-
-import bzrlib
-if hasattr(bzrlib, "initialize"):
- bzrlib.initialize()
-
-import bzrlib.plugin
-bzrlib.plugin.load_plugins()
-
-import bzrlib.generate_ids
-import bzrlib.transport
-import bzrlib.errors
-import bzrlib.ui
-import bzrlib.urlutils
-import bzrlib.branch
import sys
-import os
-import json
-import re
-import StringIO
-import atexit, shutil, hashlib, urlparse, subprocess
-
-NAME_RE = re.compile('^([^<>]+)')
-AUTHOR_RE = re.compile('^([^<>]+?)? ?[<>]([^<>]*)(?:$|>)')
-EMAIL_RE = re.compile(r'([^ \t<>]+@[^ \t<>]+)')
-RAW_AUTHOR_RE = re.compile('^(\w+) (.+)? <(.*)> (\d+) ([+-]\d+)')
-
-def die(msg, *args):
- sys.stderr.write('ERROR: %s\n' % (msg % args))
- sys.exit(1)
-
-def warn(msg, *args):
- sys.stderr.write('WARNING: %s\n' % (msg % args))
-
-def gittz(tz):
- return '%+03d%02d' % (tz / 3600, tz % 3600 / 60)
-
-def get_config(config):
- cmd = ['git', 'config', '--get', config]
- process = subprocess.Popen(cmd, stdout=subprocess.PIPE)
- output, _ = process.communicate()
- return output
-
-class Marks:
-
- def __init__(self, path):
- self.path = path
- self.tips = {}
- self.marks = {}
- self.rev_marks = {}
- self.last_mark = 0
- self.load()
-
- def load(self):
- if not os.path.exists(self.path):
- return
-
- tmp = json.load(open(self.path))
- self.tips = tmp['tips']
- self.marks = tmp['marks']
- self.last_mark = tmp['last-mark']
-
- for rev, mark in self.marks.iteritems():
- self.rev_marks[mark] = rev
-
- def dict(self):
- return { 'tips': self.tips, 'marks': self.marks, 'last-mark' : self.last_mark }
-
- def store(self):
- json.dump(self.dict(), open(self.path, 'w'))
-
- def __str__(self):
- return str(self.dict())
-
- def from_rev(self, rev):
- return self.marks[rev]
-
- def to_rev(self, mark):
- return str(self.rev_marks[mark])
-
- def next_mark(self):
- self.last_mark += 1
- return self.last_mark
-
- def get_mark(self, rev):
- self.last_mark += 1
- self.marks[rev] = self.last_mark
- return self.last_mark
-
- def is_marked(self, rev):
- return rev in self.marks
-
- def new_mark(self, rev, mark):
- self.marks[rev] = mark
- self.rev_marks[mark] = rev
- self.last_mark = mark
-
- def get_tip(self, branch):
- try:
- return str(self.tips[branch])
- except KeyError:
- return None
-
- def set_tip(self, branch, tip):
- self.tips[branch] = tip
-
-class Parser:
-
- def __init__(self, repo):
- self.repo = repo
- self.line = self.get_line()
-
- def get_line(self):
- return sys.stdin.readline().strip()
-
- def __getitem__(self, i):
- return self.line.split()[i]
-
- def check(self, word):
- return self.line.startswith(word)
-
- def each_block(self, separator):
- while self.line != separator:
- yield self.line
- self.line = self.get_line()
-
- def __iter__(self):
- return self.each_block('')
-
- def next(self):
- self.line = self.get_line()
- if self.line == 'done':
- self.line = None
-
- def get_mark(self):
- i = self.line.index(':') + 1
- return int(self.line[i:])
-
- def get_data(self):
- if not self.check('data'):
- return None
- i = self.line.index(' ') + 1
- size = int(self.line[i:])
- return sys.stdin.read(size)
-
- def get_author(self):
- m = RAW_AUTHOR_RE.match(self.line)
- if not m:
- return None
- _, name, email, date, tz = m.groups()
- name = name.decode('utf-8')
- committer = '%s <%s>' % (name, email)
- tz = int(tz)
- tz = ((tz / 100) * 3600) + ((tz % 100) * 60)
- return (committer, int(date), tz)
-
-def rev_to_mark(rev):
- return marks.from_rev(rev)
-
-def mark_to_rev(mark):
- return marks.to_rev(mark)
-
-def fixup_user(user):
- name = mail = None
- user = user.replace('"', '')
- m = AUTHOR_RE.match(user)
- if m:
- name = m.group(1)
- mail = m.group(2).strip()
- else:
- m = EMAIL_RE.match(user)
- if m:
- mail = m.group(1)
- else:
- m = NAME_RE.match(user)
- if m:
- name = m.group(1).strip()
-
- if not name:
- name = 'unknown'
- if not mail:
- mail = 'Unknown'
-
- return '%s <%s>' % (name, mail)
-
-def get_filechanges(cur, prev):
- modified = {}
- removed = {}
-
- changes = cur.changes_from(prev)
-
- def u(s):
- return s.encode('utf-8')
-
- for path, fid, kind in changes.added:
- modified[u(path)] = fid
- for path, fid, kind in changes.removed:
- removed[u(path)] = None
- for path, fid, kind, mod, _ in changes.modified:
- modified[u(path)] = fid
- for oldpath, newpath, fid, kind, mod, _ in changes.renamed:
- removed[u(oldpath)] = None
- if kind == 'directory':
- lst = cur.list_files(from_dir=newpath, recursive=True)
- for path, file_class, kind, fid, entry in lst:
- if kind != 'directory':
- modified[u(newpath + '/' + path)] = fid
- else:
- modified[u(newpath)] = fid
-
- return modified, removed
-
-def export_files(tree, files):
- final = []
- for path, fid in files.iteritems():
- kind = tree.kind(fid)
-
- h = tree.get_file_sha1(fid)
-
- if kind == 'symlink':
- d = tree.get_symlink_target(fid)
- mode = '120000'
- elif kind == 'file':
-
- if tree.is_executable(fid):
- mode = '100755'
- else:
- mode = '100644'
-
- # is the blob already exported?
- if h in filenodes:
- mark = filenodes[h]
- final.append((mode, mark, path))
- continue
-
- d = tree.get_file_text(fid)
- elif kind == 'directory':
- continue
- else:
- die("Unhandled kind '%s' for path '%s'" % (kind, path))
-
- mark = marks.next_mark()
- filenodes[h] = mark
-
- print "blob"
- print "mark :%u" % mark
- print "data %d" % len(d)
- print d
-
- final.append((mode, mark, path))
-
- return final
-
-def export_branch(repo, name):
- ref = '%s/heads/%s' % (prefix, name)
- tip = marks.get_tip(name)
-
- branch = get_remote_branch(name)
- repo = branch.repository
-
- branch.lock_read()
- revs = branch.iter_merge_sorted_revisions(None, tip, 'exclude', 'forward')
- try:
- tip_revno = branch.revision_id_to_revno(tip)
- last_revno, _ = branch.last_revision_info()
- total = last_revno - tip_revno
- except bzrlib.errors.NoSuchRevision:
- tip_revno = 0
- total = 0
-
- for revid, _, seq, _ in revs:
-
- if marks.is_marked(revid):
- continue
-
- rev = repo.get_revision(revid)
- revno = seq[0]
-
- parents = rev.parent_ids
- time = rev.timestamp
- tz = rev.timezone
- committer = rev.committer.encode('utf-8')
- committer = "%s %u %s" % (fixup_user(committer), time, gittz(tz))
- authors = rev.get_apparent_authors()
- if authors:
- author = authors[0].encode('utf-8')
- author = "%s %u %s" % (fixup_user(author), time, gittz(tz))
- else:
- author = committer
- msg = rev.message.encode('utf-8')
-
- msg += '\n'
-
- if len(parents) == 0:
- parent = bzrlib.revision.NULL_REVISION
- else:
- parent = parents[0]
-
- cur_tree = repo.revision_tree(revid)
- prev = repo.revision_tree(parent)
- modified, removed = get_filechanges(cur_tree, prev)
-
- modified_final = export_files(cur_tree, modified)
-
- if len(parents) == 0:
- print 'reset %s' % ref
-
- print "commit %s" % ref
- print "mark :%d" % (marks.get_mark(revid))
- print "author %s" % (author)
- print "committer %s" % (committer)
- print "data %d" % (len(msg))
- print msg
-
- for i, p in enumerate(parents):
- try:
- m = rev_to_mark(p)
- except KeyError:
- # ghost?
- continue
- if i == 0:
- print "from :%s" % m
- else:
- print "merge :%s" % m
-
- for f in removed:
- print "D %s" % (f,)
- for f in modified_final:
- print "M %s :%u %s" % f
- print
-
- if len(seq) > 1:
- # let's skip branch revisions from the progress report
- continue
-
- progress = (revno - tip_revno)
- if (progress % 100 == 0):
- if total:
- print "progress revision %d '%s' (%d/%d)" % (revno, name, progress, total)
- else:
- print "progress revision %d '%s' (%d)" % (revno, name, progress)
-
- branch.unlock()
-
- revid = branch.last_revision()
-
- # make sure the ref is updated
- print "reset %s" % ref
- print "from :%u" % rev_to_mark(revid)
- print
-
- marks.set_tip(name, revid)
-
-def export_tag(repo, name):
- ref = '%s/tags/%s' % (prefix, name)
- print "reset %s" % ref
- print "from :%u" % rev_to_mark(tags[name])
- print
-
-def do_import(parser):
- repo = parser.repo
- path = os.path.join(dirname, 'marks-git')
-
- print "feature done"
- if os.path.exists(path):
- print "feature import-marks=%s" % path
- print "feature export-marks=%s" % path
- print "feature force"
- sys.stdout.flush()
-
- while parser.check('import'):
- ref = parser[1]
- if ref.startswith('refs/heads/'):
- name = ref[len('refs/heads/'):]
- export_branch(repo, name)
- if ref.startswith('refs/tags/'):
- name = ref[len('refs/tags/'):]
- export_tag(repo, name)
- parser.next()
-
- print 'done'
-
- sys.stdout.flush()
-
-def parse_blob(parser):
- parser.next()
- mark = parser.get_mark()
- parser.next()
- data = parser.get_data()
- blob_marks[mark] = data
- parser.next()
-
-class CustomTree():
-
- def __init__(self, branch, revid, parents, files):
- self.updates = {}
- self.branch = branch
-
- def copy_tree(revid):
- files = files_cache[revid] = {}
- branch.lock_read()
- tree = branch.repository.revision_tree(revid)
- try:
- for path, entry in tree.iter_entries_by_dir():
- files[path] = [entry.file_id, None]
- finally:
- branch.unlock()
- return files
-
- if len(parents) == 0:
- self.base_id = bzrlib.revision.NULL_REVISION
- self.base_files = {}
- else:
- self.base_id = parents[0]
- self.base_files = files_cache.get(self.base_id, None)
- if not self.base_files:
- self.base_files = copy_tree(self.base_id)
-
- self.files = files_cache[revid] = self.base_files.copy()
- self.rev_files = {}
-
- for path, data in self.files.iteritems():
- fid, mark = data
- self.rev_files[fid] = [path, mark]
-
- for path, f in files.iteritems():
- fid, mark = self.files.get(path, [None, None])
- if not fid:
- fid = bzrlib.generate_ids.gen_file_id(path)
- f['path'] = path
- self.rev_files[fid] = [path, mark]
- self.updates[fid] = f
-
- def last_revision(self):
- return self.base_id
-
- def iter_changes(self):
- changes = []
-
- def get_parent(dirname, basename):
- parent_fid, mark = self.base_files.get(dirname, [None, None])
- if parent_fid:
- return parent_fid
- parent_fid, mark = self.files.get(dirname, [None, None])
- if parent_fid:
- return parent_fid
- if basename == '':
- return None
- fid = bzrlib.generate_ids.gen_file_id(path)
- add_entry(fid, dirname, 'directory')
- return fid
-
- def add_entry(fid, path, kind, mode=None):
- dirname, basename = os.path.split(path)
- parent_fid = get_parent(dirname, basename)
-
- executable = False
- if mode == '100755':
- executable = True
- elif mode == '120000':
- kind = 'symlink'
-
- change = (fid,
- (None, path),
- True,
- (False, True),
- (None, parent_fid),
- (None, basename),
- (None, kind),
- (None, executable))
- self.files[path] = [change[0], None]
- changes.append(change)
-
- def update_entry(fid, path, kind, mode=None):
- dirname, basename = os.path.split(path)
- parent_fid = get_parent(dirname, basename)
-
- executable = False
- if mode == '100755':
- executable = True
- elif mode == '120000':
- kind = 'symlink'
-
- change = (fid,
- (path, path),
- True,
- (True, True),
- (None, parent_fid),
- (None, basename),
- (None, kind),
- (None, executable))
- self.files[path] = [change[0], None]
- changes.append(change)
-
- def remove_entry(fid, path, kind):
- dirname, basename = os.path.split(path)
- parent_fid = get_parent(dirname, basename)
- change = (fid,
- (path, None),
- True,
- (True, False),
- (parent_fid, None),
- (None, None),
- (None, None),
- (None, None))
- del self.files[path]
- changes.append(change)
-
- for fid, f in self.updates.iteritems():
- path = f['path']
-
- if 'deleted' in f:
- remove_entry(fid, path, 'file')
- continue
-
- if path in self.base_files:
- update_entry(fid, path, 'file', f['mode'])
- else:
- add_entry(fid, path, 'file', f['mode'])
-
- self.files[path][1] = f['mark']
- self.rev_files[fid][1] = f['mark']
-
- return changes
-
- def get_content(self, file_id):
- path, mark = self.rev_files[file_id]
- if mark:
- return blob_marks[mark]
-
- # last resort
- tree = self.branch.repository.revision_tree(self.base_id)
- return tree.get_file_text(file_id)
-
- def get_file_with_stat(self, file_id, path=None):
- content = self.get_content(file_id)
- return (StringIO.StringIO(content), None)
-
- def get_symlink_target(self, file_id):
- return self.get_content(file_id)
-
- def id2path(self, file_id):
- path, mark = self.rev_files[file_id]
- return path
-
-def c_style_unescape(string):
- if string[0] == string[-1] == '"':
- return string.decode('string-escape')[1:-1]
- return string
-
-def parse_commit(parser):
- parents = []
-
- ref = parser[1]
- parser.next()
-
- if ref.startswith('refs/heads/'):
- name = ref[len('refs/heads/'):]
- branch = get_remote_branch(name)
- else:
- die('unknown ref')
-
- commit_mark = parser.get_mark()
- parser.next()
- author = parser.get_author()
- parser.next()
- committer = parser.get_author()
- parser.next()
- data = parser.get_data()
- parser.next()
- if parser.check('from'):
- parents.append(parser.get_mark())
- parser.next()
- while parser.check('merge'):
- parents.append(parser.get_mark())
- parser.next()
-
- # fast-export adds an extra newline
- if data[-1] == '\n':
- data = data[:-1]
-
- files = {}
-
- for line in parser:
- if parser.check('M'):
- t, m, mark_ref, path = line.split(' ', 3)
- mark = int(mark_ref[1:])
- f = { 'mode' : m, 'mark' : mark }
- elif parser.check('D'):
- t, path = line.split(' ', 1)
- f = { 'deleted' : True }
- else:
- die('Unknown file command: %s' % line)
- path = c_style_unescape(path).decode('utf-8')
- files[path] = f
-
- committer, date, tz = committer
- author, _, _ = author
- parents = [mark_to_rev(p) for p in parents]
- revid = bzrlib.generate_ids.gen_revision_id(committer, date)
- props = {}
- props['branch-nick'] = branch.nick
- props['authors'] = author
-
- mtree = CustomTree(branch, revid, parents, files)
- changes = mtree.iter_changes()
-
- branch.lock_write()
- try:
- builder = branch.get_commit_builder(parents, None, date, tz, committer, props, revid)
- try:
- list(builder.record_iter_changes(mtree, mtree.last_revision(), changes))
- builder.finish_inventory()
- builder.commit(data.decode('utf-8', 'replace'))
- except Exception, e:
- builder.abort()
- raise
- finally:
- branch.unlock()
-
- parsed_refs[ref] = revid
- marks.new_mark(revid, commit_mark)
-
-def parse_reset(parser):
- ref = parser[1]
- parser.next()
-
- # ugh
- if parser.check('commit'):
- parse_commit(parser)
- return
- if not parser.check('from'):
- return
- from_mark = parser.get_mark()
- parser.next()
-
- parsed_refs[ref] = mark_to_rev(from_mark)
-
-def do_export(parser):
- parser.next()
-
- for line in parser.each_block('done'):
- if parser.check('blob'):
- parse_blob(parser)
- elif parser.check('commit'):
- parse_commit(parser)
- elif parser.check('reset'):
- parse_reset(parser)
- elif parser.check('tag'):
- pass
- elif parser.check('feature'):
- pass
- else:
- die('unhandled export command: %s' % line)
-
- for ref, revid in parsed_refs.iteritems():
- if ref.startswith('refs/heads/'):
- name = ref[len('refs/heads/'):]
- branch = get_remote_branch(name)
- branch.generate_revision_history(revid, marks.get_tip(name))
-
- if name in peers:
- peer = bzrlib.branch.Branch.open(peers[name],
- possible_transports=transports)
- try:
- peer.bzrdir.push_branch(branch, revision_id=revid,
- overwrite=force)
- except bzrlib.errors.DivergedBranches:
- print "error %s non-fast forward" % ref
- continue
-
- try:
- wt = branch.bzrdir.open_workingtree()
- wt.update()
- except bzrlib.errors.NoWorkingTree:
- pass
- elif ref.startswith('refs/tags/'):
- # TODO: implement tag push
- print "error %s pushing tags not supported" % ref
- continue
- else:
- # transport-helper/fast-export bugs
- continue
-
- print "ok %s" % ref
-
- print
-
-def do_capabilities(parser):
- print "import"
- print "export"
- print "refspec refs/heads/*:%s/heads/*" % prefix
- print "refspec refs/tags/*:%s/tags/*" % prefix
-
- path = os.path.join(dirname, 'marks-git')
-
- if os.path.exists(path):
- print "*import-marks %s" % path
- print "*export-marks %s" % path
-
- print "option"
- print
-
-class InvalidOptionValue(Exception):
- pass
-
-def get_bool_option(val):
- if val == 'true':
- return True
- elif val == 'false':
- return False
- else:
- raise InvalidOptionValue()
-
-def do_option(parser):
- global force
- opt, val = parser[1:3]
- try:
- if opt == 'force':
- force = get_bool_option(val)
- print 'ok'
- else:
- print 'unsupported'
- except InvalidOptionValue:
- print "error '%s' is not a valid value for option '%s'" % (val, opt)
-
-def ref_is_valid(name):
- return not True in [c in name for c in '~^: \\']
-
-def do_list(parser):
- master_branch = None
-
- for name in branches:
- if not master_branch:
- master_branch = name
- print "? refs/heads/%s" % name
-
- branch = get_remote_branch(master_branch)
- branch.lock_read()
- for tag, revid in branch.tags.get_tag_dict().items():
- try:
- branch.revision_id_to_dotted_revno(revid)
- except bzrlib.errors.NoSuchRevision:
- continue
- if not ref_is_valid(tag):
- continue
- print "? refs/tags/%s" % tag
- tags[tag] = revid
- branch.unlock()
-
- print "@refs/heads/%s HEAD" % master_branch
- print
-
-def clone(path, remote_branch):
- try:
- bdir = bzrlib.bzrdir.BzrDir.create(path, possible_transports=transports)
- except bzrlib.errors.AlreadyControlDirError:
- bdir = bzrlib.bzrdir.BzrDir.open(path, possible_transports=transports)
- repo = bdir.find_repository()
- repo.fetch(remote_branch.repository)
- return remote_branch.sprout(bdir, repository=repo)
-
-def get_remote_branch(name):
- remote_branch = bzrlib.branch.Branch.open(branches[name],
- possible_transports=transports)
- if isinstance(remote_branch.bzrdir.root_transport, bzrlib.transport.local.LocalTransport):
- return remote_branch
-
- branch_path = os.path.join(dirname, 'clone', name)
-
- try:
- branch = bzrlib.branch.Branch.open(branch_path,
- possible_transports=transports)
- except bzrlib.errors.NotBranchError:
- # clone
- branch = clone(branch_path, remote_branch)
- else:
- # pull
- try:
- branch.pull(remote_branch, overwrite=True)
- except bzrlib.errors.DivergedBranches:
- # use remote branch for now
- return remote_branch
-
- return branch
-
-def find_branches(repo):
- transport = repo.bzrdir.root_transport
-
- for fn in transport.iter_files_recursive():
- if not fn.endswith('.bzr/branch-format'):
- continue
-
- name = subdir = fn[:-len('/.bzr/branch-format')]
- name = name if name != '' else 'master'
- name = name.replace('/', '+')
-
- try:
- cur = transport.clone(subdir)
- branch = bzrlib.branch.Branch.open_from_transport(cur)
- except bzrlib.errors.NotBranchError:
- continue
- else:
- yield name, branch.base
-
-def get_repo(url, alias):
- normal_url = bzrlib.urlutils.normalize_url(url)
- origin = bzrlib.bzrdir.BzrDir.open(url, possible_transports=transports)
- is_local = isinstance(origin.transport, bzrlib.transport.local.LocalTransport)
-
- shared_path = os.path.join(gitdir, 'bzr')
- try:
- shared_dir = bzrlib.bzrdir.BzrDir.open(shared_path,
- possible_transports=transports)
- except bzrlib.errors.NotBranchError:
- shared_dir = bzrlib.bzrdir.BzrDir.create(shared_path,
- possible_transports=transports)
- try:
- shared_repo = shared_dir.open_repository()
- except bzrlib.errors.NoRepositoryPresent:
- shared_repo = shared_dir.create_repository(shared=True)
-
- if not is_local:
- clone_path = os.path.join(dirname, 'clone')
- if not os.path.exists(clone_path):
- os.mkdir(clone_path)
- else:
- # check and remove old organization
- try:
- bdir = bzrlib.bzrdir.BzrDir.open(clone_path,
- possible_transports=transports)
- bdir.destroy_repository()
- except bzrlib.errors.NotBranchError:
- pass
- except bzrlib.errors.NoRepositoryPresent:
- pass
-
- wanted = get_config('remote.%s.bzr-branches' % alias).rstrip().split(', ')
- # stupid python
- wanted = [e for e in wanted if e]
- if not wanted:
- wanted = get_config('remote-bzr.branches').rstrip().split(', ')
- # stupid python
- wanted = [e for e in wanted if e]
-
- if not wanted:
- try:
- repo = origin.open_repository()
- if not repo.bzrdir.root_transport.listable():
- # this repository is not usable for us
- raise bzrlib.errors.NoRepositoryPresent(repo.bzrdir)
- except bzrlib.errors.NoRepositoryPresent:
- wanted = ['master']
-
- if wanted:
- def list_wanted(url, wanted):
- for name in wanted:
- subdir = name if name != 'master' else ''
- yield name, bzrlib.urlutils.join(url, subdir)
-
- branch_list = list_wanted(url, wanted)
- else:
- branch_list = find_branches(repo)
-
- for name, url in branch_list:
- if not is_local:
- peers[name] = url
- branches[name] = url
-
- return origin
-
-def fix_path(alias, orig_url):
- url = urlparse.urlparse(orig_url, 'file')
- if url.scheme != 'file' or os.path.isabs(url.path):
- return
- abs_url = urlparse.urljoin("%s/" % os.getcwd(), orig_url)
- cmd = ['git', 'config', 'remote.%s.url' % alias, "bzr::%s" % abs_url]
- subprocess.call(cmd)
-
-def main(args):
- global marks, prefix, gitdir, dirname
- global tags, filenodes
- global blob_marks
- global parsed_refs
- global files_cache
- global is_tmp
- global branches, peers
- global transports
- global force
-
- marks = None
- is_tmp = False
- gitdir = os.environ.get('GIT_DIR', None)
-
- if len(args) < 3:
- die('Not enough arguments.')
-
- if not gitdir:
- die('GIT_DIR not set')
-
- alias = args[1]
- url = args[2]
-
- tags = {}
- filenodes = {}
- blob_marks = {}
- parsed_refs = {}
- files_cache = {}
- branches = {}
- peers = {}
- transports = []
- force = False
-
- if alias[5:] == url:
- is_tmp = True
- alias = hashlib.sha1(alias).hexdigest()
-
- prefix = 'refs/bzr/%s' % alias
- dirname = os.path.join(gitdir, 'bzr', alias)
-
- if not is_tmp:
- fix_path(alias, url)
-
- if not os.path.exists(dirname):
- os.makedirs(dirname)
-
- if hasattr(bzrlib.ui.ui_factory, 'be_quiet'):
- bzrlib.ui.ui_factory.be_quiet(True)
-
- repo = get_repo(url, alias)
-
- marks_path = os.path.join(dirname, 'marks-int')
- marks = Marks(marks_path)
-
- parser = Parser(repo)
- for line in parser:
- if parser.check('capabilities'):
- do_capabilities(parser)
- elif parser.check('list'):
- do_list(parser)
- elif parser.check('import'):
- do_import(parser)
- elif parser.check('export'):
- do_export(parser)
- elif parser.check('option'):
- do_option(parser)
- else:
- die('unhandled command: %s' % line)
- sys.stdout.flush()
-def bye():
- if not marks:
- return
- if not is_tmp:
- marks.store()
- else:
- shutil.rmtree(dirname)
+sys.stderr.write('WARNING: git-remote-bzr is now maintained independently.\n')
+sys.stderr.write('WARNING: For more information visit https://github.com/felipec/git-remote-bzr\n')
-atexit.register(bye)
-sys.exit(main(sys.argv))
+sys.stderr.write('''WARNING:
+WARNING: You can pick a directory on your $PATH and download it, e.g.:
+WARNING: $ wget -O $HOME/bin/git-remote-bzr \\
+WARNING: https://raw.github.com/felipec/git-remote-bzr/master/git-remote-bzr
+WARNING: $ chmod +x $HOME/bin/git-remote-bzr
+''')