streamclone: include obsstore file into stream bundle if client can read it

mercurial/bundle2.py

@@ -1,2315 +1,2322 b''
 # bundle2.py - generic container format to transmit arbitrary data.
 #
 # Copyright 2013 Facebook, Inc.
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 """Handling of the new bundle2 format
 
 The goal of bundle2 is to act as an atomically packet to transmit a set of
 payloads in an application agnostic way. It consist in a sequence of "parts"
 that will be handed to and processed by the application layer.
 
 
 General format architecture
 ===========================
 
 The format is architectured as follow
 
  - magic string
  - stream level parameters
  - payload parts (any number)
  - end of stream marker.
 
 the Binary format
 ============================
 
 All numbers are unsigned and big-endian.
 
 stream level parameters
 ------------------------
 
 Binary format is as follow
 
 :params size: int32
 
   The total number of Bytes used by the parameters
 
 :params value: arbitrary number of Bytes
 
   A blob of `params size` containing the serialized version of all stream level
   parameters.
 
   The blob contains a space separated list of parameters. Parameters with value
   are stored in the form `<name>=<value>`. Both name and value are urlquoted.
 
   Empty name are obviously forbidden.
 
   Name MUST start with a letter. If this first letter is lower case, the
   parameter is advisory and can be safely ignored. However when the first
   letter is capital, the parameter is mandatory and the bundling process MUST
   stop if he is not able to proceed it.
 
   Stream parameters use a simple textual format for two main reasons:
 
   - Stream level parameters should remain simple and we want to discourage any
     crazy usage.
   - Textual data allow easy human inspection of a bundle2 header in case of
     troubles.
 
   Any Applicative level options MUST go into a bundle2 part instead.
 
 Payload part
 ------------------------
 
 Binary format is as follow
 
 :header size: int32
 
   The total number of Bytes used by the part header. When the header is empty
   (size = 0) this is interpreted as the end of stream marker.
 
 :header:
 
     The header defines how to interpret the part. It contains two piece of
     data: the part type, and the part parameters.
 
     The part type is used to route an application level handler, that can
     interpret payload.
 
     Part parameters are passed to the application level handler.  They are
     meant to convey information that will help the application level object to
     interpret the part payload.
 
     The binary format of the header is has follow
 
     :typesize: (one byte)
 
     :parttype: alphanumerical part name (restricted to [a-zA-Z0-9_:-]*)
 
     :partid: A 32bits integer (unique in the bundle) that can be used to refer
              to this part.
 
     :parameters:
 
         Part's parameter may have arbitrary content, the binary structure is::
 
             <mandatory-count><advisory-count><param-sizes><param-data>
 
         :mandatory-count: 1 byte, number of mandatory parameters
 
         :advisory-count:  1 byte, number of advisory parameters
 
         :param-sizes:
 
             N couple of bytes, where N is the total number of parameters. Each
             couple contains (<size-of-key>, <size-of-value) for one parameter.
 
         :param-data:
 
             A blob of bytes from which each parameter key and value can be
             retrieved using the list of size couples stored in the previous
             field.
 
             Mandatory parameters comes first, then the advisory ones.
 
             Each parameter's key MUST be unique within the part.
 
 :payload:
 
     payload is a series of `<chunksize><chunkdata>`.
 
     `chunksize` is an int32, `chunkdata` are plain bytes (as much as
     `chunksize` says)` The payload part is concluded by a zero size chunk.
 
     The current implementation always produces either zero or one chunk.
     This is an implementation limitation that will ultimately be lifted.
 
     `chunksize` can be negative to trigger special case processing. No such
     processing is in place yet.
 
 Bundle processing
 ============================
 
 Each part is processed in order using a "part handler". Handler are registered
 for a certain part type.
 
 The matching of a part to its handler is case insensitive. The case of the
 part type is used to know if a part is mandatory or advisory. If the Part type
 contains any uppercase char it is considered mandatory. When no handler is
 known for a Mandatory part, the process is aborted and an exception is raised.
 If the part is advisory and no handler is known, the part is ignored. When the
 process is aborted, the full bundle is still read from the stream to keep the
 channel usable. But none of the part read from an abort are processed. In the
 future, dropping the stream may become an option for channel we do not care to
 preserve.
 """
 
 from __future__ import absolute_import, division
 
 import collections
 import errno
 import os
 import re
 import string
 import struct
 import sys
 
 from .i18n import _
 from . import (
     bookmarks,
     changegroup,
     encoding,
     error,
     node as nodemod,
     obsolete,
     phases,
     pushkey,
     pycompat,
     streamclone,
     tags,
     url,
     util,
 )
 from .utils import (
     stringutil,
 )
 
 urlerr = util.urlerr
 urlreq = util.urlreq
 
 _pack = struct.pack
 _unpack = struct.unpack
 
 _fstreamparamsize = '>i'
 _fpartheadersize = '>i'
 _fparttypesize = '>B'
 _fpartid = '>I'
 _fpayloadsize = '>i'
 _fpartparamcount = '>BB'
 
 preferedchunksize = 32768
 
 _parttypeforbidden = re.compile('[^a-zA-Z0-9_:-]')
 
 def outdebug(ui, message):
     """debug regarding output stream (bundling)"""
     if ui.configbool('devel', 'bundle2.debug'):
         ui.debug('bundle2-output: %s\n' % message)
 
 def indebug(ui, message):
     """debug on input stream (unbundling)"""
     if ui.configbool('devel', 'bundle2.debug'):
         ui.debug('bundle2-input: %s\n' % message)
 
 def validateparttype(parttype):
     """raise ValueError if a parttype contains invalid character"""
     if _parttypeforbidden.search(parttype):
         raise ValueError(parttype)
 
 def _makefpartparamsizes(nbparams):
     """return a struct format to read part parameter sizes
 
     The number parameters is variable so we need to build that format
     dynamically.
     """
     return '>'+('BB'*nbparams)
 
 parthandlermapping = {}
 
 def parthandler(parttype, params=()):
     """decorator that register a function as a bundle2 part handler
 
     eg::
 
         @parthandler('myparttype', ('mandatory', 'param', 'handled'))
         def myparttypehandler(...):
             '''process a part of type "my part".'''
             ...
     """
     validateparttype(parttype)
     def _decorator(func):
         lparttype = parttype.lower() # enforce lower case matching.
         assert lparttype not in parthandlermapping
         parthandlermapping[lparttype] = func
         func.params = frozenset(params)
         return func
     return _decorator
 
 class unbundlerecords(object):
     """keep record of what happens during and unbundle
 
     New records are added using `records.add('cat', obj)`. Where 'cat' is a
     category of record and obj is an arbitrary object.
 
     `records['cat']` will return all entries of this category 'cat'.
 
     Iterating on the object itself will yield `('category', obj)` tuples
     for all entries.
 
     All iterations happens in chronological order.
     """
 
     def __init__(self):
         self._categories = {}
         self._sequences = []
         self._replies = {}
 
     def add(self, category, entry, inreplyto=None):
         """add a new record of a given category.
 
         The entry can then be retrieved in the list returned by
         self['category']."""
         self._categories.setdefault(category, []).append(entry)
         self._sequences.append((category, entry))
         if inreplyto is not None:
             self.getreplies(inreplyto).add(category, entry)
 
     def getreplies(self, partid):
         """get the records that are replies to a specific part"""
         return self._replies.setdefault(partid, unbundlerecords())
 
     def __getitem__(self, cat):
         return tuple(self._categories.get(cat, ()))
 
     def __iter__(self):
         return iter(self._sequences)
 
     def __len__(self):
         return len(self._sequences)
 
     def __nonzero__(self):
         return bool(self._sequences)
 
     __bool__ = __nonzero__
 
 class bundleoperation(object):
     """an object that represents a single bundling process
 
     Its purpose is to carry unbundle-related objects and states.
 
     A new object should be created at the beginning of each bundle processing.
     The object is to be returned by the processing function.
 
     The object has very little content now it will ultimately contain:
     * an access to the repo the bundle is applied to,
     * a ui object,
     * a way to retrieve a transaction to add changes to the repo,
     * a way to record the result of processing each part,
     * a way to construct a bundle response when applicable.
     """
 
     def __init__(self, repo, transactiongetter, captureoutput=True, source=''):
         self.repo = repo
         self.ui = repo.ui
         self.records = unbundlerecords()
         self.reply = None
         self.captureoutput = captureoutput
         self.hookargs = {}
         self._gettransaction = transactiongetter
         # carries value that can modify part behavior
         self.modes = {}
         self.source = source
 
     def gettransaction(self):
         transaction = self._gettransaction()
 
         if self.hookargs:
             # the ones added to the transaction supercede those added
             # to the operation.
             self.hookargs.update(transaction.hookargs)
             transaction.hookargs = self.hookargs
 
         # mark the hookargs as flushed.  further attempts to add to
         # hookargs will result in an abort.
         self.hookargs = None
 
         return transaction
 
     def addhookargs(self, hookargs):
         if self.hookargs is None:
             raise error.ProgrammingError('attempted to add hookargs to '
                                          'operation after transaction started')
         self.hookargs.update(hookargs)
 
 class TransactionUnavailable(RuntimeError):
     pass
 
 def _notransaction():
     """default method to get a transaction while processing a bundle
 
     Raise an exception to highlight the fact that no transaction was expected
     to be created"""
     raise TransactionUnavailable()
 
 def applybundle(repo, unbundler, tr, source, url=None, **kwargs):
     # transform me into unbundler.apply() as soon as the freeze is lifted
     if isinstance(unbundler, unbundle20):
         tr.hookargs['bundle2'] = '1'
         if source is not None and 'source' not in tr.hookargs:
             tr.hookargs['source'] = source
         if url is not None and 'url' not in tr.hookargs:
             tr.hookargs['url'] = url
         return processbundle(repo, unbundler, lambda: tr, source=source)
     else:
         # the transactiongetter won't be used, but we might as well set it
         op = bundleoperation(repo, lambda: tr, source=source)
         _processchangegroup(op, unbundler, tr, source, url, **kwargs)
         return op
 
 class partiterator(object):
     def __init__(self, repo, op, unbundler):
         self.repo = repo
         self.op = op
         self.unbundler = unbundler
         self.iterator = None
         self.count = 0
         self.current = None
 
     def __enter__(self):
         def func():
             itr = enumerate(self.unbundler.iterparts())
             for count, p in itr:
                 self.count = count
                 self.current = p
                 yield p
                 p.consume()
                 self.current = None
         self.iterator = func()
         return self.iterator
 
     def __exit__(self, type, exc, tb):
         if not self.iterator:
             return
 
         # Only gracefully abort in a normal exception situation. User aborts
         # like Ctrl+C throw a KeyboardInterrupt which is not a base Exception,
         # and should not gracefully cleanup.
         if isinstance(exc, Exception):
             # Any exceptions seeking to the end of the bundle at this point are
             # almost certainly related to the underlying stream being bad.
             # And, chances are that the exception we're handling is related to
             # getting in that bad state. So, we swallow the seeking error and
             # re-raise the original error.
             seekerror = False
             try:
                 if self.current:
                     # consume the part content to not corrupt the stream.
                     self.current.consume()
 
                 for part in self.iterator:
                     # consume the bundle content
                     part.consume()
             except Exception:
                 seekerror = True
 
             # Small hack to let caller code distinguish exceptions from bundle2
             # processing from processing the old format. This is mostly needed
             # to handle different return codes to unbundle according to the type
             # of bundle. We should probably clean up or drop this return code
             # craziness in a future version.
             exc.duringunbundle2 = True
             salvaged = []
             replycaps = None
             if self.op.reply is not None:
                 salvaged = self.op.reply.salvageoutput()
                 replycaps = self.op.reply.capabilities
             exc._replycaps = replycaps
             exc._bundle2salvagedoutput = salvaged
 
             # Re-raising from a variable loses the original stack. So only use
             # that form if we need to.
             if seekerror:
                 raise exc
 
         self.repo.ui.debug('bundle2-input-bundle: %i parts total\n' %
                            self.count)
 
 def processbundle(repo, unbundler, transactiongetter=None, op=None, source=''):
     """This function process a bundle, apply effect to/from a repo
 
     It iterates over each part then searches for and uses the proper handling
     code to process the part. Parts are processed in order.
 
     Unknown Mandatory part will abort the process.
 
     It is temporarily possible to provide a prebuilt bundleoperation to the
     function. This is used to ensure output is properly propagated in case of
     an error during the unbundling. This output capturing part will likely be
     reworked and this ability will probably go away in the process.
     """
     if op is None:
         if transactiongetter is None:
             transactiongetter = _notransaction
         op = bundleoperation(repo, transactiongetter, source=source)
     # todo:
     # - replace this is a init function soon.
     # - exception catching
     unbundler.params
     if repo.ui.debugflag:
         msg = ['bundle2-input-bundle:']
         if unbundler.params:
             msg.append(' %i params' % len(unbundler.params))
         if op._gettransaction is None or op._gettransaction is _notransaction:
             msg.append(' no-transaction')
         else:
             msg.append(' with-transaction')
         msg.append('\n')
         repo.ui.debug(''.join(msg))
 
     processparts(repo, op, unbundler)
 
     return op
 
 def processparts(repo, op, unbundler):
     with partiterator(repo, op, unbundler) as parts:
         for part in parts:
             _processpart(op, part)
 
 def _processchangegroup(op, cg, tr, source, url, **kwargs):
     ret = cg.apply(op.repo, tr, source, url, **kwargs)
     op.records.add('changegroup', {
         'return': ret,
     })
     return ret
 
 def _gethandler(op, part):
     status = 'unknown' # used by debug output
     try:
         handler = parthandlermapping.get(part.type)
         if handler is None:
             status = 'unsupported-type'
             raise error.BundleUnknownFeatureError(parttype=part.type)
         indebug(op.ui, 'found a handler for part %s' % part.type)
         unknownparams = part.mandatorykeys - handler.params
         if unknownparams:
             unknownparams = list(unknownparams)
             unknownparams.sort()
             status = 'unsupported-params (%s)' % ', '.join(unknownparams)
             raise error.BundleUnknownFeatureError(parttype=part.type,
                                                   params=unknownparams)
         status = 'supported'
     except error.BundleUnknownFeatureError as exc:
         if part.mandatory: # mandatory parts
             raise
         indebug(op.ui, 'ignoring unsupported advisory part %s' % exc)
         return # skip to part processing
     finally:
         if op.ui.debugflag:
             msg = ['bundle2-input-part: "%s"' % part.type]
             if not part.mandatory:
                 msg.append(' (advisory)')
             nbmp = len(part.mandatorykeys)
             nbap = len(part.params) - nbmp
             if nbmp or nbap:
                 msg.append(' (params:')
                 if nbmp:
                     msg.append(' %i mandatory' % nbmp)
                 if nbap:
                     msg.append(' %i advisory' % nbmp)
                 msg.append(')')
             msg.append(' %s\n' % status)
             op.ui.debug(''.join(msg))
 
     return handler
 
 def _processpart(op, part):
     """process a single part from a bundle
 
     The part is guaranteed to have been fully consumed when the function exits
     (even if an exception is raised)."""
     handler = _gethandler(op, part)
     if handler is None:
         return
 
     # handler is called outside the above try block so that we don't
     # risk catching KeyErrors from anything other than the
     # parthandlermapping lookup (any KeyError raised by handler()
     # itself represents a defect of a different variety).
     output = None
     if op.captureoutput and op.reply is not None:
         op.ui.pushbuffer(error=True, subproc=True)
         output = ''
     try:
         handler(op, part)
     finally:
         if output is not None:
             output = op.ui.popbuffer()
         if output:
             outpart = op.reply.newpart('output', data=output,
                                        mandatory=False)
             outpart.addparam(
                 'in-reply-to', pycompat.bytestr(part.id), mandatory=False)
 
 def decodecaps(blob):
     """decode a bundle2 caps bytes blob into a dictionary
 
     The blob is a list of capabilities (one per line)
     Capabilities may have values using a line of the form::
 
         capability=value1,value2,value3
 
     The values are always a list."""
     caps = {}
     for line in blob.splitlines():
         if not line:
             continue
         if '=' not in line:
             key, vals = line, ()
         else:
             key, vals = line.split('=', 1)
             vals = vals.split(',')
         key = urlreq.unquote(key)
         vals = [urlreq.unquote(v) for v in vals]
         caps[key] = vals
     return caps
 
 def encodecaps(caps):
     """encode a bundle2 caps dictionary into a bytes blob"""
     chunks = []
     for ca in sorted(caps):
         vals = caps[ca]
         ca = urlreq.quote(ca)
         vals = [urlreq.quote(v) for v in vals]
         if vals:
             ca = "%s=%s" % (ca, ','.join(vals))
         chunks.append(ca)
     return '\n'.join(chunks)
 
 bundletypes = {
     "": ("", 'UN'),       # only when using unbundle on ssh and old http servers
                           # since the unification ssh accepts a header but there
                           # is no capability signaling it.
     "HG20": (), # special-cased below
     "HG10UN": ("HG10UN", 'UN'),
     "HG10BZ": ("HG10", 'BZ'),
     "HG10GZ": ("HG10GZ", 'GZ'),
 }
 
 # hgweb uses this list to communicate its preferred type
 bundlepriority = ['HG10GZ', 'HG10BZ', 'HG10UN']
 
 class bundle20(object):
     """represent an outgoing bundle2 container
 
     Use the `addparam` method to add stream level parameter. and `newpart` to
     populate it. Then call `getchunks` to retrieve all the binary chunks of
     data that compose the bundle2 container."""
 
     _magicstring = 'HG20'
 
     def __init__(self, ui, capabilities=()):
         self.ui = ui
         self._params = []
         self._parts = []
         self.capabilities = dict(capabilities)
         self._compengine = util.compengines.forbundletype('UN')
         self._compopts = None
         # If compression is being handled by a consumer of the raw
         # data (e.g. the wire protocol), unsetting this flag tells
         # consumers that the bundle is best left uncompressed.
         self.prefercompressed = True
 
     def setcompression(self, alg, compopts=None):
         """setup core part compression to <alg>"""
         if alg in (None, 'UN'):
             return
         assert not any(n.lower() == 'compression' for n, v in self._params)
         self.addparam('Compression', alg)
         self._compengine = util.compengines.forbundletype(alg)
         self._compopts = compopts
 
     @property
     def nbparts(self):
         """total number of parts added to the bundler"""
         return len(self._parts)
 
     # methods used to defines the bundle2 content
     def addparam(self, name, value=None):
         """add a stream level parameter"""
         if not name:
             raise error.ProgrammingError(b'empty parameter name')
         if name[0:1] not in pycompat.bytestr(string.ascii_letters):
             raise error.ProgrammingError(b'non letter first character: %s'
                                          % name)
         self._params.append((name, value))
 
     def addpart(self, part):
         """add a new part to the bundle2 container
 
         Parts contains the actual applicative payload."""
         assert part.id is None
         part.id = len(self._parts) # very cheap counter
         self._parts.append(part)
 
     def newpart(self, typeid, *args, **kwargs):
         """create a new part and add it to the containers
 
         As the part is directly added to the containers. For now, this means
         that any failure to properly initialize the part after calling
         ``newpart`` should result in a failure of the whole bundling process.
 
         You can still fall back to manually create and add if you need better
         control."""
         part = bundlepart(typeid, *args, **kwargs)
         self.addpart(part)
         return part
 
     # methods used to generate the bundle2 stream
     def getchunks(self):
         if self.ui.debugflag:
             msg = ['bundle2-output-bundle: "%s",' % self._magicstring]
             if self._params:
                 msg.append(' (%i params)' % len(self._params))
             msg.append(' %i parts total\n' % len(self._parts))
             self.ui.debug(''.join(msg))
         outdebug(self.ui, 'start emission of %s stream' % self._magicstring)
         yield self._magicstring
         param = self._paramchunk()
         outdebug(self.ui, 'bundle parameter: %s' % param)
         yield _pack(_fstreamparamsize, len(param))
         if param:
             yield param
         for chunk in self._compengine.compressstream(self._getcorechunk(),
                                                      self._compopts):
             yield chunk
 
     def _paramchunk(self):
         """return a encoded version of all stream parameters"""
         blocks = []
         for par, value in self._params:
             par = urlreq.quote(par)
             if value is not None:
                 value = urlreq.quote(value)
                 par = '%s=%s' % (par, value)
             blocks.append(par)
         return ' '.join(blocks)
 
     def _getcorechunk(self):
         """yield chunk for the core part of the bundle
 
         (all but headers and parameters)"""
         outdebug(self.ui, 'start of parts')
         for part in self._parts:
             outdebug(self.ui, 'bundle part: "%s"' % part.type)
             for chunk in part.getchunks(ui=self.ui):
                 yield chunk
         outdebug(self.ui, 'end of bundle')
         yield _pack(_fpartheadersize, 0)
 
 
     def salvageoutput(self):
         """return a list with a copy of all output parts in the bundle
 
         This is meant to be used during error handling to make sure we preserve
         server output"""
         salvaged = []
         for part in self._parts:
             if part.type.startswith('output'):
                 salvaged.append(part.copy())
         return salvaged
 
 
 class unpackermixin(object):
     """A mixin to extract bytes and struct data from a stream"""
 
     def __init__(self, fp):
         self._fp = fp
 
     def _unpack(self, format):
         """unpack this struct format from the stream
 
         This method is meant for internal usage by the bundle2 protocol only.
         They directly manipulate the low level stream including bundle2 level
         instruction.
 
         Do not use it to implement higher-level logic or methods."""
         data = self._readexact(struct.calcsize(format))
         return _unpack(format, data)
 
     def _readexact(self, size):
         """read exactly <size> bytes from the stream
 
         This method is meant for internal usage by the bundle2 protocol only.
         They directly manipulate the low level stream including bundle2 level
         instruction.
 
         Do not use it to implement higher-level logic or methods."""
         return changegroup.readexactly(self._fp, size)
 
 def getunbundler(ui, fp, magicstring=None):
     """return a valid unbundler object for a given magicstring"""
     if magicstring is None:
         magicstring = changegroup.readexactly(fp, 4)
     magic, version = magicstring[0:2], magicstring[2:4]
     if magic != 'HG':
         ui.debug(
             "error: invalid magic: %r (version %r), should be 'HG'\n"
             % (magic, version))
         raise error.Abort(_('not a Mercurial bundle'))
     unbundlerclass = formatmap.get(version)
     if unbundlerclass is None:
         raise error.Abort(_('unknown bundle version %s') % version)
     unbundler = unbundlerclass(ui, fp)
     indebug(ui, 'start processing of %s stream' % magicstring)
     return unbundler
 
 class unbundle20(unpackermixin):
     """interpret a bundle2 stream
 
     This class is fed with a binary stream and yields parts through its
     `iterparts` methods."""
 
     _magicstring = 'HG20'
 
     def __init__(self, ui, fp):
         """If header is specified, we do not read it out of the stream."""
         self.ui = ui
         self._compengine = util.compengines.forbundletype('UN')
         self._compressed = None
         super(unbundle20, self).__init__(fp)
 
     @util.propertycache
     def params(self):
         """dictionary of stream level parameters"""
         indebug(self.ui, 'reading bundle2 stream parameters')
         params = {}
         paramssize = self._unpack(_fstreamparamsize)[0]
         if paramssize < 0:
             raise error.BundleValueError('negative bundle param size: %i'
                                          % paramssize)
         if paramssize:
             params = self._readexact(paramssize)
             params = self._processallparams(params)
         return params
 
     def _processallparams(self, paramsblock):
         """"""
         params = util.sortdict()
         for p in paramsblock.split(' '):
             p = p.split('=', 1)
             p = [urlreq.unquote(i) for i in p]
             if len(p) < 2:
                 p.append(None)
             self._processparam(*p)
             params[p[0]] = p[1]
         return params
 
 
     def _processparam(self, name, value):
         """process a parameter, applying its effect if needed
 
         Parameter starting with a lower case letter are advisory and will be
         ignored when unknown.  Those starting with an upper case letter are
         mandatory and will this function will raise a KeyError when unknown.
 
         Note: no option are currently supported. Any input will be either
               ignored or failing.
         """
         if not name:
             raise ValueError(r'empty parameter name')
         if name[0:1] not in pycompat.bytestr(string.ascii_letters):
             raise ValueError(r'non letter first character: %s' % name)
         try:
             handler = b2streamparamsmap[name.lower()]
         except KeyError:
             if name[0:1].islower():
                 indebug(self.ui, "ignoring unknown parameter %s" % name)
             else:
                 raise error.BundleUnknownFeatureError(params=(name,))
         else:
             handler(self, name, value)
 
     def _forwardchunks(self):
         """utility to transfer a bundle2 as binary
 
         This is made necessary by the fact the 'getbundle' command over 'ssh'
         have no way to know then the reply end, relying on the bundle to be
         interpreted to know its end. This is terrible and we are sorry, but we
         needed to move forward to get general delta enabled.
         """
         yield self._magicstring
         assert 'params' not in vars(self)
         paramssize = self._unpack(_fstreamparamsize)[0]
         if paramssize < 0:
             raise error.BundleValueError('negative bundle param size: %i'
                                          % paramssize)
         yield _pack(_fstreamparamsize, paramssize)
         if paramssize:
             params = self._readexact(paramssize)
             self._processallparams(params)
             yield params
             assert self._compengine.bundletype()[1] == 'UN'
         # From there, payload might need to be decompressed
         self._fp = self._compengine.decompressorreader(self._fp)
         emptycount = 0
         while emptycount < 2:
             # so we can brainlessly loop
             assert _fpartheadersize == _fpayloadsize
             size = self._unpack(_fpartheadersize)[0]
             yield _pack(_fpartheadersize, size)
             if size:
                 emptycount = 0
             else:
                 emptycount += 1
                 continue
             if size == flaginterrupt:
                 continue
             elif size < 0:
                 raise error.BundleValueError('negative chunk size: %i')
             yield self._readexact(size)
 
 
     def iterparts(self, seekable=False):
         """yield all parts contained in the stream"""
         cls = seekableunbundlepart if seekable else unbundlepart
         # make sure param have been loaded
         self.params
         # From there, payload need to be decompressed
         self._fp = self._compengine.decompressorreader(self._fp)
         indebug(self.ui, 'start extraction of bundle2 parts')
         headerblock = self._readpartheader()
         while headerblock is not None:
             part = cls(self.ui, headerblock, self._fp)
             yield part
             # Ensure part is fully consumed so we can start reading the next
             # part.
             part.consume()
 
             headerblock = self._readpartheader()
         indebug(self.ui, 'end of bundle2 stream')
 
     def _readpartheader(self):
         """reads a part header size and return the bytes blob
 
         returns None if empty"""
         headersize = self._unpack(_fpartheadersize)[0]
         if headersize < 0:
             raise error.BundleValueError('negative part header size: %i'
                                          % headersize)
         indebug(self.ui, 'part header size: %i' % headersize)
         if headersize:
             return self._readexact(headersize)
         return None
 
     def compressed(self):
         self.params # load params
         return self._compressed
 
     def close(self):
         """close underlying file"""
         if util.safehasattr(self._fp, 'close'):
             return self._fp.close()
 
 formatmap = {'20': unbundle20}
 
 b2streamparamsmap = {}
 
 def b2streamparamhandler(name):
     """register a handler for a stream level parameter"""
     def decorator(func):
         assert name not in formatmap
         b2streamparamsmap[name] = func
         return func
     return decorator
 
 @b2streamparamhandler('compression')
 def processcompression(unbundler, param, value):
     """read compression parameter and install payload decompression"""
     if value not in util.compengines.supportedbundletypes:
         raise error.BundleUnknownFeatureError(params=(param,),
                                               values=(value,))
     unbundler._compengine = util.compengines.forbundletype(value)
     if value is not None:
         unbundler._compressed = True
 
 class bundlepart(object):
     """A bundle2 part contains application level payload
 
     The part `type` is used to route the part to the application level
     handler.
 
     The part payload is contained in ``part.data``. It could be raw bytes or a
     generator of byte chunks.
 
     You can add parameters to the part using the ``addparam`` method.
     Parameters can be either mandatory (default) or advisory. Remote side
     should be able to safely ignore the advisory ones.
 
     Both data and parameters cannot be modified after the generation has begun.
     """
 
     def __init__(self, parttype, mandatoryparams=(), advisoryparams=(),
                  data='', mandatory=True):
         validateparttype(parttype)
         self.id = None
         self.type = parttype
         self._data = data
         self._mandatoryparams = list(mandatoryparams)
         self._advisoryparams = list(advisoryparams)
         # checking for duplicated entries
         self._seenparams = set()
         for pname, __ in self._mandatoryparams + self._advisoryparams:
             if pname in self._seenparams:
                 raise error.ProgrammingError('duplicated params: %s' % pname)
             self._seenparams.add(pname)
         # status of the part's generation:
         # - None: not started,
         # - False: currently generated,
         # - True: generation done.
         self._generated = None
         self.mandatory = mandatory
 
     def __repr__(self):
         cls = "%s.%s" % (self.__class__.__module__, self.__class__.__name__)
         return ('<%s object at %x; id: %s; type: %s; mandatory: %s>'
                 % (cls, id(self), self.id, self.type, self.mandatory))
 
     def copy(self):
         """return a copy of the part
 
         The new part have the very same content but no partid assigned yet.
         Parts with generated data cannot be copied."""
         assert not util.safehasattr(self.data, 'next')
         return self.__class__(self.type, self._mandatoryparams,
                               self._advisoryparams, self._data, self.mandatory)
 
     # methods used to defines the part content
     @property
     def data(self):
         return self._data
 
     @data.setter
     def data(self, data):
         if self._generated is not None:
             raise error.ReadOnlyPartError('part is being generated')
         self._data = data
 
     @property
     def mandatoryparams(self):
         # make it an immutable tuple to force people through ``addparam``
         return tuple(self._mandatoryparams)
 
     @property
     def advisoryparams(self):
         # make it an immutable tuple to force people through ``addparam``
         return tuple(self._advisoryparams)
 
     def addparam(self, name, value='', mandatory=True):
         """add a parameter to the part
 
         If 'mandatory' is set to True, the remote handler must claim support
         for this parameter or the unbundling will be aborted.
 
         The 'name' and 'value' cannot exceed 255 bytes each.
         """
         if self._generated is not None:
             raise error.ReadOnlyPartError('part is being generated')
         if name in self._seenparams:
             raise ValueError('duplicated params: %s' % name)
         self._seenparams.add(name)
         params = self._advisoryparams
         if mandatory:
             params = self._mandatoryparams
         params.append((name, value))
 
     # methods used to generates the bundle2 stream
     def getchunks(self, ui):
         if self._generated is not None:
             raise error.ProgrammingError('part can only be consumed once')
         self._generated = False
 
         if ui.debugflag:
             msg = ['bundle2-output-part: "%s"' % self.type]
             if not self.mandatory:
                 msg.append(' (advisory)')
             nbmp = len(self.mandatoryparams)
             nbap = len(self.advisoryparams)
             if nbmp or nbap:
                 msg.append(' (params:')
                 if nbmp:
                     msg.append(' %i mandatory' % nbmp)
                 if nbap:
                     msg.append(' %i advisory' % nbmp)
                 msg.append(')')
             if not self.data:
                 msg.append(' empty payload')
             elif (util.safehasattr(self.data, 'next')
                   or util.safehasattr(self.data, '__next__')):
                 msg.append(' streamed payload')
             else:
                 msg.append(' %i bytes payload' % len(self.data))
             msg.append('\n')
             ui.debug(''.join(msg))
 
         #### header
         if self.mandatory:
             parttype = self.type.upper()
         else:
             parttype = self.type.lower()
         outdebug(ui, 'part %s: "%s"' % (pycompat.bytestr(self.id), parttype))
         ## parttype
         header = [_pack(_fparttypesize, len(parttype)),
                   parttype, _pack(_fpartid, self.id),
                  ]
         ## parameters
         # count
         manpar = self.mandatoryparams
         advpar = self.advisoryparams
         header.append(_pack(_fpartparamcount, len(manpar), len(advpar)))
         # size
         parsizes = []
         for key, value in manpar:
             parsizes.append(len(key))
             parsizes.append(len(value))
         for key, value in advpar:
             parsizes.append(len(key))
             parsizes.append(len(value))
         paramsizes = _pack(_makefpartparamsizes(len(parsizes) // 2), *parsizes)
         header.append(paramsizes)
         # key, value
         for key, value in manpar:
             header.append(key)
             header.append(value)
         for key, value in advpar:
             header.append(key)
             header.append(value)
         ## finalize header
         try:
             headerchunk = ''.join(header)
         except TypeError:
             raise TypeError(r'Found a non-bytes trying to '
                             r'build bundle part header: %r' % header)
         outdebug(ui, 'header chunk size: %i' % len(headerchunk))
         yield _pack(_fpartheadersize, len(headerchunk))
         yield headerchunk
         ## payload
         try:
             for chunk in self._payloadchunks():
                 outdebug(ui, 'payload chunk size: %i' % len(chunk))
                 yield _pack(_fpayloadsize, len(chunk))
                 yield chunk
         except GeneratorExit:
             # GeneratorExit means that nobody is listening for our
             # results anyway, so just bail quickly rather than trying
             # to produce an error part.
             ui.debug('bundle2-generatorexit\n')
             raise
         except BaseException as exc:
             bexc = stringutil.forcebytestr(exc)
             # backup exception data for later
             ui.debug('bundle2-input-stream-interrupt: encoding exception %s'
                      % bexc)
             tb = sys.exc_info()[2]
             msg = 'unexpected error: %s' % bexc
             interpart = bundlepart('error:abort', [('message', msg)],
                                    mandatory=False)
             interpart.id = 0
             yield _pack(_fpayloadsize, -1)
             for chunk in interpart.getchunks(ui=ui):
                 yield chunk
             outdebug(ui, 'closing payload chunk')
             # abort current part payload
             yield _pack(_fpayloadsize, 0)
             pycompat.raisewithtb(exc, tb)
         # end of payload
         outdebug(ui, 'closing payload chunk')
         yield _pack(_fpayloadsize, 0)
         self._generated = True
 
     def _payloadchunks(self):
         """yield chunks of a the part payload
 
         Exists to handle the different methods to provide data to a part."""
         # we only support fixed size data now.
         # This will be improved in the future.
         if (util.safehasattr(self.data, 'next')
             or util.safehasattr(self.data, '__next__')):
             buff = util.chunkbuffer(self.data)
             chunk = buff.read(preferedchunksize)
             while chunk:
                 yield chunk
                 chunk = buff.read(preferedchunksize)
         elif len(self.data):
             yield self.data
 
 
 flaginterrupt = -1
 
 class interrupthandler(unpackermixin):
     """read one part and process it with restricted capability
 
     This allows to transmit exception raised on the producer size during part
     iteration while the consumer is reading a part.
 
     Part processed in this manner only have access to a ui object,"""
 
     def __init__(self, ui, fp):
         super(interrupthandler, self).__init__(fp)
         self.ui = ui
 
     def _readpartheader(self):
         """reads a part header size and return the bytes blob
 
         returns None if empty"""
         headersize = self._unpack(_fpartheadersize)[0]
         if headersize < 0:
             raise error.BundleValueError('negative part header size: %i'
                                          % headersize)
         indebug(self.ui, 'part header size: %i\n' % headersize)
         if headersize:
             return self._readexact(headersize)
         return None
 
     def __call__(self):
 
         self.ui.debug('bundle2-input-stream-interrupt:'
                       ' opening out of band context\n')
         indebug(self.ui, 'bundle2 stream interruption, looking for a part.')
         headerblock = self._readpartheader()
         if headerblock is None:
             indebug(self.ui, 'no part found during interruption.')
             return
         part = unbundlepart(self.ui, headerblock, self._fp)
         op = interruptoperation(self.ui)
         hardabort = False
         try:
             _processpart(op, part)
         except (SystemExit, KeyboardInterrupt):
             hardabort = True
             raise
         finally:
             if not hardabort:
                 part.consume()
         self.ui.debug('bundle2-input-stream-interrupt:'
                       ' closing out of band context\n')
 
 class interruptoperation(object):
     """A limited operation to be use by part handler during interruption
 
     It only have access to an ui object.
     """
 
     def __init__(self, ui):
         self.ui = ui
         self.reply = None
         self.captureoutput = False
 
     @property
     def repo(self):
         raise error.ProgrammingError('no repo access from stream interruption')
 
     def gettransaction(self):
         raise TransactionUnavailable('no repo access from stream interruption')
 
 def decodepayloadchunks(ui, fh):
     """Reads bundle2 part payload data into chunks.
 
     Part payload data consists of framed chunks. This function takes
     a file handle and emits those chunks.
     """
     dolog = ui.configbool('devel', 'bundle2.debug')
     debug = ui.debug
 
     headerstruct = struct.Struct(_fpayloadsize)
     headersize = headerstruct.size
     unpack = headerstruct.unpack
 
     readexactly = changegroup.readexactly
     read = fh.read
 
     chunksize = unpack(readexactly(fh, headersize))[0]
     indebug(ui, 'payload chunk size: %i' % chunksize)
 
     # changegroup.readexactly() is inlined below for performance.
     while chunksize:
         if chunksize >= 0:
             s = read(chunksize)
             if len(s) < chunksize:
                 raise error.Abort(_('stream ended unexpectedly '
                                     ' (got %d bytes, expected %d)') %
                                   (len(s), chunksize))
 
             yield s
         elif chunksize == flaginterrupt:
             # Interrupt "signal" detected. The regular stream is interrupted
             # and a bundle2 part follows. Consume it.
             interrupthandler(ui, fh)()
         else:
             raise error.BundleValueError(
                 'negative payload chunk size: %s' % chunksize)
 
         s = read(headersize)
         if len(s) < headersize:
             raise error.Abort(_('stream ended unexpectedly '
                                 ' (got %d bytes, expected %d)') %
                               (len(s), chunksize))
 
         chunksize = unpack(s)[0]
 
         # indebug() inlined for performance.
         if dolog:
             debug('bundle2-input: payload chunk size: %i\n' % chunksize)
 
 class unbundlepart(unpackermixin):
     """a bundle part read from a bundle"""
 
     def __init__(self, ui, header, fp):
         super(unbundlepart, self).__init__(fp)
         self._seekable = (util.safehasattr(fp, 'seek') and
                           util.safehasattr(fp, 'tell'))
         self.ui = ui
         # unbundle state attr
         self._headerdata = header
         self._headeroffset = 0
         self._initialized = False
         self.consumed = False
         # part data
         self.id = None
         self.type = None
         self.mandatoryparams = None
         self.advisoryparams = None
         self.params = None
         self.mandatorykeys = ()
         self._readheader()
         self._mandatory = None
         self._pos = 0
 
     def _fromheader(self, size):
         """return the next <size> byte from the header"""
         offset = self._headeroffset
         data = self._headerdata[offset:(offset + size)]
         self._headeroffset = offset + size
         return data
 
     def _unpackheader(self, format):
         """read given format from header
 
         This automatically compute the size of the format to read."""
         data = self._fromheader(struct.calcsize(format))
         return _unpack(format, data)
 
     def _initparams(self, mandatoryparams, advisoryparams):
         """internal function to setup all logic related parameters"""
         # make it read only to prevent people touching it by mistake.
         self.mandatoryparams = tuple(mandatoryparams)
         self.advisoryparams  = tuple(advisoryparams)
         # user friendly UI
         self.params = util.sortdict(self.mandatoryparams)
         self.params.update(self.advisoryparams)
         self.mandatorykeys = frozenset(p[0] for p in mandatoryparams)
 
     def _readheader(self):
         """read the header and setup the object"""
         typesize = self._unpackheader(_fparttypesize)[0]
         self.type = self._fromheader(typesize)
         indebug(self.ui, 'part type: "%s"' % self.type)
         self.id = self._unpackheader(_fpartid)[0]
         indebug(self.ui, 'part id: "%s"' % pycompat.bytestr(self.id))
         # extract mandatory bit from type
         self.mandatory = (self.type != self.type.lower())
         self.type = self.type.lower()
         ## reading parameters
         # param count
         mancount, advcount = self._unpackheader(_fpartparamcount)
         indebug(self.ui, 'part parameters: %i' % (mancount + advcount))
         # param size
         fparamsizes = _makefpartparamsizes(mancount + advcount)
         paramsizes = self._unpackheader(fparamsizes)
         # make it a list of couple again
         paramsizes = list(zip(paramsizes[::2], paramsizes[1::2]))
         # split mandatory from advisory
         mansizes = paramsizes[:mancount]
         advsizes = paramsizes[mancount:]
         # retrieve param value
         manparams = []
         for key, value in mansizes:
             manparams.append((self._fromheader(key), self._fromheader(value)))
         advparams = []
         for key, value in advsizes:
             advparams.append((self._fromheader(key), self._fromheader(value)))
         self._initparams(manparams, advparams)
         ## part payload
         self._payloadstream = util.chunkbuffer(self._payloadchunks())
         # we read the data, tell it
         self._initialized = True
 
     def _payloadchunks(self):
         """Generator of decoded chunks in the payload."""
         return decodepayloadchunks(self.ui, self._fp)
 
     def consume(self):
         """Read the part payload until completion.
 
         By consuming the part data, the underlying stream read offset will
         be advanced to the next part (or end of stream).
         """
         if self.consumed:
             return
 
         chunk = self.read(32768)
         while chunk:
             self._pos += len(chunk)
             chunk = self.read(32768)
 
     def read(self, size=None):
         """read payload data"""
         if not self._initialized:
             self._readheader()
         if size is None:
             data = self._payloadstream.read()
         else:
             data = self._payloadstream.read(size)
         self._pos += len(data)
         if size is None or len(data) < size:
             if not self.consumed and self._pos:
                 self.ui.debug('bundle2-input-part: total payload size %i\n'
                               % self._pos)
             self.consumed = True
         return data
 
 class seekableunbundlepart(unbundlepart):
     """A bundle2 part in a bundle that is seekable.
 
     Regular ``unbundlepart`` instances can only be read once. This class
     extends ``unbundlepart`` to enable bi-directional seeking within the
     part.
 
     Bundle2 part data consists of framed chunks. Offsets when seeking
     refer to the decoded data, not the offsets in the underlying bundle2
     stream.
 
     To facilitate quickly seeking within the decoded data, instances of this
     class maintain a mapping between offsets in the underlying stream and
     the decoded payload. This mapping will consume memory in proportion
     to the number of chunks within the payload (which almost certainly
     increases in proportion with the size of the part).
     """
     def __init__(self, ui, header, fp):
         # (payload, file) offsets for chunk starts.
         self._chunkindex = []
 
         super(seekableunbundlepart, self).__init__(ui, header, fp)
 
     def _payloadchunks(self, chunknum=0):
         '''seek to specified chunk and start yielding data'''
         if len(self._chunkindex) == 0:
             assert chunknum == 0, 'Must start with chunk 0'
             self._chunkindex.append((0, self._tellfp()))
         else:
             assert chunknum < len(self._chunkindex), \
                    'Unknown chunk %d' % chunknum
             self._seekfp(self._chunkindex[chunknum][1])
 
         pos = self._chunkindex[chunknum][0]
 
         for chunk in decodepayloadchunks(self.ui, self._fp):
             chunknum += 1
             pos += len(chunk)
             if chunknum == len(self._chunkindex):
                 self._chunkindex.append((pos, self._tellfp()))
 
             yield chunk
 
     def _findchunk(self, pos):
         '''for a given payload position, return a chunk number and offset'''
         for chunk, (ppos, fpos) in enumerate(self._chunkindex):
             if ppos == pos:
                 return chunk, 0
             elif ppos > pos:
                 return chunk - 1, pos - self._chunkindex[chunk - 1][0]
         raise ValueError('Unknown chunk')
 
     def tell(self):
         return self._pos
 
     def seek(self, offset, whence=os.SEEK_SET):
         if whence == os.SEEK_SET:
             newpos = offset
         elif whence == os.SEEK_CUR:
             newpos = self._pos + offset
         elif whence == os.SEEK_END:
             if not self.consumed:
                 # Can't use self.consume() here because it advances self._pos.
                 chunk = self.read(32768)
                 while chunk:
                     chunk = self.read(32768)
             newpos = self._chunkindex[-1][0] - offset
         else:
             raise ValueError('Unknown whence value: %r' % (whence,))
 
         if newpos > self._chunkindex[-1][0] and not self.consumed:
             # Can't use self.consume() here because it advances self._pos.
             chunk = self.read(32768)
             while chunk:
                 chunk = self.read(32668)
 
         if not 0 <= newpos <= self._chunkindex[-1][0]:
             raise ValueError('Offset out of range')
 
         if self._pos != newpos:
             chunk, internaloffset = self._findchunk(newpos)
             self._payloadstream = util.chunkbuffer(self._payloadchunks(chunk))
             adjust = self.read(internaloffset)
             if len(adjust) != internaloffset:
                 raise error.Abort(_('Seek failed\n'))
             self._pos = newpos
 
     def _seekfp(self, offset, whence=0):
         """move the underlying file pointer
 
         This method is meant for internal usage by the bundle2 protocol only.
         They directly manipulate the low level stream including bundle2 level
         instruction.
 
         Do not use it to implement higher-level logic or methods."""
         if self._seekable:
             return self._fp.seek(offset, whence)
         else:
             raise NotImplementedError(_('File pointer is not seekable'))
 
     def _tellfp(self):
         """return the file offset, or None if file is not seekable
 
         This method is meant for internal usage by the bundle2 protocol only.
         They directly manipulate the low level stream including bundle2 level
         instruction.
 
         Do not use it to implement higher-level logic or methods."""
         if self._seekable:
             try:
                 return self._fp.tell()
             except IOError as e:
                 if e.errno == errno.ESPIPE:
                     self._seekable = False
                 else:
                     raise
         return None
 
 # These are only the static capabilities.
 # Check the 'getrepocaps' function for the rest.
 capabilities = {'HG20': (),
                 'bookmarks': (),
                 'error': ('abort', 'unsupportedcontent', 'pushraced',
                           'pushkey'),
                 'listkeys': (),
                 'pushkey': (),
                 'digests': tuple(sorted(util.DIGESTS.keys())),
                 'remote-changegroup': ('http', 'https'),
                 'hgtagsfnodes': (),
                 'rev-branch-cache': (),
                 'phases': ('heads',),
                 'stream': ('v2',),
                }
 
 def getrepocaps(repo, allowpushback=False, role=None):
     """return the bundle2 capabilities for a given repo
 
     Exists to allow extensions (like evolution) to mutate the capabilities.
 
     The returned value is used for servers advertising their capabilities as
     well as clients advertising their capabilities to servers as part of
     bundle2 requests. The ``role`` argument specifies which is which.
     """
     if role not in ('client', 'server'):
         raise error.ProgrammingError('role argument must be client or server')
 
     caps = capabilities.copy()
     caps['changegroup'] = tuple(sorted(
         changegroup.supportedincomingversions(repo)))
     if obsolete.isenabled(repo, obsolete.exchangeopt):
         supportedformat = tuple('V%i' % v for v in obsolete.formats)
         caps['obsmarkers'] = supportedformat
     if allowpushback:
         caps['pushback'] = ()
     cpmode = repo.ui.config('server', 'concurrent-push-mode')
     if cpmode == 'check-related':
         caps['checkheads'] = ('related',)
     if 'phases' in repo.ui.configlist('devel', 'legacy.exchange'):
         caps.pop('phases')
 
     # Don't advertise stream clone support in server mode if not configured.
     if role == 'server':
         streamsupported = repo.ui.configbool('server', 'uncompressed',
                                              untrusted=True)
         featuresupported = repo.ui.configbool('server', 'bundle2.stream')
 
         if not streamsupported or not featuresupported:
             caps.pop('stream')
     # Else always advertise support on client, because payload support
     # should always be advertised.
 
     return caps
 
 def bundle2caps(remote):
     """return the bundle capabilities of a peer as dict"""
     raw = remote.capable('bundle2')
     if not raw and raw != '':
         return {}
     capsblob = urlreq.unquote(remote.capable('bundle2'))
     return decodecaps(capsblob)
 
 def obsmarkersversion(caps):
     """extract the list of supported obsmarkers versions from a bundle2caps dict
     """
     obscaps = caps.get('obsmarkers', ())
     return [int(c[1:]) for c in obscaps if c.startswith('V')]
 
 def writenewbundle(ui, repo, source, filename, bundletype, outgoing, opts,
                    vfs=None, compression=None, compopts=None):
     if bundletype.startswith('HG10'):
         cg = changegroup.makechangegroup(repo, outgoing, '01', source)
         return writebundle(ui, cg, filename, bundletype, vfs=vfs,
                            compression=compression, compopts=compopts)
     elif not bundletype.startswith('HG20'):
         raise error.ProgrammingError('unknown bundle type: %s' % bundletype)
 
     caps = {}
     if 'obsolescence' in opts:
         caps['obsmarkers'] = ('V1',)
     bundle = bundle20(ui, caps)
     bundle.setcompression(compression, compopts)
     _addpartsfromopts(ui, repo, bundle, source, outgoing, opts)
     chunkiter = bundle.getchunks()
 
     return changegroup.writechunks(ui, chunkiter, filename, vfs=vfs)
 
 def _addpartsfromopts(ui, repo, bundler, source, outgoing, opts):
     # We should eventually reconcile this logic with the one behind
     # 'exchange.getbundle2partsgenerator'.
     #
     # The type of input from 'getbundle' and 'writenewbundle' are a bit
     # different right now. So we keep them separated for now for the sake of
     # simplicity.
 
     # we might not always want a changegroup in such bundle, for example in
     # stream bundles
     if opts.get('changegroup', True):
         cgversion = opts.get('cg.version')
         if cgversion is None:
             cgversion = changegroup.safeversion(repo)
         cg = changegroup.makechangegroup(repo, outgoing, cgversion, source)
         part = bundler.newpart('changegroup', data=cg.getchunks())
         part.addparam('version', cg.version)
         if 'clcount' in cg.extras:
             part.addparam('nbchanges', '%d' % cg.extras['clcount'],
                           mandatory=False)
         if opts.get('phases') and repo.revs('%ln and secret()',
                                             outgoing.missingheads):
             part.addparam('targetphase', '%d' % phases.secret, mandatory=False)
 
     if opts.get('streamv2', False):
         addpartbundlestream2(bundler, repo, stream=True)
 
     if opts.get('tagsfnodescache', True):
         addparttagsfnodescache(repo, bundler, outgoing)
 
     if opts.get('revbranchcache', True):
         addpartrevbranchcache(repo, bundler, outgoing)
 
     if opts.get('obsolescence', False):
         obsmarkers = repo.obsstore.relevantmarkers(outgoing.missing)
         buildobsmarkerspart(bundler, obsmarkers)
 
     if opts.get('phases', False):
         headsbyphase = phases.subsetphaseheads(repo, outgoing.missing)
         phasedata = phases.binaryencode(headsbyphase)
         bundler.newpart('phase-heads', data=phasedata)
 
 def addparttagsfnodescache(repo, bundler, outgoing):
     # we include the tags fnode cache for the bundle changeset
     # (as an optional parts)
     cache = tags.hgtagsfnodescache(repo.unfiltered())
     chunks = []
 
     # .hgtags fnodes are only relevant for head changesets. While we could
     # transfer values for all known nodes, there will likely be little to
     # no benefit.
     #
     # We don't bother using a generator to produce output data because
     # a) we only have 40 bytes per head and even esoteric numbers of heads
     # consume little memory (1M heads is 40MB) b) we don't want to send the
     # part if we don't have entries and knowing if we have entries requires
     # cache lookups.
     for node in outgoing.missingheads:
         # Don't compute missing, as this may slow down serving.
         fnode = cache.getfnode(node, computemissing=False)
         if fnode is not None:
             chunks.extend([node, fnode])
 
     if chunks:
         bundler.newpart('hgtagsfnodes', data=''.join(chunks))
 
 def addpartrevbranchcache(repo, bundler, outgoing):
     # we include the rev branch cache for the bundle changeset
     # (as an optional parts)
     cache = repo.revbranchcache()
     cl = repo.unfiltered().changelog
     branchesdata = collections.defaultdict(lambda: (set(), set()))
     for node in outgoing.missing:
         branch, close = cache.branchinfo(cl.rev(node))
         branchesdata[branch][close].add(node)
 
     def generate():
         for branch, (nodes, closed) in sorted(branchesdata.items()):
             utf8branch = encoding.fromlocal(branch)
             yield rbcstruct.pack(len(utf8branch), len(nodes), len(closed))
             yield utf8branch
             for n in sorted(nodes):
                 yield n
             for n in sorted(closed):
                 yield n
 
     bundler.newpart('cache:rev-branch-cache', data=generate(),
                     mandatory=False)
 
 def _formatrequirementsspec(requirements):
     return urlreq.quote(','.join(sorted(requirements)))
 
 def _formatrequirementsparams(requirements):
     requirements = _formatrequirementsspec(requirements)
     params = "%s%s" % (urlreq.quote("requirements="), requirements)
     return params
 
 def addpartbundlestream2(bundler, repo, **kwargs):
     if not kwargs.get(r'stream', False):
         return
 
     if not streamclone.allowservergeneration(repo):
         raise error.Abort(_('stream data requested but server does not allow '
                             'this feature'),
                           hint=_('well-behaved clients should not be '
                                  'requesting stream data from servers not '
                                  'advertising it; the client may be buggy'))
 
     # Stream clones don't compress well. And compression undermines a
     # goal of stream clones, which is to be fast. Communicate the desire
     # to avoid compression to consumers of the bundle.
     bundler.prefercompressed = False
 
     # get the includes and excludes
     includepats = kwargs.get(r'includepats')
     excludepats = kwargs.get(r'excludepats')
 
     narrowstream = repo.ui.configbool('experimental.server',
                                       'stream-narrow-clones')
 
     if (includepats or excludepats) and not narrowstream:
         raise error.Abort(_('server does not support narrow stream clones'))
 
+    includeobsmarkers = False
+    if repo.obsstore:
+        remoteversions = obsmarkersversion(bundler.capabilities)
+        if repo.obsstore._version in remoteversions:
+            includeobsmarkers = True
+
     filecount, bytecount, it = streamclone.generatev2(repo, includepats,
-                                                      excludepats)
+                                                      excludepats,
+                                                      includeobsmarkers)
     requirements = _formatrequirementsspec(repo.requirements)
     part = bundler.newpart('stream2', data=it)
     part.addparam('bytecount', '%d' % bytecount, mandatory=True)
     part.addparam('filecount', '%d' % filecount, mandatory=True)
     part.addparam('requirements', requirements, mandatory=True)
 
 def buildobsmarkerspart(bundler, markers):
     """add an obsmarker part to the bundler with <markers>
 
     No part is created if markers is empty.
     Raises ValueError if the bundler doesn't support any known obsmarker format.
     """
     if not markers:
         return None
 
     remoteversions = obsmarkersversion(bundler.capabilities)
     version = obsolete.commonversion(remoteversions)
     if version is None:
         raise ValueError('bundler does not support common obsmarker format')
     stream = obsolete.encodemarkers(markers, True, version=version)
     return bundler.newpart('obsmarkers', data=stream)
 
 def writebundle(ui, cg, filename, bundletype, vfs=None, compression=None,
                 compopts=None):
     """Write a bundle file and return its filename.
 
     Existing files will not be overwritten.
     If no filename is specified, a temporary file is created.
     bz2 compression can be turned off.
     The bundle file will be deleted in case of errors.
     """
 
     if bundletype == "HG20":
         bundle = bundle20(ui)
         bundle.setcompression(compression, compopts)
         part = bundle.newpart('changegroup', data=cg.getchunks())
         part.addparam('version', cg.version)
         if 'clcount' in cg.extras:
             part.addparam('nbchanges', '%d' % cg.extras['clcount'],
                           mandatory=False)
         chunkiter = bundle.getchunks()
     else:
         # compression argument is only for the bundle2 case
         assert compression is None
         if cg.version != '01':
             raise error.Abort(_('old bundle types only supports v1 '
                                 'changegroups'))
         header, comp = bundletypes[bundletype]
         if comp not in util.compengines.supportedbundletypes:
             raise error.Abort(_('unknown stream compression type: %s')
                               % comp)
         compengine = util.compengines.forbundletype(comp)
         def chunkiter():
             yield header
             for chunk in compengine.compressstream(cg.getchunks(), compopts):
                 yield chunk
         chunkiter = chunkiter()
 
     # parse the changegroup data, otherwise we will block
     # in case of sshrepo because we don't know the end of the stream
     return changegroup.writechunks(ui, chunkiter, filename, vfs=vfs)
 
 def combinechangegroupresults(op):
     """logic to combine 0 or more addchangegroup results into one"""
     results = [r.get('return', 0)
                for r in op.records['changegroup']]
     changedheads = 0
     result = 1
     for ret in results:
         # If any changegroup result is 0, return 0
         if ret == 0:
             result = 0
             break
         if ret < -1:
             changedheads += ret + 1
         elif ret > 1:
             changedheads += ret - 1
     if changedheads > 0:
         result = 1 + changedheads
     elif changedheads < 0:
         result = -1 + changedheads
     return result
 
 @parthandler('changegroup', ('version', 'nbchanges', 'treemanifest',
                              'targetphase'))
 def handlechangegroup(op, inpart):
     """apply a changegroup part on the repo
 
     This is a very early implementation that will massive rework before being
     inflicted to any end-user.
     """
     from . import localrepo
 
     tr = op.gettransaction()
     unpackerversion = inpart.params.get('version', '01')
     # We should raise an appropriate exception here
     cg = changegroup.getunbundler(unpackerversion, inpart, None)
     # the source and url passed here are overwritten by the one contained in
     # the transaction.hookargs argument. So 'bundle2' is a placeholder
     nbchangesets = None
     if 'nbchanges' in inpart.params:
         nbchangesets = int(inpart.params.get('nbchanges'))
     if ('treemanifest' in inpart.params and
         'treemanifest' not in op.repo.requirements):
         if len(op.repo.changelog) != 0:
             raise error.Abort(_(
                 "bundle contains tree manifests, but local repo is "
                 "non-empty and does not use tree manifests"))
         op.repo.requirements.add('treemanifest')
         op.repo.svfs.options = localrepo.resolvestorevfsoptions(
             op.repo.ui, op.repo.requirements, op.repo.features)
         op.repo._writerequirements()
     extrakwargs = {}
     targetphase = inpart.params.get('targetphase')
     if targetphase is not None:
         extrakwargs[r'targetphase'] = int(targetphase)
     ret = _processchangegroup(op, cg, tr, 'bundle2', 'bundle2',
                               expectedtotal=nbchangesets, **extrakwargs)
     if op.reply is not None:
         # This is definitely not the final form of this
         # return. But one need to start somewhere.
         part = op.reply.newpart('reply:changegroup', mandatory=False)
         part.addparam(
             'in-reply-to', pycompat.bytestr(inpart.id), mandatory=False)
         part.addparam('return', '%i' % ret, mandatory=False)
     assert not inpart.read()
 
 _remotechangegroupparams = tuple(['url', 'size', 'digests'] +
     ['digest:%s' % k for k in util.DIGESTS.keys()])
 @parthandler('remote-changegroup', _remotechangegroupparams)
 def handleremotechangegroup(op, inpart):
     """apply a bundle10 on the repo, given an url and validation information
 
     All the information about the remote bundle to import are given as
     parameters. The parameters include:
       - url: the url to the bundle10.
       - size: the bundle10 file size. It is used to validate what was
         retrieved by the client matches the server knowledge about the bundle.
       - digests: a space separated list of the digest types provided as
         parameters.
       - digest:<digest-type>: the hexadecimal representation of the digest with
         that name. Like the size, it is used to validate what was retrieved by
         the client matches what the server knows about the bundle.
 
     When multiple digest types are given, all of them are checked.
     """
     try:
         raw_url = inpart.params['url']
     except KeyError:
         raise error.Abort(_('remote-changegroup: missing "%s" param') % 'url')
     parsed_url = util.url(raw_url)
     if parsed_url.scheme not in capabilities['remote-changegroup']:
         raise error.Abort(_('remote-changegroup does not support %s urls') %
             parsed_url.scheme)
 
     try:
         size = int(inpart.params['size'])
     except ValueError:
         raise error.Abort(_('remote-changegroup: invalid value for param "%s"')
             % 'size')
     except KeyError:
         raise error.Abort(_('remote-changegroup: missing "%s" param') % 'size')
 
     digests = {}
     for typ in inpart.params.get('digests', '').split():
         param = 'digest:%s' % typ
         try:
             value = inpart.params[param]
         except KeyError:
             raise error.Abort(_('remote-changegroup: missing "%s" param') %
                 param)
         digests[typ] = value
 
     real_part = util.digestchecker(url.open(op.ui, raw_url), size, digests)
 
     tr = op.gettransaction()
     from . import exchange
     cg = exchange.readbundle(op.repo.ui, real_part, raw_url)
     if not isinstance(cg, changegroup.cg1unpacker):
         raise error.Abort(_('%s: not a bundle version 1.0') %
             util.hidepassword(raw_url))
     ret = _processchangegroup(op, cg, tr, 'bundle2', 'bundle2')
     if op.reply is not None:
         # This is definitely not the final form of this
         # return. But one need to start somewhere.
         part = op.reply.newpart('reply:changegroup')
         part.addparam(
             'in-reply-to', pycompat.bytestr(inpart.id), mandatory=False)
         part.addparam('return', '%i' % ret, mandatory=False)
     try:
         real_part.validate()
     except error.Abort as e:
         raise error.Abort(_('bundle at %s is corrupted:\n%s') %
                           (util.hidepassword(raw_url), bytes(e)))
     assert not inpart.read()
 
 @parthandler('reply:changegroup', ('return', 'in-reply-to'))
 def handlereplychangegroup(op, inpart):
     ret = int(inpart.params['return'])
     replyto = int(inpart.params['in-reply-to'])
     op.records.add('changegroup', {'return': ret}, replyto)
 
 @parthandler('check:bookmarks')
 def handlecheckbookmarks(op, inpart):
     """check location of bookmarks
 
     This part is to be used to detect push race regarding bookmark, it
     contains binary encoded (bookmark, node) tuple. If the local state does
     not marks the one in the part, a PushRaced exception is raised
     """
     bookdata = bookmarks.binarydecode(inpart)
 
     msgstandard = ('remote repository changed while pushing - please try again '
                    '(bookmark "%s" move from %s to %s)')
     msgmissing = ('remote repository changed while pushing - please try again '
                   '(bookmark "%s" is missing, expected %s)')
     msgexist = ('remote repository changed while pushing - please try again '
                 '(bookmark "%s" set on %s, expected missing)')
     for book, node in bookdata:
         currentnode = op.repo._bookmarks.get(book)
         if currentnode != node:
             if node is None:
                 finalmsg = msgexist % (book, nodemod.short(currentnode))
             elif currentnode is None:
                 finalmsg = msgmissing % (book, nodemod.short(node))
             else:
                 finalmsg = msgstandard % (book, nodemod.short(node),
                                           nodemod.short(currentnode))
             raise error.PushRaced(finalmsg)
 
 @parthandler('check:heads')
 def handlecheckheads(op, inpart):
     """check that head of the repo did not change
 
     This is used to detect a push race when using unbundle.
     This replaces the "heads" argument of unbundle."""
     h = inpart.read(20)
     heads = []
     while len(h) == 20:
         heads.append(h)
         h = inpart.read(20)
     assert not h
     # Trigger a transaction so that we are guaranteed to have the lock now.
     if op.ui.configbool('experimental', 'bundle2lazylocking'):
         op.gettransaction()
     if sorted(heads) != sorted(op.repo.heads()):
         raise error.PushRaced('remote repository changed while pushing - '
                               'please try again')
 
 @parthandler('check:updated-heads')
 def handlecheckupdatedheads(op, inpart):
     """check for race on the heads touched by a push
 
     This is similar to 'check:heads' but focus on the heads actually updated
     during the push. If other activities happen on unrelated heads, it is
     ignored.
 
     This allow server with high traffic to avoid push contention as long as
     unrelated parts of the graph are involved."""
     h = inpart.read(20)
     heads = []
     while len(h) == 20:
         heads.append(h)
         h = inpart.read(20)
     assert not h
     # trigger a transaction so that we are guaranteed to have the lock now.
     if op.ui.configbool('experimental', 'bundle2lazylocking'):
         op.gettransaction()
 
     currentheads = set()
     for ls in op.repo.branchmap().itervalues():
         currentheads.update(ls)
 
     for h in heads:
         if h not in currentheads:
             raise error.PushRaced('remote repository changed while pushing - '
                                   'please try again')
 
 @parthandler('check:phases')
 def handlecheckphases(op, inpart):
     """check that phase boundaries of the repository did not change
 
     This is used to detect a push race.
     """
     phasetonodes = phases.binarydecode(inpart)
     unfi = op.repo.unfiltered()
     cl = unfi.changelog
     phasecache = unfi._phasecache
     msg = ('remote repository changed while pushing - please try again '
            '(%s is %s expected %s)')
     for expectedphase, nodes in enumerate(phasetonodes):
         for n in nodes:
             actualphase = phasecache.phase(unfi, cl.rev(n))
             if actualphase != expectedphase:
                 finalmsg = msg % (nodemod.short(n),
                                   phases.phasenames[actualphase],
                                   phases.phasenames[expectedphase])
                 raise error.PushRaced(finalmsg)
 
 @parthandler('output')
 def handleoutput(op, inpart):
     """forward output captured on the server to the client"""
     for line in inpart.read().splitlines():
         op.ui.status(_('remote: %s\n') % line)
 
 @parthandler('replycaps')
 def handlereplycaps(op, inpart):
     """Notify that a reply bundle should be created
 
     The payload contains the capabilities information for the reply"""
     caps = decodecaps(inpart.read())
     if op.reply is None:
         op.reply = bundle20(op.ui, caps)
 
 class AbortFromPart(error.Abort):
     """Sub-class of Abort that denotes an error from a bundle2 part."""
 
 @parthandler('error:abort', ('message', 'hint'))
 def handleerrorabort(op, inpart):
     """Used to transmit abort error over the wire"""
     raise AbortFromPart(inpart.params['message'],
                         hint=inpart.params.get('hint'))
 
 @parthandler('error:pushkey', ('namespace', 'key', 'new', 'old', 'ret',
                                'in-reply-to'))
 def handleerrorpushkey(op, inpart):
     """Used to transmit failure of a mandatory pushkey over the wire"""
     kwargs = {}
     for name in ('namespace', 'key', 'new', 'old', 'ret'):
         value = inpart.params.get(name)
         if value is not None:
             kwargs[name] = value
     raise error.PushkeyFailed(inpart.params['in-reply-to'],
                               **pycompat.strkwargs(kwargs))
 
 @parthandler('error:unsupportedcontent', ('parttype', 'params'))
 def handleerrorunsupportedcontent(op, inpart):
     """Used to transmit unknown content error over the wire"""
     kwargs = {}
     parttype = inpart.params.get('parttype')
     if parttype is not None:
         kwargs['parttype'] = parttype
     params = inpart.params.get('params')
     if params is not None:
         kwargs['params'] = params.split('\0')
 
     raise error.BundleUnknownFeatureError(**pycompat.strkwargs(kwargs))
 
 @parthandler('error:pushraced', ('message',))
 def handleerrorpushraced(op, inpart):
     """Used to transmit push race error over the wire"""
     raise error.ResponseError(_('push failed:'), inpart.params['message'])
 
 @parthandler('listkeys', ('namespace',))
 def handlelistkeys(op, inpart):
     """retrieve pushkey namespace content stored in a bundle2"""
     namespace = inpart.params['namespace']
     r = pushkey.decodekeys(inpart.read())
     op.records.add('listkeys', (namespace, r))
 
 @parthandler('pushkey', ('namespace', 'key', 'old', 'new'))
 def handlepushkey(op, inpart):
     """process a pushkey request"""
     dec = pushkey.decode
     namespace = dec(inpart.params['namespace'])
     key = dec(inpart.params['key'])
     old = dec(inpart.params['old'])
     new = dec(inpart.params['new'])
     # Grab the transaction to ensure that we have the lock before performing the
     # pushkey.
     if op.ui.configbool('experimental', 'bundle2lazylocking'):
         op.gettransaction()
     ret = op.repo.pushkey(namespace, key, old, new)
     record = {'namespace': namespace,
               'key': key,
               'old': old,
               'new': new}
     op.records.add('pushkey', record)
     if op.reply is not None:
         rpart = op.reply.newpart('reply:pushkey')
         rpart.addparam(
             'in-reply-to', pycompat.bytestr(inpart.id), mandatory=False)
         rpart.addparam('return', '%i' % ret, mandatory=False)
     if inpart.mandatory and not ret:
         kwargs = {}
         for key in ('namespace', 'key', 'new', 'old', 'ret'):
             if key in inpart.params:
                 kwargs[key] = inpart.params[key]
         raise error.PushkeyFailed(partid='%d' % inpart.id,
                                   **pycompat.strkwargs(kwargs))
 
 @parthandler('bookmarks')
 def handlebookmark(op, inpart):
     """transmit bookmark information
 
     The part contains binary encoded bookmark information.
 
     The exact behavior of this part can be controlled by the 'bookmarks' mode
     on the bundle operation.
 
     When mode is 'apply' (the default) the bookmark information is applied as
     is to the unbundling repository. Make sure a 'check:bookmarks' part is
     issued earlier to check for push races in such update. This behavior is
     suitable for pushing.
 
     When mode is 'records', the information is recorded into the 'bookmarks'
     records of the bundle operation. This behavior is suitable for pulling.
     """
     changes = bookmarks.binarydecode(inpart)
 
     pushkeycompat = op.repo.ui.configbool('server', 'bookmarks-pushkey-compat')
     bookmarksmode = op.modes.get('bookmarks', 'apply')
 
     if bookmarksmode == 'apply':
         tr = op.gettransaction()
         bookstore = op.repo._bookmarks
         if pushkeycompat:
             allhooks = []
             for book, node in changes:
                 hookargs = tr.hookargs.copy()
                 hookargs['pushkeycompat'] = '1'
                 hookargs['namespace'] = 'bookmarks'
                 hookargs['key'] = book
                 hookargs['old'] = nodemod.hex(bookstore.get(book, ''))
                 hookargs['new'] = nodemod.hex(node if node is not None else '')
                 allhooks.append(hookargs)
 
             for hookargs in allhooks:
                 op.repo.hook('prepushkey', throw=True,
                              **pycompat.strkwargs(hookargs))
 
         bookstore.applychanges(op.repo, op.gettransaction(), changes)
 
         if pushkeycompat:
             def runhook():
                 for hookargs in allhooks:
                     op.repo.hook('pushkey', **pycompat.strkwargs(hookargs))
             op.repo._afterlock(runhook)
 
     elif bookmarksmode == 'records':
         for book, node in changes:
             record = {'bookmark': book, 'node': node}
             op.records.add('bookmarks', record)
     else:
         raise error.ProgrammingError('unkown bookmark mode: %s' % bookmarksmode)
 
 @parthandler('phase-heads')
 def handlephases(op, inpart):
     """apply phases from bundle part to repo"""
     headsbyphase = phases.binarydecode(inpart)
     phases.updatephases(op.repo.unfiltered(), op.gettransaction, headsbyphase)
 
 @parthandler('reply:pushkey', ('return', 'in-reply-to'))
 def handlepushkeyreply(op, inpart):
     """retrieve the result of a pushkey request"""
     ret = int(inpart.params['return'])
     partid = int(inpart.params['in-reply-to'])
     op.records.add('pushkey', {'return': ret}, partid)
 
 @parthandler('obsmarkers')
 def handleobsmarker(op, inpart):
     """add a stream of obsmarkers to the repo"""
     tr = op.gettransaction()
     markerdata = inpart.read()
     if op.ui.config('experimental', 'obsmarkers-exchange-debug'):
         op.ui.write(('obsmarker-exchange: %i bytes received\n')
                     % len(markerdata))
     # The mergemarkers call will crash if marker creation is not enabled.
     # we want to avoid this if the part is advisory.
     if not inpart.mandatory and op.repo.obsstore.readonly:
         op.repo.ui.debug('ignoring obsolescence markers, feature not enabled\n')
         return
     new = op.repo.obsstore.mergemarkers(tr, markerdata)
     op.repo.invalidatevolatilesets()
     if new:
         op.repo.ui.status(_('%i new obsolescence markers\n') % new)
     op.records.add('obsmarkers', {'new': new})
     if op.reply is not None:
         rpart = op.reply.newpart('reply:obsmarkers')
         rpart.addparam(
             'in-reply-to', pycompat.bytestr(inpart.id), mandatory=False)
         rpart.addparam('new', '%i' % new, mandatory=False)
 
 
 @parthandler('reply:obsmarkers', ('new', 'in-reply-to'))
 def handleobsmarkerreply(op, inpart):
     """retrieve the result of a pushkey request"""
     ret = int(inpart.params['new'])
     partid = int(inpart.params['in-reply-to'])
     op.records.add('obsmarkers', {'new': ret}, partid)
 
 @parthandler('hgtagsfnodes')
 def handlehgtagsfnodes(op, inpart):
     """Applies .hgtags fnodes cache entries to the local repo.
 
     Payload is pairs of 20 byte changeset nodes and filenodes.
     """
     # Grab the transaction so we ensure that we have the lock at this point.
     if op.ui.configbool('experimental', 'bundle2lazylocking'):
         op.gettransaction()
     cache = tags.hgtagsfnodescache(op.repo.unfiltered())
 
     count = 0
     while True:
         node = inpart.read(20)
         fnode = inpart.read(20)
         if len(node) < 20 or len(fnode) < 20:
             op.ui.debug('ignoring incomplete received .hgtags fnodes data\n')
             break
         cache.setfnode(node, fnode)
         count += 1
 
     cache.write()
     op.ui.debug('applied %i hgtags fnodes cache entries\n' % count)
 
 rbcstruct = struct.Struct('>III')
 
 @parthandler('cache:rev-branch-cache')
 def handlerbc(op, inpart):
     """receive a rev-branch-cache payload and update the local cache
 
     The payload is a series of data related to each branch
 
     1) branch name length
     2) number of open heads
     3) number of closed heads
     4) open heads nodes
     5) closed heads nodes
     """
     total = 0
     rawheader = inpart.read(rbcstruct.size)
     cache = op.repo.revbranchcache()
     cl = op.repo.unfiltered().changelog
     while rawheader:
         header = rbcstruct.unpack(rawheader)
         total += header[1] + header[2]
         utf8branch = inpart.read(header[0])
         branch = encoding.tolocal(utf8branch)
         for x in pycompat.xrange(header[1]):
             node = inpart.read(20)
             rev = cl.rev(node)
             cache.setdata(branch, rev, node, False)
         for x in pycompat.xrange(header[2]):
             node = inpart.read(20)
             rev = cl.rev(node)
             cache.setdata(branch, rev, node, True)
         rawheader = inpart.read(rbcstruct.size)
     cache.write()
 
 @parthandler('pushvars')
 def bundle2getvars(op, part):
     '''unbundle a bundle2 containing shellvars on the server'''
     # An option to disable unbundling on server-side for security reasons
     if op.ui.configbool('push', 'pushvars.server'):
         hookargs = {}
         for key, value in part.advisoryparams:
             key = key.upper()
             # We want pushed variables to have USERVAR_ prepended so we know
             # they came from the --pushvar flag.
             key = "USERVAR_" + key
             hookargs[key] = value
         op.addhookargs(hookargs)
 
 @parthandler('stream2', ('requirements', 'filecount', 'bytecount'))
 def handlestreamv2bundle(op, part):
 
     requirements = urlreq.unquote(part.params['requirements']).split(',')
     filecount = int(part.params['filecount'])
     bytecount = int(part.params['bytecount'])
 
     repo = op.repo
     if len(repo):
         msg = _('cannot apply stream clone to non empty repository')
         raise error.Abort(msg)
 
     repo.ui.debug('applying stream bundle\n')
     streamclone.applybundlev2(repo, part, filecount, bytecount,
                               requirements)
 
 def widen_bundle(repo, oldmatcher, newmatcher, common, known, cgversion,
                  ellipses):
     """generates bundle2 for widening a narrow clone
 
     repo is the localrepository instance
     oldmatcher matches what the client already has
     newmatcher matches what the client needs (including what it already has)
     common is set of common heads between server and client
     known is a set of revs known on the client side (used in ellipses)
     cgversion is the changegroup version to send
     ellipses is boolean value telling whether to send ellipses data or not
 
     returns bundle2 of the data required for extending
     """
     bundler = bundle20(repo.ui)
     commonnodes = set()
     cl = repo.changelog
     for r in repo.revs("::%ln", common):
         commonnodes.add(cl.node(r))
     if commonnodes:
         # XXX: we should only send the filelogs (and treemanifest). user
         # already has the changelog and manifest
         packer = changegroup.getbundler(cgversion, repo,
                                         oldmatcher=oldmatcher,
                                         matcher=newmatcher,
                                         fullnodes=commonnodes)
         cgdata = packer.generate(set([nodemod.nullid]), list(commonnodes),
                                  False, 'narrow_widen', changelog=False)
 
         part = bundler.newpart('changegroup', data=cgdata)
         part.addparam('version', cgversion)
         if 'treemanifest' in repo.requirements:
             part.addparam('treemanifest', '1')
 
     return bundler

mercurial/streamclone.py

@@ -1,660 +1,663 b''
 # streamclone.py - producing and consuming streaming repository data
 #
 # Copyright 2015 Gregory Szorc <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import contextlib
 import os
 import struct
 
 from .i18n import _
 from . import (
     branchmap,
     cacheutil,
     error,
     narrowspec,
     phases,
     pycompat,
     repository,
     store,
     util,
 )
 
 def canperformstreamclone(pullop, bundle2=False):
     """Whether it is possible to perform a streaming clone as part of pull.
 
     ``bundle2`` will cause the function to consider stream clone through
     bundle2 and only through bundle2.
 
     Returns a tuple of (supported, requirements). ``supported`` is True if
     streaming clone is supported and False otherwise. ``requirements`` is
     a set of repo requirements from the remote, or ``None`` if stream clone
     isn't supported.
     """
     repo = pullop.repo
     remote = pullop.remote
 
     bundle2supported = False
     if pullop.canusebundle2:
         if 'v2' in pullop.remotebundle2caps.get('stream', []):
             bundle2supported = True
         # else
             # Server doesn't support bundle2 stream clone or doesn't support
             # the versions we support. Fall back and possibly allow legacy.
 
     # Ensures legacy code path uses available bundle2.
     if bundle2supported and not bundle2:
         return False, None
     # Ensures bundle2 doesn't try to do a stream clone if it isn't supported.
     elif bundle2 and not bundle2supported:
         return False, None
 
     # Streaming clone only works on empty repositories.
     if len(repo):
         return False, None
 
     # Streaming clone only works if all data is being requested.
     if pullop.heads:
         return False, None
 
     streamrequested = pullop.streamclonerequested
 
     # If we don't have a preference, let the server decide for us. This
     # likely only comes into play in LANs.
     if streamrequested is None:
         # The server can advertise whether to prefer streaming clone.
         streamrequested = remote.capable('stream-preferred')
 
     if not streamrequested:
         return False, None
 
     # In order for stream clone to work, the client has to support all the
     # requirements advertised by the server.
     #
     # The server advertises its requirements via the "stream" and "streamreqs"
     # capability. "stream" (a value-less capability) is advertised if and only
     # if the only requirement is "revlogv1." Else, the "streamreqs" capability
     # is advertised and contains a comma-delimited list of requirements.
     requirements = set()
     if remote.capable('stream'):
         requirements.add('revlogv1')
     else:
         streamreqs = remote.capable('streamreqs')
         # This is weird and shouldn't happen with modern servers.
         if not streamreqs:
             pullop.repo.ui.warn(_(
                 'warning: stream clone requested but server has them '
                 'disabled\n'))
             return False, None
 
         streamreqs = set(streamreqs.split(','))
         # Server requires something we don't support. Bail.
         missingreqs = streamreqs - repo.supportedformats
         if missingreqs:
             pullop.repo.ui.warn(_(
                 'warning: stream clone requested but client is missing '
                 'requirements: %s\n') % ', '.join(sorted(missingreqs)))
             pullop.repo.ui.warn(
                 _('(see https://www.mercurial-scm.org/wiki/MissingRequirement '
                   'for more information)\n'))
             return False, None
         requirements = streamreqs
 
     return True, requirements
 
 def maybeperformlegacystreamclone(pullop):
     """Possibly perform a legacy stream clone operation.
 
     Legacy stream clones are performed as part of pull but before all other
     operations.
 
     A legacy stream clone will not be performed if a bundle2 stream clone is
     supported.
     """
     from . import localrepo
 
     supported, requirements = canperformstreamclone(pullop)
 
     if not supported:
         return
 
     repo = pullop.repo
     remote = pullop.remote
 
     # Save remote branchmap. We will use it later to speed up branchcache
     # creation.
     rbranchmap = None
     if remote.capable('branchmap'):
         with remote.commandexecutor() as e:
             rbranchmap = e.callcommand('branchmap', {}).result()
 
     repo.ui.status(_('streaming all changes\n'))
 
     with remote.commandexecutor() as e:
         fp = e.callcommand('stream_out', {}).result()
 
     # TODO strictly speaking, this code should all be inside the context
     # manager because the context manager is supposed to ensure all wire state
     # is flushed when exiting. But the legacy peers don't do this, so it
     # doesn't matter.
     l = fp.readline()
     try:
         resp = int(l)
     except ValueError:
         raise error.ResponseError(
             _('unexpected response from remote server:'), l)
     if resp == 1:
         raise error.Abort(_('operation forbidden by server'))
     elif resp == 2:
         raise error.Abort(_('locking the remote repository failed'))
     elif resp != 0:
         raise error.Abort(_('the server sent an unknown error code'))
 
     l = fp.readline()
     try:
         filecount, bytecount = map(int, l.split(' ', 1))
     except (ValueError, TypeError):
         raise error.ResponseError(
             _('unexpected response from remote server:'), l)
 
     with repo.lock():
         consumev1(repo, fp, filecount, bytecount)
 
         # new requirements = old non-format requirements +
         #                    new format-related remote requirements
         # requirements from the streamed-in repository
         repo.requirements = requirements | (
                 repo.requirements - repo.supportedformats)
         repo.svfs.options = localrepo.resolvestorevfsoptions(
             repo.ui, repo.requirements, repo.features)
         repo._writerequirements()
 
         if rbranchmap:
             branchmap.replacecache(repo, rbranchmap)
 
         repo.invalidate()
 
 def allowservergeneration(repo):
     """Whether streaming clones are allowed from the server."""
     if repository.REPO_FEATURE_STREAM_CLONE not in repo.features:
         return False
 
     if not repo.ui.configbool('server', 'uncompressed', untrusted=True):
         return False
 
     # The way stream clone works makes it impossible to hide secret changesets.
     # So don't allow this by default.
     secret = phases.hassecret(repo)
     if secret:
         return repo.ui.configbool('server', 'uncompressedallowsecret')
 
     return True
 
 # This is it's own function so extensions can override it.
 def _walkstreamfiles(repo, matcher=None):
     return repo.store.walk(matcher)
 
 def generatev1(repo):
     """Emit content for version 1 of a streaming clone.
 
     This returns a 3-tuple of (file count, byte size, data iterator).
 
     The data iterator consists of N entries for each file being transferred.
     Each file entry starts as a line with the file name and integer size
     delimited by a null byte.
 
     The raw file data follows. Following the raw file data is the next file
     entry, or EOF.
 
     When used on the wire protocol, an additional line indicating protocol
     success will be prepended to the stream. This function is not responsible
     for adding it.
 
     This function will obtain a repository lock to ensure a consistent view of
     the store is captured. It therefore may raise LockError.
     """
     entries = []
     total_bytes = 0
     # Get consistent snapshot of repo, lock during scan.
     with repo.lock():
         repo.ui.debug('scanning\n')
         for name, ename, size in _walkstreamfiles(repo):
             if size:
                 entries.append((name, size))
                 total_bytes += size
 
     repo.ui.debug('%d files, %d bytes to transfer\n' %
                   (len(entries), total_bytes))
 
     svfs = repo.svfs
     debugflag = repo.ui.debugflag
 
     def emitrevlogdata():
         for name, size in entries:
             if debugflag:
                 repo.ui.debug('sending %s (%d bytes)\n' % (name, size))
             # partially encode name over the wire for backwards compat
             yield '%s\0%d\n' % (store.encodedir(name), size)
             # auditing at this stage is both pointless (paths are already
             # trusted by the local repo) and expensive
             with svfs(name, 'rb', auditpath=False) as fp:
                 if size <= 65536:
                     yield fp.read(size)
                 else:
                     for chunk in util.filechunkiter(fp, limit=size):
                         yield chunk
 
     return len(entries), total_bytes, emitrevlogdata()
 
 def generatev1wireproto(repo):
     """Emit content for version 1 of streaming clone suitable for the wire.
 
     This is the data output from ``generatev1()`` with 2 header lines. The
     first line indicates overall success. The 2nd contains the file count and
     byte size of payload.
 
     The success line contains "0" for success, "1" for stream generation not
     allowed, and "2" for error locking the repository (possibly indicating
     a permissions error for the server process).
     """
     if not allowservergeneration(repo):
         yield '1\n'
         return
 
     try:
         filecount, bytecount, it = generatev1(repo)
     except error.LockError:
         yield '2\n'
         return
 
     # Indicates successful response.
     yield '0\n'
     yield '%d %d\n' % (filecount, bytecount)
     for chunk in it:
         yield chunk
 
 def generatebundlev1(repo, compression='UN'):
     """Emit content for version 1 of a stream clone bundle.
 
     The first 4 bytes of the output ("HGS1") denote this as stream clone
     bundle version 1.
 
     The next 2 bytes indicate the compression type. Only "UN" is currently
     supported.
 
     The next 16 bytes are two 64-bit big endian unsigned integers indicating
     file count and byte count, respectively.
 
     The next 2 bytes is a 16-bit big endian unsigned short declaring the length
     of the requirements string, including a trailing \0. The following N bytes
     are the requirements string, which is ASCII containing a comma-delimited
     list of repo requirements that are needed to support the data.
 
     The remaining content is the output of ``generatev1()`` (which may be
     compressed in the future).
 
     Returns a tuple of (requirements, data generator).
     """
     if compression != 'UN':
         raise ValueError('we do not support the compression argument yet')
 
     requirements = repo.requirements & repo.supportedformats
     requires = ','.join(sorted(requirements))
 
     def gen():
         yield 'HGS1'
         yield compression
 
         filecount, bytecount, it = generatev1(repo)
         repo.ui.status(_('writing %d bytes for %d files\n') %
                          (bytecount, filecount))
 
         yield struct.pack('>QQ', filecount, bytecount)
         yield struct.pack('>H', len(requires) + 1)
         yield requires + '\0'
 
         # This is where we'll add compression in the future.
         assert compression == 'UN'
 
         progress = repo.ui.makeprogress(_('bundle'), total=bytecount,
                                         unit=_('bytes'))
         progress.update(0)
 
         for chunk in it:
             progress.increment(step=len(chunk))
             yield chunk
 
         progress.complete()
 
     return requirements, gen()
 
 def consumev1(repo, fp, filecount, bytecount):
     """Apply the contents from version 1 of a streaming clone file handle.
 
     This takes the output from "stream_out" and applies it to the specified
     repository.
 
     Like "stream_out," the status line added by the wire protocol is not
     handled by this function.
     """
     with repo.lock():
         repo.ui.status(_('%d files to transfer, %s of data\n') %
                        (filecount, util.bytecount(bytecount)))
         progress = repo.ui.makeprogress(_('clone'), total=bytecount,
                                         unit=_('bytes'))
         progress.update(0)
         start = util.timer()
 
         # TODO: get rid of (potential) inconsistency
         #
         # If transaction is started and any @filecache property is
         # changed at this point, it causes inconsistency between
         # in-memory cached property and streamclone-ed file on the
         # disk. Nested transaction prevents transaction scope "clone"
         # below from writing in-memory changes out at the end of it,
         # even though in-memory changes are discarded at the end of it
         # regardless of transaction nesting.
         #
         # But transaction nesting can't be simply prohibited, because
         # nesting occurs also in ordinary case (e.g. enabling
         # clonebundles).
 
         with repo.transaction('clone'):
             with repo.svfs.backgroundclosing(repo.ui, expectedcount=filecount):
                 for i in pycompat.xrange(filecount):
                     # XXX doesn't support '\n' or '\r' in filenames
                     l = fp.readline()
                     try:
                         name, size = l.split('\0', 1)
                         size = int(size)
                     except (ValueError, TypeError):
                         raise error.ResponseError(
                             _('unexpected response from remote server:'), l)
                     if repo.ui.debugflag:
                         repo.ui.debug('adding %s (%s)\n' %
                                       (name, util.bytecount(size)))
                     # for backwards compat, name was partially encoded
                     path = store.decodedir(name)
                     with repo.svfs(path, 'w', backgroundclose=True) as ofp:
                         for chunk in util.filechunkiter(fp, limit=size):
                             progress.increment(step=len(chunk))
                             ofp.write(chunk)
 
             # force @filecache properties to be reloaded from
             # streamclone-ed file at next access
             repo.invalidate(clearfilecache=True)
 
         elapsed = util.timer() - start
         if elapsed <= 0:
             elapsed = 0.001
         progress.complete()
         repo.ui.status(_('transferred %s in %.1f seconds (%s/sec)\n') %
                        (util.bytecount(bytecount), elapsed,
                         util.bytecount(bytecount / elapsed)))
 
 def readbundle1header(fp):
     compression = fp.read(2)
     if compression != 'UN':
         raise error.Abort(_('only uncompressed stream clone bundles are '
             'supported; got %s') % compression)
 
     filecount, bytecount = struct.unpack('>QQ', fp.read(16))
     requireslen = struct.unpack('>H', fp.read(2))[0]
     requires = fp.read(requireslen)
 
     if not requires.endswith('\0'):
         raise error.Abort(_('malformed stream clone bundle: '
                             'requirements not properly encoded'))
 
     requirements = set(requires.rstrip('\0').split(','))
 
     return filecount, bytecount, requirements
 
 def applybundlev1(repo, fp):
     """Apply the content from a stream clone bundle version 1.
 
     We assume the 4 byte header has been read and validated and the file handle
     is at the 2 byte compression identifier.
     """
     if len(repo):
         raise error.Abort(_('cannot apply stream clone bundle on non-empty '
                             'repo'))
 
     filecount, bytecount, requirements = readbundle1header(fp)
     missingreqs = requirements - repo.supportedformats
     if missingreqs:
         raise error.Abort(_('unable to apply stream clone: '
                             'unsupported format: %s') %
                             ', '.join(sorted(missingreqs)))
 
     consumev1(repo, fp, filecount, bytecount)
 
 class streamcloneapplier(object):
     """Class to manage applying streaming clone bundles.
 
     We need to wrap ``applybundlev1()`` in a dedicated type to enable bundle
     readers to perform bundle type-specific functionality.
     """
     def __init__(self, fh):
         self._fh = fh
 
     def apply(self, repo):
         return applybundlev1(repo, self._fh)
 
 # type of file to stream
 _fileappend = 0 # append only file
 _filefull = 1   # full snapshot file
 
 # Source of the file
 _srcstore = 's' # store (svfs)
 _srccache = 'c' # cache (cache)
 
 # This is it's own function so extensions can override it.
 def _walkstreamfullstorefiles(repo):
     """list snapshot file from the store"""
     fnames = []
     if not repo.publishing():
         fnames.append('phaseroots')
     return fnames
 
 def _filterfull(entry, copy, vfsmap):
     """actually copy the snapshot files"""
     src, name, ftype, data = entry
     if ftype != _filefull:
         return entry
     return (src, name, ftype, copy(vfsmap[src].join(name)))
 
 @contextlib.contextmanager
 def maketempcopies():
     """return a function to temporary copy file"""
     files = []
     try:
         def copy(src):
             fd, dst = pycompat.mkstemp()
             os.close(fd)
             files.append(dst)
             util.copyfiles(src, dst, hardlink=True)
             return dst
         yield copy
     finally:
         for tmp in files:
             util.tryunlink(tmp)
 
 def _makemap(repo):
     """make a (src -> vfs) map for the repo"""
     vfsmap = {
         _srcstore: repo.svfs,
         _srccache: repo.cachevfs,
     }
     # we keep repo.vfs out of the on purpose, ther are too many danger there
     # (eg: .hg/hgrc)
     assert repo.vfs not in vfsmap.values()
 
     return vfsmap
 
 def _emit2(repo, entries, totalfilesize):
     """actually emit the stream bundle"""
     vfsmap = _makemap(repo)
     progress = repo.ui.makeprogress(_('bundle'), total=totalfilesize,
                                     unit=_('bytes'))
     progress.update(0)
     with maketempcopies() as copy, progress:
         # copy is delayed until we are in the try
         entries = [_filterfull(e, copy, vfsmap) for e in entries]
         yield None # this release the lock on the repository
         seen = 0
 
         for src, name, ftype, data in entries:
             vfs = vfsmap[src]
             yield src
             yield util.uvarintencode(len(name))
             if ftype == _fileappend:
                 fp = vfs(name)
                 size = data
             elif ftype == _filefull:
                 fp = open(data, 'rb')
                 size = util.fstat(fp).st_size
             try:
                 yield util.uvarintencode(size)
                 yield name
                 if size <= 65536:
                     chunks = (fp.read(size),)
                 else:
                     chunks = util.filechunkiter(fp, limit=size)
                 for chunk in chunks:
                     seen += len(chunk)
                     progress.update(seen)
                     yield chunk
             finally:
                 fp.close()
 
-def generatev2(repo, includes, excludes):
+def generatev2(repo, includes, excludes, includeobsmarkers):
     """Emit content for version 2 of a streaming clone.
 
     the data stream consists the following entries:
     1) A char representing the file destination (eg: store or cache)
     2) A varint containing the length of the filename
     3) A varint containing the length of file data
     4) N bytes containing the filename (the internal, store-agnostic form)
     5) N bytes containing the file data
 
     Returns a 3-tuple of (file count, file size, data iterator).
     """
 
     # temporarily raise error until we add storage level logic
     if includes or excludes:
         raise error.Abort(_("server does not support narrow stream clones"))
 
     with repo.lock():
 
         entries = []
         totalfilesize = 0
 
         matcher = None
         if includes or excludes:
             matcher = narrowspec.match(repo.root, includes, excludes)
 
         repo.ui.debug('scanning\n')
         for name, ename, size in _walkstreamfiles(repo, matcher):
             if size:
                 entries.append((_srcstore, name, _fileappend, size))
                 totalfilesize += size
         for name in _walkstreamfullstorefiles(repo):
             if repo.svfs.exists(name):
                 totalfilesize += repo.svfs.lstat(name).st_size
                 entries.append((_srcstore, name, _filefull, None))
+        if includeobsmarkers and repo.svfs.exists('obsstore'):
+            totalfilesize += repo.svfs.lstat('obsstore').st_size
+            entries.append((_srcstore, 'obsstore', _filefull, None))
         for name in cacheutil.cachetocopy(repo):
             if repo.cachevfs.exists(name):
                 totalfilesize += repo.cachevfs.lstat(name).st_size
                 entries.append((_srccache, name, _filefull, None))
 
         chunks = _emit2(repo, entries, totalfilesize)
         first = next(chunks)
         assert first is None
 
     return len(entries), totalfilesize, chunks
 
 @contextlib.contextmanager
 def nested(*ctxs):
     this = ctxs[0]
     rest = ctxs[1:]
     with this:
         if rest:
             with nested(*rest):
                 yield
         else:
             yield
 
 def consumev2(repo, fp, filecount, filesize):
     """Apply the contents from a version 2 streaming clone.
 
     Data is read from an object that only needs to provide a ``read(size)``
     method.
     """
     with repo.lock():
         repo.ui.status(_('%d files to transfer, %s of data\n') %
                        (filecount, util.bytecount(filesize)))
 
         start = util.timer()
         progress = repo.ui.makeprogress(_('clone'), total=filesize,
                                         unit=_('bytes'))
         progress.update(0)
 
         vfsmap = _makemap(repo)
 
         with repo.transaction('clone'):
             ctxs = (vfs.backgroundclosing(repo.ui)
                     for vfs in vfsmap.values())
             with nested(*ctxs):
                 for i in range(filecount):
                     src = util.readexactly(fp, 1)
                     vfs = vfsmap[src]
                     namelen = util.uvarintdecodestream(fp)
                     datalen = util.uvarintdecodestream(fp)
 
                     name = util.readexactly(fp, namelen)
 
                     if repo.ui.debugflag:
                         repo.ui.debug('adding [%s] %s (%s)\n' %
                                       (src, name, util.bytecount(datalen)))
 
                     with vfs(name, 'w') as ofp:
                         for chunk in util.filechunkiter(fp, limit=datalen):
                             progress.increment(step=len(chunk))
                             ofp.write(chunk)
 
             # force @filecache properties to be reloaded from
             # streamclone-ed file at next access
             repo.invalidate(clearfilecache=True)
 
         elapsed = util.timer() - start
         if elapsed <= 0:
             elapsed = 0.001
         repo.ui.status(_('transferred %s in %.1f seconds (%s/sec)\n') %
                        (util.bytecount(progress.pos), elapsed,
                         util.bytecount(progress.pos / elapsed)))
         progress.complete()
 
 def applybundlev2(repo, fp, filecount, filesize, requirements):
     from . import localrepo
 
     missingreqs = [r for r in requirements if r not in repo.supported]
     if missingreqs:
         raise error.Abort(_('unable to apply stream clone: '
                             'unsupported format: %s') %
                           ', '.join(sorted(missingreqs)))
 
     consumev2(repo, fp, filecount, filesize)
 
     # new requirements = old non-format requirements +
     #                    new format-related remote requirements
     # requirements from the streamed-in repository
     repo.requirements = set(requirements) | (
             repo.requirements - repo.supportedformats)
     repo.svfs.options = localrepo.resolvestorevfsoptions(
         repo.ui, repo.requirements, repo.features)
     repo._writerequirements()

tests/test-clone-uncompressed.t

@@ -1,516 +1,561 b''
 #require serve no-reposimplestore no-chg
 
 #testcases stream-legacy stream-bundle2
 
 #if stream-legacy
   $ cat << EOF >> $HGRCPATH
   > [server]
   > bundle2.stream = no
   > EOF
 #endif
 
 Initialize repository
 the status call is to check for issue5130
 
   $ hg init server
   $ cd server
   $ touch foo
   $ hg -q commit -A -m initial
   >>> for i in range(1024):
   ...     with open(str(i), 'wb') as fh:
   ...         fh.write(b"%d" % i) and None
   $ hg -q commit -A -m 'add a lot of files'
   $ hg st
   $ hg --config server.uncompressed=false serve -p $HGPORT -d --pid-file=hg.pid
   $ cat hg.pid > $DAEMON_PIDS
   $ cd ..
 
 Cannot stream clone when server.uncompressed is set
 
   $ get-with-headers.py $LOCALIP:$HGPORT '?cmd=stream_out'
   200 Script output follows
   
   1
 
 #if stream-legacy
   $ hg debugcapabilities http://localhost:$HGPORT
   Main capabilities:
     batch
     branchmap
     $USUAL_BUNDLE2_CAPS_SERVER$
     changegroupsubset
     compression=$BUNDLE2_COMPRESSIONS$
     getbundle
     httpheader=1024
     httpmediatype=0.1rx,0.1tx,0.2tx
     known
     lookup
     pushkey
     unbundle=HG10GZ,HG10BZ,HG10UN
     unbundlehash
   Bundle2 capabilities:
     HG20
     bookmarks
     changegroup
       01
       02
     digests
       md5
       sha1
       sha512
     error
       abort
       unsupportedcontent
       pushraced
       pushkey
     hgtagsfnodes
     listkeys
     phases
       heads
     pushkey
     remote-changegroup
       http
       https
     rev-branch-cache
 
   $ hg clone --stream -U http://localhost:$HGPORT server-disabled
   warning: stream clone requested but server has them disabled
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 2 changesets with 1025 changes to 1025 files
   new changesets 96ee1d7354c4:c17445101a72
 
   $ get-with-headers.py $LOCALIP:$HGPORT '?cmd=getbundle' content-type --bodyfile body --hgproto 0.2 --requestheader "x-hgarg-1=bundlecaps=HG20%2Cbundle2%3DHG20%250Abookmarks%250Achangegroup%253D01%252C02%250Adigests%253Dmd5%252Csha1%252Csha512%250Aerror%253Dabort%252Cunsupportedcontent%252Cpushraced%252Cpushkey%250Ahgtagsfnodes%250Alistkeys%250Aphases%253Dheads%250Apushkey%250Aremote-changegroup%253Dhttp%252Chttps&cg=0&common=0000000000000000000000000000000000000000&heads=c17445101a72edac06facd130d14808dfbd5c7c2&stream=1"
   200 Script output follows
   content-type: application/mercurial-0.2
   
 
   $ f --size body --hexdump --bytes 100
   body: size=232
   0000: 04 6e 6f 6e 65 48 47 32 30 00 00 00 00 00 00 00 |.noneHG20.......|
   0010: cf 0b 45 52 52 4f 52 3a 41 42 4f 52 54 00 00 00 |..ERROR:ABORT...|
   0020: 00 01 01 07 3c 04 72 6d 65 73 73 61 67 65 73 74 |....<.rmessagest|
   0030: 72 65 61 6d 20 64 61 74 61 20 72 65 71 75 65 73 |ream data reques|
   0040: 74 65 64 20 62 75 74 20 73 65 72 76 65 72 20 64 |ted but server d|
   0050: 6f 65 73 20 6e 6f 74 20 61 6c 6c 6f 77 20 74 68 |oes not allow th|
   0060: 69 73 20 66                                     |is f|
 
 #endif
 #if stream-bundle2
   $ hg debugcapabilities http://localhost:$HGPORT
   Main capabilities:
     batch
     branchmap
     $USUAL_BUNDLE2_CAPS_SERVER$
     changegroupsubset
     compression=$BUNDLE2_COMPRESSIONS$
     getbundle
     httpheader=1024
     httpmediatype=0.1rx,0.1tx,0.2tx
     known
     lookup
     pushkey
     unbundle=HG10GZ,HG10BZ,HG10UN
     unbundlehash
   Bundle2 capabilities:
     HG20
     bookmarks
     changegroup
       01
       02
     digests
       md5
       sha1
       sha512
     error
       abort
       unsupportedcontent
       pushraced
       pushkey
     hgtagsfnodes
     listkeys
     phases
       heads
     pushkey
     remote-changegroup
       http
       https
     rev-branch-cache
 
   $ hg clone --stream -U http://localhost:$HGPORT server-disabled
   warning: stream clone requested but server has them disabled
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 2 changesets with 1025 changes to 1025 files
   new changesets 96ee1d7354c4:c17445101a72
 
   $ get-with-headers.py $LOCALIP:$HGPORT '?cmd=getbundle' content-type --bodyfile body --hgproto 0.2 --requestheader "x-hgarg-1=bundlecaps=HG20%2Cbundle2%3DHG20%250Abookmarks%250Achangegroup%253D01%252C02%250Adigests%253Dmd5%252Csha1%252Csha512%250Aerror%253Dabort%252Cunsupportedcontent%252Cpushraced%252Cpushkey%250Ahgtagsfnodes%250Alistkeys%250Aphases%253Dheads%250Apushkey%250Aremote-changegroup%253Dhttp%252Chttps&cg=0&common=0000000000000000000000000000000000000000&heads=c17445101a72edac06facd130d14808dfbd5c7c2&stream=1"
   200 Script output follows
   content-type: application/mercurial-0.2
   
 
   $ f --size body --hexdump --bytes 100
   body: size=232
   0000: 04 6e 6f 6e 65 48 47 32 30 00 00 00 00 00 00 00 |.noneHG20.......|
   0010: cf 0b 45 52 52 4f 52 3a 41 42 4f 52 54 00 00 00 |..ERROR:ABORT...|
   0020: 00 01 01 07 3c 04 72 6d 65 73 73 61 67 65 73 74 |....<.rmessagest|
   0030: 72 65 61 6d 20 64 61 74 61 20 72 65 71 75 65 73 |ream data reques|
   0040: 74 65 64 20 62 75 74 20 73 65 72 76 65 72 20 64 |ted but server d|
   0050: 6f 65 73 20 6e 6f 74 20 61 6c 6c 6f 77 20 74 68 |oes not allow th|
   0060: 69 73 20 66                                     |is f|
 
 #endif
 
   $ killdaemons.py
   $ cd server
   $ hg serve -p $HGPORT -d --pid-file=hg.pid
   $ cat hg.pid > $DAEMON_PIDS
   $ cd ..
 
 Basic clone
 
 #if stream-legacy
   $ hg clone --stream -U http://localhost:$HGPORT clone1
   streaming all changes
   1027 files to transfer, 96.3 KB of data
   transferred 96.3 KB in * seconds (*/sec) (glob)
   searching for changes
   no changes found
 #endif
 #if stream-bundle2
   $ hg clone --stream -U http://localhost:$HGPORT clone1
   streaming all changes
   1030 files to transfer, 96.4 KB of data
   transferred 96.4 KB in * seconds (* */sec) (glob)
 
   $ ls -1 clone1/.hg/cache
   branch2-served
   rbc-names-v1
   rbc-revs-v1
 #endif
 
 getbundle requests with stream=1 are uncompressed
 
   $ get-with-headers.py $LOCALIP:$HGPORT '?cmd=getbundle' content-type --bodyfile body --hgproto '0.1 0.2 comp=zlib,none' --requestheader "x-hgarg-1=bundlecaps=HG20%2Cbundle2%3DHG20%250Abookmarks%250Achangegroup%253D01%252C02%250Adigests%253Dmd5%252Csha1%252Csha512%250Aerror%253Dabort%252Cunsupportedcontent%252Cpushraced%252Cpushkey%250Ahgtagsfnodes%250Alistkeys%250Aphases%253Dheads%250Apushkey%250Aremote-changegroup%253Dhttp%252Chttps&cg=0&common=0000000000000000000000000000000000000000&heads=c17445101a72edac06facd130d14808dfbd5c7c2&stream=1"
   200 Script output follows
   content-type: application/mercurial-0.2
   
 
   $ f --size --hex --bytes 256 body
   body: size=112230
   0000: 04 6e 6f 6e 65 48 47 32 30 00 00 00 00 00 00 00 |.noneHG20.......|
   0010: 70 07 53 54 52 45 41 4d 32 00 00 00 00 03 00 09 |p.STREAM2.......|
   0020: 05 09 04 0c 35 62 79 74 65 63 6f 75 6e 74 39 38 |....5bytecount98|
   0030: 37 35 38 66 69 6c 65 63 6f 75 6e 74 31 30 33 30 |758filecount1030|
   0040: 72 65 71 75 69 72 65 6d 65 6e 74 73 64 6f 74 65 |requirementsdote|
   0050: 6e 63 6f 64 65 25 32 43 66 6e 63 61 63 68 65 25 |ncode%2Cfncache%|
   0060: 32 43 67 65 6e 65 72 61 6c 64 65 6c 74 61 25 32 |2Cgeneraldelta%2|
   0070: 43 72 65 76 6c 6f 67 76 31 25 32 43 73 74 6f 72 |Crevlogv1%2Cstor|
   0080: 65 00 00 80 00 73 08 42 64 61 74 61 2f 30 2e 69 |e....s.Bdata/0.i|
   0090: 00 03 00 01 00 00 00 00 00 00 00 02 00 00 00 01 |................|
   00a0: 00 00 00 00 00 00 00 01 ff ff ff ff ff ff ff ff |................|
   00b0: 80 29 63 a0 49 d3 23 87 bf ce fe 56 67 92 67 2c |.)c.I.#....Vg.g,|
   00c0: 69 d1 ec 39 00 00 00 00 00 00 00 00 00 00 00 00 |i..9............|
   00d0: 75 30 73 08 42 64 61 74 61 2f 31 2e 69 00 03 00 |u0s.Bdata/1.i...|
   00e0: 01 00 00 00 00 00 00 00 02 00 00 00 01 00 00 00 |................|
   00f0: 00 00 00 00 01 ff ff ff ff ff ff ff ff f9 76 da |..............v.|
 
 --uncompressed is an alias to --stream
 
 #if stream-legacy
   $ hg clone --uncompressed -U http://localhost:$HGPORT clone1-uncompressed
   streaming all changes
   1027 files to transfer, 96.3 KB of data
   transferred 96.3 KB in * seconds (*/sec) (glob)
   searching for changes
   no changes found
 #endif
 #if stream-bundle2
   $ hg clone --uncompressed -U http://localhost:$HGPORT clone1-uncompressed
   streaming all changes
   1030 files to transfer, 96.4 KB of data
   transferred 96.4 KB in * seconds (* */sec) (glob)
 #endif
 
 Clone with background file closing enabled
 
 #if stream-legacy
   $ hg --debug --config worker.backgroundclose=true --config worker.backgroundcloseminfilecount=1 clone --stream -U http://localhost:$HGPORT clone-background | grep -v adding
   using http://localhost:$HGPORT/
   sending capabilities command
   sending branchmap command
   streaming all changes
   sending stream_out command
   1027 files to transfer, 96.3 KB of data
   starting 4 threads for background file closing
   updating the branch cache
   transferred 96.3 KB in * seconds (*/sec) (glob)
   query 1; heads
   sending batch command
   searching for changes
   all remote heads known locally
   no changes found
   sending getbundle command
   bundle2-input-bundle: with-transaction
   bundle2-input-part: "listkeys" (params: 1 mandatory) supported
   bundle2-input-part: "phase-heads" supported
   bundle2-input-part: total payload size 24
   bundle2-input-bundle: 1 parts total
   checking for updated bookmarks
   (sent 5 HTTP requests and * bytes; received * bytes in responses) (glob)
 #endif
 #if stream-bundle2
   $ hg --debug --config worker.backgroundclose=true --config worker.backgroundcloseminfilecount=1 clone --stream -U http://localhost:$HGPORT clone-background | grep -v adding
   using http://localhost:$HGPORT/
   sending capabilities command
   query 1; heads
   sending batch command
   streaming all changes
   sending getbundle command
   bundle2-input-bundle: with-transaction
   bundle2-input-part: "stream2" (params: 3 mandatory) supported
   applying stream bundle
   1030 files to transfer, 96.4 KB of data
   starting 4 threads for background file closing
   starting 4 threads for background file closing
   updating the branch cache
   transferred 96.4 KB in * seconds (* */sec) (glob)
   bundle2-input-part: total payload size 112077
   bundle2-input-part: "listkeys" (params: 1 mandatory) supported
   bundle2-input-bundle: 1 parts total
   checking for updated bookmarks
   (sent 3 HTTP requests and * bytes; received * bytes in responses) (glob)
 #endif
 
 Cannot stream clone when there are secret changesets
 
   $ hg -R server phase --force --secret -r tip
   $ hg clone --stream -U http://localhost:$HGPORT secret-denied
   warning: stream clone requested but server has them disabled
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 1 changesets with 1 changes to 1 files
   new changesets 96ee1d7354c4
 
   $ killdaemons.py
 
 Streaming of secrets can be overridden by server config
 
   $ cd server
   $ hg serve --config server.uncompressedallowsecret=true -p $HGPORT -d --pid-file=hg.pid
   $ cat hg.pid > $DAEMON_PIDS
   $ cd ..
 
 #if stream-legacy
   $ hg clone --stream -U http://localhost:$HGPORT secret-allowed
   streaming all changes
   1027 files to transfer, 96.3 KB of data
   transferred 96.3 KB in * seconds (*/sec) (glob)
   searching for changes
   no changes found
 #endif
 #if stream-bundle2
   $ hg clone --stream -U http://localhost:$HGPORT secret-allowed
   streaming all changes
   1030 files to transfer, 96.4 KB of data
   transferred 96.4 KB in * seconds (* */sec) (glob)
 #endif
 
   $ killdaemons.py
 
 Verify interaction between preferuncompressed and secret presence
 
   $ cd server
   $ hg serve --config server.preferuncompressed=true -p $HGPORT -d --pid-file=hg.pid
   $ cat hg.pid > $DAEMON_PIDS
   $ cd ..
 
   $ hg clone -U http://localhost:$HGPORT preferuncompressed-secret
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 1 changesets with 1 changes to 1 files
   new changesets 96ee1d7354c4
 
   $ killdaemons.py
 
 Clone not allowed when full bundles disabled and can't serve secrets
 
   $ cd server
   $ hg serve --config server.disablefullbundle=true -p $HGPORT -d --pid-file=hg.pid
   $ cat hg.pid > $DAEMON_PIDS
   $ cd ..
 
   $ hg clone --stream http://localhost:$HGPORT secret-full-disabled
   warning: stream clone requested but server has them disabled
   requesting all changes
   remote: abort: server has pull-based clones disabled
   abort: pull failed on remote
   (remove --pull if specified or upgrade Mercurial)
   [255]
 
 Local stream clone with secrets involved
 (This is just a test over behavior: if you have access to the repo's files,
 there is no security so it isn't important to prevent a clone here.)
 
   $ hg clone -U --stream server local-secret
   warning: stream clone requested but server has them disabled
   requesting all changes
   adding changesets
   adding manifests
   adding file changes
   added 1 changesets with 1 changes to 1 files
   new changesets 96ee1d7354c4
 
 Stream clone while repo is changing:
 
   $ mkdir changing
   $ cd changing
 
 extension for delaying the server process so we reliably can modify the repo
 while cloning
 
   $ cat > delayer.py <<EOF
   > import time
   > from mercurial import extensions, vfs
   > def __call__(orig, self, path, *args, **kwargs):
   >     if path == 'data/f1.i':
   >         time.sleep(2)
   >     return orig(self, path, *args, **kwargs)
   > extensions.wrapfunction(vfs.vfs, '__call__', __call__)
   > EOF
 
 prepare repo with small and big file to cover both code paths in emitrevlogdata
 
   $ hg init repo
   $ touch repo/f1
   $ $TESTDIR/seq.py 50000 > repo/f2
   $ hg -R repo ci -Aqm "0"
   $ hg serve -R repo -p $HGPORT1 -d --pid-file=hg.pid --config extensions.delayer=delayer.py
   $ cat hg.pid >> $DAEMON_PIDS
 
 clone while modifying the repo between stating file with write lock and
 actually serving file content
 
   $ hg clone -q --stream -U http://localhost:$HGPORT1 clone &
   $ sleep 1
   $ echo >> repo/f1
   $ echo >> repo/f2
   $ hg -R repo ci -m "1"
   $ wait
   $ hg -R clone id
   000000000000
   $ cd ..
 
 Stream repository with bookmarks
 --------------------------------
 
 (revert introduction of secret changeset)
 
   $ hg -R server phase --draft 'secret()'
 
 add a bookmark
 
   $ hg -R server bookmark -r tip some-bookmark
 
 clone it
 
 #if stream-legacy
   $ hg clone --stream http://localhost:$HGPORT with-bookmarks
   streaming all changes
   1027 files to transfer, 96.3 KB of data
   transferred 96.3 KB in * seconds (*) (glob)
   searching for changes
   no changes found
   updating to branch default
   1025 files updated, 0 files merged, 0 files removed, 0 files unresolved
 #endif
 #if stream-bundle2
   $ hg clone --stream http://localhost:$HGPORT with-bookmarks
   streaming all changes
   1033 files to transfer, 96.6 KB of data
   transferred 96.6 KB in * seconds (* */sec) (glob)
   updating to branch default
   1025 files updated, 0 files merged, 0 files removed, 0 files unresolved
 #endif
   $ hg -R with-bookmarks bookmarks
      some-bookmark             1:c17445101a72
 
 Stream repository with phases
 -----------------------------
 
 Clone as publishing
 
   $ hg -R server phase -r 'all()'
   0: draft
   1: draft
 
 #if stream-legacy
   $ hg clone --stream http://localhost:$HGPORT phase-publish
   streaming all changes
   1027 files to transfer, 96.3 KB of data
   transferred 96.3 KB in * seconds (*) (glob)
   searching for changes
   no changes found
   updating to branch default
   1025 files updated, 0 files merged, 0 files removed, 0 files unresolved
 #endif
 #if stream-bundle2
   $ hg clone --stream http://localhost:$HGPORT phase-publish
   streaming all changes
   1033 files to transfer, 96.6 KB of data
   transferred 96.6 KB in * seconds (* */sec) (glob)
   updating to branch default
   1025 files updated, 0 files merged, 0 files removed, 0 files unresolved
 #endif
   $ hg -R phase-publish phase -r 'all()'
   0: public
   1: public
 
 Clone as non publishing
 
   $ cat << EOF >> server/.hg/hgrc
   > [phases]
   > publish = False
   > EOF
   $ killdaemons.py
   $ hg -R server serve -p $HGPORT -d --pid-file=hg.pid
   $ cat hg.pid > $DAEMON_PIDS
 
 #if stream-legacy
 
 With v1 of the stream protocol, changeset are always cloned as public. It make
 stream v1 unsuitable for non-publishing repository.
 
   $ hg clone --stream http://localhost:$HGPORT phase-no-publish
   streaming all changes
   1027 files to transfer, 96.3 KB of data
   transferred 96.3 KB in * seconds (*) (glob)
   searching for changes
   no changes found
   updating to branch default
   1025 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ hg -R phase-no-publish phase -r 'all()'
   0: public
   1: public
 #endif
 #if stream-bundle2
   $ hg clone --stream http://localhost:$HGPORT phase-no-publish
   streaming all changes
   1034 files to transfer, 96.7 KB of data
   transferred 96.7 KB in * seconds (* */sec) (glob)
   updating to branch default
   1025 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ hg -R phase-no-publish phase -r 'all()'
   0: draft
   1: draft
 #endif
 
   $ killdaemons.py
+
+#if stream-legacy
+
+With v1 of the stream protocol, changeset are always cloned as public. There's
+no obsolescence markers exchange in stream v1.
+
+#endif
+#if stream-bundle2
+
+Stream repository with obsolescence
+-----------------------------------
+
+Clone non-publishing with obsolescence
+
+  $ cat >> $HGRCPATH << EOF
+  > [experimental]
+  > evolution=all
+  > EOF
+
+  $ cd server
+  $ echo foo > foo
+  $ hg -q commit -m 'about to be pruned'
+  $ hg debugobsolete `hg log -r . -T '{node}'` -d '0 0' -u test --record-parents
+  obsoleted 1 changesets
+  $ hg up null -q
+  $ hg log -T '{rev}: {phase}\n'
+  1: draft
+  0: draft
+  $ hg serve -p $HGPORT -d --pid-file=hg.pid
+  $ cat hg.pid > $DAEMON_PIDS
+  $ cd ..
+
+  $ hg clone -U --stream http://localhost:$HGPORT with-obsolescence
+  streaming all changes
+  1035 files to transfer, 97.1 KB of data
+  transferred 97.1 KB in * seconds (* */sec) (glob)
+  $ hg -R with-obsolescence log -T '{rev}: {phase}\n'
+  1: draft
+  0: draft
+  $ hg debugobsolete -R with-obsolescence
+  50382b884f66690b7045cac93a540cba4d4c906f 0 {c17445101a72edac06facd130d14808dfbd5c7c2} (Thu Jan 01 00:00:00 1970 +0000) {'user': 'test'}
+
+  $ killdaemons.py
+
+#endif