wireproto: implement batching on peer executor interface
This is a bit more complicated than non-batch requests because we
need to buffer sends until the last request arrives *and* we need
to support resolving futures as data arrives from the remote.
In a classical concurrent.futures executor model, the future
"starts" as soon as it is submitted. However, we have nothing to
start until the last command is submitted.
If we did nothing, calling result() would deadlock, since the future
hasn't "started." So in the case where we queue the command, we return
a special future type whose result() will trigger sendcommands().
This eliminates the deadlock potential. It also serves as a check
against callers who may be calling result() prematurely, as it will
prevent any subsequent callcommands() from working. This behavior
is slightly annoying and a bit restrictive. But it's the world
that half duplex connections forces on us.
In order to support streaming responses, we were previously using
a generator. But with a futures-based API, we're using futures
and not generators. So in order to get streaming, we need a
background thread to read data from the server.
The approach taken in this patch is to leverage the ThreadPoolExecutor
from concurrent.futures for managing a background thread. We create
an executor and future that resolves when all response data is
processed (or an error occurs). When exiting the context manager,
we wait on that background reading before returning.
I was hoping we could manually spin up a threading.Thread and this
would be simple. But I ran into a few deadlocks when implementing.
After looking at the source code to concurrent.futures, I figured
it would just be easier to use a ThreadPoolExecutor than implement
all the code needed to manually manage a thread.
To prove this works, a use of the batch API in discovery has been
updated.
Differential Revision: https://phab.mercurial-scm.org/D3269
# wireprotov1peer.py - Client-side functionality for wire protocol version 1.
#
# Copyright 2005-2010 Matt Mackall <mpm@selenic.com>
#
# 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 hashlib
import sys
import weakref
from .i18n import _
from .node import (
bin,
)
from .thirdparty.zope import (
interface as zi,
)
from . import (
bundle2,
changegroup as changegroupmod,
encoding,
error,
pushkey as pushkeymod,
pycompat,
repository,
util,
wireprototypes,
)
urlreq = util.urlreq
def batchable(f):
'''annotation for batchable methods
Such methods must implement a coroutine as follows:
@batchable
def sample(self, one, two=None):
# Build list of encoded arguments suitable for your wire protocol:
encargs = [('one', encode(one),), ('two', encode(two),)]
# Create future for injection of encoded result:
encresref = future()
# Return encoded arguments and future:
yield encargs, encresref
# Assuming the future to be filled with the result from the batched
# request now. Decode it:
yield decode(encresref.value)
The decorator returns a function which wraps this coroutine as a plain
method, but adds the original method as an attribute called "batchable",
which is used by remotebatch to split the call into separate encoding and
decoding phases.
'''
def plain(*args, **opts):
batchable = f(*args, **opts)
encargsorres, encresref = next(batchable)
if not encresref:
return encargsorres # a local result in this case
self = args[0]
cmd = pycompat.bytesurl(f.__name__) # ensure cmd is ascii bytestr
encresref.set(self._submitone(cmd, encargsorres))
return next(batchable)
setattr(plain, 'batchable', f)
return plain
class future(object):
'''placeholder for a value to be set later'''
def set(self, value):
if util.safehasattr(self, 'value'):
raise error.RepoError("future is already set")
self.value = value
class batcher(object):
'''base class for batches of commands submittable in a single request
All methods invoked on instances of this class are simply queued and
return a a future for the result. Once you call submit(), all the queued
calls are performed and the results set in their respective futures.
'''
def __init__(self):
self.calls = []
def __getattr__(self, name):
def call(*args, **opts):
resref = future()
# Please don't invent non-ascii method names, or you will
# give core hg a very sad time.
self.calls.append((name.encode('ascii'), args, opts, resref,))
return resref
return call
def submit(self):
raise NotImplementedError()
class iterbatcher(batcher):
def submit(self):
raise NotImplementedError()
def results(self):
raise NotImplementedError()
class remoteiterbatcher(iterbatcher):
def __init__(self, remote):
super(remoteiterbatcher, self).__init__()
self._remote = remote
def __getattr__(self, name):
# Validate this method is batchable, since submit() only supports
# batchable methods.
fn = getattr(self._remote, name)
if not getattr(fn, 'batchable', None):
raise error.ProgrammingError('Attempted to batch a non-batchable '
'call to %r' % name)
return super(remoteiterbatcher, self).__getattr__(name)
def submit(self):
"""Break the batch request into many patch calls and pipeline them.
This is mostly valuable over http where request sizes can be
limited, but can be used in other places as well.
"""
# 2-tuple of (command, arguments) that represents what will be
# sent over the wire.
requests = []
# 4-tuple of (command, final future, @batchable generator, remote
# future).
results = []
for command, args, opts, finalfuture in self.calls:
mtd = getattr(self._remote, command)
batchable = mtd.batchable(mtd.__self__, *args, **opts)
commandargs, fremote = next(batchable)
assert fremote
requests.append((command, commandargs))
results.append((command, finalfuture, batchable, fremote))
if requests:
self._resultiter = self._remote._submitbatch(requests)
self._results = results
def results(self):
for command, finalfuture, batchable, remotefuture in self._results:
# Get the raw result, set it in the remote future, feed it
# back into the @batchable generator so it can be decoded, and
# set the result on the final future to this value.
remoteresult = next(self._resultiter)
remotefuture.set(remoteresult)
finalfuture.set(next(batchable))
# Verify our @batchable generators only emit 2 values.
try:
next(batchable)
except StopIteration:
pass
else:
raise error.ProgrammingError('%s @batchable generator emitted '
'unexpected value count' % command)
yield finalfuture.value
def encodebatchcmds(req):
"""Return a ``cmds`` argument value for the ``batch`` command."""
escapearg = wireprototypes.escapebatcharg
cmds = []
for op, argsdict in req:
# Old servers didn't properly unescape argument names. So prevent
# the sending of argument names that may not be decoded properly by
# servers.
assert all(escapearg(k) == k for k in argsdict)
args = ','.join('%s=%s' % (escapearg(k), escapearg(v))
for k, v in argsdict.iteritems())
cmds.append('%s %s' % (op, args))
return ';'.join(cmds)
class unsentfuture(pycompat.futures.Future):
"""A Future variation to represent an unsent command.
Because we buffer commands and don't submit them immediately, calling
``result()`` on an unsent future could deadlock. Futures for buffered
commands are represented by this type, which wraps ``result()`` to
call ``sendcommands()``.
"""
def result(self, timeout=None):
if self.done():
return pycompat.futures.Future.result(self, timeout)
self._peerexecutor.sendcommands()
# This looks like it will infinitely recurse. However,
# sendcommands() should modify __class__. This call serves as a check
# on that.
return self.result(timeout)
@zi.implementer(repository.ipeercommandexecutor)
class peerexecutor(object):
def __init__(self, peer):
self._peer = peer
self._sent = False
self._closed = False
self._calls = []
self._futures = weakref.WeakSet()
self._responseexecutor = None
self._responsef = None
def __enter__(self):
return self
def __exit__(self, exctype, excvalee, exctb):
self.close()
def callcommand(self, command, args):
if self._sent:
raise error.ProgrammingError('callcommand() cannot be used '
'after commands are sent')
if self._closed:
raise error.ProgrammingError('callcommand() cannot be used '
'after close()')
# Commands are dispatched through methods on the peer.
fn = getattr(self._peer, pycompat.sysstr(command), None)
if not fn:
raise error.ProgrammingError(
'cannot call command %s: method of same name not available '
'on peer' % command)
# Commands are either batchable or they aren't. If a command
# isn't batchable, we send it immediately because the executor
# can no longer accept new commands after a non-batchable command.
# If a command is batchable, we queue it for later. But we have
# to account for the case of a non-batchable command arriving after
# a batchable one and refuse to service it.
def addcall():
f = pycompat.futures.Future()
self._futures.add(f)
self._calls.append((command, args, fn, f))
return f
if getattr(fn, 'batchable', False):
f = addcall()
# But since we don't issue it immediately, we wrap its result()
# to trigger sending so we avoid deadlocks.
f.__class__ = unsentfuture
f._peerexecutor = self
else:
if self._calls:
raise error.ProgrammingError(
'%s is not batchable and cannot be called on a command '
'executor along with other commands' % command)
f = addcall()
# Non-batchable commands can never coexist with another command
# in this executor. So send the command immediately.
self.sendcommands()
return f
def sendcommands(self):
if self._sent:
return
if not self._calls:
return
self._sent = True
# Unhack any future types so caller seens a clean type and to break
# cycle between us and futures.
for f in self._futures:
if isinstance(f, unsentfuture):
f.__class__ = pycompat.futures.Future
f._peerexecutor = None
calls = self._calls
# Mainly to destroy references to futures.
self._calls = None
# Simple case of a single command. We call it synchronously.
if len(calls) == 1:
command, args, fn, f = calls[0]
# Future was cancelled. Ignore it.
if not f.set_running_or_notify_cancel():
return
try:
result = fn(**pycompat.strkwargs(args))
except Exception:
f.set_exception_info(*sys.exc_info()[1:])
else:
f.set_result(result)
return
# Batch commands are a bit harder. First, we have to deal with the
# @batchable coroutine. That's a bit annoying. Furthermore, we also
# need to preserve streaming. i.e. it should be possible for the
# futures to resolve as data is coming in off the wire without having
# to wait for the final byte of the final response. We do this by
# spinning up a thread to read the responses.
requests = []
states = []
for command, args, fn, f in calls:
# Future was cancelled. Ignore it.
if not f.set_running_or_notify_cancel():
continue
try:
batchable = fn.batchable(fn.__self__,
**pycompat.strkwargs(args))
except Exception:
f.set_exception_info(*sys.exc_info()[1:])
return
# Encoded arguments and future holding remote result.
try:
encodedargs, fremote = next(batchable)
except Exception:
f.set_exception_info(*sys.exc_info()[1:])
return
requests.append((command, encodedargs))
states.append((command, f, batchable, fremote))
if not requests:
return
# This will emit responses in order they were executed.
wireresults = self._peer._submitbatch(requests)
# The use of a thread pool executor here is a bit weird for something
# that only spins up a single thread. However, thread management is
# hard and it is easy to encounter race conditions, deadlocks, etc.
# concurrent.futures already solves these problems and its thread pool
# executor has minimal overhead. So we use it.
self._responseexecutor = pycompat.futures.ThreadPoolExecutor(1)
self._responsef = self._responseexecutor.submit(self._readbatchresponse,
states, wireresults)
def close(self):
self.sendcommands()
if self._closed:
return
self._closed = True
if not self._responsef:
return
# We need to wait on our in-flight response and then shut down the
# executor once we have a result.
try:
self._responsef.result()
finally:
self._responseexecutor.shutdown(wait=True)
self._responsef = None
self._responseexecutor = None
# If any of our futures are still in progress, mark them as
# errored. Otherwise a result() could wait indefinitely.
for f in self._futures:
if not f.done():
f.set_exception(error.ResponseError(
_('unfulfilled batch command response')))
self._futures = None
def _readbatchresponse(self, states, wireresults):
# Executes in a thread to read data off the wire.
for command, f, batchable, fremote in states:
# Grab raw result off the wire and teach the internal future
# about it.
remoteresult = next(wireresults)
fremote.set(remoteresult)
# And ask the coroutine to decode that value.
try:
result = next(batchable)
except Exception:
f.set_exception_info(*sys.exc_info()[1:])
else:
f.set_result(result)
class wirepeer(repository.legacypeer):
"""Client-side interface for communicating with a peer repository.
Methods commonly call wire protocol commands of the same name.
See also httppeer.py and sshpeer.py for protocol-specific
implementations of this interface.
"""
def commandexecutor(self):
return peerexecutor(self)
# Begin of ipeercommands interface.
def iterbatch(self):
return remoteiterbatcher(self)
@batchable
def lookup(self, key):
self.requirecap('lookup', _('look up remote revision'))
f = future()
yield {'key': encoding.fromlocal(key)}, f
d = f.value
success, data = d[:-1].split(" ", 1)
if int(success):
yield bin(data)
else:
self._abort(error.RepoError(data))
@batchable
def heads(self):
f = future()
yield {}, f
d = f.value
try:
yield wireprototypes.decodelist(d[:-1])
except ValueError:
self._abort(error.ResponseError(_("unexpected response:"), d))
@batchable
def known(self, nodes):
f = future()
yield {'nodes': wireprototypes.encodelist(nodes)}, f
d = f.value
try:
yield [bool(int(b)) for b in d]
except ValueError:
self._abort(error.ResponseError(_("unexpected response:"), d))
@batchable
def branchmap(self):
f = future()
yield {}, f
d = f.value
try:
branchmap = {}
for branchpart in d.splitlines():
branchname, branchheads = branchpart.split(' ', 1)
branchname = encoding.tolocal(urlreq.unquote(branchname))
branchheads = wireprototypes.decodelist(branchheads)
branchmap[branchname] = branchheads
yield branchmap
except TypeError:
self._abort(error.ResponseError(_("unexpected response:"), d))
@batchable
def listkeys(self, namespace):
if not self.capable('pushkey'):
yield {}, None
f = future()
self.ui.debug('preparing listkeys for "%s"\n' % namespace)
yield {'namespace': encoding.fromlocal(namespace)}, f
d = f.value
self.ui.debug('received listkey for "%s": %i bytes\n'
% (namespace, len(d)))
yield pushkeymod.decodekeys(d)
@batchable
def pushkey(self, namespace, key, old, new):
if not self.capable('pushkey'):
yield False, None
f = future()
self.ui.debug('preparing pushkey for "%s:%s"\n' % (namespace, key))
yield {'namespace': encoding.fromlocal(namespace),
'key': encoding.fromlocal(key),
'old': encoding.fromlocal(old),
'new': encoding.fromlocal(new)}, f
d = f.value
d, output = d.split('\n', 1)
try:
d = bool(int(d))
except ValueError:
raise error.ResponseError(
_('push failed (unexpected response):'), d)
for l in output.splitlines(True):
self.ui.status(_('remote: '), l)
yield d
def stream_out(self):
return self._callstream('stream_out')
def getbundle(self, source, **kwargs):
kwargs = pycompat.byteskwargs(kwargs)
self.requirecap('getbundle', _('look up remote changes'))
opts = {}
bundlecaps = kwargs.get('bundlecaps') or set()
for key, value in kwargs.iteritems():
if value is None:
continue
keytype = wireprototypes.GETBUNDLE_ARGUMENTS.get(key)
if keytype is None:
raise error.ProgrammingError(
'Unexpectedly None keytype for key %s' % key)
elif keytype == 'nodes':
value = wireprototypes.encodelist(value)
elif keytype == 'csv':
value = ','.join(value)
elif keytype == 'scsv':
value = ','.join(sorted(value))
elif keytype == 'boolean':
value = '%i' % bool(value)
elif keytype != 'plain':
raise KeyError('unknown getbundle option type %s'
% keytype)
opts[key] = value
f = self._callcompressable("getbundle", **pycompat.strkwargs(opts))
if any((cap.startswith('HG2') for cap in bundlecaps)):
return bundle2.getunbundler(self.ui, f)
else:
return changegroupmod.cg1unpacker(f, 'UN')
def unbundle(self, cg, heads, url):
'''Send cg (a readable file-like object representing the
changegroup to push, typically a chunkbuffer object) to the
remote server as a bundle.
When pushing a bundle10 stream, return an integer indicating the
result of the push (see changegroup.apply()).
When pushing a bundle20 stream, return a bundle20 stream.
`url` is the url the client thinks it's pushing to, which is
visible to hooks.
'''
if heads != ['force'] and self.capable('unbundlehash'):
heads = wireprototypes.encodelist(
['hashed', hashlib.sha1(''.join(sorted(heads))).digest()])
else:
heads = wireprototypes.encodelist(heads)
if util.safehasattr(cg, 'deltaheader'):
# this a bundle10, do the old style call sequence
ret, output = self._callpush("unbundle", cg, heads=heads)
if ret == "":
raise error.ResponseError(
_('push failed:'), output)
try:
ret = int(ret)
except ValueError:
raise error.ResponseError(
_('push failed (unexpected response):'), ret)
for l in output.splitlines(True):
self.ui.status(_('remote: '), l)
else:
# bundle2 push. Send a stream, fetch a stream.
stream = self._calltwowaystream('unbundle', cg, heads=heads)
ret = bundle2.getunbundler(self.ui, stream)
return ret
# End of ipeercommands interface.
# Begin of ipeerlegacycommands interface.
def branches(self, nodes):
n = wireprototypes.encodelist(nodes)
d = self._call("branches", nodes=n)
try:
br = [tuple(wireprototypes.decodelist(b)) for b in d.splitlines()]
return br
except ValueError:
self._abort(error.ResponseError(_("unexpected response:"), d))
def between(self, pairs):
batch = 8 # avoid giant requests
r = []
for i in xrange(0, len(pairs), batch):
n = " ".join([wireprototypes.encodelist(p, '-')
for p in pairs[i:i + batch]])
d = self._call("between", pairs=n)
try:
r.extend(l and wireprototypes.decodelist(l) or []
for l in d.splitlines())
except ValueError:
self._abort(error.ResponseError(_("unexpected response:"), d))
return r
def changegroup(self, nodes, kind):
n = wireprototypes.encodelist(nodes)
f = self._callcompressable("changegroup", roots=n)
return changegroupmod.cg1unpacker(f, 'UN')
def changegroupsubset(self, bases, heads, kind):
self.requirecap('changegroupsubset', _('look up remote changes'))
bases = wireprototypes.encodelist(bases)
heads = wireprototypes.encodelist(heads)
f = self._callcompressable("changegroupsubset",
bases=bases, heads=heads)
return changegroupmod.cg1unpacker(f, 'UN')
# End of ipeerlegacycommands interface.
def _submitbatch(self, req):
"""run batch request <req> on the server
Returns an iterator of the raw responses from the server.
"""
ui = self.ui
if ui.debugflag and ui.configbool('devel', 'debug.peer-request'):
ui.debug('devel-peer-request: batched-content\n')
for op, args in req:
msg = 'devel-peer-request: - %s (%d arguments)\n'
ui.debug(msg % (op, len(args)))
unescapearg = wireprototypes.unescapebatcharg
rsp = self._callstream("batch", cmds=encodebatchcmds(req))
chunk = rsp.read(1024)
work = [chunk]
while chunk:
while ';' not in chunk and chunk:
chunk = rsp.read(1024)
work.append(chunk)
merged = ''.join(work)
while ';' in merged:
one, merged = merged.split(';', 1)
yield unescapearg(one)
chunk = rsp.read(1024)
work = [merged, chunk]
yield unescapearg(''.join(work))
def _submitone(self, op, args):
return self._call(op, **pycompat.strkwargs(args))
def debugwireargs(self, one, two, three=None, four=None, five=None):
# don't pass optional arguments left at their default value
opts = {}
if three is not None:
opts[r'three'] = three
if four is not None:
opts[r'four'] = four
return self._call('debugwireargs', one=one, two=two, **opts)
def _call(self, cmd, **args):
"""execute <cmd> on the server
The command is expected to return a simple string.
returns the server reply as a string."""
raise NotImplementedError()
def _callstream(self, cmd, **args):
"""execute <cmd> on the server
The command is expected to return a stream. Note that if the
command doesn't return a stream, _callstream behaves
differently for ssh and http peers.
returns the server reply as a file like object.
"""
raise NotImplementedError()
def _callcompressable(self, cmd, **args):
"""execute <cmd> on the server
The command is expected to return a stream.
The stream may have been compressed in some implementations. This
function takes care of the decompression. This is the only difference
with _callstream.
returns the server reply as a file like object.
"""
raise NotImplementedError()
def _callpush(self, cmd, fp, **args):
"""execute a <cmd> on server
The command is expected to be related to a push. Push has a special
return method.
returns the server reply as a (ret, output) tuple. ret is either
empty (error) or a stringified int.
"""
raise NotImplementedError()
def _calltwowaystream(self, cmd, fp, **args):
"""execute <cmd> on server
The command will send a stream to the server and get a stream in reply.
"""
raise NotImplementedError()
def _abort(self, exception):
"""clearly abort the wire protocol connection and raise the exception
"""
raise NotImplementedError()