##// END OF EJS Templates
copy: add flag for disabling copy tracing...
copy: add flag for disabling copy tracing Copy tracing can be up to 80% of rebase time when rebasing stacks of commits in large repos (hundreds of thousands of files). This provides the option of turning off the majority of copy tracing. It does not turn off _forwardcopies() since that is used to carry copy information inside a commit across a rebase. This will affect the situation where a user edits a file, then rebases on top of commits that have moved that file. The move will not be detected and the user will have to manually resolve the issue (possibly by redoing the rebase with this flag off). The reason to have a flag instead of trying to fix the actual copy tracing performance is that copy tracing is fundamentally an O(number of files in the repo) operation. In order to know if file X in the rebase source was copied anywhere, we have to walk the filelog for every new file that exists in the rebase destination (i.e. a file in the destination that is not in the common ancestor). Without an index that lets us trace forward (i.e. from file Y in the common ancestor forward to the rebase destination), it will never be an O(number of changes in my branch) operation. In mozilla-central, rebasing a 3 commit stack across 20,000 revs goes from 39s to 11s.

File last commit:

r25992:2d76f8a2 default
r26013:38f92d12 default
Show More
worker.py
165 lines | 4.4 KiB | text/x-python | PythonLexer
Bryan O'Sullivan
worker: count the number of CPUs...
r18635 # worker.py - master-slave parallelism support
#
# 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.
Gregory Szorc
worker: use absolute_import
r25992 from __future__ import absolute_import
import errno
import os
import signal
import sys
import threading
from .i18n import _
from . import util
Bryan O'Sullivan
worker: count the number of CPUs...
r18635
def countcpus():
'''try to count the number of CPUs on the system'''
# posix
try:
n = int(os.sysconf('SC_NPROCESSORS_ONLN'))
if n > 0:
return n
except (AttributeError, ValueError):
pass
# windows
try:
n = int(os.environ['NUMBER_OF_PROCESSORS'])
if n > 0:
return n
except (KeyError, ValueError):
pass
return 1
Bryan O'Sullivan
worker: estimate whether it's worth running a task in parallel...
r18636
def _numworkers(ui):
s = ui.config('worker', 'numcpus')
if s:
try:
n = int(s)
if n >= 1:
return n
except ValueError:
raise util.Abort(_('number of cpus must be an integer'))
return min(max(countcpus(), 4), 32)
if os.name == 'posix':
_startupcost = 0.01
else:
_startupcost = 1e30
def worthwhile(ui, costperop, nops):
'''try to determine whether the benefit of multiple processes can
outweigh the cost of starting them'''
linear = costperop * nops
workers = _numworkers(ui)
benefit = linear - (_startupcost * workers + linear / workers)
return benefit >= 0.15
Bryan O'Sullivan
worker: partition a list (of tasks) into equal-sized chunks
r18637
Bryan O'Sullivan
worker: allow a function to be run in multiple worker processes...
r18638 def worker(ui, costperarg, func, staticargs, args):
'''run a function, possibly in parallel in multiple worker
processes.
returns a progress iterator
costperarg - cost of a single task
func - function to run
staticargs - arguments to pass to every invocation of the function
args - arguments to split into chunks, to pass to individual
workers
'''
if worthwhile(ui, costperarg, len(args)):
return _platformworker(ui, func, staticargs, args)
return func(*staticargs + (args,))
def _posixworker(ui, func, staticargs, args):
rfd, wfd = os.pipe()
workers = _numworkers(ui)
Bryan O'Sullivan
worker: fix a race in SIGINT handling...
r18708 oldhandler = signal.getsignal(signal.SIGINT)
signal.signal(signal.SIGINT, signal.SIG_IGN)
Bryan O'Sullivan
worker: handle worker failures more aggressively...
r18709 pids, problem = [], [0]
Bryan O'Sullivan
worker: allow a function to be run in multiple worker processes...
r18638 for pargs in partition(args, workers):
pid = os.fork()
if pid == 0:
Bryan O'Sullivan
worker: fix a race in SIGINT handling...
r18708 signal.signal(signal.SIGINT, oldhandler)
Bryan O'Sullivan
worker: allow a function to be run in multiple worker processes...
r18638 try:
os.close(rfd)
for i, item in func(*(staticargs + (pargs,))):
os.write(wfd, '%d %s\n' % (i, item))
os._exit(0)
except KeyboardInterrupt:
os._exit(255)
Matt Mackall
worker: properly report errors from worker processes (issue3982)
r19408 # other exceptions are allowed to propagate, we rely
# on lock.py's pid checks to avoid release callbacks
Bryan O'Sullivan
worker: handle worker failures more aggressively...
r18709 pids.append(pid)
pids.reverse()
Bryan O'Sullivan
worker: allow a function to be run in multiple worker processes...
r18638 os.close(wfd)
fp = os.fdopen(rfd, 'rb', 0)
Bryan O'Sullivan
worker: handle worker failures more aggressively...
r18709 def killworkers():
# if one worker bails, there's no good reason to wait for the rest
for p in pids:
try:
os.kill(p, signal.SIGTERM)
Gregory Szorc
global: mass rewrite to use modern exception syntax...
r25660 except OSError as err:
Bryan O'Sullivan
worker: handle worker failures more aggressively...
r18709 if err.errno != errno.ESRCH:
raise
def waitforworkers():
Mads Kiilerich
cleanup: avoid _ for local unused tmp variables - that is reserved for i18n...
r22199 for _pid in pids:
Bryan O'Sullivan
worker: handle worker failures more aggressively...
r18709 st = _exitstatus(os.wait()[1])
Matt Mackall
worker: check problem state correctly (issue3982)...
r19406 if st and not problem[0]:
Bryan O'Sullivan
worker: handle worker failures more aggressively...
r18709 problem[0] = st
killworkers()
t = threading.Thread(target=waitforworkers)
t.start()
Bryan O'Sullivan
worker: allow a function to be run in multiple worker processes...
r18638 def cleanup():
signal.signal(signal.SIGINT, oldhandler)
Bryan O'Sullivan
worker: handle worker failures more aggressively...
r18709 t.join()
status = problem[0]
if status:
if status < 0:
os.kill(os.getpid(), -status)
sys.exit(status)
Bryan O'Sullivan
worker: allow a function to be run in multiple worker processes...
r18638 try:
for line in fp:
l = line.split(' ', 1)
yield int(l[0]), l[1][:-1]
except: # re-raises
Bryan O'Sullivan
worker: handle worker failures more aggressively...
r18709 killworkers()
Bryan O'Sullivan
worker: allow a function to be run in multiple worker processes...
r18638 cleanup()
raise
cleanup()
Bryan O'Sullivan
worker: on error, exit similarly to the first failing worker...
r18707 def _posixexitstatus(code):
'''convert a posix exit status into the same form returned by
os.spawnv
returns None if the process was stopped instead of exiting'''
if os.WIFEXITED(code):
return os.WEXITSTATUS(code)
elif os.WIFSIGNALED(code):
return -os.WTERMSIG(code)
Bryan O'Sullivan
worker: allow a function to be run in multiple worker processes...
r18638 if os.name != 'nt':
_platformworker = _posixworker
Bryan O'Sullivan
worker: on error, exit similarly to the first failing worker...
r18707 _exitstatus = _posixexitstatus
Bryan O'Sullivan
worker: allow a function to be run in multiple worker processes...
r18638
Bryan O'Sullivan
worker: partition a list (of tasks) into equal-sized chunks
r18637 def partition(lst, nslices):
'''partition a list into N slices of equal size'''
n = len(lst)
chunk, slop = n / nslices, n % nslices
end = 0
for i in xrange(nslices):
start = end
end = start + chunk
if slop:
end += 1
slop -= 1
yield lst[start:end]