mirror of
https://github.com/facebook/sapling.git
synced 2024-10-12 17:58:27 +03:00
9dc21f8d0b
Summary: D13853115 adds `edenscm/` to `sys.path` and code still uses `import mercurial`. That has nasty problems if both `import mercurial` and `import edenscm.mercurial` are used, because Python would think `mercurial.foo` and `edenscm.mercurial.foo` are different modules so code like `try: ... except mercurial.error.Foo: ...`, or `isinstance(x, mercurial.foo.Bar)` would fail to handle the `edenscm.mercurial` version. There are also some module-level states (ex. `extensions._extensions`) that would cause trouble if they have multiple versions in a single process. Change imports to use the `edenscm` so ideally the `mercurial` is no longer imported at all. Add checks in extensions.py to catch unexpected extensions importing modules from the old (wrong) locations when running tests. Reviewed By: phillco Differential Revision: D13868981 fbshipit-source-id: f4e2513766957fd81d85407994f7521a08e4de48
111 lines
4.1 KiB
Python
111 lines
4.1 KiB
Python
# Copyright 2018 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.
|
|
|
|
from __future__ import absolute_import
|
|
|
|
import itertools
|
|
import random
|
|
|
|
from edenscm.mercurial import context
|
|
|
|
|
|
# TODO: make these configs
|
|
MAX_FILES_PER_COMMIT = 10000
|
|
DESIRED_FILES_PER_COMMIT = 6
|
|
MAX_EDITS_PER_FILE = 6
|
|
FILE_DELETION_CHANCE = 6 # percent
|
|
ADD_DELETE_RATIO = 3
|
|
DELETION_MAX_SIZE = 2000
|
|
|
|
BLACKLIST = [".hgdirsync", ".hgtags"]
|
|
|
|
|
|
class randomeditsgenerator(object):
|
|
def __init__(self, ctx):
|
|
"""``ctx`` is used to build a set of paths"""
|
|
self.ui = ctx.repo().ui
|
|
self.pathchars = "abcdefghijklmnopqrstuvwxyz"
|
|
self.dirs = self.makedirs()
|
|
self.fnames = self.makefilenames()
|
|
|
|
def makedirs(self):
|
|
dircount = self.ui.configint("repogenerator", "filenamedircount", 3)
|
|
|
|
# We generate all path combinations up-front and then shuffle them to
|
|
# try and distribute the paths. (`` getrandompath`` only selects from
|
|
# a prorated slice of this list based on generation progress, so if it
|
|
# were not randomized, initial edis would all be in a/a/*.)
|
|
#
|
|
# One downside to this approach: third-level directories are still too
|
|
# sparse initially.
|
|
paths = list(itertools.product(self.pathchars, repeat=dircount))
|
|
random.shuffle(paths)
|
|
return paths
|
|
|
|
def makefilenames(self):
|
|
leaflen = self.ui.configint("repogenerator", "filenameleaflength", 3)
|
|
|
|
# Unlike with dirs, there's no upside to randomizing, so keep
|
|
# alphabetical for simplicity.
|
|
return list(itertools.product(self.pathchars, repeat=leaflen))
|
|
|
|
def getrandompath(self, wctx):
|
|
# Limit the dictionary of directory names initially, and expand them
|
|
# over time (based on our goal) to mimic the organic growth of
|
|
# directories and projects.
|
|
maxdir = max(1, int(len(self.dirs) * self.getcompletionratio(wctx)))
|
|
dirparts = random.sample(self.dirs[0:maxdir], 1)[0]
|
|
|
|
# Same thing but for filenames.
|
|
maxfname = max(1, int(len(self.fnames) * self.getcompletionratio(wctx)))
|
|
fnparts = random.sample(self.fnames[0:maxfname], 1)[0]
|
|
|
|
return "/".join(dirparts) + "/" + "".join(fnparts)
|
|
|
|
def getcompletionratio(self, wctx):
|
|
tiprev = wctx.p1().rev() + 1 # rev is 0-based
|
|
goalrev = self.ui.configint("repogenerator", "numcommits", 10000)
|
|
return float(tiprev) / goalrev
|
|
|
|
def makerandomedits(self, wctx):
|
|
i = 0
|
|
while i < DESIRED_FILES_PER_COMMIT:
|
|
path = self.getrandompath(wctx)
|
|
existingdata = ""
|
|
if isinstance(wctx, context.workingctx):
|
|
path = path.encode("ascii", "ignore")
|
|
|
|
if wctx[path].exists():
|
|
if random.randrange(0, 100) <= FILE_DELETION_CHANCE:
|
|
wctx[path].remove()
|
|
else:
|
|
existingdata = wctx[path].data()
|
|
else:
|
|
existingdata = "new file\n"
|
|
|
|
if len(existingdata) > 0:
|
|
for _ in range(0, random.randrange(1, MAX_EDITS_PER_FILE)):
|
|
if len(existingdata) <= 1:
|
|
existingdata = "new file"
|
|
break
|
|
if random.randrange(0, 10) > ADD_DELETE_RATIO:
|
|
idx = random.randrange(0, len(existingdata))
|
|
existingdata = (
|
|
existingdata[:idx]
|
|
+ "/* random data: %s */\n" % random.randrange(0, 900000)
|
|
+ existingdata[idx:]
|
|
)
|
|
else:
|
|
length = random.randrange(
|
|
0, min(len(existingdata) - 1, DELETION_MAX_SIZE)
|
|
)
|
|
idx = random.randrange(0, len(existingdata) - length)
|
|
existingdata = (
|
|
existingdata[:idx] + existingdata[(idx + length) :]
|
|
)
|
|
|
|
wctx[path].write(existingdata, "")
|
|
i += 1
|