##// END OF EJS Templates
xdiff: add a preprocessing step that trims files...
xdiff: add a preprocessing step that trims files xdiff has a `xdl_trim_ends` step that removes common lines, unmatchable lines. That is in theory good, but happens too late - after splitting, hashing, and adjusting the hash values so they are unique. Those splitting, hashing and adjusting hash values steps could have noticeable overhead. Diffing two large files with minor (one-line-ish) changes are not uncommon. In that case, the raw performance of those preparation steps seriously matter. Even allocating an O(N) array and storing line offsets to it is expensive. Therefore my previous attempts [1] [2] cannot be good enough since they do not remove the O(N) array assignment. This patch adds a preprocessing step - `xdl_trim_files` that runs before other preprocessing steps. It counts common prefix and suffix and lines in them (needed for displaying line number), without doing anything else. Testing with a crafted large (169MB) file, with minor change: ``` open('a','w').write(''.join('%s\n' % (i % 100000) for i in xrange(30000000) if i != 6000000)) open('b','w').write(''.join('%s\n' % (i % 100000) for i in xrange(30000000) if i != 6003000)) ``` Running xdiff by a simple binary [3], this patch improves the xdiff perf by more than 10x for the above case: ``` # xdiff before this patch 2.41s user 1.13s system 98% cpu 3.592 total # xdiff after this patch 0.14s user 0.16s system 98% cpu 0.309 total # gnu diffutils 0.12s user 0.15s system 98% cpu 0.272 total # (best of 20 runs) ``` It's still slightly slower than GNU diffutils. But it's pretty close now. Testing with real repo data: For the whole repo, this patch makes xdiff 25% faster: ``` # hg perfbdiff --count 100 --alldata -c d334afc585e2 --blocks [--xdiff] # xdiff, after ! wall 0.058861 comb 0.050000 user 0.050000 sys 0.000000 (best of 100) # xdiff, before ! wall 0.077816 comb 0.080000 user 0.080000 sys 0.000000 (best of 91) # bdiff ! wall 0.117473 comb 0.120000 user 0.120000 sys 0.000000 (best of 67) ``` For files that are long (ex. commands.py), the speedup is more than 3x, very significant: ``` # hg perfbdiff --count 3000 --blocks commands.py.i 1 [--xdiff] # xdiff, after ! wall 0.690583 comb 0.690000 user 0.690000 sys 0.000000 (best of 12) # xdiff, before ! wall 2.240361 comb 2.210000 user 2.210000 sys 0.000000 (best of 4) # bdiff ! wall 2.469852 comb 2.440000 user 2.440000 sys 0.000000 (best of 4) ``` [1]: https://phab.mercurial-scm.org/D2631 [2]: https://phab.mercurial-scm.org/D2634 [3]: ``` // Code to run xdiff from command line. No proper error handling. #include <stdlib.h> #include <unistd.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include "mercurial/thirdparty/xdiff/xdiff.h" #define ensure(x) if (!(x)) exit(255); mmfile_t readfile(const char *path) { struct stat st; int fd = open(path, O_RDONLY); fstat(fd, &st); mmfile_t file = { malloc(st.st_size), st.st_size }; ensure(read(fd, file.ptr, st.st_size) == st.st_size); close(fd); return file; } int main(int argc, char const *argv[]) { mmfile_t a = readfile(argv[1]), b = readfile(argv[2]); xpparam_t xpp = {0}; xdemitconf_t xecfg = {0}; xdemitcb_t ecb = {0}; xdl_diff(&a, &b, &xpp, &xecfg, &ecb); return 0; } ``` Differential Revision: https://phab.mercurial-scm.org/D2686

File last commit:

r34648:dacfcdd8 default
r36838:f33a87cf default
Show More
osutil.py
102 lines | 3.5 KiB | text/x-python | PythonLexer
# osutil.py - CFFI version of osutil.c
#
# Copyright 2016 Maciej Fijalkowski <fijall@gmail.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 os
import stat as statmod
from ..pure.osutil import *
from .. import (
pycompat,
)
if pycompat.isdarwin:
from . import _osutil
ffi = _osutil.ffi
lib = _osutil.lib
listdir_batch_size = 4096
# tweakable number, only affects performance, which chunks
# of bytes do we get back from getattrlistbulk
attrkinds = [None] * 20 # we need the max no for enum VXXX, 20 is plenty
attrkinds[lib.VREG] = statmod.S_IFREG
attrkinds[lib.VDIR] = statmod.S_IFDIR
attrkinds[lib.VLNK] = statmod.S_IFLNK
attrkinds[lib.VBLK] = statmod.S_IFBLK
attrkinds[lib.VCHR] = statmod.S_IFCHR
attrkinds[lib.VFIFO] = statmod.S_IFIFO
attrkinds[lib.VSOCK] = statmod.S_IFSOCK
class stat_res(object):
def __init__(self, st_mode, st_mtime, st_size):
self.st_mode = st_mode
self.st_mtime = st_mtime
self.st_size = st_size
tv_sec_ofs = ffi.offsetof("struct timespec", "tv_sec")
buf = ffi.new("char[]", listdir_batch_size)
def listdirinternal(dfd, req, stat, skip):
ret = []
while True:
r = lib.getattrlistbulk(dfd, req, buf, listdir_batch_size, 0)
if r == 0:
break
if r == -1:
raise OSError(ffi.errno, os.strerror(ffi.errno))
cur = ffi.cast("val_attrs_t*", buf)
for i in range(r):
lgt = cur.length
assert lgt == ffi.cast('uint32_t*', cur)[0]
ofs = cur.name_info.attr_dataoffset
str_lgt = cur.name_info.attr_length
base_ofs = ffi.offsetof('val_attrs_t', 'name_info')
name = str(ffi.buffer(ffi.cast("char*", cur) + base_ofs + ofs,
str_lgt - 1))
tp = attrkinds[cur.obj_type]
if name == "." or name == "..":
continue
if skip == name and tp == statmod.S_ISDIR:
return []
if stat:
mtime = cur.mtime.tv_sec
mode = (cur.accessmask & ~lib.S_IFMT)| tp
ret.append((name, tp, stat_res(st_mode=mode, st_mtime=mtime,
st_size=cur.datalength)))
else:
ret.append((name, tp))
cur = ffi.cast("val_attrs_t*", int(ffi.cast("intptr_t", cur))
+ lgt)
return ret
def listdir(path, stat=False, skip=None):
req = ffi.new("struct attrlist*")
req.bitmapcount = lib.ATTR_BIT_MAP_COUNT
req.commonattr = (lib.ATTR_CMN_RETURNED_ATTRS |
lib.ATTR_CMN_NAME |
lib.ATTR_CMN_OBJTYPE |
lib.ATTR_CMN_ACCESSMASK |
lib.ATTR_CMN_MODTIME)
req.fileattr = lib.ATTR_FILE_DATALENGTH
dfd = lib.open(path, lib.O_RDONLY, 0)
if dfd == -1:
raise OSError(ffi.errno, os.strerror(ffi.errno))
try:
ret = listdirinternal(dfd, req, stat, skip)
finally:
try:
lib.close(dfd)
except BaseException:
pass # we ignore all the errors from closing, not
# much we can do about that
return ret