Mercurial > hg
view mercurial/cffi/osutil.py @ 39872:733db72f0f54
revlog: move revision verification out of verify
File revision verification is performing low-level checks of file
storage, namely that flags are appropriate and revision data can
be resolved.
Since these checks are somewhat revlog-specific and may not
be appropriate for alternate storage backends, this commit moves
those checks from verify.py to revlog.py.
Because we're now emitting warnings/errors that apply to specific
revisions, we taught the iverifyproblem interface to expose the
problematic node and to report this node in verify output. This
was necessary to prevent unwanted test changes.
After this change, revlog.verifyintegrity() and file verify code
in verify.py both iterate over revisions and resolve their fulltext.
But they do so in separate loops. (verify.py needs to resolve
fulltexts as part of calling renamed() - at least when using revlogs.)
This should add overhead.
But on the mozilla-unified repo:
$ hg verify
before: time: real 700.640 secs (user 585.520+0.000 sys 23.480+0.000)
after: time: real 682.380 secs (user 570.370+0.000 sys 22.240+0.000)
I'm not sure what's going on. Maybe avoiding the filelog attribute
proxies shaved off enough time to offset the losses? Maybe fulltext
resolution has less overhead than I thought?
I've left a comment indicating the potential for optimization. But
because it doesn't produce a performance regression on a large
repository, I'm not going to worry about it.
Differential Revision: https://phab.mercurial-scm.org/D4745
author | Gregory Szorc <gregory.szorc@gmail.com> |
---|---|
date | Mon, 24 Sep 2018 11:27:47 -0700 |
parents | dacfcdd8b94e |
children | 2372284d9457 |
line wrap: on
line source
# 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