Mercurial > hg-stable
view mercurial/filelog.py @ 30451:41a8106789ca
util: implement zstd compression engine
Now that zstd is vendored and being built (in some configurations), we
can implement a compression engine for zstd!
The zstd engine is a little different from existing engines. Because
it may not always be present, we have to defer load the module in case
importing it fails. We facilitate this via a cached property that holds
a reference to the module or None. The "available" method is
implemented to reflect reality.
The zstd engine declares its ability to handle bundles using the
"zstd" human name and the "ZS" internal name. The latter was chosen
because internal names are 2 characters (by only convention I think)
and "ZS" seems reasonable.
The engine, like others, supports specifying the compression level.
However, there are no consumers of this API that yet pass in that
argument. I have plans to change that, so stay tuned.
Since all we need to do to support bundle generation with a new
compression engine is implement and register the compression engine,
bundle generation with zstd "just works!" Tests demonstrating this
have been added.
How does performance of zstd for bundle generation compare? On the
mozilla-unified repo, `hg bundle --all -t <engine>-v2` yields the
following on my i7-6700K on Linux:
engine CPU time bundle size vs orig size throughput
none 97.0s 4,054,405,584 100.0% 41.8 MB/s
bzip2 (l=9) 393.6s 975,343,098 24.0% 10.3 MB/s
gzip (l=6) 184.0s 1,140,533,074 28.1% 22.0 MB/s
zstd (l=1) 108.2s 1,119,434,718 27.6% 37.5 MB/s
zstd (l=2) 111.3s 1,078,328,002 26.6% 36.4 MB/s
zstd (l=3) 113.7s 1,011,823,727 25.0% 35.7 MB/s
zstd (l=4) 116.0s 1,008,965,888 24.9% 35.0 MB/s
zstd (l=5) 121.0s 977,203,148 24.1% 33.5 MB/s
zstd (l=6) 131.7s 927,360,198 22.9% 30.8 MB/s
zstd (l=7) 139.0s 912,808,505 22.5% 29.2 MB/s
zstd (l=12) 198.1s 854,527,714 21.1% 20.5 MB/s
zstd (l=18) 681.6s 789,750,690 19.5% 5.9 MB/s
On compression, zstd for bundle generation delivers:
* better compression than gzip with significantly less CPU utilization
* better than bzip2 compression ratios while still being significantly
faster than gzip
* ability to aggressively tune compression level to achieve
significantly smaller bundles
That last point is important. With clone bundles, a server can
pre-generate a bundle file, upload it to a static file server, and
redirect clients to transparently download it during clone. The server
could choose to produce a zstd bundle with the highest compression
settings possible. This would take a very long time - a magnitude
longer than a typical zstd bundle generation - but the result would
be hundreds of megabytes smaller! For the clone volume we do at
Mozilla, this could translate to petabytes of bandwidth savings
per year and faster clones (due to smaller transfer size).
I don't have detailed numbers to report on decompression. However,
zstd decompression is fast: >1 GB/s output throughput on this machine,
even through the Python bindings. And it can do that regardless of the
compression level of the input. By the time you have enough data to
worry about overhead of decompression, you have plenty of other things
to worry about performance wise.
zstd is wins all around. I can't wait to implement support for it
on the wire protocol and in revlogs.
author | Gregory Szorc <gregory.szorc@gmail.com> |
---|---|
date | Fri, 11 Nov 2016 01:10:07 -0800 |
parents | 34bd1a5eef5b |
children | be5b2098a817 |
line wrap: on
line source
# filelog.py - file history class for mercurial # # Copyright 2005-2007 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 re import struct from . import ( error, mdiff, revlog, ) _mdre = re.compile('\1\n') def parsemeta(text): """return (metadatadict, keylist, metadatasize)""" # text can be buffer, so we can't use .startswith or .index if text[:2] != '\1\n': return None, None s = _mdre.search(text, 2).start() mtext = text[2:s] meta = {} for l in mtext.splitlines(): k, v = l.split(": ", 1) meta[k] = v return meta, (s + 2) def packmeta(meta, text): keys = sorted(meta.iterkeys()) metatext = "".join("%s: %s\n" % (k, meta[k]) for k in keys) return "\1\n%s\1\n%s" % (metatext, text) def _censoredtext(text): m, offs = parsemeta(text) return m and "censored" in m class filelog(revlog.revlog): def __init__(self, opener, path): super(filelog, self).__init__(opener, "/".join(("data", path + ".i"))) def read(self, node): t = self.revision(node) if not t.startswith('\1\n'): return t s = t.index('\1\n', 2) return t[s + 2:] def add(self, text, meta, transaction, link, p1=None, p2=None): if meta or text.startswith('\1\n'): text = packmeta(meta, text) return self.addrevision(text, transaction, link, p1, p2) def renamed(self, node): if self.parents(node)[0] != revlog.nullid: return False t = self.revision(node) m = parsemeta(t)[0] if m and "copy" in m: return (m["copy"], revlog.bin(m["copyrev"])) return False def size(self, rev): """return the size of a given revision""" # for revisions with renames, we have to go the slow way node = self.node(rev) if self.renamed(node): return len(self.read(node)) if self.iscensored(rev): return 0 # XXX if self.read(node).startswith("\1\n"), this returns (size+4) return super(filelog, self).size(rev) def cmp(self, node, text): """compare text with a given file revision returns True if text is different than what is stored. """ t = text if text.startswith('\1\n'): t = '\1\n\1\n' + text samehashes = not super(filelog, self).cmp(node, t) if samehashes: return False # censored files compare against the empty file if self.iscensored(self.rev(node)): return text != '' # renaming a file produces a different hash, even if the data # remains unchanged. Check if it's the case (slow): if self.renamed(node): t2 = self.read(node) return t2 != text return True def checkhash(self, text, p1, p2, node, rev=None): try: super(filelog, self).checkhash(text, p1, p2, node, rev=rev) except error.RevlogError: if _censoredtext(text): raise error.CensoredNodeError(self.indexfile, node, text) raise def iscensored(self, rev): """Check if a file revision is censored.""" return self.flags(rev) & revlog.REVIDX_ISCENSORED def _peek_iscensored(self, baserev, delta, flush): """Quickly check if a delta produces a censored revision.""" # Fragile heuristic: unless new file meta keys are added alphabetically # preceding "censored", all censored revisions are prefixed by # "\1\ncensored:". A delta producing such a censored revision must be a # full-replacement delta, so we inspect the first and only patch in the # delta for this prefix. hlen = struct.calcsize(">lll") if len(delta) <= hlen: return False oldlen = self.rawsize(baserev) newlen = len(delta) - hlen if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): return False add = "\1\ncensored:" addlen = len(add) return newlen >= addlen and delta[hlen:hlen + addlen] == add