revlog: linearize created changegroups in generaldelta revlogs
This greatly improves the speed of the bundling process, and often reduces the
bundle size considerably. (Although if the repository is already ordered, this
has little effect on both time and bundle size.)
For non-generaldelta clients, the reduced bundle size translates to a reduced
repository size, similar to shrinking the revlogs (which uses the exact same
algorithm). For generaldelta clients the difference is minor.
When the new bundle format comes, reordering will not be necessary since we
can then store the deltaparent relationsships directly. The eventual default
behavior for clients and servers is presented in the table below, where "new"
implies support for GD as well as the new bundle format:
old client new client
old server old bundle, no reorder old bundle, no reorder
new server, non-GD old bundle, no reorder[1] old bundle, no reorder[2]
new server, GD old bundle, reorder[3] new bundle, no reorder[4]
[1] reordering is expensive on the server in this case, skip it
[2] client can choose to do its own redelta here
[3] reordering is needed because otherwise the pull does a lot of extra
work on the server
[4] reordering isn't needed because client can get deltabase in bundle
format
Currently, the default is to reorder on GD-servers, and not otherwise. A new
setting, bundle.reorder, has been added to override the default reordering
behavior. It can be set to either 'auto' (the default), or any true or false
value as a standard boolean setting, to either force the reordering on or off
regardless of generaldelta.
Some timing data from a relatively branch test repository follows. All
bundling is done with --all --type none options.
Non-generaldelta, non-shrunk repo:
-----------------------------------
Size: 276M
Without reorder (default):
Bundle time: 14.4 seconds
Bundle size: 939M
With reorder:
Bundle time: 1 minute, 29.3 seconds
Bundle size: 381M
Generaldelta, non-shrunk repo:
-----------------------------------
Size: 87M
Without reorder:
Bundle time: 2 minutes, 1.4 seconds
Bundle size: 939M
With reorder (default):
Bundle time: 25.5 seconds
Bundle size: 381M
--- a/mercurial/localrepo.py Wed May 18 23:11:34 2011 +0200
+++ b/mercurial/localrepo.py Wed May 18 23:26:26 2011 +0200
@@ -1544,18 +1544,23 @@
return fstate[1][x]
bundler = changegroup.bundle10(lookup)
+ reorder = self.ui.config('bundle', 'reorder', 'auto')
+ if reorder == 'auto':
+ reorder = None
+ else:
+ reorder = util.parsebool(reorder)
def gengroup():
# Create a changenode group generator that will call our functions
# back to lookup the owning changenode and collect information.
- for chunk in cl.group(csets, bundler):
+ for chunk in cl.group(csets, bundler, reorder=reorder):
yield chunk
self.ui.progress(_('bundling'), None)
# Create a generator for the manifestnodes that calls our lookup
# and data collection functions back.
count[0] = 0
- for chunk in mf.group(prune(mf, mfs), bundler):
+ for chunk in mf.group(prune(mf, mfs), bundler, reorder=reorder):
yield chunk
self.ui.progress(_('bundling'), None)
@@ -1572,7 +1577,7 @@
first = True
for chunk in filerevlog.group(prune(filerevlog, fstate[1]),
- bundler):
+ bundler, reorder=reorder):
if first:
if chunk == bundler.close():
break
@@ -1640,17 +1645,22 @@
return cl.node(revlog.linkrev(revlog.rev(x)))
bundler = changegroup.bundle10(lookup)
+ reorder = self.ui.config('bundle', 'reorder', 'auto')
+ if reorder == 'auto':
+ reorder = None
+ else:
+ reorder = util.parsebool(reorder)
def gengroup():
'''yield a sequence of changegroup chunks (strings)'''
# construct a list of all changed files
- for chunk in cl.group(nodes, bundler):
+ for chunk in cl.group(nodes, bundler, reorder=reorder):
yield chunk
self.ui.progress(_('bundling'), None)
count[0] = 0
- for chunk in mf.group(gennodelst(mf), bundler):
+ for chunk in mf.group(gennodelst(mf), bundler, reorder=reorder):
yield chunk
self.ui.progress(_('bundling'), None)
@@ -1661,7 +1671,8 @@
raise util.Abort(_("empty or missing revlog for %s") % fname)
fstate[0] = fname
first = True
- for chunk in filerevlog.group(gennodelst(filerevlog), bundler):
+ for chunk in filerevlog.group(gennodelst(filerevlog), bundler,
+ reorder=reorder):
if first:
if chunk == bundler.close():
break
--- a/mercurial/revlog.py Wed May 18 23:11:34 2011 +0200
+++ b/mercurial/revlog.py Wed May 18 23:26:26 2011 +0200
@@ -14,7 +14,7 @@
# import stuff from node for others to import from revlog
from node import bin, hex, nullid, nullrev, short #@UnusedImport
from i18n import _
-import ancestor, mdiff, parsers, error, util
+import ancestor, mdiff, parsers, error, util, dagutil
import struct, zlib, errno
_pack = struct.pack
@@ -1086,7 +1086,7 @@
self._basecache = (curr, chainbase)
return node
- def group(self, nodelist, bundler):
+ def group(self, nodelist, bundler, reorder=None):
"""Calculate a delta group, yielding a sequence of changegroup chunks
(strings).
@@ -1098,7 +1098,14 @@
changegroup starts with a full revision.
"""
- revs = sorted([self.rev(n) for n in nodelist])
+ # for generaldelta revlogs, we linearize the revs; this will both be
+ # much quicker and generate a much smaller bundle
+ if (self._generaldelta and reorder is not False) or reorder:
+ dag = dagutil.revlogdag(self)
+ revs = set(self.rev(n) for n in nodelist)
+ revs = dag.linearize(revs)
+ else:
+ revs = sorted([self.rev(n) for n in nodelist])
# if we don't have any revisions touched by these changesets, bail
if not revs: