sparse-revlog: implement algorithm to write sparse delta chains (
issue5480)
The classic behavior of revlog._isgooddeltainfo is to consider the span size
of the whole delta chain, and limit it to 4 * textlen.
Once sparse-revlog writing is allowed (and enforced with a requirement),
revlog._isgooddeltainfo considers the span of the largest chunk as the
distance used in the verification, instead of using the span of the whole
delta chain.
In order to compute the span of the largest chunk, we need to slice into
chunks a chain with the new revision at the top of the revlog, and take the
maximal span of these chunks. The sparse read density is a parameter to the
slicing, as it will stop when the global read density reaches this threshold.
For instance, a density of 50% means that 2 of 4 read bytes are actually used
for the reconstruction of the revision (the others are part of other chains).
This allows a new revision to be potentially stored with a diff against
another revision anywhere in the history, instead of forcing it in the last 4
* textlen. The result is a much better compression on repositories that have
many concurrent branches. Here are a comparison between using deltas from
current upstream (aggressive-merge-deltas on by default) and deltas from a
sparse-revlog
Comparison of `.hg/store/` size:
mercurial (6.74% merges):
before: 46,831,873 bytes
after: 46,795,992 bytes (no relevant change)
pypy (8.30% merges):
before: 333,524,651 bytes
after: 308,417,511 bytes -8%
netbeans (34.21% merges):
before: 1,141,847,554 bytes
after: 1,131,093,161 bytes -1%
mozilla-central (4.84% merges):
before: 2,344,248,850 bytes
after: 2,328,459,258 bytes -1%
large-private-repo-A (merge 19.73%)
before: 41,510,550,163 bytes
after: 8,121,763,428 bytes -80%
large-private-repo-B (23.77%)
before: 58,702,221,709 bytes
after: 8,351,588,828 bytes -76%
Comparison of `00manifest.d` size:
mercurial (6.74% merges):
before: 6,143,044 bytes
after: 6,107,163 bytes
pypy (8.30% merges):
before: 52,941,780 bytes
after: 27,834,082 bytes -48%
netbeans (34.21% merges):
before: 130,088,982 bytes
after: 119,337,636 bytes -10%
mozilla-central (4.84% merges):
before: 215,096,339 bytes
after: 199,496,863 bytes -8%
large-private-repo-A (merge 19.73%)
before: 33,725,285,081 bytes
after: 390,302,545 bytes -99%
large-private-repo-B (23.77%)
before: 49,457,701,645 bytes
after: 1,366,752,187 bytes -97%
The better delta chains provide a performance boost in relevant repositories:
pypy, bundling 1000 revisions:
before: 1.670s
after: 1.149s -31%
Unbundling got a bit slower. probably because the sparse algorithm is still
pure
python.
pypy, unbundling 1000 revisions:
before: 4.062s
after: 4.507s +10%
Performance of bundle/unbundle in repository with few concurrent branches (eg:
mercurial) are unaffected.
No significant differences have been noticed then timing `hg push` and `hg
pull` locally. More state timings are being gathered.
Same as for aggressive-merge-delta, better delta comes with longer delta
chains. Longer chains have a performance impact. For example. The length of
the chain needed to get the manifest of pypy's tip moves from 82 item to 1929
items. This moves the restore time from 3.88ms to 11.3ms.
Delta chain length is an independent issue that affects repository without
this changes. It will be dealt with independently.
No significant differences have been observed on repositories where
`sparse-revlog` have not much effect (mercurial, unity, netbeans). On pypy,
small differences have been observed on some operation affected by delta chain
building and retrieval.
pypy, perfmanifest
before: 0.006162s
after: 0.017899s +190%
pypy, commit:
before: 0.382
after: 0.376 -1%
pypy, status:
before: 0.157
after: 0.168 +7%
More comprehensive and stable timing comparisons are in progress.
#require chg
$ cp $HGRCPATH $HGRCPATH.orig
init repo
$ chg init foo
$ cd foo
ill-formed config
$ chg status
$ echo '=brokenconfig' >> $HGRCPATH
$ chg status
hg: parse error at * (glob)
[255]
$ cp $HGRCPATH.orig $HGRCPATH
long socket path
$ sockpath=$TESTTMP/this/path/should/be/longer/than/one-hundred-and-seven/characters/where/107/is/the/typical/size/limit/of/unix-domain-socket
$ mkdir -p $sockpath
$ bakchgsockname=$CHGSOCKNAME
$ CHGSOCKNAME=$sockpath/server
$ export CHGSOCKNAME
$ chg root
$TESTTMP/foo
$ rm -rf $sockpath
$ CHGSOCKNAME=$bakchgsockname
$ export CHGSOCKNAME
$ cd ..
editor
------
$ cat >> pushbuffer.py <<EOF
> def reposetup(ui, repo):
> repo.ui.pushbuffer(subproc=True)
> EOF
$ chg init editor
$ cd editor
by default, system() should be redirected to the client:
$ touch foo
$ CHGDEBUG= HGEDITOR=cat chg ci -Am channeled --edit 2>&1 \
> | egrep "HG:|run 'cat"
chg: debug: * run 'cat "*"' at '$TESTTMP/editor' (glob)
HG: Enter commit message. Lines beginning with 'HG:' are removed.
HG: Leave message empty to abort commit.
HG: --
HG: user: test
HG: branch 'default'
HG: added foo
but no redirection should be made if output is captured:
$ touch bar
$ CHGDEBUG= HGEDITOR=cat chg ci -Am bufferred --edit \
> --config extensions.pushbuffer="$TESTTMP/pushbuffer.py" 2>&1 \
> | egrep "HG:|run 'cat"
[1]
check that commit commands succeeded:
$ hg log -T '{rev}:{desc}\n'
1:bufferred
0:channeled
$ cd ..
pager
-----
$ cat >> fakepager.py <<EOF
> import sys
> for line in sys.stdin:
> sys.stdout.write('paged! %r\n' % line)
> EOF
enable pager extension globally, but spawns the master server with no tty:
$ chg init pager
$ cd pager
$ cat >> $HGRCPATH <<EOF
> [extensions]
> pager =
> [pager]
> pager = $PYTHON $TESTTMP/fakepager.py
> EOF
$ chg version > /dev/null
$ touch foo
$ chg ci -qAm foo
pager should be enabled if the attached client has a tty:
$ chg log -l1 -q --config ui.formatted=True
paged! '0:1f7b0de80e11\n'
$ chg log -l1 -q --config ui.formatted=False
0:1f7b0de80e11
chg waits for pager if runcommand raises
$ cat > $TESTTMP/crash.py <<EOF
> from mercurial import registrar
> cmdtable = {}
> command = registrar.command(cmdtable)
> @command(b'crash')
> def pagercrash(ui, repo, *pats, **opts):
> ui.write('going to crash\n')
> raise Exception('.')
> EOF
$ cat > $TESTTMP/fakepager.py <<EOF
> from __future__ import absolute_import
> import sys
> import time
> for line in iter(sys.stdin.readline, ''):
> if 'crash' in line: # only interested in lines containing 'crash'
> # if chg exits when pager is sleeping (incorrectly), the output
> # will be captured by the next test case
> time.sleep(1)
> sys.stdout.write('crash-pager: %s' % line)
> EOF
$ cat >> .hg/hgrc <<EOF
> [extensions]
> crash = $TESTTMP/crash.py
> EOF
$ chg crash --pager=on --config ui.formatted=True 2>/dev/null
crash-pager: going to crash
[255]
$ cd ..
server lifecycle
----------------
chg server should be restarted on code change, and old server will shut down
automatically. In this test, we use the following time parameters:
- "sleep 1" to make mtime different
- "sleep 2" to notice mtime change (polling interval is 1 sec)
set up repository with an extension:
$ chg init extreload
$ cd extreload
$ touch dummyext.py
$ cat <<EOF >> .hg/hgrc
> [extensions]
> dummyext = dummyext.py
> EOF
isolate socket directory for stable result:
$ OLDCHGSOCKNAME=$CHGSOCKNAME
$ mkdir chgsock
$ CHGSOCKNAME=`pwd`/chgsock/server
warm up server:
$ CHGDEBUG= chg log 2>&1 | egrep 'instruction|start'
chg: debug: * start cmdserver at $TESTTMP/extreload/chgsock/server.* (glob)
new server should be started if extension modified:
$ sleep 1
$ touch dummyext.py
$ CHGDEBUG= chg log 2>&1 | egrep 'instruction|start'
chg: debug: * instruction: unlink $TESTTMP/extreload/chgsock/server-* (glob)
chg: debug: * instruction: reconnect (glob)
chg: debug: * start cmdserver at $TESTTMP/extreload/chgsock/server.* (glob)
old server will shut down, while new server should still be reachable:
$ sleep 2
$ CHGDEBUG= chg log 2>&1 | (egrep 'instruction|start' || true)
socket file should never be unlinked by old server:
(simulates unowned socket by updating mtime, which makes sure server exits
at polling cycle)
$ ls chgsock/server-*
chgsock/server-* (glob)
$ touch chgsock/server-*
$ sleep 2
$ ls chgsock/server-*
chgsock/server-* (glob)
since no server is reachable from socket file, new server should be started:
(this test makes sure that old server shut down automatically)
$ CHGDEBUG= chg log 2>&1 | egrep 'instruction|start'
chg: debug: * start cmdserver at $TESTTMP/extreload/chgsock/server.* (glob)
shut down servers and restore environment:
$ rm -R chgsock
$ CHGSOCKNAME=$OLDCHGSOCKNAME
$ cd ..