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 svn15
$ SVNREPOPATH=`pwd`/svn-repo
#if windows
$ SVNREPOURL=file:///`$PYTHON -c "import urllib, sys; sys.stdout.write(urllib.quote(sys.argv[1]))" "$SVNREPOPATH"`
#else
$ SVNREPOURL=file://`$PYTHON -c "import urllib, sys; sys.stdout.write(urllib.quote(sys.argv[1]))" "$SVNREPOPATH"`
#endif
$ filter_svn_output () {
> egrep -v 'Committing|Transmitting|Updating|(^$)' || true
> }
create subversion repo
$ WCROOT="`pwd`/svn-wc"
$ svnadmin create svn-repo
$ svn co "$SVNREPOURL" svn-wc
Checked out revision 0.
$ cd svn-wc
$ mkdir src
$ echo alpha > src/alpha
$ svn add src
A src
A src/alpha
$ mkdir externals
$ echo other > externals/other
$ svn add externals
A externals
A externals/other
$ svn ci -qm 'Add alpha'
$ svn up -q
$ echo "externals -r1 $SVNREPOURL/externals" > extdef
$ svn propset -F extdef svn:externals src
property 'svn:externals' set on 'src'
$ svn ci -qm 'Setting externals'
$ cd ..
create hg repo
$ mkdir sub
$ cd sub
$ hg init t
$ cd t
first revision, no sub
$ echo a > a
$ hg ci -Am0
adding a
add first svn sub with leading whitespaces
$ echo "s = [svn] $SVNREPOURL/src" >> .hgsub
$ echo "subdir/s = [svn] $SVNREPOURL/src" >> .hgsub
$ svn co --quiet "$SVNREPOURL"/src s
$ mkdir subdir
$ svn co --quiet "$SVNREPOURL"/src subdir/s
$ hg add .hgsub
svn subrepo is disabled by default
$ hg ci -m1
abort: svn subrepos not allowed
(see 'hg help config.subrepos' for details)
[255]
so enable it
$ cat >> $HGRCPATH <<EOF
> [subrepos]
> svn:allowed = true
> EOF
$ hg ci -m1
make sure we avoid empty commits (issue2445)
$ hg sum
parent: 1:* tip (glob)
1
branch: default
commit: (clean)
update: (current)
phases: 2 draft
$ hg ci -moops
nothing changed
[1]
debugsub
$ hg debugsub
path s
source file:/*/$TESTTMP/svn-repo/src (glob)
revision 2
path subdir/s
source file:/*/$TESTTMP/svn-repo/src (glob)
revision 2
change file in svn and hg, commit
$ echo a >> a
$ echo alpha >> s/alpha
$ hg sum
parent: 1:* tip (glob)
1
branch: default
commit: 1 modified, 1 subrepos
update: (current)
phases: 2 draft
$ hg commit --subrepos -m 'Message!' | filter_svn_output
committing subrepository s
Sending*s/alpha (glob)
Committed revision 3.
Fetching external item into '*s/externals'* (glob)
External at revision 1.
At revision 3.
$ hg debugsub
path s
source file:/*/$TESTTMP/svn-repo/src (glob)
revision 3
path subdir/s
source file:/*/$TESTTMP/svn-repo/src (glob)
revision 2
missing svn file, commit should fail
$ rm s/alpha
$ hg commit --subrepos -m 'abort on missing file'
committing subrepository s
abort: cannot commit missing svn entries (in subrepository "s")
[255]
$ svn revert s/alpha > /dev/null
add an unrelated revision in svn and update the subrepo to without
bringing any changes.
$ svn mkdir "$SVNREPOURL/unrelated" -qm 'create unrelated'
$ svn up -q s
$ hg sum
parent: 2:* tip (glob)
Message!
branch: default
commit: (clean)
update: (current)
phases: 3 draft
$ echo a > s/a
should be empty despite change to s/a
$ hg st
add a commit from svn
$ cd "$WCROOT/src"
$ svn up -q
$ echo xyz >> alpha
$ svn propset svn:mime-type 'text/xml' alpha
property 'svn:mime-type' set on 'alpha'
$ svn ci -qm 'amend a from svn'
$ cd ../../sub/t
this commit from hg will fail
$ echo zzz >> s/alpha
$ (hg ci --subrepos -m 'amend alpha from hg' 2>&1; echo "[$?]") | grep -vi 'out of date'
committing subrepository s
abort: svn:*Commit failed (details follow): (glob)
[255]
$ svn revert -q s/alpha
this commit fails because of meta changes
$ svn propset svn:mime-type 'text/html' s/alpha
property 'svn:mime-type' set on 's/alpha'
$ (hg ci --subrepos -m 'amend alpha from hg' 2>&1; echo "[$?]") | grep -vi 'out of date'
committing subrepository s
abort: svn:*Commit failed (details follow): (glob)
[255]
$ svn revert -q s/alpha
this commit fails because of externals changes
$ echo zzz > s/externals/other
$ hg ci --subrepos -m 'amend externals from hg'
committing subrepository s
abort: cannot commit svn externals (in subrepository "s")
[255]
$ hg diff --subrepos -r 1:2 | grep -v diff
--- a/.hgsubstate Thu Jan 01 00:00:00 1970 +0000
+++ b/.hgsubstate Thu Jan 01 00:00:00 1970 +0000
@@ -1,2 +1,2 @@
-2 s
+3 s
2 subdir/s
--- a/a Thu Jan 01 00:00:00 1970 +0000
+++ b/a Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +1,2 @@
a
+a
$ svn revert -q s/externals/other
this commit fails because of externals meta changes
$ svn propset svn:mime-type 'text/html' s/externals/other
property 'svn:mime-type' set on 's/externals/other'
$ hg ci --subrepos -m 'amend externals from hg'
committing subrepository s
abort: cannot commit svn externals (in subrepository "s")
[255]
$ svn revert -q s/externals/other
clone
$ cd ..
$ hg clone t tc
updating to branch default
A tc/s/alpha
U tc/s
Fetching external item into 'tc/s/externals'* (glob)
A tc/s/externals/other
Checked out external at revision 1.
Checked out revision 3.
A tc/subdir/s/alpha
U tc/subdir/s
Fetching external item into 'tc/subdir/s/externals'* (glob)
A tc/subdir/s/externals/other
Checked out external at revision 1.
Checked out revision 2.
3 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cd tc
debugsub in clone
$ hg debugsub
path s
source file:/*/$TESTTMP/svn-repo/src (glob)
revision 3
path subdir/s
source file:/*/$TESTTMP/svn-repo/src (glob)
revision 2
verify subrepo is contained within the repo directory
$ $PYTHON -c "import os.path; print os.path.exists('s')"
True
update to nullrev (must delete the subrepo)
$ hg up null
0 files updated, 0 files merged, 3 files removed, 0 files unresolved
$ ls
Check hg update --clean
$ cd "$TESTTMP/sub/t"
$ cd s
$ echo c0 > alpha
$ echo c1 > f1
$ echo c1 > f2
$ svn add f1 -q
$ svn status | sort
? * a (glob)
? * f2 (glob)
A * f1 (glob)
M * alpha (glob)
Performing status on external item at 'externals'* (glob)
X * externals (glob)
$ cd ../..
$ hg -R t update -C
Fetching external item into 't/s/externals'* (glob)
Checked out external at revision 1.
Checked out revision 3.
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cd t/s
$ svn status | sort
? * a (glob)
? * f1 (glob)
? * f2 (glob)
Performing status on external item at 'externals'* (glob)
X * externals (glob)
Sticky subrepositories, no changes
$ cd "$TESTTMP/sub/t"
$ hg id -n
2
$ cd s
$ svnversion
3
$ cd ..
$ hg update 1
U *s/alpha (glob)
Fetching external item into '*s/externals'* (glob)
Checked out external at revision 1.
Checked out revision 2.
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg id -n
1
$ cd s
$ svnversion
2
$ cd ..
Sticky subrepositories, file changes
$ touch s/f1
$ cd s
$ svn add f1
A f1
$ cd ..
$ hg id -n
1+
$ cd s
$ svnversion
2M
$ cd ..
$ hg update tip
subrepository s diverged (local revision: 2, remote revision: 3)
(M)erge, keep (l)ocal [working copy] or keep (r)emote [destination]? m
subrepository sources for s differ
use (l)ocal source (2) or (r)emote source (3)? l
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg id -n
2+
$ cd s
$ svnversion
2M
$ cd ..
$ hg update --clean tip
U *s/alpha (glob)
Fetching external item into '*s/externals'* (glob)
Checked out external at revision 1.
Checked out revision 3.
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
Sticky subrepository, revision updates
$ hg id -n
2
$ cd s
$ svnversion
3
$ cd ..
$ cd s
$ svn update -qr 1
$ cd ..
$ hg update 1
subrepository s diverged (local revision: 3, remote revision: 2)
(M)erge, keep (l)ocal [working copy] or keep (r)emote [destination]? m
subrepository sources for s differ (in checked out version)
use (l)ocal source (1) or (r)emote source (2)? l
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg id -n
1+
$ cd s
$ svnversion
1
$ cd ..
Sticky subrepository, file changes and revision updates
$ touch s/f1
$ cd s
$ svn add f1
A f1
$ svnversion
1M
$ cd ..
$ hg id -n
1+
$ hg update tip
subrepository s diverged (local revision: 3, remote revision: 3)
(M)erge, keep (l)ocal [working copy] or keep (r)emote [destination]? m
subrepository sources for s differ
use (l)ocal source (1) or (r)emote source (3)? l
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg id -n
2+
$ cd s
$ svnversion
1M
$ cd ..
Sticky repository, update --clean
$ hg update --clean tip | grep -v 's[/\]externals[/\]other'
U *s/alpha (glob)
U *s (glob)
Fetching external item into '*s/externals'* (glob)
Checked out external at revision 1.
Checked out revision 3.
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg id -n
2
$ cd s
$ svnversion
3
$ cd ..
Test subrepo already at intended revision:
$ cd s
$ svn update -qr 2
$ cd ..
$ hg update 1
subrepository s diverged (local revision: 3, remote revision: 2)
(M)erge, keep (l)ocal [working copy] or keep (r)emote [destination]? m
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg id -n
1+
$ cd s
$ svnversion
2
$ cd ..
Test case where subversion would fail to update the subrepo because there
are unknown directories being replaced by tracked ones (happens with rebase).
$ cd "$WCROOT/src"
$ mkdir dir
$ echo epsilon.py > dir/epsilon.py
$ svn add dir
A dir
A dir/epsilon.py
$ svn ci -qm 'Add dir/epsilon.py'
$ cd ../..
$ hg init rebaserepo
$ cd rebaserepo
$ svn co -r5 --quiet "$SVNREPOURL"/src s
$ echo "s = [svn] $SVNREPOURL/src" >> .hgsub
$ hg add .hgsub
$ hg ci -m addsub
$ echo a > a
$ hg add .
adding a
$ hg ci -m adda
$ hg up 0
0 files updated, 0 files merged, 1 files removed, 0 files unresolved
$ svn up -qr6 s
$ hg ci -m updatesub
created new head
$ echo pyc > s/dir/epsilon.pyc
$ hg up 1
D *s/dir (glob)
Fetching external item into '*s/externals'* (glob)
Checked out external at revision 1.
Checked out revision 5.
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg up -q 2
Modify one of the externals to point to a different path so we can
test having obstructions when switching branches on checkout:
$ hg checkout tip
0 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ echo "obstruct = [svn] $SVNREPOURL/externals" >> .hgsub
$ svn co -r5 --quiet "$SVNREPOURL"/externals obstruct
$ hg commit -m 'Start making obstructed working copy'
$ hg book other
$ hg co -r 'p1(tip)'
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
(leaving bookmark other)
$ echo "obstruct = [svn] $SVNREPOURL/src" >> .hgsub
$ svn co -r5 --quiet "$SVNREPOURL"/src obstruct
$ hg commit -m 'Other branch which will be obstructed'
created new head
Switching back to the head where we have another path mapped to the
same subrepo should work if the subrepo is clean.
$ hg co other
A *obstruct/other (glob)
Checked out revision 1.
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
(activating bookmark other)
This is surprising, but is also correct based on the current code:
$ echo "updating should (maybe) fail" > obstruct/other
$ hg co tip
abort: uncommitted changes
(commit or update --clean to discard changes)
[255]
Point to a Subversion branch which has since been deleted and recreated
First, create that condition in the repository.
$ hg ci --subrepos -m cleanup | filter_svn_output
committing subrepository obstruct
Sending obstruct/other
Committed revision 7.
At revision 7.
$ svn mkdir -qm "baseline" $SVNREPOURL/trunk
$ svn copy -qm "initial branch" $SVNREPOURL/trunk $SVNREPOURL/branch
$ svn co --quiet "$SVNREPOURL"/branch tempwc
$ cd tempwc
$ echo "something old" > somethingold
$ svn add somethingold
A somethingold
$ svn ci -qm 'Something old'
$ svn rm -qm "remove branch" $SVNREPOURL/branch
$ svn copy -qm "recreate branch" $SVNREPOURL/trunk $SVNREPOURL/branch
$ svn up -q
$ echo "something new" > somethingnew
$ svn add somethingnew
A somethingnew
$ svn ci -qm 'Something new'
$ cd ..
$ rm -rf tempwc
$ svn co "$SVNREPOURL/branch"@10 recreated
A recreated/somethingold
Checked out revision 10.
$ echo "recreated = [svn] $SVNREPOURL/branch" >> .hgsub
$ hg ci -m addsub
$ cd recreated
$ svn up -q
$ cd ..
$ hg ci -m updatesub
$ hg up -r-2
D *recreated/somethingnew (glob)
A *recreated/somethingold (glob)
Checked out revision 10.
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
(leaving bookmark other)
$ test -f recreated/somethingold
Test archive
$ hg archive -S ../archive-all --debug --config progress.debug=true
archiving: 0/2 files (0.00%)
archiving: .hgsub 1/2 files (50.00%)
archiving: .hgsubstate 2/2 files (100.00%)
archiving (obstruct): 0/1 files (0.00%)
archiving (obstruct): 1/1 files (100.00%)
archiving (recreated): 0/1 files (0.00%)
archiving (recreated): 1/1 files (100.00%)
archiving (s): 0/2 files (0.00%)
archiving (s): 1/2 files (50.00%)
archiving (s): 2/2 files (100.00%)
$ hg archive -S ../archive-exclude --debug --config progress.debug=true -X **old
archiving: 0/2 files (0.00%)
archiving: .hgsub 1/2 files (50.00%)
archiving: .hgsubstate 2/2 files (100.00%)
archiving (obstruct): 0/1 files (0.00%)
archiving (obstruct): 1/1 files (100.00%)
archiving (recreated): 0 files
archiving (s): 0/2 files (0.00%)
archiving (s): 1/2 files (50.00%)
archiving (s): 2/2 files (100.00%)
$ find ../archive-exclude | sort
../archive-exclude
../archive-exclude/.hg_archival.txt
../archive-exclude/.hgsub
../archive-exclude/.hgsubstate
../archive-exclude/obstruct
../archive-exclude/obstruct/other
../archive-exclude/s
../archive-exclude/s/alpha
../archive-exclude/s/dir
../archive-exclude/s/dir/epsilon.py
Test forgetting files, not implemented in svn subrepo, used to
traceback
$ hg forget 'notafile*'
notafile*: $ENOENT$
[1]
Test a subrepo referencing a just moved svn path. Last commit rev will
be different from the revision, and the path will be different as
well.
$ cd "$WCROOT"
$ svn up > /dev/null
$ mkdir trunk/subdir branches
$ echo a > trunk/subdir/a
$ svn add trunk/subdir branches
A trunk/subdir
A trunk/subdir/a
A branches
$ svn ci -qm addsubdir
$ svn cp -qm branchtrunk $SVNREPOURL/trunk $SVNREPOURL/branches/somebranch
$ cd ..
$ hg init repo2
$ cd repo2
$ svn co $SVNREPOURL/branches/somebranch/subdir
A subdir/a
Checked out revision 15.
$ echo "subdir = [svn] $SVNREPOURL/branches/somebranch/subdir" > .hgsub
$ hg add .hgsub
$ hg ci -m addsub
$ hg up null
0 files updated, 0 files merged, 2 files removed, 0 files unresolved
$ hg up
A *subdir/a (glob)
Checked out revision 15.
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cd ..
Test sanitizing ".hg/hgrc" in subrepo
$ cd sub/t
$ hg update -q -C tip
$ cd s
$ mkdir .hg
$ echo '.hg/hgrc in svn repo' > .hg/hgrc
$ mkdir -p sub/.hg
$ echo 'sub/.hg/hgrc in svn repo' > sub/.hg/hgrc
$ svn add .hg sub
A .hg
A .hg/hgrc
A sub
A sub/.hg
A sub/.hg/hgrc
$ svn ci -qm 'add .hg/hgrc to be sanitized at hg update'
$ svn up -q
$ cd ..
$ hg commit -S -m 'commit with svn revision including .hg/hgrc'
$ grep ' s$' .hgsubstate
16 s
$ cd ..
$ hg -R tc pull -u -q 2>&1 | sort
warning: removing potentially hostile 'hgrc' in '$TESTTMP/sub/tc/s/.hg'
warning: removing potentially hostile 'hgrc' in '$TESTTMP/sub/tc/s/sub/.hg'
$ cd tc
$ grep ' s$' .hgsubstate
16 s
$ test -f s/.hg/hgrc
[1]
$ test -f s/sub/.hg/hgrc
[1]
Test that sanitizing is omitted in meta data area:
$ mkdir s/.svn/.hg
$ echo '.hg/hgrc in svn metadata area' > s/.svn/.hg/hgrc
$ hg update -q -C '.^1'
$ cd ../..
SEC: test for ssh exploit
$ hg init ssh-vuln
$ cd ssh-vuln
$ echo "s = [svn]$SVNREPOURL/src" >> .hgsub
$ svn co --quiet "$SVNREPOURL"/src s
$ hg add .hgsub
$ hg ci -m1
$ echo "s = [svn]svn+ssh://-oProxyCommand=touch%20owned%20nested" > .hgsub
$ hg ci -m2
$ cd ..
$ hg clone ssh-vuln ssh-vuln-clone
updating to branch default
abort: potentially unsafe url: 'svn+ssh://-oProxyCommand=touch owned nested' (in subrepository "s")
[255]
also check that a percent encoded '-' (%2D) doesn't work
$ cd ssh-vuln
$ echo "s = [svn]svn+ssh://%2DoProxyCommand=touch%20owned%20nested" > .hgsub
$ hg ci -m3
$ cd ..
$ rm -r ssh-vuln-clone
$ hg clone ssh-vuln ssh-vuln-clone
updating to branch default
abort: potentially unsafe url: 'svn+ssh://-oProxyCommand=touch owned nested' (in subrepository "s")
[255]
also check that hiding the attack in the username doesn't work:
$ cd ssh-vuln
$ echo "s = [svn]svn+ssh://%2DoProxyCommand=touch%20owned%20foo@example.com/nested" > .hgsub
$ hg ci -m3
$ cd ..
$ rm -r ssh-vuln-clone
$ hg clone ssh-vuln ssh-vuln-clone
updating to branch default
abort: potentially unsafe url: 'svn+ssh://-oProxyCommand=touch owned foo@example.com/nested' (in subrepository "s")
[255]