changelog: disable delta chains
This patch disables delta chains on changelogs. After this patch, new
entries on changelogs - including existing changelogs - will be stored
as the fulltext of that data (likely compressed). No delta computation
will be performed.
An overview of delta chains and data justifying this change follows.
Revlogs try to store entries as a delta against a previous entry (either
a parent revision in the case of generaldelta or the previous physical
revision when not using generaldelta). Most of the time this is the
correct thing to do: it frequently results in less CPU usage and smaller
storage.
Delta chains are most effective when the base revision being deltad
against is similar to the current data. This tends to occur naturally
for manifests and file data, since only small parts of each tend to
change with each revision. Changelogs, however, are a different story.
Changelog entries represent changesets/commits. And unless commits in a
repository are homogonous (same author, changing same files, similar
commit messages, etc), a delta from one entry to the next tends to be
relatively large compared to the size of the entry. This means that
delta chains tend to be short. How short? Here is the full vs delta
revision breakdown on some real world repos:
Repo % Full % Delta Max Length
hg 45.8 54.2 6
mozilla-central 42.4 57.6 8
mozilla-unified 42.5 57.5 17
pypy 46.1 53.9 6
python-zstandard 46.1 53.9 3
(I threw in python-zstandard as an example of a repo that is homogonous.
It contains a small Python project with changes all from the same
author.)
Contrast this with the manifest revlog for these repos, where 99+% of
revisions are deltas and delta chains run into the thousands.
So delta chains aren't as useful on changelogs. But even a short delta
chain may provide benefits. Let's measure that.
Delta chains may require less CPU to read revisions if the CPU time
spent reading smaller deltas is less than the CPU time used to
decompress larger individual entries. We can measure this via
`hg perfrevlog -c -d 1` to iterate a revlog to resolve each revision's
fulltext. Here are the results of that command on a repo using delta
chains in its changelog and on a repo without delta chains:
hg (forward)
! wall 0.407008 comb 0.410000 user 0.410000 sys 0.000000 (best of 25)
! wall 0.390061 comb 0.390000 user 0.390000 sys 0.000000 (best of 26)
hg (reverse)
! wall 0.515221 comb 0.520000 user 0.520000 sys 0.000000 (best of 19)
! wall 0.400018 comb 0.400000 user 0.390000 sys 0.010000 (best of 25)
mozilla-central (forward)
! wall 4.508296 comb 4.490000 user 4.490000 sys 0.000000 (best of 3)
! wall 4.370222 comb 4.370000 user 4.350000 sys 0.020000 (best of 3)
mozilla-central (reverse)
! wall 5.758995 comb 5.760000 user 5.720000 sys 0.040000 (best of 3)
! wall 4.346503 comb 4.340000 user 4.320000 sys 0.020000 (best of 3)
mozilla-unified (forward)
! wall 4.957088 comb 4.950000 user 4.940000 sys 0.010000 (best of 3)
! wall 4.660528 comb 4.650000 user 4.630000 sys 0.020000 (best of 3)
mozilla-unified (reverse)
! wall 6.119827 comb 6.110000 user 6.090000 sys 0.020000 (best of 3)
! wall 4.675136 comb 4.670000 user 4.670000 sys 0.000000 (best of 3)
pypy (forward)
! wall 1.231122 comb 1.240000 user 1.230000 sys 0.010000 (best of 8)
! wall 1.164896 comb 1.160000 user 1.160000 sys 0.000000 (best of 9)
pypy (reverse)
! wall 1.467049 comb 1.460000 user 1.460000 sys 0.000000 (best of 7)
! wall 1.160200 comb 1.170000 user 1.160000 sys 0.010000 (best of 9)
The data clearly shows that it takes less wall and CPU time to resolve
revisions when there are no delta chains in the changelogs, regardless
of the direction of traversal. Furthermore, not using a delta chain
means that fulltext resolution in reverse is as fast as iterating
forward. So not using delta chains on the changelog is a clear CPU win
for reading operations.
An example of a user-visible operation showing this speed-up is revset
evaluation. Here are results for
`hg perfrevset 'author(gps) or author(mpm)'`:
hg
! wall 1.655506 comb 1.660000 user 1.650000 sys 0.010000 (best of 6)
! wall 1.612723 comb 1.610000 user 1.600000 sys 0.010000 (best of 7)
mozilla-central
! wall 17.629826 comb 17.640000 user 17.600000 sys 0.040000 (best of 3)
! wall 17.311033 comb 17.300000 user 17.260000 sys 0.040000 (best of 3)
What about 00changelog.i size?
Repo Delta Chains No Delta Chains
hg 7,033,250 6,976,771
mozilla-central 82,978,748 81,574,623
mozilla-unified 88,112,349 86,702,162
pypy 20,740,699 20,659,741
The data shows that removing delta chains from the changelog makes the
changelog smaller.
Delta chains are also used during changegroup generation. This
operation essentially converts a series of revisions to one large
delta chain. And changegroup generation is smart: if the delta in
the revlog matches what the changegroup is emitting, it will reuse
the delta instead of recalculating it. We can measure the impact
removing changelog delta chains has on changegroup generation via
`hg perfchangegroupchangelog`:
hg
! wall 1.589245 comb 1.590000 user 1.590000 sys 0.000000 (best of 7)
! wall 1.788060 comb 1.790000 user 1.790000 sys 0.000000 (best of 6)
mozilla-central
! wall 17.382585 comb 17.380000 user 17.340000 sys 0.040000 (best of 3)
! wall 20.161357 comb 20.160000 user 20.120000 sys 0.040000 (best of 3)
mozilla-unified
! wall 18.722839 comb 18.720000 user 18.680000 sys 0.040000 (best of 3)
! wall 21.168075 comb 21.170000 user 21.130000 sys 0.040000 (best of 3)
pypy
! wall 4.828317 comb 4.830000 user 4.820000 sys 0.010000 (best of 3)
! wall 5.415455 comb 5.420000 user 5.410000 sys 0.010000 (best of 3)
The data shows eliminating delta chains makes the changelog part of
changegroup generation slower. This is expected since we now have to
compute deltas for revisions where we could recycle the delta before.
It is worth putting this regression into context of overall changegroup
times. Here is the rough total CPU time spent in changegroup generation
for various repos while using delta chains on the changelog:
Repo CPU Time (s) CPU Time w/ compression
hg 4.50 7.05
mozilla-central 111.1 222.0
pypy 28.68 75.5
Before compression, removing delta chains from the changegroup adds
~4.4% overhead to hg changegroup generation, 1.3% to mozilla-central,
and 2.0% to pypy. When you factor in zlib compression, these percentages
are roughly divided by 2.
While the increased CPU usage for changegroup generation is unfortunate,
I think it is acceptable because the percentage is small, server
operators (those likely impacted most by this) have other mechanisms
to mitigate CPU consumption (namely reducing zlib compression level and
pre-generated clone bundles), and because there is room to optimize this
in the future. For example, we could use the nullid as the base revision,
effectively encoding the full revision for each entry in the changegroup.
When doing this, `hg perfchangegroupchangelog` nearly halves:
mozilla-unified
! wall 21.168075 comb 21.170000 user 21.130000 sys 0.040000 (best of 3)
! wall 11.196461 comb 11.200000 user 11.190000 sys 0.010000 (best of 3)
This looks very promising as a future optimization opportunity.
It's worth that the changes in test-acl.t to the changegroup part size.
This is because revision 6 in the changegroup had a delta chain of
length 2 before and after this patch the base revision is nullrev.
When the base revision is nullrev, cg2packer.deltaparent() hardcodes
the *previous* revision from the changegroup as the delta parent.
This caused the delta in the changegroup to switch base revisions,
the delta to change, and the size to change accordingly. While the
size increased in this case, I think sizes will remain the same
on average, as the delta base for changelog revisions doesn't matter
too much (as this patch shows). So, I don't consider this a regression.
$ cat >> $HGRCPATH <<EOF
> [extensions]
> censor=
> EOF
$ cp $HGRCPATH $HGRCPATH.orig
Create repo with unimpeachable content
$ hg init r
$ cd r
$ echo 'Initially untainted file' > target
$ echo 'Normal file here' > bystander
$ hg add target bystander
$ hg ci -m init
Clone repo so we can test pull later
$ cd ..
$ hg clone r rpull
updating to branch default
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cd r
Introduce content which will ultimately require censorship. Name the first
censored node C1, second C2, and so on
$ echo 'Tainted file' > target
$ echo 'Passwords: hunter2' >> target
$ hg ci -m taint target
$ C1=`hg id --debug -i`
$ echo 'hunter3' >> target
$ echo 'Normal file v2' > bystander
$ hg ci -m moretaint target bystander
$ C2=`hg id --debug -i`
Add a new sanitized versions to correct our mistake. Name the first head H1,
the second head H2, and so on
$ echo 'Tainted file is now sanitized' > target
$ hg ci -m sanitized target
$ H1=`hg id --debug -i`
$ hg update -r $C2
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ echo 'Tainted file now super sanitized' > target
$ hg ci -m 'super sanitized' target
created new head
$ H2=`hg id --debug -i`
Verify target contents before censorship at each revision
$ hg cat -r $H1 target
Tainted file is now sanitized
$ hg cat -r $H2 target
Tainted file now super sanitized
$ hg cat -r $C2 target
Tainted file
Passwords: hunter2
hunter3
$ hg cat -r $C1 target
Tainted file
Passwords: hunter2
$ hg cat -r 0 target
Initially untainted file
Try to censor revision with too large of a tombstone message
$ hg censor -r $C1 -t 'blah blah blah blah blah blah blah blah bla' target
abort: censor tombstone must be no longer than censored data
[255]
Censor revision with 2 offenses
(this also tests file pattern matching: path relative to cwd case)
$ mkdir -p foo/bar/baz
$ hg --cwd foo/bar/baz censor -r $C2 -t "remove password" ../../../target
$ hg cat -r $H1 target
Tainted file is now sanitized
$ hg cat -r $H2 target
Tainted file now super sanitized
$ hg cat -r $C2 target
abort: censored node: 1e0247a9a4b7
(set censor.policy to ignore errors)
[255]
$ hg cat -r $C1 target
Tainted file
Passwords: hunter2
$ hg cat -r 0 target
Initially untainted file
Censor revision with 1 offense
(this also tests file pattern matching: with 'path:' scheme)
$ hg --cwd foo/bar/baz censor -r $C1 path:target
$ hg cat -r $H1 target
Tainted file is now sanitized
$ hg cat -r $H2 target
Tainted file now super sanitized
$ hg cat -r $C2 target
abort: censored node: 1e0247a9a4b7
(set censor.policy to ignore errors)
[255]
$ hg cat -r $C1 target
abort: censored node: 613bc869fceb
(set censor.policy to ignore errors)
[255]
$ hg cat -r 0 target
Initially untainted file
Can only checkout target at uncensored revisions, -X is workaround for --all
$ hg revert -r $C2 target
abort: censored node: 1e0247a9a4b7
(set censor.policy to ignore errors)
[255]
$ hg revert -r $C1 target
abort: censored node: 613bc869fceb
(set censor.policy to ignore errors)
[255]
$ hg revert -r $C1 --all
reverting bystander
reverting target
abort: censored node: 613bc869fceb
(set censor.policy to ignore errors)
[255]
$ hg revert -r $C1 --all -X target
$ cat target
Tainted file now super sanitized
$ hg revert -r 0 --all
reverting target
$ cat target
Initially untainted file
$ hg revert -r $H2 --all
reverting bystander
reverting target
$ cat target
Tainted file now super sanitized
Uncensored file can be viewed at any revision
$ hg cat -r $H1 bystander
Normal file v2
$ hg cat -r $C2 bystander
Normal file v2
$ hg cat -r $C1 bystander
Normal file here
$ hg cat -r 0 bystander
Normal file here
Can update to children of censored revision
$ hg update -r $H1
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat target
Tainted file is now sanitized
$ hg update -r $H2
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat target
Tainted file now super sanitized
Set censor policy to abort in trusted $HGRC so hg verify fails
$ cp $HGRCPATH.orig $HGRCPATH
$ cat >> $HGRCPATH <<EOF
> [censor]
> policy = abort
> EOF
Repo fails verification due to censorship
$ hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
target@1: censored file data
target@2: censored file data
2 files, 5 changesets, 7 total revisions
2 integrity errors encountered!
(first damaged changeset appears to be 1)
[1]
Cannot update to revision with censored data
$ hg update -r $C2
abort: censored node: 1e0247a9a4b7
(set censor.policy to ignore errors)
[255]
$ hg update -r $C1
abort: censored node: 613bc869fceb
(set censor.policy to ignore errors)
[255]
$ hg update -r 0
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg update -r $H2
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
Set censor policy to ignore in trusted $HGRC so hg verify passes
$ cp $HGRCPATH.orig $HGRCPATH
$ cat >> $HGRCPATH <<EOF
> [censor]
> policy = ignore
> EOF
Repo passes verification with warnings with explicit config
$ hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
2 files, 5 changesets, 7 total revisions
May update to revision with censored data with explicit config
$ hg update -r $C2
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat target
$ hg update -r $C1
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat target
$ hg update -r 0
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat target
Initially untainted file
$ hg update -r $H2
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat target
Tainted file now super sanitized
Can merge in revision with censored data. Test requires one branch of history
with the file censored, but we can't censor at a head, so advance H1.
$ hg update -r $H1
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ C3=$H1
$ echo 'advanced head H1' > target
$ hg ci -m 'advance head H1' target
$ H1=`hg id --debug -i`
$ hg censor -r $C3 target
$ hg update -r $H2
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg merge -r $C3
merging target
0 files updated, 1 files merged, 0 files removed, 0 files unresolved
(branch merge, don't forget to commit)
Revisions present in repository heads may not be censored
$ hg update -C -r $H2
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg censor -r $H2 target
abort: cannot censor file in heads (78a8fc215e79)
(clean/delete and commit first)
[255]
$ echo 'twiddling thumbs' > bystander
$ hg ci -m 'bystander commit'
$ H2=`hg id --debug -i`
$ hg censor -r "$H2^" target
abort: cannot censor file in heads (efbe78065929)
(clean/delete and commit first)
[255]
Cannot censor working directory
$ echo 'seriously no passwords' > target
$ hg ci -m 'extend second head arbitrarily' target
$ H2=`hg id --debug -i`
$ hg update -r "$H2^"
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg censor -r . target
abort: cannot censor working directory
(clean/delete/update first)
[255]
$ hg update -r $H2
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
Can re-add file after being deleted + censored
$ C4=$H2
$ hg rm target
$ hg ci -m 'delete target so it may be censored'
$ H2=`hg id --debug -i`
$ hg censor -r $C4 target
$ hg cat -r $C4 target
$ hg cat -r "$H2^^" target
Tainted file now super sanitized
$ echo 'fresh start' > target
$ hg add target
$ hg ci -m reincarnated target
$ H2=`hg id --debug -i`
$ hg cat -r $H2 target
fresh start
$ hg cat -r "$H2^" target
target: no such file in rev 452ec1762369
[1]
$ hg cat -r $C4 target
$ hg cat -r "$H2^^^" target
Tainted file now super sanitized
Can censor after revlog has expanded to no longer permit inline storage
$ for x in `python $TESTDIR/seq.py 0 50000`
> do
> echo "Password: hunter$x" >> target
> done
$ hg ci -m 'add 100k passwords'
$ H2=`hg id --debug -i`
$ C5=$H2
$ hg revert -r "$H2^" target
$ hg ci -m 'cleaned 100k passwords'
$ H2=`hg id --debug -i`
$ hg censor -r $C5 target
$ hg cat -r $C5 target
$ hg cat -r $H2 target
fresh start
Repo with censored nodes can be cloned and cloned nodes are censored
$ cd ..
$ hg clone r rclone
updating to branch default
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cd rclone
$ hg cat -r $H1 target
advanced head H1
$ hg cat -r $H2~5 target
Tainted file now super sanitized
$ hg cat -r $C2 target
$ hg cat -r $C1 target
$ hg cat -r 0 target
Initially untainted file
$ hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
2 files, 12 changesets, 13 total revisions
Repo cloned before tainted content introduced can pull censored nodes
$ cd ../rpull
$ hg cat -r tip target
Initially untainted file
$ hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
2 files, 1 changesets, 2 total revisions
$ hg pull -r $H1 -r $H2
pulling from $TESTTMP/r (glob)
searching for changes
adding changesets
adding manifests
adding file changes
added 11 changesets with 11 changes to 2 files (+1 heads)
(run 'hg heads' to see heads, 'hg merge' to merge)
$ hg update 4
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat target
Tainted file now super sanitized
$ hg cat -r $H1 target
advanced head H1
$ hg cat -r $H2~5 target
Tainted file now super sanitized
$ hg cat -r $C2 target
$ hg cat -r $C1 target
$ hg cat -r 0 target
Initially untainted file
$ hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
2 files, 12 changesets, 13 total revisions
Censored nodes can be pushed if they censor previously unexchanged nodes
$ echo 'Passwords: hunter2hunter2' > target
$ hg ci -m 're-add password from clone' target
created new head
$ H3=`hg id --debug -i`
$ REV=$H3
$ echo 'Re-sanitized; nothing to see here' > target
$ hg ci -m 're-sanitized' target
$ H2=`hg id --debug -i`
$ CLEANREV=$H2
$ hg cat -r $REV target
Passwords: hunter2hunter2
$ hg censor -r $REV target
$ hg cat -r $REV target
$ hg cat -r $CLEANREV target
Re-sanitized; nothing to see here
$ hg push -f -r $H2
pushing to $TESTTMP/r (glob)
searching for changes
adding changesets
adding manifests
adding file changes
added 2 changesets with 2 changes to 1 files (+1 heads)
$ cd ../r
$ hg cat -r $REV target
$ hg cat -r $CLEANREV target
Re-sanitized; nothing to see here
$ hg update $CLEANREV
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat target
Re-sanitized; nothing to see here
Censored nodes can be bundled up and unbundled in another repo
$ hg bundle --base 0 ../pwbundle
13 changesets found
$ cd ../rclone
$ hg unbundle ../pwbundle
adding changesets
adding manifests
adding file changes
added 2 changesets with 2 changes to 2 files (+1 heads)
(run 'hg heads .' to see heads, 'hg merge' to merge)
$ hg cat -r $REV target
$ hg cat -r $CLEANREV target
Re-sanitized; nothing to see here
$ hg update $CLEANREV
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat target
Re-sanitized; nothing to see here
$ hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
2 files, 14 changesets, 15 total revisions
Censored nodes can be imported on top of censored nodes, consecutively
$ hg init ../rimport
$ hg bundle --base 1 ../rimport/splitbundle
12 changesets found
$ cd ../rimport
$ hg pull -r $H1 -r $H2 ../r
pulling from ../r
adding changesets
adding manifests
adding file changes
added 8 changesets with 10 changes to 2 files (+1 heads)
(run 'hg heads' to see heads, 'hg merge' to merge)
$ hg unbundle splitbundle
adding changesets
adding manifests
adding file changes
added 6 changesets with 5 changes to 2 files (+1 heads)
(run 'hg heads .' to see heads, 'hg merge' to merge)
$ hg update $H2
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat target
Re-sanitized; nothing to see here
$ hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
2 files, 14 changesets, 15 total revisions
$ cd ../r
Can import bundle where first revision of a file is censored
$ hg init ../rinit
$ hg censor -r 0 target
$ hg bundle -r 0 --base null ../rinit/initbundle
1 changesets found
$ cd ../rinit
$ hg unbundle initbundle
adding changesets
adding manifests
adding file changes
added 1 changesets with 2 changes to 2 files
(run 'hg update' to get a working copy)
$ hg cat -r 0 target