view tests/test-lfconvert.t @ 30155:b7a966ce89ed

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.

  $ USERCACHE="$TESTTMP/cache"; export USERCACHE
  $ mkdir "${USERCACHE}"
  $ cat >> $HGRCPATH <<EOF
  > [format]
  > usegeneraldelta=yes
  > [extensions]
  > largefiles =
  > share =
  > strip =
  > convert =
  > [largefiles]
  > minsize = 0.5
  > patterns = **.other
  >     **.dat
  > usercache=${USERCACHE}
  > EOF

"lfconvert" works
  $ hg init bigfile-repo
  $ cd bigfile-repo
  $ cat >> .hg/hgrc <<EOF
  > [extensions]
  > largefiles = !
  > EOF
  $ mkdir sub
  $ dd if=/dev/zero bs=1k count=256 > large 2> /dev/null
  $ dd if=/dev/zero bs=1k count=256 > large2 2> /dev/null
  $ echo normal > normal1
  $ echo alsonormal > sub/normal2
  $ dd if=/dev/zero bs=1k count=10 > sub/maybelarge.dat 2> /dev/null
  $ hg addremove
  adding large
  adding large2
  adding normal1
  adding sub/maybelarge.dat
  adding sub/normal2
  $ hg commit -m"add large, normal1" large normal1
  $ hg commit -m"add sub/*" sub

Test tag parsing
  $ cat >> .hgtags <<EOF
  > IncorrectlyFormattedTag!
  > invalidhash sometag
  > 0123456789abcdef anothertag
  > EOF
  $ hg add .hgtags
  $ hg commit -m"add large2" large2 .hgtags

Test link+rename largefile codepath
  $ [ -d .hg/largefiles ] && echo fail || echo pass
  pass
  $ cd ..
  $ hg lfconvert --size 0.2 bigfile-repo largefiles-repo
  initializing destination largefiles-repo
  skipping incorrectly formatted tag IncorrectlyFormattedTag!
  skipping incorrectly formatted id invalidhash
  no mapping for id 0123456789abcdef
#if symlink
  $ hg --cwd bigfile-repo rename large2 large3
  $ ln -sf large bigfile-repo/large3
  $ hg --cwd bigfile-repo commit -m"make large2 a symlink" large2 large3
  $ hg lfconvert --size 0.2 bigfile-repo largefiles-repo-symlink
  initializing destination largefiles-repo-symlink
  skipping incorrectly formatted tag IncorrectlyFormattedTag!
  skipping incorrectly formatted id invalidhash
  no mapping for id 0123456789abcdef
  abort: renamed/copied largefile large3 becomes symlink
  [255]
#endif
  $ cd bigfile-repo
  $ hg strip --no-backup 2
  0 files updated, 0 files merged, 2 files removed, 0 files unresolved
  $ cd ..
  $ rm -rf largefiles-repo largefiles-repo-symlink

  $ hg lfconvert --size 0.2 bigfile-repo largefiles-repo
  initializing destination largefiles-repo

"lfconvert" converts content correctly
  $ cd largefiles-repo
  $ hg up
  getting changed largefiles
  2 largefiles updated, 0 removed
  4 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ hg locate
  .hglf/large
  .hglf/sub/maybelarge.dat
  normal1
  sub/normal2
  $ cat normal1
  normal
  $ cat sub/normal2
  alsonormal
  $ md5sum.py large sub/maybelarge.dat
  ec87a838931d4d5d2e94a04644788a55  large
  1276481102f218c981e0324180bafd9f  sub/maybelarge.dat

"lfconvert" adds 'largefiles' to .hg/requires.
  $ cat .hg/requires
  dotencode
  fncache
  generaldelta
  largefiles
  revlogv1
  store

"lfconvert" includes a newline at the end of the standin files.
  $ cat .hglf/large .hglf/sub/maybelarge.dat
  2e000fa7e85759c7f4c254d4d9c33ef481e459a7
  34e163be8e43c5631d8b92e9c43ab0bf0fa62b9c
  $ cd ..

add some changesets to rename/remove/merge
  $ cd bigfile-repo
  $ hg mv -q sub stuff
  $ hg commit -m"rename sub/ to stuff/"
  $ hg update -q 1
  $ echo blah >> normal3
  $ echo blah >> sub/normal2
  $ echo blah >> sub/maybelarge.dat
  $ md5sum.py sub/maybelarge.dat
  1dd0b99ff80e19cff409702a1d3f5e15  sub/maybelarge.dat
  $ hg commit -A -m"add normal3, modify sub/*"
  adding normal3
  created new head
  $ hg rm large normal3
  $ hg commit -q -m"remove large, normal3"
  $ hg merge
  merging sub/maybelarge.dat and stuff/maybelarge.dat to stuff/maybelarge.dat
  merging sub/normal2 and stuff/normal2 to stuff/normal2
  warning: $TESTTMP/bigfile-repo/stuff/maybelarge.dat looks like a binary file. (glob)
  warning: conflicts while merging stuff/maybelarge.dat! (edit, then use 'hg resolve --mark')
  0 files updated, 1 files merged, 0 files removed, 1 files unresolved
  use 'hg resolve' to retry unresolved file merges or 'hg update -C .' to abandon
  [1]
  $ hg cat -r . sub/maybelarge.dat > stuff/maybelarge.dat
  $ hg resolve -m stuff/maybelarge.dat
  (no more unresolved files)
  $ hg commit -m"merge"
  $ hg log -G --template "{rev}:{node|short}  {desc|firstline}\n"
  @    5:4884f215abda  merge
  |\
  | o  4:7285f817b77e  remove large, normal3
  | |
  | o  3:67e3892e3534  add normal3, modify sub/*
  | |
  o |  2:c96c8beb5d56  rename sub/ to stuff/
  |/
  o  1:020c65d24e11  add sub/*
  |
  o  0:117b8328f97a  add large, normal1
  
  $ cd ..

lfconvert with rename, merge, and remove
  $ rm -rf largefiles-repo
  $ hg lfconvert --size 0.2 bigfile-repo largefiles-repo
  initializing destination largefiles-repo
  $ cd largefiles-repo
  $ hg log -G --template "{rev}:{node|short}  {desc|firstline}\n"
  o    5:8e05f5f2b77e  merge
  |\
  | o  4:a5a02de7a8e4  remove large, normal3
  | |
  | o  3:55759520c76f  add normal3, modify sub/*
  | |
  o |  2:261ad3f3f037  rename sub/ to stuff/
  |/
  o  1:334e5237836d  add sub/*
  |
  o  0:d4892ec57ce2  add large, normal1
  
  $ hg locate -r 2
  .hglf/large
  .hglf/stuff/maybelarge.dat
  normal1
  stuff/normal2
  $ hg locate -r 3
  .hglf/large
  .hglf/sub/maybelarge.dat
  normal1
  normal3
  sub/normal2
  $ hg locate -r 4
  .hglf/sub/maybelarge.dat
  normal1
  sub/normal2
  $ hg locate -r 5
  .hglf/stuff/maybelarge.dat
  normal1
  stuff/normal2
  $ hg update
  getting changed largefiles
  1 largefiles updated, 0 removed
  3 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ cat stuff/normal2
  alsonormal
  blah
  $ md5sum.py stuff/maybelarge.dat
  1dd0b99ff80e19cff409702a1d3f5e15  stuff/maybelarge.dat
  $ cat .hglf/stuff/maybelarge.dat
  76236b6a2c6102826c61af4297dd738fb3b1de38
  $ cd ..

"lfconvert" error cases
  $ hg lfconvert http://localhost/foo foo
  abort: http://localhost/foo is not a local Mercurial repo
  [255]
  $ hg lfconvert foo ssh://localhost/foo
  abort: ssh://localhost/foo is not a local Mercurial repo
  [255]
  $ hg lfconvert nosuchrepo foo
  abort: repository nosuchrepo not found!
  [255]
  $ hg share -q -U bigfile-repo shared
  $ printf 'bogus' > shared/.hg/sharedpath
  $ hg lfconvert shared foo
  abort: .hg/sharedpath points to nonexistent directory $TESTTMP/bogus! (glob)
  [255]
  $ hg lfconvert bigfile-repo largefiles-repo
  initializing destination largefiles-repo
  abort: repository largefiles-repo already exists!
  [255]

add another largefile to the new largefiles repo
  $ cd largefiles-repo
  $ dd if=/dev/zero bs=1k count=1k > anotherlarge 2> /dev/null
  $ hg add --lfsize=1 anotherlarge
  $ hg commit -m "add anotherlarge (should be a largefile)"
  $ cat .hglf/anotherlarge
  3b71f43ff30f4b15b5cd85dd9e95ebc7e84eb5a3
  $ hg tag mytag
  $ cd ..

round-trip: converting back to a normal (non-largefiles) repo with
"lfconvert --to-normal" should give the same as ../bigfile-repo
  $ cd largefiles-repo
  $ hg lfconvert --to-normal . ../normal-repo
  initializing destination ../normal-repo
  0 additional largefiles cached
  scanning source...
  sorting...
  converting...
  7 add large, normal1
  6 add sub/*
  5 rename sub/ to stuff/
  4 add normal3, modify sub/*
  3 remove large, normal3
  2 merge
  1 add anotherlarge (should be a largefile)
  0 Added tag mytag for changeset abacddda7028
  $ cd ../normal-repo
  $ cat >> .hg/hgrc <<EOF
  > [extensions]
  > largefiles = !
  > EOF

  $ hg log -G --template "{rev}:{node|short}  {desc|firstline}\n"
  o  7:b5fedc110b9d  Added tag mytag for changeset 867ab992ecf4
  |
  o  6:867ab992ecf4  add anotherlarge (should be a largefile)
  |
  o    5:4884f215abda  merge
  |\
  | o  4:7285f817b77e  remove large, normal3
  | |
  | o  3:67e3892e3534  add normal3, modify sub/*
  | |
  o |  2:c96c8beb5d56  rename sub/ to stuff/
  |/
  o  1:020c65d24e11  add sub/*
  |
  o  0:117b8328f97a  add large, normal1
  
  $ hg update
  5 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ hg locate
  .hgtags
  anotherlarge
  normal1
  stuff/maybelarge.dat
  stuff/normal2
  $ [ -d .hg/largefiles ] && echo fail || echo pass
  pass

  $ cd ..

Clearing the usercache ensures that commitctx doesn't try to cache largefiles
from the working dir on a convert.
  $ rm "${USERCACHE}"/*
  $ hg convert largefiles-repo
  assuming destination largefiles-repo-hg
  initializing destination largefiles-repo-hg repository
  scanning source...
  sorting...
  converting...
  7 add large, normal1
  6 add sub/*
  5 rename sub/ to stuff/
  4 add normal3, modify sub/*
  3 remove large, normal3
  2 merge
  1 add anotherlarge (should be a largefile)
  0 Added tag mytag for changeset abacddda7028

  $ hg -R largefiles-repo-hg log -G --template "{rev}:{node|short}  {desc|firstline}\n"
  o  7:2f08f66459b7  Added tag mytag for changeset 17126745edfd
  |
  o  6:17126745edfd  add anotherlarge (should be a largefile)
  |
  o    5:9cc5aa7204f0  merge
  |\
  | o  4:a5a02de7a8e4  remove large, normal3
  | |
  | o  3:55759520c76f  add normal3, modify sub/*
  | |
  o |  2:261ad3f3f037  rename sub/ to stuff/
  |/
  o  1:334e5237836d  add sub/*
  |
  o  0:d4892ec57ce2  add large, normal1
  
Verify will fail (for now) if the usercache is purged before converting, since
largefiles are not cached in the converted repo's local store by the conversion
process.
  $ cd largefiles-repo-hg
  $ cat >> .hg/hgrc <<EOF
  > [experimental]
  > evolution=createmarkers
  > EOF
  $ hg debugobsolete `hg log -r tip -T "{node}"`
  $ cd ..

  $ hg -R largefiles-repo-hg verify --large --lfa
  checking changesets
  checking manifests
  crosschecking files in changesets and manifests
  checking files
  9 files, 8 changesets, 13 total revisions
  searching 7 changesets for largefiles
  changeset 0:d4892ec57ce2: large references missing $TESTTMP/largefiles-repo-hg/.hg/largefiles/2e000fa7e85759c7f4c254d4d9c33ef481e459a7 (glob)
  changeset 1:334e5237836d: sub/maybelarge.dat references missing $TESTTMP/largefiles-repo-hg/.hg/largefiles/34e163be8e43c5631d8b92e9c43ab0bf0fa62b9c (glob)
  changeset 2:261ad3f3f037: stuff/maybelarge.dat references missing $TESTTMP/largefiles-repo-hg/.hg/largefiles/34e163be8e43c5631d8b92e9c43ab0bf0fa62b9c (glob)
  changeset 3:55759520c76f: sub/maybelarge.dat references missing $TESTTMP/largefiles-repo-hg/.hg/largefiles/76236b6a2c6102826c61af4297dd738fb3b1de38 (glob)
  changeset 5:9cc5aa7204f0: stuff/maybelarge.dat references missing $TESTTMP/largefiles-repo-hg/.hg/largefiles/76236b6a2c6102826c61af4297dd738fb3b1de38 (glob)
  changeset 6:17126745edfd: anotherlarge references missing $TESTTMP/largefiles-repo-hg/.hg/largefiles/3b71f43ff30f4b15b5cd85dd9e95ebc7e84eb5a3 (glob)
  verified existence of 6 revisions of 4 largefiles
  [1]
  $ hg -R largefiles-repo-hg showconfig paths
  [1]


Avoid a traceback if a largefile isn't available (issue3519)

Ensure the largefile can be cached in the source if necessary
  $ hg clone -U largefiles-repo issue3519
  $ rm -f "${USERCACHE}"/*
  $ hg lfconvert --to-normal issue3519 normalized3519
  initializing destination normalized3519
  4 additional largefiles cached
  scanning source...
  sorting...
  converting...
  7 add large, normal1
  6 add sub/*
  5 rename sub/ to stuff/
  4 add normal3, modify sub/*
  3 remove large, normal3
  2 merge
  1 add anotherlarge (should be a largefile)
  0 Added tag mytag for changeset abacddda7028

Ensure the abort message is useful if a largefile is entirely unavailable
  $ rm -rf normalized3519
  $ rm "${USERCACHE}"/*
  $ rm issue3519/.hg/largefiles/*
  $ rm largefiles-repo/.hg/largefiles/*
  $ hg lfconvert --to-normal issue3519 normalized3519
  initializing destination normalized3519
  anotherlarge: largefile 3b71f43ff30f4b15b5cd85dd9e95ebc7e84eb5a3 not available from file:/*/$TESTTMP/largefiles-repo (glob)
  stuff/maybelarge.dat: largefile 76236b6a2c6102826c61af4297dd738fb3b1de38 not available from file:/*/$TESTTMP/largefiles-repo (glob)
  stuff/maybelarge.dat: largefile 76236b6a2c6102826c61af4297dd738fb3b1de38 not available from file:/*/$TESTTMP/largefiles-repo (glob)
  sub/maybelarge.dat: largefile 76236b6a2c6102826c61af4297dd738fb3b1de38 not available from file:/*/$TESTTMP/largefiles-repo (glob)
  large: largefile 2e000fa7e85759c7f4c254d4d9c33ef481e459a7 not available from file:/*/$TESTTMP/largefiles-repo (glob)
  sub/maybelarge.dat: largefile 76236b6a2c6102826c61af4297dd738fb3b1de38 not available from file:/*/$TESTTMP/largefiles-repo (glob)
  large: largefile 2e000fa7e85759c7f4c254d4d9c33ef481e459a7 not available from file:/*/$TESTTMP/largefiles-repo (glob)
  stuff/maybelarge.dat: largefile 34e163be8e43c5631d8b92e9c43ab0bf0fa62b9c not available from file:/*/$TESTTMP/largefiles-repo (glob)
  large: largefile 2e000fa7e85759c7f4c254d4d9c33ef481e459a7 not available from file:/*/$TESTTMP/largefiles-repo (glob)
  sub/maybelarge.dat: largefile 34e163be8e43c5631d8b92e9c43ab0bf0fa62b9c not available from file:/*/$TESTTMP/largefiles-repo (glob)
  large: largefile 2e000fa7e85759c7f4c254d4d9c33ef481e459a7 not available from file:/*/$TESTTMP/largefiles-repo (glob)
  0 additional largefiles cached
  11 largefiles failed to download
  abort: all largefiles must be present locally
  [255]