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.
test that a commit clears the merge state.
$ hg init repo
$ cd repo
$ echo foo > file1
$ echo foo > file2
$ hg commit -Am 'add files'
adding file1
adding file2
$ echo bar >> file1
$ echo bar >> file2
$ hg commit -Am 'append bar to files'
create a second head with conflicting edits
$ hg up -C 0
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ echo baz >> file1
$ echo baz >> file2
$ hg commit -Am 'append baz to files'
created new head
create a third head with no conflicting edits
$ hg up -qC 0
$ echo foo > file3
$ hg commit -Am 'add non-conflicting file'
adding file3
created new head
failing merge
$ hg up -qC 2
$ hg merge --tool=internal:fail 1
0 files updated, 0 files merged, 0 files removed, 2 files unresolved
use 'hg resolve' to retry unresolved file merges or 'hg merge --abort' to abandon
[1]
resolve -l should contain unresolved entries
$ hg resolve -l
U file1
U file2
$ hg resolve -l --no-status
file1
file2
resolving an unknown path should emit a warning, but not for -l
$ hg resolve -m does-not-exist
arguments do not match paths that need resolving
$ hg resolve -l does-not-exist
tell users how they could have used resolve
$ mkdir nested
$ cd nested
$ hg resolve -m file1
arguments do not match paths that need resolving
(try: hg resolve -m path:file1)
$ hg resolve -m file1 filez
arguments do not match paths that need resolving
(try: hg resolve -m path:file1 path:filez)
$ hg resolve -m path:file1 path:filez
$ hg resolve -l
R file1
U file2
$ hg resolve -m filez file2
arguments do not match paths that need resolving
(try: hg resolve -m path:filez path:file2)
$ hg resolve -m path:filez path:file2
(no more unresolved files)
$ hg resolve -l
R file1
R file2
cleanup
$ hg resolve -u
$ cd ..
$ rmdir nested
don't allow marking or unmarking driver-resolved files
$ cat > $TESTTMP/markdriver.py << EOF
> '''mark and unmark files as driver-resolved'''
> from mercurial import (
> merge,
> pycompat,
> registrar,
> scmutil,
> )
> cmdtable = {}
> command = registrar.command(cmdtable)
> @command(b'markdriver',
> [(b'u', b'unmark', None, b'')],
> b'FILE...')
> def markdriver(ui, repo, *pats, **opts):
> wlock = repo.wlock()
> opts = pycompat.byteskwargs(opts)
> try:
> ms = merge.mergestate.read(repo)
> m = scmutil.match(repo[None], pats, opts)
> for f in ms:
> if not m(f):
> continue
> if not opts[b'unmark']:
> ms.mark(f, b'd')
> else:
> ms.mark(f, b'u')
> ms.commit()
> finally:
> wlock.release()
> EOF
$ hg --config extensions.markdriver=$TESTTMP/markdriver.py markdriver file1
$ hg resolve --list
D file1
U file2
$ hg resolve --mark file1
not marking file1 as it is driver-resolved
this should not print out file1
$ hg resolve --mark --all
(no more unresolved files -- run "hg resolve --all" to conclude)
$ hg resolve --mark 'glob:file*'
(no more unresolved files -- run "hg resolve --all" to conclude)
$ hg resolve --list
D file1
R file2
$ hg resolve --unmark file1
not unmarking file1 as it is driver-resolved
(no more unresolved files -- run "hg resolve --all" to conclude)
$ hg resolve --unmark --all
$ hg resolve --list
D file1
U file2
$ hg --config extensions.markdriver=$TESTTMP/markdriver.py markdriver --unmark file1
$ hg resolve --list
U file1
U file2
resolve the failure
$ echo resolved > file1
$ hg resolve -m file1
resolve -l should show resolved file as resolved
$ hg resolve -l
R file1
U file2
$ hg resolve -l -Tjson
[
{
"path": "file1",
"status": "R"
},
{
"path": "file2",
"status": "U"
}
]
$ hg resolve -l -T '{path} {status} {p1rev} {p2rev}\n'
file1 R 2 1
file2 U 2 1
resolve -m without paths should mark all resolved
$ hg resolve -m
(no more unresolved files)
$ hg commit -m 'resolved'
resolve -l should be empty after commit
$ hg resolve -l
$ hg resolve -l -Tjson
[
]
resolve --all should abort when no merge in progress
$ hg resolve --all
abort: resolve command not applicable when not merging
[255]
resolve -m should abort when no merge in progress
$ hg resolve -m
abort: resolve command not applicable when not merging
[255]
can not update or merge when there are unresolved conflicts
$ hg up -qC 0
$ echo quux >> file1
$ hg up 1
merging file1
warning: conflicts while merging file1! (edit, then use 'hg resolve --mark')
1 files updated, 0 files merged, 0 files removed, 1 files unresolved
use 'hg resolve' to retry unresolved file merges
[1]
$ hg up 0
abort: outstanding merge conflicts
[255]
$ hg merge 2
abort: outstanding merge conflicts
[255]
$ hg merge --force 2
abort: outstanding merge conflicts
[255]
set up conflict-free merge
$ hg up -qC 3
$ hg merge 1
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
(branch merge, don't forget to commit)
resolve --all should do nothing in merge without conflicts
$ hg resolve --all
(no more unresolved files)
resolve -m should do nothing in merge without conflicts
$ hg resolve -m
(no more unresolved files)
get back to conflicting state
$ hg up -qC 2
$ hg merge --tool=internal:fail 1
0 files updated, 0 files merged, 0 files removed, 2 files unresolved
use 'hg resolve' to retry unresolved file merges or 'hg merge --abort' to abandon
[1]
resolve without arguments should suggest --all
$ hg resolve
abort: no files or directories specified
(use --all to re-merge all unresolved files)
[255]
resolve --all should re-merge all unresolved files
$ hg resolve --all
merging file1
merging file2
warning: conflicts while merging file1! (edit, then use 'hg resolve --mark')
warning: conflicts while merging file2! (edit, then use 'hg resolve --mark')
[1]
$ cat file1.orig
foo
baz
$ cat file2.orig
foo
baz
.orig files should exists where specified
$ hg resolve --all --verbose --config 'ui.origbackuppath=.hg/origbackups'
merging file1
creating directory: $TESTTMP/repo/.hg/origbackups
merging file2
warning: conflicts while merging file1! (edit, then use 'hg resolve --mark')
warning: conflicts while merging file2! (edit, then use 'hg resolve --mark')
[1]
$ ls .hg/origbackups
file1
file2
$ grep '<<<' file1 > /dev/null
$ grep '<<<' file2 > /dev/null
resolve <file> should re-merge file
$ echo resolved > file1
$ hg resolve -q file1
warning: conflicts while merging file1! (edit, then use 'hg resolve --mark')
[1]
$ grep '<<<' file1 > /dev/null
test .orig behavior with resolve
$ hg resolve -q file1 --tool "sh -c 'f --dump \"$TESTTMP/repo/file1.orig\"'"
$TESTTMP/repo/file1.orig:
>>>
foo
baz
<<<
resolve <file> should do nothing if 'file' was marked resolved
$ echo resolved > file1
$ hg resolve -m file1
$ hg resolve -q file1
$ cat file1
resolved
insert unsupported advisory merge record
$ hg --config extensions.fakemergerecord=$TESTDIR/fakemergerecord.py fakemergerecord -x
$ hg debugmergestate
* version 2 records
local: 57653b9f834a4493f7240b0681efcb9ae7cab745
other: dc77451844e37f03f5c559e3b8529b2b48d381d1
labels:
local: working copy
other: merge rev
unrecognized entry: x advisory record
file extras: file1 (ancestorlinknode = 99726c03216e233810a2564cbc0adfe395007eac)
file: file1 (record type "F", state "r", hash 60b27f004e454aca81b0480209cce5081ec52390)
local path: file1 (flags "")
ancestor path: file1 (node 2ed2a3912a0b24502043eae84ee4b279c18b90dd)
other path: file1 (node 6f4310b00b9a147241b071a60c28a650827fb03d)
file extras: file2 (ancestorlinknode = 99726c03216e233810a2564cbc0adfe395007eac)
file: file2 (record type "F", state "u", hash cb99b709a1978bd205ab9dfd4c5aaa1fc91c7523)
local path: file2 (flags "")
ancestor path: file2 (node 2ed2a3912a0b24502043eae84ee4b279c18b90dd)
other path: file2 (node 6f4310b00b9a147241b071a60c28a650827fb03d)
$ hg resolve -l
R file1
U file2
insert unsupported mandatory merge record
$ hg --config extensions.fakemergerecord=$TESTDIR/fakemergerecord.py fakemergerecord -X
$ hg debugmergestate
* version 2 records
local: 57653b9f834a4493f7240b0681efcb9ae7cab745
other: dc77451844e37f03f5c559e3b8529b2b48d381d1
labels:
local: working copy
other: merge rev
file extras: file1 (ancestorlinknode = 99726c03216e233810a2564cbc0adfe395007eac)
file: file1 (record type "F", state "r", hash 60b27f004e454aca81b0480209cce5081ec52390)
local path: file1 (flags "")
ancestor path: file1 (node 2ed2a3912a0b24502043eae84ee4b279c18b90dd)
other path: file1 (node 6f4310b00b9a147241b071a60c28a650827fb03d)
file extras: file2 (ancestorlinknode = 99726c03216e233810a2564cbc0adfe395007eac)
file: file2 (record type "F", state "u", hash cb99b709a1978bd205ab9dfd4c5aaa1fc91c7523)
local path: file2 (flags "")
ancestor path: file2 (node 2ed2a3912a0b24502043eae84ee4b279c18b90dd)
other path: file2 (node 6f4310b00b9a147241b071a60c28a650827fb03d)
unrecognized entry: X mandatory record
$ hg resolve -l
abort: unsupported merge state records: X
(see https://mercurial-scm.org/wiki/MergeStateRecords for more information)
[255]
$ hg resolve -ma
abort: unsupported merge state records: X
(see https://mercurial-scm.org/wiki/MergeStateRecords for more information)
[255]
$ hg summary
warning: merge state has unsupported record types: X
parent: 2:57653b9f834a
append baz to files
parent: 1:dc77451844e3
append bar to files
branch: default
commit: 2 modified, 2 unknown (merge)
update: 2 new changesets (update)
phases: 5 draft
update --clean shouldn't abort on unsupported records
$ hg up -qC 1
$ hg debugmergestate
no merge state found
test crashed merge with empty mergestate
$ mkdir .hg/merge
$ touch .hg/merge/state
resolve -l should be empty
$ hg resolve -l
$ cd ..