view tests/test-uncommit.t @ 40326:fed697fa1734

sqlitestore: file storage backend using SQLite

This commit provides an extension which uses SQLite to store file
data (as opposed to revlogs).

As the inline documentation describes, there are still several
aspects to the extension that are incomplete. But it's a start.
The extension does support basic clone, checkout, and commit
workflows, which makes it suitable for simple use cases.

One notable missing feature is support for "bundlerepos." This is
probably responsible for the most test failures when the extension
is activated as part of the test suite.

All revision data is stored in SQLite. Data is stored as zstd
compressed chunks (default if zstd is available), zlib compressed
chunks (default if zstd is not available), or raw chunks (if
configured or if a compressed delta is not smaller than the raw
delta). This makes things very similar to revlogs.

Unlike revlogs, the extension doesn't yet enforce a limit on delta
chain length. This is an obvious limitation and should be addressed.
This is somewhat mitigated by the use of zstd, which is much faster
than zlib to decompress.

There is a dedicated table for storing deltas. Deltas are stored
by the SHA-1 hash of their uncompressed content. The "fileindex" table
has columns that reference the delta for each revision and the base
delta that delta should be applied against. A recursive SQL query
is used to resolve the delta chain along with the delta data.

By storing deltas by hash, we are able to de-duplicate delta storage!
With revlogs, the same deltas in different revlogs would result in
duplicate storage of that delta. In this scheme, inserting the
duplicate delta is a no-op and delta chains simply reference the
existing delta.

When initially implementing this extension, I did not have
content-indexed deltas and deltas could be duplicated across files
(just like revlogs). When I implemented content-indexed deltas, the
size of the SQLite database for a full clone of mozilla-unified
dropped:

before: 2,554,261,504 bytes
after:  2,488,754,176 bytes

Surprisingly, this is still larger than the bytes size of revlog
files:

revlog files: 2,104,861,230 bytes
du -b:        2,254,381,614

I would have expected storage to be smaller since we're not limiting
delta chain length and since we're using zstd instead of zlib. I
suspect the SQLite indexes and per-column overhead account for the
bulk of the differences. (Keep in mind that revlog uses a 64-byte
packed struct for revision index data and deltas are stored without
padding. Aside from the 12 unused bytes in the 32 byte node field,
revlogs are pretty efficient.) Another source of overhead is file
name storage. With revlogs, file names are stored in the filesystem.
But with SQLite, we need to store file names in the database. This is
roughly equivalent to the size of the fncache file, which for the
mozilla-unified repository is ~34MB.

Since the SQLite database isn't append-only and since delta chains
can reference any delta, this opens some interesting possibilities.
For example, we could store deltas in reverse, such that fulltexts
are stored for newer revisions and deltas are applied to reconstruct
older revisions. This is likely a more optimal storage strategy for
version control, as new data tends to be more frequently accessed
than old data. We would obviously need wire protocol support for
transferring revision data from newest to oldest. And we would
probably need some kind of mechanism for "re-encoding" stores. But
it should be doable.

This extension is very much experimental quality. There are a handful
of features that don't work. It probably isn't suitable for day-to-day
use. But it could be used in limited cases (e.g. read-only checkouts
like in CI). And it is also a good proving ground for alternate
storage backends. As we continue to define interfaces for all things
storage, it will be useful to have a viable alternate storage backend
to see how things shake out in practice.

test-storage.py passes on Python 2 and introduces no new test failures on
Python 3. Having the storage-level unit tests has proved to be insanely
useful when developing this extension. Those tests caught numerous bugs
during development and I'm convinced this style of testing is the way
forward for ensuring alternate storage backends work as intended. Of
course, test coverage isn't close to what it needs to be. But it is
a start. And what coverage we have gives me confidence that basic store
functionality is implemented properly.

Differential Revision: https://phab.mercurial-scm.org/D4928

Test uncommit - set up the config

  $ cat >> $HGRCPATH <<EOF
  > [experimental]
  > evolution.createmarkers=True
  > evolution.allowunstable=True
  > [extensions]
  > uncommit =
  > drawdag=$TESTDIR/drawdag.py
  > EOF

Build up a repo

  $ hg init repo
  $ cd repo
  $ hg bookmark foo

Help for uncommit

  $ hg help uncommit
  hg uncommit [OPTION]... [FILE]...
  
  uncommit part or all of a local changeset
  
      This command undoes the effect of a local commit, returning the affected
      files to their uncommitted state. This means that files modified or
      deleted in the changeset will be left unchanged, and so will remain
      modified in the working directory.
  
      If no files are specified, the commit will be pruned, unless --keep is
      given.
  
  (use 'hg help -e uncommit' to show help for the uncommit extension)
  
  options ([+] can be repeated):
  
      --keep                allow an empty commit after uncommiting
   -I --include PATTERN [+] include names matching the given patterns
   -X --exclude PATTERN [+] exclude names matching the given patterns
  
  (some details hidden, use --verbose to show complete help)

Uncommit with no commits should fail

  $ hg uncommit
  abort: cannot uncommit null changeset
  (no changeset checked out)
  [255]

Create some commits

  $ touch files
  $ hg add files
  $ for i in a ab abc abcd abcde; do echo $i > files; echo $i > file-$i; hg add file-$i; hg commit -m "added file-$i"; done
  $ ls
  file-a
  file-ab
  file-abc
  file-abcd
  file-abcde
  files

  $ hg log -G -T '{rev}:{node} {desc}' --hidden
  @  4:6c4fd43ed714e7fcd8adbaa7b16c953c2e985b60 added file-abcde
  |
  o  3:6db330d65db434145c0b59d291853e9a84719b24 added file-abcd
  |
  o  2:abf2df566fc193b3ac34d946e63c1583e4d4732b added file-abc
  |
  o  1:69a232e754b08d568c4899475faf2eb44b857802 added file-ab
  |
  o  0:3004d2d9b50883c1538fc754a3aeb55f1b4084f6 added file-a
  
Simple uncommit off the top, also moves bookmark

  $ hg bookmark
   * foo                       4:6c4fd43ed714
  $ hg uncommit
  $ hg status
  M files
  A file-abcde
  $ hg bookmark
   * foo                       3:6db330d65db4

  $ hg log -G -T '{rev}:{node} {desc}' --hidden
  x  4:6c4fd43ed714e7fcd8adbaa7b16c953c2e985b60 added file-abcde
  |
  @  3:6db330d65db434145c0b59d291853e9a84719b24 added file-abcd
  |
  o  2:abf2df566fc193b3ac34d946e63c1583e4d4732b added file-abc
  |
  o  1:69a232e754b08d568c4899475faf2eb44b857802 added file-ab
  |
  o  0:3004d2d9b50883c1538fc754a3aeb55f1b4084f6 added file-a
  

Recommit

  $ hg commit -m 'new change abcde'
  $ hg status
  $ hg heads -T '{rev}:{node} {desc}'
  5:0c07a3ccda771b25f1cb1edbd02e683723344ef1 new change abcde (no-eol)

Uncommit of non-existent and unchanged files has no effect
  $ hg uncommit nothinghere
  nothing to uncommit
  [1]
  $ hg status
  $ hg uncommit file-abc
  nothing to uncommit
  [1]
  $ hg status

Try partial uncommit, also moves bookmark

  $ hg bookmark
   * foo                       5:0c07a3ccda77
  $ hg uncommit files
  $ hg status
  M files
  $ hg bookmark
   * foo                       6:3727deee06f7
  $ hg heads -T '{rev}:{node} {desc}'
  6:3727deee06f72f5ffa8db792ee299cf39e3e190b new change abcde (no-eol)
  $ hg log -r . -p -T '{rev}:{node} {desc}'
  6:3727deee06f72f5ffa8db792ee299cf39e3e190b new change abcdediff -r 6db330d65db4 -r 3727deee06f7 file-abcde
  --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
  +++ b/file-abcde	Thu Jan 01 00:00:00 1970 +0000
  @@ -0,0 +1,1 @@
  +abcde
  
  $ hg log -G -T '{rev}:{node} {desc}' --hidden
  @  6:3727deee06f72f5ffa8db792ee299cf39e3e190b new change abcde
  |
  | x  5:0c07a3ccda771b25f1cb1edbd02e683723344ef1 new change abcde
  |/
  | x  4:6c4fd43ed714e7fcd8adbaa7b16c953c2e985b60 added file-abcde
  |/
  o  3:6db330d65db434145c0b59d291853e9a84719b24 added file-abcd
  |
  o  2:abf2df566fc193b3ac34d946e63c1583e4d4732b added file-abc
  |
  o  1:69a232e754b08d568c4899475faf2eb44b857802 added file-ab
  |
  o  0:3004d2d9b50883c1538fc754a3aeb55f1b4084f6 added file-a
  
  $ hg commit -m 'update files for abcde'

Uncommit with dirty state

  $ echo "foo" >> files
  $ cat files
  abcde
  foo
  $ hg status
  M files
  $ hg uncommit
  abort: uncommitted changes
  [255]
  $ hg uncommit files
  $ cat files
  abcde
  foo
  $ hg commit --amend -m "files abcde + foo"

Testing the 'experimental.uncommitondirtywdir' config

  $ echo "bar" >> files
  $ hg uncommit
  abort: uncommitted changes
  [255]
  $ hg uncommit --config experimental.uncommitondirtywdir=True
  $ hg commit -m "files abcde + foo"

Uncommit in the middle of a stack, does not move bookmark

  $ hg checkout '.^^^'
  1 files updated, 0 files merged, 2 files removed, 0 files unresolved
  (leaving bookmark foo)
  $ hg log -r . -p -T '{rev}:{node} {desc}'
  2:abf2df566fc193b3ac34d946e63c1583e4d4732b added file-abcdiff -r 69a232e754b0 -r abf2df566fc1 file-abc
  --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
  +++ b/file-abc	Thu Jan 01 00:00:00 1970 +0000
  @@ -0,0 +1,1 @@
  +abc
  diff -r 69a232e754b0 -r abf2df566fc1 files
  --- a/files	Thu Jan 01 00:00:00 1970 +0000
  +++ b/files	Thu Jan 01 00:00:00 1970 +0000
  @@ -1,1 +1,1 @@
  -ab
  +abc
  
  $ hg bookmark
     foo                       10:48e5bd7cd583
  $ hg uncommit
  3 new orphan changesets
  $ hg status
  M files
  A file-abc
  $ hg heads -T '{rev}:{node} {desc}'
  10:48e5bd7cd583eb24164ef8b89185819c84c96ed7 files abcde + foo (no-eol)
  $ hg bookmark
     foo                       10:48e5bd7cd583
  $ hg commit -m 'new abc'
  created new head

Partial uncommit in the middle, does not move bookmark

  $ hg checkout '.^'
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  $ hg log -r . -p -T '{rev}:{node} {desc}'
  1:69a232e754b08d568c4899475faf2eb44b857802 added file-abdiff -r 3004d2d9b508 -r 69a232e754b0 file-ab
  --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
  +++ b/file-ab	Thu Jan 01 00:00:00 1970 +0000
  @@ -0,0 +1,1 @@
  +ab
  diff -r 3004d2d9b508 -r 69a232e754b0 files
  --- a/files	Thu Jan 01 00:00:00 1970 +0000
  +++ b/files	Thu Jan 01 00:00:00 1970 +0000
  @@ -1,1 +1,1 @@
  -a
  +ab
  
  $ hg bookmark
     foo                       10:48e5bd7cd583
  $ hg uncommit file-ab
  1 new orphan changesets
  $ hg status
  A file-ab

  $ hg heads -T '{rev}:{node} {desc}\n'
  12:8eb87968f2edb7f27f27fe676316e179de65fff6 added file-ab
  11:5dc89ca4486f8a88716c5797fa9f498d13d7c2e1 new abc
  10:48e5bd7cd583eb24164ef8b89185819c84c96ed7 files abcde + foo

  $ hg bookmark
     foo                       10:48e5bd7cd583
  $ hg commit -m 'update ab'
  $ hg status
  $ hg heads -T '{rev}:{node} {desc}\n'
  13:f21039c59242b085491bb58f591afc4ed1c04c09 update ab
  11:5dc89ca4486f8a88716c5797fa9f498d13d7c2e1 new abc
  10:48e5bd7cd583eb24164ef8b89185819c84c96ed7 files abcde + foo

  $ hg log -G -T '{rev}:{node} {desc}' --hidden
  @  13:f21039c59242b085491bb58f591afc4ed1c04c09 update ab
  |
  o  12:8eb87968f2edb7f27f27fe676316e179de65fff6 added file-ab
  |
  | *  11:5dc89ca4486f8a88716c5797fa9f498d13d7c2e1 new abc
  | |
  | | *  10:48e5bd7cd583eb24164ef8b89185819c84c96ed7 files abcde + foo
  | | |
  | | | x  9:8a6b58c173ca6a2e3745d8bd86698718d664bc6c files abcde + foo
  | | |/
  | | | x  8:39ad452c7f684a55d161c574340c5766c4569278 update files for abcde
  | | |/
  | | | x  7:0977fa602c2fd7d8427ed4e7ee15ea13b84c9173 update files for abcde
  | | |/
  | | *  6:3727deee06f72f5ffa8db792ee299cf39e3e190b new change abcde
  | | |
  | | | x  5:0c07a3ccda771b25f1cb1edbd02e683723344ef1 new change abcde
  | | |/
  | | | x  4:6c4fd43ed714e7fcd8adbaa7b16c953c2e985b60 added file-abcde
  | | |/
  | | *  3:6db330d65db434145c0b59d291853e9a84719b24 added file-abcd
  | | |
  | | x  2:abf2df566fc193b3ac34d946e63c1583e4d4732b added file-abc
  | |/
  | x  1:69a232e754b08d568c4899475faf2eb44b857802 added file-ab
  |/
  o  0:3004d2d9b50883c1538fc754a3aeb55f1b4084f6 added file-a
  
Uncommit with draft parent

  $ hg uncommit
  $ hg phase -r .
  12: draft
  $ hg commit -m 'update ab again'

Phase is preserved

  $ hg uncommit --keep --config phases.new-commit=secret
  $ hg phase -r .
  15: draft
  $ hg commit --amend -m 'update ab again'

Uncommit with public parent

  $ hg phase -p "::.^"
  $ hg uncommit
  $ hg phase -r .
  12: public

Partial uncommit with public parent

  $ echo xyz > xyz
  $ hg add xyz
  $ hg commit -m "update ab and add xyz"
  $ hg uncommit xyz
  $ hg status
  A xyz
  $ hg phase -r .
  18: draft
  $ hg phase -r ".^"
  12: public

Uncommit leaving an empty changeset

  $ cd $TESTTMP
  $ hg init repo1
  $ cd repo1
  $ hg debugdrawdag <<'EOS'
  > Q
  > |
  > P
  > EOS
  $ hg up Q -q
  $ hg uncommit --keep
  $ hg log -G -T '{desc} FILES: {files}'
  @  Q FILES:
  |
  | x  Q FILES: Q
  |/
  o  P FILES: P
  
  $ hg status
  A Q

  $ cd ..
  $ rm -rf repo1

Testing uncommit while merge

  $ hg init repo2
  $ cd repo2

Create some history

  $ touch a
  $ hg add a
  $ for i in 1 2 3; do echo $i > a; hg commit -m "a $i"; done
  $ hg checkout 0
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ touch b
  $ hg add b
  $ for i in 1 2 3; do echo $i > b; hg commit -m "b $i"; done
  created new head
  $ hg log -G -T '{rev}:{node} {desc}' --hidden
  @  5:2cd56cdde163ded2fbb16ba2f918c96046ab0bf2 b 3
  |
  o  4:c3a0d5bb3b15834ffd2ef9ef603e93ec65cf2037 b 2
  |
  o  3:49bb009ca26078726b8870f1edb29fae8f7618f5 b 1
  |
  | o  2:990982b7384266e691f1bc08ca36177adcd1c8a9 a 3
  | |
  | o  1:24d38e3cf160c7b6f5ffe82179332229886a6d34 a 2
  |/
  o  0:ea4e33293d4d274a2ba73150733c2612231f398c a 1
  

Add and expect uncommit to fail on both merge working dir and merge changeset

  $ hg merge 2
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)

  $ hg uncommit
  abort: outstanding uncommitted merge
  [255]

  $ hg uncommit --config experimental.uncommitondirtywdir=True
  abort: cannot uncommit while merging
  [255]

  $ hg status
  M a
  $ hg commit -m 'merge a and b'

  $ hg uncommit
  abort: cannot uncommit merge changeset
  [255]

  $ hg status
  $ hg log -G -T '{rev}:{node} {desc}' --hidden
  @    6:c03b9c37bc67bf504d4912061cfb527b47a63c6e merge a and b
  |\
  | o  5:2cd56cdde163ded2fbb16ba2f918c96046ab0bf2 b 3
  | |
  | o  4:c3a0d5bb3b15834ffd2ef9ef603e93ec65cf2037 b 2
  | |
  | o  3:49bb009ca26078726b8870f1edb29fae8f7618f5 b 1
  | |
  o |  2:990982b7384266e691f1bc08ca36177adcd1c8a9 a 3
  | |
  o |  1:24d38e3cf160c7b6f5ffe82179332229886a6d34 a 2
  |/
  o  0:ea4e33293d4d274a2ba73150733c2612231f398c a 1