view tests/test-rebase-detach.t @ 46607:e9901d01d135

revlog: add a mechanism to verify expected file position before appending

If someone uses `hg debuglocks`, or some non-hg process writes to the .hg
directory without respecting the locks, or if the repo's on a networked
filesystem, it's possible for the revlog code to write out corrupted data.

The form of this corruption can vary depending on what data was written and how
that happened. We are in the "networked filesystem" case (though I've had users
also do this to themselves with the "`hg debuglocks`" scenario), and most often
see this with the changelog. What ends up happening is we produce two items
(let's call them rev1 and rev2) in the .i file that have the same linkrev,
baserev, and offset into the .d file, while the data in the .d file is appended
properly. rev2's compressed_size is accurate for rev2, but when we go to
decompress the data in the .d file, we use the offset that's recorded in the
index file, which is the same as rev1, and attempt to decompress
rev2.compressed_size bytes of rev1's data. This usually does not succeed. :)

When using inline data, this also fails, though I haven't investigated why too
closely. This shows up as a "patch decode" error. I believe what's happening
there is that we're basically ignoring the offset field, getting the data
properly, but since baserev != rev, it thinks this is a delta based on rev
(instead of a full text) and can't actually apply it as such.

For now, I'm going to make this an optional component and default it to entirely
off. I may increase the default severity of this in the future, once I've
enabled it for my users and we gain more experience with it. Luckily, most of my
users have a versioned filesystem and can roll back to before the corruption has
been written, it's just a hassle to do so and not everyone knows how (so it's a
support burden). Users on other filesystems will not have that luxury, and this
can cause them to have a corrupted repository that they are unlikely to know how
to resolve, and they'll see this as a data-loss event. Refusing to create the
corruption is a much better user experience.

This mechanism is not perfect. There may be false-negatives (racy writes that
are not detected). There should not be any false-positives (non-racy writes that
are detected as such). This is not a mechanism that makes putting a repo on a
networked filesystem "safe" or "supported", just *less* likely to cause
corruption.

Differential Revision: https://phab.mercurial-scm.org/D9952

  $ cat >> $HGRCPATH <<EOF
  > [extensions]
  > rebase=
  > 
  > [alias]
  > tglog = log -G --template "{rev}: {node|short} '{desc}'\n"
  > 
  > [extensions]
  > drawdag=$TESTDIR/drawdag.py
  > EOF

Rebasing D onto B detaching from C (one commit):

  $ hg init a1
  $ cd a1

  $ hg debugdrawdag <<EOF
  > D
  > |
  > C B
  > |/
  > A
  > EOF
  $ hg phase --force --secret D

  $ hg rebase -s D -d B
  rebasing 3:e7b3f00ed42e D tip "D"
  saved backup bundle to $TESTTMP/a1/.hg/strip-backup/e7b3f00ed42e-6f368371-rebase.hg

  $ hg log -G --template "{rev}:{phase} '{desc}' {branches}\n"
  o  3:secret 'D'
  |
  | o  2:draft 'C'
  | |
  o |  1:draft 'B'
  |/
  o  0:draft 'A'
  
  $ hg manifest --rev tip
  A
  B
  D

  $ cd ..


Rebasing D onto B detaching from C (two commits):

  $ hg init a2
  $ cd a2

  $ hg debugdrawdag <<EOF
  > E
  > |
  > D
  > |
  > C B
  > |/
  > A
  > EOF

  $ hg rebase -s D -d B
  rebasing 3:e7b3f00ed42e D "D"
  rebasing 4:69a34c08022a E tip "E"
  saved backup bundle to $TESTTMP/a2/.hg/strip-backup/e7b3f00ed42e-a2ec7cea-rebase.hg

  $ hg tglog
  o  4: ee79e0744528 'E'
  |
  o  3: 10530e1d72d9 'D'
  |
  | o  2: dc0947a82db8 'C'
  | |
  o |  1: 112478962961 'B'
  |/
  o  0: 426bada5c675 'A'
  
  $ hg manifest --rev tip
  A
  B
  D
  E

  $ cd ..

Rebasing C onto B using detach (same as not using it):

  $ hg init a3
  $ cd a3

  $ hg debugdrawdag <<EOF
  > D
  > |
  > C B
  > |/
  > A
  > EOF

  $ hg rebase -s C -d B
  rebasing 2:dc0947a82db8 C "C"
  rebasing 3:e7b3f00ed42e D tip "D"
  saved backup bundle to $TESTTMP/a3/.hg/strip-backup/dc0947a82db8-b8481714-rebase.hg

  $ hg tglog
  o  3: 7375f3dbfb0f 'D'
  |
  o  2: bbfdd6cb49aa 'C'
  |
  o  1: 112478962961 'B'
  |
  o  0: 426bada5c675 'A'
  
  $ hg manifest --rev tip
  A
  B
  C
  D

  $ cd ..


Rebasing D onto B detaching from C and collapsing:

  $ hg init a4
  $ cd a4

  $ hg debugdrawdag <<EOF
  > E
  > |
  > D
  > |
  > C B
  > |/
  > A
  > EOF
  $ hg phase --force --secret E

  $ hg rebase --collapse -s D -d B
  rebasing 3:e7b3f00ed42e D "D"
  rebasing 4:69a34c08022a E tip "E"
  saved backup bundle to $TESTTMP/a4/.hg/strip-backup/e7b3f00ed42e-a2ec7cea-rebase.hg

  $ hg  log -G --template "{rev}:{phase} '{desc}' {branches}\n"
  o  3:secret 'Collapsed revision
  |  * D
  |  * E'
  | o  2:draft 'C'
  | |
  o |  1:draft 'B'
  |/
  o  0:draft 'A'
  
  $ hg manifest --rev tip
  A
  B
  D
  E

  $ cd ..

Rebasing across null as ancestor
  $ hg init a5
  $ cd a5

  $ hg debugdrawdag <<EOF
  > E
  > |
  > D
  > |
  > C
  > |
  > A B
  > EOF

  $ hg rebase -s C -d B
  rebasing 2:dc0947a82db8 C "C"
  rebasing 3:e7b3f00ed42e D "D"
  rebasing 4:69a34c08022a E tip "E"
  saved backup bundle to $TESTTMP/a5/.hg/strip-backup/dc0947a82db8-3eefec98-rebase.hg

  $ hg tglog
  o  4: e3d0c70d606d 'E'
  |
  o  3: e9153d36a1af 'D'
  |
  o  2: a7ac28b870a8 'C'
  |
  o  1: fc2b737bb2e5 'B'
  
  o  0: 426bada5c675 'A'
  
  $ hg rebase -d 1 -s 3
  rebasing 3:e9153d36a1af "D"
  rebasing 4:e3d0c70d606d tip "E"
  saved backup bundle to $TESTTMP/a5/.hg/strip-backup/e9153d36a1af-db7388ed-rebase.hg
  $ hg tglog
  o  4: 2c24e540eccd 'E'
  |
  o  3: 73f786ed52ff 'D'
  |
  | o  2: a7ac28b870a8 'C'
  |/
  o  1: fc2b737bb2e5 'B'
  
  o  0: 426bada5c675 'A'
  
  $ cd ..

Verify that target is not selected as external rev (issue3085)

  $ hg init a6
  $ cd a6

  $ hg debugdrawdag <<EOF
  > H
  > | G
  > |/|
  > F E
  > |/
  > A
  > EOF
  $ hg up -q G

  $ echo "I" >> E
  $ hg ci -m "I"
  $ hg tag --local I
  $ hg merge H
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ hg ci -m "Merge"
  $ echo "J" >> F
  $ hg ci -m "J"
  $ hg tglog
  @  7: c6aaf0d259c0 'J'
  |
  o    6: 0cfbc7e8faaf 'Merge'
  |\
  | o  5: b92d164ad3cb 'I'
  | |
  o |  4: 4ea5b230dea3 'H'
  | |
  | o  3: c6001eacfde5 'G'
  |/|
  o |  2: 8908a377a434 'F'
  | |
  | o  1: 7fb047a69f22 'E'
  |/
  o  0: 426bada5c675 'A'
  
  $ hg rebase -s I -d H --collapse --config ui.merge=internal:other
  rebasing 5:b92d164ad3cb I "I"
  rebasing 6:0cfbc7e8faaf "Merge"
  rebasing 7:c6aaf0d259c0 tip "J"
  saved backup bundle to $TESTTMP/a6/.hg/strip-backup/b92d164ad3cb-88fd7ab7-rebase.hg

  $ hg tglog
  @  5: 65079693dac4 'Collapsed revision
  |  * I
  |  * Merge
  |  * J'
  o  4: 4ea5b230dea3 'H'
  |
  | o  3: c6001eacfde5 'G'
  |/|
  o |  2: 8908a377a434 'F'
  | |
  | o  1: 7fb047a69f22 'E'
  |/
  o  0: 426bada5c675 'A'
  

  $ hg log --rev tip
  changeset:   5:65079693dac4
  tag:         tip
  user:        test
  date:        Thu Jan 01 00:00:00 1970 +0000
  summary:     Collapsed revision
  

  $ cd ..

Ensure --continue restores a correct state (issue3046) and phase:
  $ hg init a7
  $ cd a7

  $ hg debugdrawdag <<EOF
  > C B
  > |/
  > A
  > EOF
  $ hg up -q C
  $ echo 'B2' > B
  $ hg ci -A -m 'B2'
  adding B
  $ hg phase --force --secret .
  $ hg rebase -s . -d B --config ui.merge=internal:merge
  rebasing 3:17b4880d2402 tip "B2"
  merging B
  warning: conflicts while merging B! (edit, then use 'hg resolve --mark')
  unresolved conflicts (see 'hg resolve', then 'hg rebase --continue')
  [240]
  $ hg resolve --all -t internal:local
  (no more unresolved files)
  continue: hg rebase --continue
  $ hg rebase -c
  rebasing 3:17b4880d2402 tip "B2"
  note: not rebasing 3:17b4880d2402 tip "B2", its destination already has all its changes
  saved backup bundle to $TESTTMP/a7/.hg/strip-backup/17b4880d2402-1ae1f6cc-rebase.hg
  $ hg  log -G --template "{rev}:{phase} '{desc}' {branches}\n"
  o  2:draft 'C'
  |
  | @  1:draft 'B'
  |/
  o  0:draft 'A'
  

  $ cd ..