view tests/test-rename-dir-merge.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

  $ hg init t
  $ cd t

  $ mkdir a
  $ echo foo > a/a
  $ echo bar > a/b
  $ hg ci -Am "0"
  adding a/a
  adding a/b

  $ hg co -C 0
  0 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ hg mv a b
  moving a/a to b/a
  moving a/b to b/b
  $ hg ci -m "1 mv a/ b/"

  $ hg co -C 0
  2 files updated, 0 files merged, 2 files removed, 0 files unresolved
  $ echo baz > a/c
  $ echo quux > a/d
  $ hg add a/c
  $ hg ci -m "2 add a/c"
  created new head

  $ hg merge --debug 1
    unmatched files in local:
     a/c
    unmatched files in other:
     b/a
     b/b
    all copies found (* = to merge, ! = divergent, % = renamed and deleted):
     on remote side:
      src: 'a/a' -> dst: 'b/a' 
      src: 'a/b' -> dst: 'b/b' 
    checking for directory renames
     discovered dir src: 'a/' -> dst: 'b/'
     pending file src: 'a/c' -> dst: 'b/c'
  resolving manifests
   branchmerge: True, force: False, partial: False
   ancestor: f9b20c0d4c51, local: ce36d17b18fb+, remote: 397f8b00a740
   a/a: other deleted -> r
  removing a/a
   a/b: other deleted -> r
  removing a/b
   b/a: remote created -> g
  getting b/a
   b/b: remote created -> g
  getting b/b
   b/c: remote directory rename - move from a/c -> dm
  moving a/c to b/c
  3 files updated, 0 files merged, 2 files removed, 0 files unresolved
  (branch merge, don't forget to commit)

  $ echo a/* b/*
  a/d b/a b/b b/c
  $ hg st -C
  M b/a
  M b/b
  A b/c
    a/c
  R a/a
  R a/b
  R a/c
  ? a/d
  $ hg ci -m "3 merge 2+1"
  $ hg debugrename b/c
  b/c renamed from a/c:354ae8da6e890359ef49ade27b68bbc361f3ca88

  $ hg co -C 1
  0 files updated, 0 files merged, 1 files removed, 0 files unresolved
  $ hg merge --debug 2
    unmatched files in local:
     b/a
     b/b
    unmatched files in other:
     a/c
    all copies found (* = to merge, ! = divergent, % = renamed and deleted):
     on local side:
      src: 'a/a' -> dst: 'b/a' 
      src: 'a/b' -> dst: 'b/b' 
    checking for directory renames
     discovered dir src: 'a/' -> dst: 'b/'
     pending file src: 'a/c' -> dst: 'b/c'
  resolving manifests
   branchmerge: True, force: False, partial: False
   ancestor: f9b20c0d4c51, local: 397f8b00a740+, remote: ce36d17b18fb
  starting 4 threads for background file closing (?)
   b/c: local directory rename - get from a/c -> dg
  getting a/c to b/c
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)

  $ echo a/* b/*
  a/d b/a b/b b/c
  $ hg st -C
  A b/c
    a/c
  ? a/d
  $ hg ci -m "4 merge 1+2"
  created new head
  $ hg debugrename b/c
  b/c renamed from a/c:354ae8da6e890359ef49ade27b68bbc361f3ca88

Local directory rename with conflicting file added in remote source directory
and untracked in local target directory.

  $ hg co -qC 1
  $ echo target > b/c
  $ hg merge 2
  b/c: untracked file differs
  abort: untracked files in working directory differ from files in requested revision
  [20]
  $ cat b/c
  target
but it should succeed if the content matches
  $ hg cat -r 2 a/c > b/c
  $ hg merge 2
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ hg st -C
  A b/c
    a/c
  ? a/d

Local directory rename with conflicting file added in remote source directory
and committed in local target directory.

  $ hg co -qC 1
  $ echo target > b/c
  $ hg add b/c
  $ hg commit -qm 'new file in target directory'
  $ hg merge 2
  merging b/c and a/c to b/c
  warning: conflicts while merging b/c! (edit, then use 'hg resolve --mark')
  0 files updated, 0 files merged, 0 files removed, 1 files unresolved
  use 'hg resolve' to retry unresolved file merges or 'hg merge --abort' to abandon
  [1]
  $ hg st -A
  M b/c
    a/c
  ? a/d
  ? b/c.orig
  C b/a
  C b/b
  $ cat b/c
  <<<<<<< working copy: f1c50ca4f127 - test: new file in target directory
  target
  =======
  baz
  >>>>>>> merge rev:    ce36d17b18fb - test: 2 add a/c
  $ rm b/c.orig

Remote directory rename with conflicting file added in remote target directory
and committed in local source directory.

  $ hg co -qC 2
  $ hg st -A
  ? a/d
  C a/a
  C a/b
  C a/c
  $ hg merge 5
  merging a/c and b/c to b/c
  warning: conflicts while merging b/c! (edit, then use 'hg resolve --mark')
  2 files updated, 0 files merged, 2 files removed, 1 files unresolved
  use 'hg resolve' to retry unresolved file merges or 'hg merge --abort' to abandon
  [1]
  $ hg st -A
  M b/a
  M b/b
  M b/c
    a/c
  R a/a
  R a/b
  R a/c
  ? a/d
  ? b/c.orig
  $ cat b/c
  <<<<<<< working copy: ce36d17b18fb - test: 2 add a/c
  baz
  =======
  target
  >>>>>>> merge rev:    f1c50ca4f127 - test: new file in target directory

Second scenario with two repos:

  $ cd ..
  $ hg init r1
  $ cd r1
  $ mkdir a
  $ echo foo > a/f
  $ hg add a
  adding a/f
  $ hg ci -m "a/f == foo"
  $ cd ..

  $ hg clone r1 r2
  updating to branch default
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ cd r2
  $ hg mv a b
  moving a/f to b/f
  $ echo foo1 > b/f
  $ hg ci -m" a -> b, b/f == foo1"
  $ cd ..

  $ cd r1
  $ mkdir a/aa
  $ echo bar > a/aa/g
  $ hg add a/aa
  adding a/aa/g
  $ hg ci -m "a/aa/g"
  $ hg pull ../r2
  pulling from ../r2
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files (+1 heads)
  new changesets 7d51ed18da25
  1 local changesets published
  (run 'hg heads' to see heads, 'hg merge' to merge)

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

  $ hg st -C
  M b/f
  A b/aa/g
    a/aa/g
  R a/aa/g
  R a/f

  $ cd ..

Test renames to separate directories

  $ hg init a
  $ cd a
  $ mkdir a
  $ touch a/s
  $ touch a/t
  $ hg ci -Am0
  adding a/s
  adding a/t

Add more files

  $ touch a/s2
  $ touch a/t2
  $ hg ci -Am1
  adding a/s2
  adding a/t2

Do moves on a branch

  $ hg up 0
  0 files updated, 0 files merged, 2 files removed, 0 files unresolved
  $ mkdir s
  $ mkdir t
  $ hg mv a/s s
  $ hg mv a/t t
  $ hg ci -Am2
  created new head
  $ hg st --copies --change .
  A s/s
    a/s
  A t/t
    a/t
  R a/s
  R a/t

Merge shouldn't move s2, t2

  $ hg merge
  2 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ hg st --copies
  M a/s2
  M a/t2

Try the merge in the other direction. It may or may not be appropriate for
status to list copies here.

  $ hg up -C 1
  4 files updated, 0 files merged, 2 files removed, 0 files unresolved
  $ hg merge
  2 files updated, 0 files merged, 2 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ hg st --copies
  M s/s
  M t/t
  R a/s
  R a/t