view tests/test-narrow-pull.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

  $ . "$TESTDIR/narrow-library.sh"

  $ hg init master
  $ cd master
  $ cat >> .hg/hgrc <<EOF
  > [narrow]
  > serveellipses=True
  > EOF
  $ for x in `$TESTDIR/seq.py 10`
  > do
  >   echo $x > "f$x"
  >   hg add "f$x"
  >   hg commit -m "Commit f$x"
  > done
  $ cd ..

narrow clone a couple files, f2 and f8

  $ hg clone --narrow ssh://user@dummy/master narrow --include "f2" --include "f8"
  requesting all changes
  adding changesets
  adding manifests
  adding file changes
  added 5 changesets with 2 changes to 2 files
  new changesets *:* (glob)
  updating to branch default
  2 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ cd narrow
  $ ls -A
  .hg
  f2
  f8
  $ cat f2 f8
  2
  8

  $ cd ..

change every upstream file twice

  $ cd master
  $ for x in `$TESTDIR/seq.py 10`
  > do
  >   echo "update#1 $x" >> "f$x"
  >   hg commit -m "Update#1 to f$x" "f$x"
  > done
  $ for x in `$TESTDIR/seq.py 10`
  > do
  >   echo "update#2 $x" >> "f$x"
  >   hg commit -m "Update#2 to f$x" "f$x"
  > done
  $ cd ..

look for incoming changes

  $ cd narrow
  $ hg incoming --limit 3
  comparing with ssh://user@dummy/master
  searching for changes
  changeset:   5:ddc055582556
  user:        test
  date:        Thu Jan 01 00:00:00 1970 +0000
  summary:     Update#1 to f1
  
  changeset:   6:f66eb5ad621d
  user:        test
  date:        Thu Jan 01 00:00:00 1970 +0000
  summary:     Update#1 to f2
  
  changeset:   7:c42ecff04e99
  user:        test
  date:        Thu Jan 01 00:00:00 1970 +0000
  summary:     Update#1 to f3
  

Interrupting the pull is safe
  $ hg --config hooks.pretxnchangegroup.bad=false pull -q
  transaction abort!
  rollback completed
  abort: pretxnchangegroup.bad hook exited with status 1
  [40]
  $ hg id
  223311e70a6f tip

pull new changes down to the narrow clone. Should get 8 new changesets: 4
relevant to the narrow spec, and 4 ellipsis nodes gluing them all together.

  $ hg pull
  pulling from ssh://user@dummy/master
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 9 changesets with 4 changes to 2 files
  new changesets *:* (glob)
  (run 'hg update' to get a working copy)
  $ hg log -T '{rev}: {desc}\n'
  13: Update#2 to f10
  12: Update#2 to f8
  11: Update#2 to f7
  10: Update#2 to f2
  9: Update#2 to f1
  8: Update#1 to f8
  7: Update#1 to f7
  6: Update#1 to f2
  5: Update#1 to f1
  4: Commit f10
  3: Commit f8
  2: Commit f7
  1: Commit f2
  0: Commit f1
  $ hg update tip
  2 files updated, 0 files merged, 0 files removed, 0 files unresolved

add a change and push it

  $ echo "update#3 2" >> f2
  $ hg commit -m "Update#3 to f2" f2
  $ hg log f2 -T '{rev}: {desc}\n'
  14: Update#3 to f2
  10: Update#2 to f2
  6: Update#1 to f2
  1: Commit f2
  $ hg push
  pushing to ssh://user@dummy/master
  searching for changes
  remote: adding changesets
  remote: adding manifests
  remote: adding file changes
  remote: added 1 changesets with 1 changes to 1 files
  $ cd ..

  $ cd master
  $ hg log f2 -T '{rev}: {desc}\n'
  30: Update#3 to f2
  21: Update#2 to f2
  11: Update#1 to f2
  1: Commit f2
  $ hg log -l 3 -T '{rev}: {desc}\n'
  30: Update#3 to f2
  29: Update#2 to f10
  28: Update#2 to f9

Can pull into repo with a single commit

  $ cd ..
  $ hg clone -q --narrow ssh://user@dummy/master narrow2 --include "f1" -r 0
  $ cd narrow2
  $ hg pull -q -r 1
  remote: abort: unexpected error: unable to resolve parent while packing '00manifest.i' 1 for changeset 0
  transaction abort!
  rollback completed
  abort: pull failed on remote
  [255]