view tests/test-contrib-testparseutil.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

  $ testparseutil="$TESTDIR"/../contrib/testparseutil.py

Internal test by doctest

  $ "$PYTHON" -m doctest "$testparseutil"

Tests for embedded python script

Typical cases

  $ "$PYTHON" "$testparseutil" -v pyembedded <<NO_CHECK_EOF
  >   >>> for f in [1, 2, 3]:
  >   ...     foo = 1
  >   >>> foo = 2
  >   $ echo "doctest is terminated by command, empty line, or comment"
  >   >>> foo = 31
  >   expected output of doctest fragment
  >   >>> foo = 32
  >   
  >   >>> foo = 33
  > 
  >   >>> foo = 34
  > comment
  >   >>> foo = 35
  > 
  >   $ "\$PYTHON" <<EOF
  >   > foo = 4
  >   > 
  >   > EOF
  >   $ cat > foo.py <<EOF
  >   > foo = 5
  >   > EOF
  >   $ cat >> foo.py <<EOF
  >   > foo = 6 # appended
  >   > EOF
  > 
  > NO_CHECK_EOF limit mark makes parsing ignore corresponded fragment
  > (this is useful to use bad code intentionally)
  > 
  >   $ "\$PYTHON" <<NO_CHECK_EOF
  >   > foo = 7 # this should be ignored at detection
  >   > NO_CHECK_EOF
  >   $ cat > foo.py <<NO_CHECK_EOF
  >   > foo = 8 # this should be ignored at detection
  >   > NO_CHECK_EOF
  > 
  > doctest fragment ended by EOF
  > 
  >   >>> foo = 9
  > NO_CHECK_EOF
  <stdin>:1: <anonymous> starts
    |for f in [1, 2, 3]:
    |    foo = 1
    |foo = 2
  <stdin>:4: <anonymous> ends
  <stdin>:5: <anonymous> starts
    |foo = 31
    |
    |foo = 32
    |
    |foo = 33
  <stdin>:10: <anonymous> ends
  <stdin>:11: <anonymous> starts
    |foo = 34
  <stdin>:12: <anonymous> ends
  <stdin>:13: <anonymous> starts
    |foo = 35
  <stdin>:14: <anonymous> ends
  <stdin>:16: <anonymous> starts
    |foo = 4
    |
  <stdin>:18: <anonymous> ends
  <stdin>:20: foo.py starts
    |foo = 5
  <stdin>:21: foo.py ends
  <stdin>:23: foo.py starts
    |foo = 6 # appended
  <stdin>:24: foo.py ends
  <stdin>:38: <anonymous> starts
    |foo = 9
  <stdin>:39: <anonymous> ends

Invalid test script

(similar test for shell script and hgrc configuration is omitted,
because this tests common base class of them)

  $ "$PYTHON" "$testparseutil" -v pyembedded <<NO_CHECK_EOF > detected
  >   $ "\$PYTHON" <<EOF
  >   > foo = 1
  > 
  >   $ "\$PYTHON" <<EOF
  >   > foo = 2
  >   $ cat > bar.py <<EOF
  >   > bar = 2 # this fragment will be detected as expected
  >   > EOF
  > 
  >   $ cat > foo.py <<EOF
  >   > foo = 3
  > NO_CHECK_EOF
  <stdin>:3: unexpected line for "heredoc python invocation"
  <stdin>:6: unexpected line for "heredoc python invocation"
  <stdin>:11: unexpected end of file for "heredoc .py file"
  [1]
  $ cat detected
  <stdin>:7: bar.py starts
    |bar = 2 # this fragment will be detected as expected
  <stdin>:8: bar.py ends

Tests for embedded shell script

  $ "$PYTHON" "$testparseutil" -v shembedded <<NO_CHECK_EOF
  >   $ cat > foo.sh <<EOF
  >   > foo = 1
  >   > 
  >   > foo = 2
  >   > EOF
  >   $ cat >> foo.sh <<EOF
  >   > foo = 3 # appended
  >   > EOF
  > 
  > NO_CHECK_EOF limit mark makes parsing ignore corresponded fragment
  > (this is useful to use bad code intentionally)
  > 
  >   $ cat > foo.sh <<NO_CHECK_EOF
  >   > # this should be ignored at detection
  >   > foo = 4
  >   > NO_CHECK_EOF
  > 
  > NO_CHECK_EOF
  <stdin>:2: foo.sh starts
    |foo = 1
    |
    |foo = 2
  <stdin>:5: foo.sh ends
  <stdin>:7: foo.sh starts
    |foo = 3 # appended
  <stdin>:8: foo.sh ends

Tests for embedded hgrc configuration

  $ "$PYTHON" "$testparseutil" -v hgrcembedded <<NO_CHECK_EOF
  >   $ cat > .hg/hgrc <<EOF
  >   > [ui]
  >   > verbose = true
  >   > 
  >   > # end of local configuration
  >   > EOF
  > 
  >   $ cat > \$HGRCPATH <<EOF
  >   > [extensions]
  >   > rebase =
  >   > # end of global configuration
  >   > EOF
  > 
  >   $ cat >> \$HGRCPATH <<EOF
  >   > # appended
  >   > [extensions]
  >   > rebase =!
  >   > EOF
  > 
  > NO_CHECK_EOF limit mark makes parsing ignore corresponded fragment
  > (this is useful to use bad code intentionally)
  > 
  >   $ cat > .hg/hgrc <<NO_CHECK_EOF
  >   > # this local configuration should be ignored at detection
  >   > [ui]
  >   > username = foo bar
  >   > NO_CHECK_EOF
  > 
  >   $ cat > \$HGRCPATH <<NO_CHECK_EOF
  >   > # this global configuration should be ignored at detection
  >   > [extensions]
  >   > foobar =
  >   > NO_CHECK_EOF
  > NO_CHECK_EOF
  <stdin>:2: .hg/hgrc starts
    |[ui]
    |verbose = true
    |
    |# end of local configuration
  <stdin>:6: .hg/hgrc ends
  <stdin>:9: $HGRCPATH starts
    |[extensions]
    |rebase =
    |# end of global configuration
  <stdin>:12: $HGRCPATH ends
  <stdin>:15: $HGRCPATH starts
    |# appended
    |[extensions]
    |rebase =!
  <stdin>:18: $HGRCPATH ends