view tests/test-clone-pull-corruption.t @ 47965:f9e6f2bb721d

rhg: Don’t compare ambiguous files one byte at a time

Even though the use of `BufReader` reduces the number of syscalls to read
the file from disk, `.bytes()` yields a separate `Result` for every byte.
Creating those results and dispatching on them is most likely costly.

Instead, this commit opts for simplicity by reading the entire file into memory
and comparing a single pair of byte strings. Note that memory already needs to
contain the entire previous contents of the file, as read from the filelog.
So with an extremely large file this doubles memory use but does not make it
grow by orders of magnitude.

At first I wrote code that still avoids reading the entire file into memory
and compares one buffer at a time with `BufReader`. Find this code below for
posterity. However its correctness is subtle. I ended up preferring the
simplicity of the obviously-correct single comparison.

```rust
let mut reader = BufReader::new(fobj);
let mut expected = &contents_in_p1[..];
loop {
    let buf = reader.fill_buf().when_reading_file(&fs_path)?;
    if buf.is_empty() {
        // Found EOF
        return Ok(expected.is_empty());
    } else if let Some(rest) = expected.drop_prefix(buf) {
        // What we read so far matches the expected content, continue reading
        let buf_len = buf.len();
        reader.consume(buf_len);
        expected = rest
    } else {
        // Found different content
        return Ok(false);
    }
}
```

Differential Revision: https://phab.mercurial-scm.org/D11412

Corrupt an hg repo with a pull started during an aborted commit
Create two repos, so that one of them can pull from the other one.

  $ hg init source
  $ cd source
  $ touch foo
  $ hg add foo
  $ hg ci -m 'add foo'
  $ hg clone . ../corrupted
  updating to branch default
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ echo >> foo
  $ hg ci -m 'change foo'

Add a hook to wait 5 seconds and then abort the commit

  $ cd ../corrupted
  $ echo "[hooks]" >> .hg/hgrc
  $ echo 'pretxncommit = sh -c "sleep 5; exit 1"' >> .hg/hgrc

start a commit...

  $ touch bar
  $ hg add bar
  $ hg ci -m 'add bar' &

... and start a pull while the commit is still running

  $ sleep 1
  $ hg pull ../source 2>/dev/null
  pulling from ../source
  transaction abort!
  rollback completed
  abort: pretxncommit hook exited with status 1
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files
  new changesets 52998019f625
  (run 'hg update' to get a working copy)

see what happened

  $ wait
  $ hg verify
  checking changesets
  checking manifests
  crosschecking files in changesets and manifests
  checking files
  checked 2 changesets with 2 changes to 1 files

  $ cd ..