Publicly export Snapshot (and minor changes)

src/db_impl.rs

@@ -279,7 +279,6 @@ impl DB {
                 save_manifest = true;
                 mem = MemTable::new(cmp.clone());
             }
-            batch.clear();
         }
 
         // Check if we can reuse the last log file.

src/db_iter.rs

@@ -33,7 +33,7 @@ pub struct DBIterator {
 }
 
 impl DBIterator {
-    pub fn new(
+    pub(crate) fn new(
         cmp: Rc<Box<dyn Cmp>>,
         vset: Shared<VersionSet>,
         iter: MergingIter,

src/disk_env.rs

@@ -161,7 +161,7 @@ impl Env for PosixDiskEnv {
             )
         } else {
             let f = locks.remove(&l.id).unwrap();
-            if f.unlock().is_err() {
+            if FileExt::unlock(&f).is_err() {
                 return err(StatusCode::LockError, &format!("unlock failed: {}", l.id));
             }
             Ok(())

src/lib.rs

@@ -109,5 +109,6 @@ pub use filter::{BloomPolicy, FilterPolicy};
 pub use mem_env::MemEnv;
 pub use options::{in_memory, CompressorList, Options};
 pub use skipmap::SkipMap;
+pub use snapshot::Snapshot;
 pub use types::LdbIterator;
 pub use write_batch::WriteBatch;

src/snapshot.rs

@@ -22,6 +22,8 @@ impl Drop for InnerSnapshot {
     }
 }
 
+/// Can be obtained from [`DB::get_snapshot`](crate::DB::get_snapshot), and used
+/// in some methods of the DB.
 #[derive(Clone)]
 pub struct Snapshot {
     inner: Rc<InnerSnapshot>,

src/write_batch.rs

@@ -11,15 +11,26 @@ const HEADER_SIZE: usize = 12;
 
 /// A WriteBatch contains entries to be written to a MemTable (for example) in a compact form.
 ///
-/// The storage format is (with the respective length in bytes)
+/// The storage format has a 12-byte header: an 8-byte little-endian sequence number, followed by
+/// a 4-byte little-endian count of entries. The header's sequence number is zero until the
+/// WriteBatch is encoded with a given sequence number.
 ///
-/// [tag: 1, keylen: ~var, key: keylen, vallen: ~var, val: vallen]
+/// After the header are entries with one of the following formats
+/// (with respective lengths in bytes, where ~var denotes a varint):
+/// - [tag: 1, keylen: ~var, key: keylen]
+/// - [tag: 1, keylen: ~var, key: keylen, vallen: ~var, val: vallen]
+///
+/// A WriteBatch entry for `delete` uses a tag of 0 and the first format (with only a key).
+///
+/// A WriteBatch entry for `put` uses a tag of 1 and the second format
+/// (that includes both a key and value).
 pub struct WriteBatch {
     entries: Vec<u8>,
 }
 
 impl WriteBatch {
-    pub(crate) fn new() -> WriteBatch {
+    /// Initializes an empty WriteBatch with only a 12-byte header, set to zero.
+    pub fn new() -> WriteBatch {
         let mut v = Vec::with_capacity(128);
         v.resize(HEADER_SIZE, 0);
 
@@ -60,9 +71,10 @@ impl WriteBatch {
         self.set_count(c + 1);
     }
 
-    /// Clear the contents of a WriteBatch.
+    /// Clear the contents of a WriteBatch, and set the 12 header bytes to 0.
     pub fn clear(&mut self) {
-        self.entries.clear()
+        self.entries.clear();
+        self.entries.resize(HEADER_SIZE, 0);
     }
 
     fn byte_size(&self) -> usize {