fix test

Signed-off-by: Little-Wallace <bupt2013211450@gmail.com>

src/storage/src/hummock/mod.rs

@@ -415,12 +415,11 @@ pub fn hit_sstable_bloom_filter(
     local_stats: &mut StoreLocalStatistic,
 ) -> bool {
     local_stats.bloom_filter_check_counts += 1;
-    let surely_not_have = sstable_info_ref.surely_not_have_hashvalue(prefix_hash);
-
-    if surely_not_have {
+    let may_exist = sstable_info_ref.may_match_hash(prefix_hash);
+    if !may_exist {
         local_stats.bloom_filter_true_negative_count += 1;
     }
-    !surely_not_have
+    may_exist
 }
 
 /// Get `user_value` from `OrderSortedUncommittedData`. If not get successful, return None.

src/storage/src/hummock/sstable/bloom.rs

@@ -83,6 +83,10 @@ impl BloomFilterReader {
         self.data.is_empty()
     }
 
+    pub fn get_raw_data(&self) -> &[u8] {
+        &self.data
+    }
+
     /// Judges whether the hash value is in the table with the given false positive rate.
     ///
     /// Note:
@@ -91,7 +95,7 @@ impl BloomFilterReader {
     ///   - if the return value is true, then the table may or may not have the user key that has
     ///     the hash actually, a.k.a. we don't know the answer.
     pub fn may_match(&self, mut h: u32) -> bool {
-        if self.k > 30 {
+        if self.k > 30 || self.k == 00 {
             // potential new encoding for short Bloom filters
             true
         } else {

src/storage/src/hummock/sstable/builder.rs

@@ -508,7 +508,7 @@ pub(super) mod tests {
         assert_eq!(table.has_bloom_filter(), with_blooms);
         for i in 0..key_count {
             let full_key = test_key_of(i);
-            assert!(!table.surely_not_have_dist_key(full_key.user_key.encode().as_slice()));
+            assert!(table.may_match(full_key.user_key.encode().as_slice()));
         }
     }
 

src/storage/src/hummock/sstable/mod.rs

@@ -147,16 +147,16 @@ impl Sstable {
         !self.filter_reader.is_empty()
     }
 
-    pub fn surely_not_have_dist_key(&self, dist_key: &[u8]) -> bool {
+    pub fn may_match(&self, dist_key: &[u8]) -> bool {
         let enable_bloom_filter: fn() -> bool = || {
             fail_point!("disable_bloom_filter", |_| false);
             true
         };
         if enable_bloom_filter() && self.has_bloom_filter() {
             let hash = xxh32::xxh32(dist_key, 0);
-            self.surely_not_have_hashvalue(hash)
+            self.may_match_hash(hash)
         } else {
-            false
+            true
         }
     }
 
@@ -166,8 +166,8 @@ impl Sstable {
     }
 
     #[inline(always)]
-    pub fn surely_not_have_hashvalue(&self, hash: u32) -> bool {
-        !self.filter_reader.may_match(hash)
+    pub fn may_match_hash(&self, hash: u32) -> bool {
+        self.filter_reader.may_match(hash)
     }
 
     pub fn block_count(&self) -> usize {

src/storage/src/hummock/sstable_store.rs

@@ -869,14 +869,22 @@ mod tests {
     async fn validate_sst(
         sstable_store: SstableStoreRef,
         info: &SstableInfo,
-        meta: SstableMeta,
+        mut meta: SstableMeta,
         x_range: Range<usize>,
     ) {
         let mut stats = StoreLocalStatistic::default();
         let holder = sstable_store.sstable(info, &mut stats).await.unwrap();
+        let mut filter_data = std::mem::take(&mut meta.bloom_filter);
+        if !filter_data.is_empty() {
+            filter_data.pop();
+        }
         assert_eq!(holder.value().meta, meta);
         let holder = sstable_store.sstable(info, &mut stats).await.unwrap();
         assert_eq!(holder.value().meta, meta);
+        assert_eq!(
+            filter_data.as_slice(),
+            holder.value().filter_reader.get_raw_data()
+        );
         let mut iter = SstableIterator::new(
             holder,
             sstable_store,

src/workspace-hack/Cargo.toml

@@ -22,6 +22,7 @@ ahash = { version = "0.8" }
 anyhow = { version = "1", features = ["backtrace"] }
 arrayvec = { version = "0.7", default-features = false, features = ["std"] }
 auto_enums = { version = "0.7", features = ["futures"] }
+base64 = { version = "0.21" }
 bytes = { version = "1", features = ["serde"] }
 chrono = { version = "0.4" }
 clap = { version = "3", features = ["derive", "env"] }
@@ -109,6 +110,7 @@ anyhow = { version = "1", features = ["backtrace"] }
 arrayvec = { version = "0.7", default-features = false, features = ["std"] }
 auto_enums = { version = "0.7", features = ["futures"] }
 auto_enums_derive = { version = "0.7", default-features = false, features = ["futures", "std"] }
+base64 = { version = "0.21" }
 bytes = { version = "1", features = ["serde"] }
 cc = { version = "1", default-features = false, features = ["parallel"] }
 chrono = { version = "0.4" }