Bump hashbrown, foldhash; Fix some compile errors in master (#3722)

# Description of Changes

There were mentions of `hashbrown` in the repo that did not go through
`spacetimedb_data_structures::map`.
This caused compile errors on master when running certain tests locally.
These have been replaced with the proper imports.

The PR also bump hashbrown to 0.16.1 and foldhash to 0.2.0.

# API and ABI breaking changes

None

# Expected complexity level and risk

2

# Testing

Covered by existing tests.

Cargo.lock

@@ -2184,6 +2184,12 @@ version = "0.1.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d9c4f5dac5e15c24eb999c26181a6ca40b39fe946cbe4c263c7209467bc83af2"
 
+[[package]]
+name = "foldhash"
+version = "0.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "77ce24cb58228fbb8aa041425bb1050850ac19177686ea6e0f41a70416f56fdb"
+
 [[package]]
 name = "foreign-types"
 version = "0.3.2"
@@ -2577,18 +2583,20 @@ checksum = "9229cfe53dfd69f0609a49f65461bd93001ea1ef889cd5529dd176593f5338a1"
 dependencies = [
  "allocator-api2",
  "equivalent",
- "foldhash",
- "rayon",
- "serde",
+ "foldhash 0.1.5",
 ]
 
 [[package]]
 name = "hashbrown"
-version = "0.16.0"
+version = "0.16.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5419bdc4f6a9207fbeba6d11b604d481addf78ecd10c11ad51e76c2f6482748d"
+checksum = "841d1cc9bed7f9236f321df977030373f4a4163ae1a7dbfe1a51a2c1a51d9100"
 dependencies = [
  "allocator-api2",
+ "equivalent",
+ "rayon",
+ "serde",
+ "serde_core",
 ]
 
 [[package]]
@@ -3156,7 +3164,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "6717a8d2a5a929a1a2eb43a12812498ed141a0bcfb7e8f7844fbdbe4303bba9f"
 dependencies = [
  "equivalent",
- "hashbrown 0.16.0",
+ "hashbrown 0.16.1",
  "serde",
  "serde_core",
 ]
@@ -4430,7 +4438,7 @@ checksum = "9c2294f53fe0d6a715ad1877300fe5432033db1e5804800f82ee84840bd09894"
 dependencies = [
  "allocator-api2",
  "bumpalo",
- "hashbrown 0.16.0",
+ "hashbrown 0.16.1",
  "oxc_data_structures",
  "oxc_estree",
  "rustc-hash",
@@ -7132,13 +7140,15 @@ dependencies = [
 name = "spacetimedb-bench"
 version = "1.9.0"
 dependencies = [
+ "ahash 0.8.12",
  "anyhow",
  "anymap",
  "byte-unit",
  "clap 4.5.50",
  "criterion",
- "foldhash",
+ "foldhash 0.2.0",
  "futures",
+ "hashbrown 0.16.1",
  "iai-callgrind",
  "iai-callgrind-macros",
  "iai-callgrind-runner",
@@ -7449,7 +7459,6 @@ dependencies = [
  "fs_extra",
  "futures",
  "futures-util",
- "hashbrown 0.15.5",
  "hex",
  "hostname",
  "http 1.3.1",
@@ -7547,7 +7556,7 @@ version = "1.9.0"
 dependencies = [
  "ahash 0.8.12",
  "crossbeam-queue",
- "hashbrown 0.15.5",
+ "hashbrown 0.16.1",
  "nohash-hasher",
  "serde",
  "smallvec",
@@ -7738,7 +7747,7 @@ version = "1.9.0"
 dependencies = [
  "decorum",
  "ethnum",
- "hashbrown 0.15.5",
+ "hashbrown 0.16.1",
  "smallvec",
 ]
 
@@ -7912,7 +7921,6 @@ dependencies = [
  "derive_more 0.99.20",
  "enum-as-inner",
  "enum-map",
- "hashbrown 0.15.5",
  "indexmap 2.12.0",
  "insta",
  "itertools 0.12.1",

Cargo.toml

@@ -179,7 +179,7 @@ env_logger = "0.10"
 ethnum = { version = "1.5.0", features = ["serde"] }
 flate2 = "1.0.24"
 flume = { version = "0.11", default-features = false, features = ["async"] }
-foldhash = "0.1.4"
+foldhash = "0.2.0"
 fs-err = "2.9.0"
 fs_extra = "1.3.0"
 fs2 = "0.4.3"
@@ -189,7 +189,7 @@ futures-util = "0.3"
 getrandom02 = { package = "getrandom", version = "0.2" }
 git2 = "0.19"
 glob = "0.3.1"
-hashbrown = { version = "0.15", default-features = false, features = ["equivalent", "inline-more"] }
+hashbrown = { version = "0.16.1", default-features = false, features = ["equivalent", "inline-more", "rayon", "serde"] }
 headers = "0.4"
 heck = "0.4"
 hex = "0.4.3"

crates/bench/Cargo.toml

@@ -52,6 +52,7 @@ spacetimedb-standalone = { path = "../standalone" }
 spacetimedb-table = { path = "../table" }
 spacetimedb-testing = { path = "../testing" }
 
+ahash.workspace = true
 anyhow.workspace = true
 anymap.workspace = true
 byte-unit.workspace = true
@@ -59,6 +60,7 @@ clap.workspace = true
 criterion.workspace = true
 futures.workspace = true
 foldhash.workspace = true
+hashbrown.workspace = true
 lazy_static.workspace = true
 log.workspace = true
 mimalloc.workspace = true

crates/bench/benches/index.rs

@@ -1,12 +1,12 @@
-use core::{hint::black_box, iter::repeat_with, time::Duration};
+use core::{hash::BuildHasherDefault, hint::black_box, iter::repeat_with, time::Duration};
 use criterion::{
     criterion_group, criterion_main,
     measurement::{Measurement as _, WallTime},
     BenchmarkGroup, Criterion,
 };
 use foldhash::{HashSet, HashSetExt};
+use hashbrown::{hash_map::Entry, HashMap};
 use itertools::Itertools as _;
-use spacetimedb_data_structures::map::{Entry, HashMap};
 use spacetimedb_sats::layout::Size;
 use spacetimedb_table::indexes::{PageIndex, PageOffset, RowPointer, SquashedOffset};
 use spacetimedb_table::table_index::unique_direct_index::UniqueDirectIndex;
@@ -191,7 +191,7 @@ impl Index for IBTree {
 }
 
 #[derive(Clone)]
-struct IAHash(HashMap<K, RowPointer>);
+struct IAHash(HashMap<K, RowPointer, BuildHasherDefault<ahash::AHasher>>);
 impl Index for IAHash {
     const NAME: &'static str = "IAHash";
     fn new() -> Self {

crates/core/Cargo.toml

@@ -60,7 +60,6 @@ flume.workspace = true
 fs2.workspace = true
 futures.workspace = true
 futures-util.workspace = true
-hashbrown = { workspace = true, features = ["rayon", "serde"] }
 hex.workspace = true
 hostname.workspace = true
 http.workspace = true

crates/core/src/subscription/delta.rs

@@ -1,5 +1,6 @@
+use crate::host::module_host::UpdatesRelValue;
 use anyhow::Result;
-use hashbrown::HashMap;
+use spacetimedb_data_structures::map::{HashCollectionExt as _, HashMap};
 use spacetimedb_execution::{Datastore, DeltaStore, Row};
 use spacetimedb_lib::metrics::ExecutionMetrics;
 use spacetimedb_primitives::ColList;
@@ -7,8 +8,6 @@ use spacetimedb_sats::product_value::InvalidFieldError;
 use spacetimedb_subscription::SubscriptionPlan;
 use spacetimedb_vm::relation::RelValue;
 
-use crate::host::module_host::UpdatesRelValue;
-
 /// Evaluate a subscription over a delta update.
 /// Returns `None` for empty updates.
 ///

crates/core/src/subscription/module_subscription_actor.rs

@@ -1253,7 +1253,6 @@ mod tests {
     use crate::subscription::query::compile_read_only_query;
     use crate::subscription::TableUpdateType;
     use core::fmt;
-    use hashbrown::HashMap;
     use itertools::Itertools;
     use pretty_assertions::assert_matches;
     use spacetimedb_client_api_messages::energy::EnergyQuanta;
@@ -1262,6 +1261,7 @@ mod tests {
         TableUpdate, Unsubscribe, UnsubscribeMulti,
     };
     use spacetimedb_commitlog::{commitlog, repo};
+    use spacetimedb_data_structures::map::{HashCollectionExt as _, HashMap};
     use spacetimedb_datastore::system_tables::{StRowLevelSecurityRow, ST_ROW_LEVEL_SECURITY_ID};
     use spacetimedb_durability::{Durability, EmptyHistory, TxOffset};
     use spacetimedb_execution::dml::MutDatastore;

crates/core/src/subscription/module_subscription_manager.rs

@@ -12,14 +12,16 @@ use crate::subscription::delta::eval_delta;
 use crate::subscription::websocket_building::BuildableWebsocketFormat;
 use crate::worker_metrics::WORKER_METRICS;
 use core::mem;
-use hashbrown::hash_map::OccupiedError;
-use hashbrown::{HashMap, HashSet};
 use parking_lot::RwLock;
 use prometheus::IntGauge;
 use spacetimedb_client_api_messages::websocket::{
     BsatnFormat, CompressableQueryUpdate, FormatSwitch, JsonFormat, QueryId, QueryUpdate, SingleQueryUpdate,
 };
-use spacetimedb_data_structures::map::{Entry, IntMap};
+use spacetimedb_data_structures::map::HashCollectionExt as _;
+use spacetimedb_data_structures::map::{
+    hash_map::{Entry, OccupiedError},
+    HashMap, HashSet, IntMap,
+};
 use spacetimedb_datastore::locking_tx_datastore::state_view::StateView;
 use spacetimedb_durability::TxOffset;
 use spacetimedb_expr::expr::CollectViews;

crates/core/src/subscription/tx.rs

@@ -1,24 +1,21 @@
+use super::module_subscription_manager::QueriedTableIndexIds;
+use itertools::Either;
+use smallvec::SmallVec;
+use spacetimedb_data_structures::map::{HashCollectionExt as _, HashMap};
+use spacetimedb_datastore::{
+    locking_tx_datastore::{state_view::StateView, TxId},
+    traits::TxData,
+};
+use spacetimedb_execution::{Datastore, DeltaStore, Row};
+use spacetimedb_lib::{query::Delta, AlgebraicValue, ProductValue};
+use spacetimedb_primitives::{IndexId, TableId};
+use spacetimedb_table::table::{IndexScanRangeIter, TableScanIter};
 use std::{
     collections::BTreeMap,
     ops::{Deref, RangeBounds},
     sync::Arc,
 };
 
-use hashbrown::HashMap;
-use itertools::Either;
-use smallvec::SmallVec;
-use spacetimedb_execution::{Datastore, DeltaStore, Row};
-use spacetimedb_lib::{query::Delta, AlgebraicValue, ProductValue};
-use spacetimedb_primitives::{IndexId, TableId};
-use spacetimedb_table::table::{IndexScanRangeIter, TableScanIter};
-
-use spacetimedb_datastore::{
-    locking_tx_datastore::{state_view::StateView, TxId},
-    traits::TxData,
-};
-
-use super::module_subscription_manager::QueriedTableIndexIds;
-
 /// If an index is defined on a set of columns,
 /// and if that index is used in a subscription query,
 /// we build the very same index for each delta table.

crates/data-structures/src/map.rs

@@ -1,23 +1,34 @@
 use core::hash::{BuildHasher, BuildHasherDefault};
-pub use hashbrown::hash_map::Entry;
 use nohash_hasher::BuildNoHashHasher;
+
+pub use hashbrown::Equivalent;
 pub use nohash_hasher::IsEnabled as ValidAsIdentityHash;
 
+pub mod hash_set {
+    pub use super::HashSet;
+    pub use hashbrown::hash_set::*;
+}
+
+pub mod hash_map {
+    pub use super::HashMap;
+    pub use hashbrown::hash_map::*;
+}
+
 pub type DefaultHashBuilder = BuildHasherDefault<ahash::AHasher>;
 // TODO(centril): expose two maps instead,
 // one `map::fast::HashMap` and one `map::ddos::HashMap`.
 // In the first case we won't care about DDoS protection at all and can use `foldhash::fast`.
 // In the lattr, we can use e.g., randomized AHash.
-pub type HashMap<K, V, S = DefaultHashBuilder> = hashbrown::HashMap<K, V, S>;
-pub type HashSet<T, S = DefaultHashBuilder> = hashbrown::HashSet<T, S>;
+pub type HashMap<K, V> = hashbrown::HashMap<K, V, DefaultHashBuilder>;
+pub type HashSet<T> = hashbrown::HashSet<T, DefaultHashBuilder>;
 
 /// A version of [`HashMap<K, V>`] using the identity hash function,
 /// which is valid for any key type that can be converted to a `u64` without truncation.
-pub type IntMap<K, V> = HashMap<K, V, BuildNoHashHasher<K>>;
+pub type IntMap<K, V> = hashbrown::HashMap<K, V, BuildNoHashHasher<K>>;
 
 /// A version of [`HashSet<K>`] using the identity hash function,
 /// which is valid for any key type that can be converted to a `u64` without truncation.
-pub type IntSet<K> = HashSet<K, BuildNoHashHasher<K>>;
+pub type IntSet<K> = hashbrown::HashSet<K, BuildNoHashHasher<K>>;
 
 pub trait HashCollectionExt {
     /// Returns a new collection with default capacity, using `S::default()` to build the hasher.
@@ -27,22 +38,22 @@ pub trait HashCollectionExt {
     fn with_capacity(capacity: usize) -> Self;
 }
 
-impl<K, V, S: BuildHasher + Default> HashCollectionExt for HashMap<K, V, S> {
+impl<K, V, S: BuildHasher + Default> HashCollectionExt for hashbrown::HashMap<K, V, S> {
     fn new() -> Self {
-        HashMap::with_hasher(S::default())
+        Self::with_hasher(S::default())
     }
 
     fn with_capacity(capacity: usize) -> Self {
-        HashMap::with_capacity_and_hasher(capacity, S::default())
+        Self::with_capacity_and_hasher(capacity, S::default())
     }
 }
 
-impl<K, S: BuildHasher + Default> HashCollectionExt for HashSet<K, S> {
+impl<K, S: BuildHasher + Default> HashCollectionExt for hashbrown::HashSet<K, S> {
     fn new() -> Self {
-        HashSet::with_hasher(S::default())
+        Self::with_hasher(S::default())
     }
 
     fn with_capacity(capacity: usize) -> Self {
-        HashSet::with_capacity_and_hasher(capacity, S::default())
+        Self::with_capacity_and_hasher(capacity, S::default())
     }
 }

crates/datastore/src/locking_tx_datastore/datastore.rs

@@ -738,8 +738,7 @@ impl TxMetrics {
         // For each table, collect the extra stats, that we don't have in `tx_data`.
         let table_stats = tx_data
             .map(|tx_data| {
-                let mut table_stats =
-                    <HashMap<_, _, _> as HashCollectionExt>::with_capacity(tx_data.num_tables_affected());
+                let mut table_stats = HashMap::with_capacity(tx_data.num_tables_affected());
                 for (table_id, _) in tx_data.table_ids_and_names() {
                     let stats = committed_state.get_table(table_id).map(|table| TableStats {
                         row_count: table.row_count,

crates/memory-usage/Cargo.toml

@@ -6,11 +6,11 @@ license-file = "LICENSE"
 description = "Provides the trait `MemoryUsage`"
 
 [features]
-hashbrown = ["dep:hashbrown"]
+hash_map = ["dep:hashbrown"]
 smallvec = ["dep:smallvec"]
 ethnum = ["dep:ethnum"]
 decorum = ["dep:decorum"]
-default = ["hashbrown", "smallvec", "ethnum", "decorum"]
+default = ["hash_map", "smallvec", "ethnum", "decorum"]
 
 [dependencies]
 hashbrown = { workspace = true, optional = true }

crates/memory-usage/src/lib.rs

@@ -103,7 +103,7 @@ impl<T: MemoryUsage> MemoryUsage for Vec<T> {
     }
 }
 
-#[cfg(feature = "hashbrown")]
+#[cfg(feature = "hash_map")]
 impl<K: MemoryUsage + Eq + core::hash::Hash, V: MemoryUsage, S: core::hash::BuildHasher> MemoryUsage
     for hashbrown::HashMap<K, V, S>
 {
@@ -112,7 +112,7 @@ impl<K: MemoryUsage + Eq + core::hash::Hash, V: MemoryUsage, S: core::hash::Buil
     }
 }
 
-#[cfg(feature = "hashbrown")]
+#[cfg(feature = "hash_map")]
 impl<K: MemoryUsage + Eq + core::hash::Hash, S: core::hash::BuildHasher> MemoryUsage for hashbrown::HashSet<K, S> {
     fn heap_usage(&self) -> usize {
         self.allocation_size() + self.iter().map(|k| k.heap_usage()).sum::<usize>()

crates/schema/Cargo.toml

@@ -27,7 +27,6 @@ unicode-normalization.workspace = true
 petgraph.workspace = true
 serde_json.workspace = true
 smallvec.workspace = true
-hashbrown.workspace = true
 enum-as-inner.workspace = true
 enum-map.workspace = true
 insta.workspace = true

crates/schema/src/auto_migrate.rs

@@ -4,7 +4,7 @@ use crate::{def::*, error::PrettyAlgebraicType, identifier::Identifier};
 use formatter::format_plan;
 use spacetimedb_data_structures::{
     error_stream::{CollectAllErrors, CombineErrors, ErrorStream},
-    map::HashSet,
+    map::{HashCollectionExt as _, HashSet},
 };
 use spacetimedb_lib::{
     db::raw_def::v9::{RawRowLevelSecurityDefV9, TableType},

crates/schema/src/def.rs

@@ -25,10 +25,10 @@ use crate::schema::{Schema, TableSchema};
 use crate::type_for_generate::{AlgebraicTypeUse, ProductTypeDef, TypespaceForGenerate};
 use deserialize::ArgsSeed;
 use enum_map::EnumMap;
-use hashbrown::{Equivalent, HashMap};
 use indexmap::IndexMap;
 use itertools::Itertools;
 use spacetimedb_data_structures::error_stream::{CollectAllErrors, CombineErrors, ErrorStream};
+use spacetimedb_data_structures::map::{Equivalent, HashMap};
 use spacetimedb_lib::db::raw_def;
 use spacetimedb_lib::db::raw_def::v9::{
     Lifecycle, RawColumnDefaultValueV9, RawConstraintDataV9, RawConstraintDefV9, RawIdentifier, RawIndexAlgorithm,
@@ -1442,7 +1442,7 @@ impl ModuleDefLookup for ViewDef {
     }
 }
 
-fn to_raw<Def, RawDef, Name, A>(data: HashMap<Name, Def, A>) -> Vec<RawDef>
+fn to_raw<Def, RawDef, Name>(data: HashMap<Name, Def>) -> Vec<RawDef>
 where
     Def: ModuleDefLookup + Into<RawDef>,
     Name: Eq + Ord + 'static,
@@ -1457,7 +1457,7 @@ mod tests {
 
     use super::*;
     use proptest::prelude::*;
-    use spacetimedb_data_structures::expect_error_matching;
+    use spacetimedb_data_structures::{expect_error_matching, map::HashCollectionExt as _};
     use spacetimedb_lib::db::raw_def::v9::RawModuleDefV9Builder;
 
     proptest! {

crates/schema/src/identifier.rs

@@ -1,6 +1,5 @@
 use crate::error::IdentifierError;
-use hashbrown::Equivalent;
-use spacetimedb_data_structures::map::HashSet;
+use spacetimedb_data_structures::map::{Equivalent, HashSet};
 use spacetimedb_sats::{de, ser};
 use std::fmt::{self, Debug, Display};
 use std::ops::Deref;

crates/schema/src/type_for_generate.rs

@@ -12,7 +12,7 @@ use petgraph::{
 use smallvec::SmallVec;
 use spacetimedb_data_structures::{
     error_stream::{CollectAllErrors, CombineErrors, ErrorStream},
-    map::{HashMap, HashSet},
+    map::{hash_set, HashMap, HashSet},
 };
 use spacetimedb_lib::{AlgebraicType, ProductTypeElement};
 use spacetimedb_sats::{
@@ -632,7 +632,7 @@ impl NodeIndexable for TypespaceForGenerateBuilder<'_> {
     }
 }
 impl<'a> IntoNodeIdentifiers for &'a TypespaceForGenerateBuilder<'a> {
-    type NodeIdentifiers = std::iter::Cloned<hashbrown::hash_set::Iter<'a, spacetimedb_sats::AlgebraicTypeRef>>;
+    type NodeIdentifiers = std::iter::Cloned<hash_set::Iter<'a, spacetimedb_sats::AlgebraicTypeRef>>;
 
     fn node_identifiers(self) -> Self::NodeIdentifiers {
         self.is_def.iter().cloned()

crates/table/Cargo.toml

@@ -31,7 +31,7 @@ blake3_pure = ["blake3/pure"]
 
 [dependencies]
 spacetimedb-data-structures = { workspace = true, features = ["memory-usage"] }
-spacetimedb-memory-usage = { workspace = true, features = ["hashbrown", "ethnum", "smallvec"] }
+spacetimedb-memory-usage = { workspace = true, features = ["hash_map", "ethnum", "smallvec"] }
 spacetimedb-primitives.workspace = true
 spacetimedb-sats = { workspace = true, features = ["blake3"] }
 spacetimedb-lib = { workspace = true, features = ["memory-usage"] }

crates/table/src/blob_store.rs

@@ -12,7 +12,7 @@
 //!   It is not optimize and is mainly intended for testing purposes.
 
 use blake3::hash;
-use spacetimedb_data_structures::map::{Entry, HashMap};
+use spacetimedb_data_structures::map::{hash_map::Entry, HashMap};
 use spacetimedb_lib::{de::Deserialize, ser::Serialize};
 use spacetimedb_memory_usage::MemoryUsage;
 

crates/table/src/pointer_map.rs

@@ -17,7 +17,7 @@ use super::indexes::{PageIndex, PageOffset, RowHash, RowPointer, SquashedOffset}
 use crate::static_assert_size;
 use core::{hint, slice};
 use spacetimedb_data_structures::map::{
-    Entry,
+    hash_map::Entry,
     IntMap, // No need to hash a hash.
 };
 use spacetimedb_sats::memory_usage::MemoryUsage;

sdks/rust/src/client_cache.rs

@@ -10,7 +10,7 @@ use bytes::Bytes;
 use core::any::type_name;
 use core::hash::Hash;
 use futures_channel::mpsc;
-use spacetimedb_data_structures::map::{DefaultHashBuilder, Entry, HashCollectionExt, HashMap};
+use spacetimedb_data_structures::map::{hash_map::Entry, HashCollectionExt, HashMap};
 use std::marker::PhantomData;
 use std::sync::Arc;
 
@@ -120,8 +120,7 @@ impl<'r, Row> TableAppliedDiff<'r, Row> {
         }
 
         // Compute the PK -> Row map for deletes.
-        let mut delete_pks =
-            <HashMap<_, _, DefaultHashBuilder> as HashCollectionExt>::with_capacity(self.deletes.len());
+        let mut delete_pks = HashMap::with_capacity(self.deletes.len());
         for (&bsatn, &row) in self.deletes.iter() {
             let pk = derive_pk(row);
             delete_pks.insert(pk, (bsatn, row));