Switch SipHash to XxHash

sffc · Sep 14, 2023 · b33aed9 · b33aed9
1 parent 44af79f
commit b33aed9
Show file tree

Hide file tree

Showing 8 changed files with 27 additions and 9 deletions.
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/components/timezone/data/data/macros/time_zone_bcp47_to_iana_v1.data.rs b/components/timezone/data/data/macros/time_zone_bcp47_to_iana_v1.data.rs
diff --git a/components/timezone/data/data/macros/time_zone_iana_to_bcp47_v1.data.rs b/components/timezone/data/data/macros/time_zone_iana_to_bcp47_v1.data.rs
diff --git a/provider/datagen/Cargo.toml b/provider/datagen/Cargo.toml
@@ -84,6 +84,7 @@ serde = { version = "1.0", default-features = false, features = ["derive", "allo
 serde_json = { version = "1.0", default-features = false, features = ["alloc"] }
 serde-aux = { version = "4.1.2", default-features = false }
 toml = "0.5"
+twox-hash = "1.6"
 zip = { version = ">=0.5, <0.7", default-features = false, features = ["deflate"] }
 
 rayon = { version = "1.5", optional = true }

diff --git a/provider/datagen/src/transform/cldr/time_zones/names.rs b/provider/datagen/src/transform/cldr/time_zones/names.rs
@@ -99,8 +99,7 @@ impl IterableDataProvider<Bcp47ToIanaMapV1Marker> for crate::DatagenProvider {
 }
 
 fn create_hasher() -> impl std::hash::Hasher {
-    #[allow(deprecated)] // use SipHasher to reduce dependency count; see #4024
-    std::hash::SipHasher::new()
+    twox_hash::XxHash64::with_seed(0)
 }
 
 fn compute_bcp47_ids_hash(bcp47_ids: &ZeroSlice<TimeZoneBcp47Id>) -> u64 {
@@ -121,7 +120,7 @@ fn test_compute_bcp47_ids_hash() {
 
     // Checksum 1: the default output of the zeroslice hashing function
     let checksum1 = compute_bcp47_ids_hash(&bcp47_ids);
-    assert_eq!(checksum1, 0x301FDFA03E1A34A4); // stability
+    assert_eq!(checksum1, 0x66FA043B31200DCB); // stability
 
     // Checksum 2: hashing of each individual element
     // (should equal 1)

diff --git a/provider/datagen/tests/data/json/time_zone/bcp47_to_iana@1/und.json b/provider/datagen/tests/data/json/time_zone/bcp47_to_iana@1/und.json
diff --git a/provider/datagen/tests/data/json/time_zone/iana_to_bcp47@1/und.json b/provider/datagen/tests/data/json/time_zone/iana_to_bcp47@1/und.json
diff --git a/provider/datagen/tests/data/postcard/fingerprints.csv b/provider/datagen/tests/data/postcard/fingerprints.csv
@@ -1937,7 +1937,7 @@ segmenter/line@1, und, 18811B, 12f61c35e42a3bc4
 segmenter/lstm/wl_auto@1, th, 72034B, c46e2e0c098c1fc1
 segmenter/sentence@1, und, 14402B, 379e46c5be8e8e5f
 segmenter/word@1, und, 14641B, d91c662e2d94f17f
-time_zone/bcp47_to_iana@1, und, 7568B, ca53028c19c920ac
+time_zone/bcp47_to_iana@1, und, 7569B, e3ddf20b1d136d6
 time_zone/exemplar_cities@1, ar, 10350B, 1c554603fa64a295
 time_zone/exemplar_cities@1, ar-EG, 10350B, 1c554603fa64a295
 time_zone/exemplar_cities@1, bn, 15146B, fc38df62995b3e8e
@@ -2010,7 +2010,7 @@ time_zone/generic_short@1, sr-Latn, 61B, da063c8d2cbb7223
 time_zone/generic_short@1, th, 21B, 70edef5aa0f7a054
 time_zone/generic_short@1, tr, 21B, 70edef5aa0f7a054
 time_zone/generic_short@1, und, 21B, 70edef5aa0f7a054
-time_zone/iana_to_bcp47@1, und, 9748B, 94d199894e028f36
+time_zone/iana_to_bcp47@1, und, 9749B, 2508f8150e633b96
 time_zone/metazone_period@1, und, 11270B, f204686d7a3e5ca0
 time_zone/specific_long@1, ar, 11362B, a886b0294ef9f145
 time_zone/specific_long@1, ar-EG, 11362B, a886b0294ef9f145