Use a nonrandom hasher when decoding NDJSON

This commit is contained in:
Kerollmops 2024-12-11 12:02:38 +01:00
parent 69c931334f
commit 93fbdc06d3
No known key found for this signature in database
GPG Key ID: F250A4C4E3AE5F5F
3 changed files with 8 additions and 5 deletions

9
Cargo.lock generated
View File

@ -3577,6 +3577,7 @@ dependencies = [
"memmap2", "memmap2",
"milli", "milli",
"roaring", "roaring",
"rustc-hash 2.1.0",
"serde", "serde",
"serde-cs", "serde-cs",
"serde_json", "serde_json",
@ -3676,7 +3677,7 @@ dependencies = [
"rhai", "rhai",
"roaring", "roaring",
"rstar", "rstar",
"rustc-hash 2.0.0", "rustc-hash 2.1.0",
"serde", "serde",
"serde_json", "serde_json",
"slice-group-by", "slice-group-by",
@ -4425,7 +4426,7 @@ dependencies = [
"bytes", "bytes",
"rand", "rand",
"ring", "ring",
"rustc-hash 2.0.0", "rustc-hash 2.1.0",
"rustls", "rustls",
"slab", "slab",
"thiserror", "thiserror",
@ -4798,9 +4799,9 @@ checksum = "08d43f7aa6b08d49f382cde6a7982047c3426db949b1424bc4b7ec9ae12c6ce2"
[[package]] [[package]]
name = "rustc-hash" name = "rustc-hash"
version = "2.0.0" version = "2.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index" source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "583034fd73374156e66797ed8e5b0d5690409c9226b22d87cb7f19821c05d152" checksum = "c7fb8039b3032c191086b10f11f319a6e99e1e82889c5cc6046f515c9db1d497"
[[package]] [[package]]
name = "rustc_version" name = "rustc_version"

View File

@ -26,6 +26,7 @@ memmap2 = "0.9.4"
milli = { path = "../milli" } milli = { path = "../milli" }
bumparaw-collections = "0.1.2" bumparaw-collections = "0.1.2"
roaring = { version = "0.10.7", features = ["serde"] } roaring = { version = "0.10.7", features = ["serde"] }
rustc-hash = "2.1.0"
serde = { version = "1.0.204", features = ["derive"] } serde = { version = "1.0.204", features = ["derive"] }
serde-cs = "0.2.4" serde-cs = "0.2.4"
serde_json = "1.0.120" serde_json = "1.0.120"

View File

@ -8,6 +8,7 @@ use bumparaw_collections::RawMap;
use memmap2::Mmap; use memmap2::Mmap;
use milli::documents::Error; use milli::documents::Error;
use milli::Object; use milli::Object;
use rustc_hash::FxBuildHasher;
use serde::de::{SeqAccess, Visitor}; use serde::de::{SeqAccess, Visitor};
use serde::{Deserialize, Deserializer}; use serde::{Deserialize, Deserializer};
use serde_json::error::Category; use serde_json::error::Category;
@ -263,7 +264,7 @@ pub fn read_ndjson(input: &File) -> Result<u64> {
match result { match result {
Ok(raw) => { Ok(raw) => {
// try to deserialize as a map // try to deserialize as a map
RawMap::from_raw_value(raw, &bump) RawMap::from_raw_value_and_hasher(raw, FxBuildHasher, &bump)
.map_err(|e| DocumentFormatError::from((PayloadType::Ndjson, e)))?; .map_err(|e| DocumentFormatError::from((PayloadType::Ndjson, e)))?;
count += 1; count += 1;
} }