Restore __deserialize_content optimization

This commit is contained in:
David Tolnay 2025-09-13 10:24:40 -07:00
parent 93f43cedaa
commit f669f16127
No known key found for this signature in database
GPG Key ID: F9BA143B95FF6D82
6 changed files with 381 additions and 338 deletions

View File

@ -11,7 +11,7 @@ use crate::de::{MapAccess, Unexpected};
#[cfg(any(feature = "std", feature = "alloc"))]
pub use self::content::{
Content, ContentDeserializer, ContentRefDeserializer, EnumDeserializer,
content_as_str, Content, ContentDeserializer, ContentRefDeserializer, EnumDeserializer,
InternallyTaggedUnitVisitor, TagContentOtherField, TagContentOtherFieldVisitor,
TagOrContentField, TagOrContentFieldVisitor, TaggedContentVisitor, UntaggedUnitVisitor,
};
@ -213,46 +213,11 @@ mod content {
self, size_hint, Deserialize, DeserializeSeed, Deserializer, EnumAccess, Expected,
IgnoredAny, MapAccess, SeqAccess, Unexpected, Visitor,
};
pub use serde_core::__private::Content;
use serde_core::__private::ContentVisitor;
/// Used from generated code to buffer the contents of the Deserializer when
/// deserializing untagged enums and internally tagged enums.
///
/// Not public API. Use serde-value instead.
#[derive(Debug, Clone)]
pub enum Content<'de> {
Bool(bool),
U8(u8),
U16(u16),
U32(u32),
U64(u64),
I8(i8),
I16(i16),
I32(i32),
I64(i64),
F32(f32),
F64(f64),
Char(char),
String(String),
Str(&'de str),
ByteBuf(Vec<u8>),
Bytes(&'de [u8]),
None,
Some(Box<Content<'de>>),
Unit,
Newtype(Box<Content<'de>>),
Seq(Vec<Content<'de>>),
Map(Vec<(Content<'de>, Content<'de>)>),
}
impl<'de> Content<'de> {
pub fn as_str(&self) -> Option<&str> {
match *self {
pub fn content_as_str<'a, 'de>(content: &'a Content<'de>) -> Option<&'a str> {
match *content {
Content::Str(x) => Some(x),
Content::String(ref x) => Some(x),
Content::Bytes(x) => str::from_utf8(x).ok(),
@ -262,8 +227,8 @@ mod content {
}
#[cold]
fn unexpected(&self) -> Unexpected {
match *self {
fn content_unexpected<'a, 'de>(content: &'a Content<'de>) -> Unexpected<'a> {
match *content {
Content::Bool(b) => Unexpected::Bool(b),
Content::U8(n) => Unexpected::Unsigned(n as u64),
Content::U16(n) => Unexpected::Unsigned(n as u64),
@ -287,252 +252,6 @@ mod content {
Content::Map(_) => Unexpected::Map,
}
}
}
impl<'de> Deserialize<'de> for Content<'de> {
fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
where
D: Deserializer<'de>,
{
// Untagged and internally tagged enums are only supported in
// self-describing formats.
let visitor = ContentVisitor { value: PhantomData };
deserializer.deserialize_any(visitor)
}
}
impl<'de, E> de::IntoDeserializer<'de, E> for Content<'de>
where
E: de::Error,
{
type Deserializer = ContentDeserializer<'de, E>;
fn into_deserializer(self) -> Self::Deserializer {
ContentDeserializer::new(self)
}
}
impl<'a, 'de, E> de::IntoDeserializer<'de, E> for &'a Content<'de>
where
E: de::Error,
{
type Deserializer = ContentRefDeserializer<'a, 'de, E>;
fn into_deserializer(self) -> Self::Deserializer {
ContentRefDeserializer::new(self)
}
}
/// Used to capture data in [`Content`] from other deserializers.
/// Cannot capture externally tagged enums, `i128` and `u128`.
struct ContentVisitor<'de> {
value: PhantomData<Content<'de>>,
}
impl<'de> ContentVisitor<'de> {
fn new() -> Self {
ContentVisitor { value: PhantomData }
}
}
impl<'de> Visitor<'de> for ContentVisitor<'de> {
type Value = Content<'de>;
fn expecting(&self, fmt: &mut fmt::Formatter) -> fmt::Result {
fmt.write_str("any value")
}
fn visit_bool<F>(self, value: bool) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::Bool(value))
}
fn visit_i8<F>(self, value: i8) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::I8(value))
}
fn visit_i16<F>(self, value: i16) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::I16(value))
}
fn visit_i32<F>(self, value: i32) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::I32(value))
}
fn visit_i64<F>(self, value: i64) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::I64(value))
}
fn visit_u8<F>(self, value: u8) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::U8(value))
}
fn visit_u16<F>(self, value: u16) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::U16(value))
}
fn visit_u32<F>(self, value: u32) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::U32(value))
}
fn visit_u64<F>(self, value: u64) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::U64(value))
}
fn visit_f32<F>(self, value: f32) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::F32(value))
}
fn visit_f64<F>(self, value: f64) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::F64(value))
}
fn visit_char<F>(self, value: char) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::Char(value))
}
fn visit_str<F>(self, value: &str) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::String(value.into()))
}
fn visit_borrowed_str<F>(self, value: &'de str) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::Str(value))
}
fn visit_string<F>(self, value: String) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::String(value))
}
fn visit_bytes<F>(self, value: &[u8]) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::ByteBuf(value.into()))
}
fn visit_borrowed_bytes<F>(self, value: &'de [u8]) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::Bytes(value))
}
fn visit_byte_buf<F>(self, value: Vec<u8>) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::ByteBuf(value))
}
fn visit_unit<F>(self) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::Unit)
}
fn visit_none<F>(self) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::None)
}
fn visit_some<D>(self, deserializer: D) -> Result<Self::Value, D::Error>
where
D: Deserializer<'de>,
{
let v = tri!(Deserialize::deserialize(deserializer));
Ok(Content::Some(Box::new(v)))
}
fn visit_newtype_struct<D>(self, deserializer: D) -> Result<Self::Value, D::Error>
where
D: Deserializer<'de>,
{
let v = tri!(Deserialize::deserialize(deserializer));
Ok(Content::Newtype(Box::new(v)))
}
fn visit_seq<V>(self, mut visitor: V) -> Result<Self::Value, V::Error>
where
V: SeqAccess<'de>,
{
let mut vec =
Vec::<Content>::with_capacity(size_hint::cautious::<Content>(visitor.size_hint()));
while let Some(e) = tri!(visitor.next_element()) {
vec.push(e);
}
Ok(Content::Seq(vec))
}
fn visit_map<V>(self, mut visitor: V) -> Result<Self::Value, V::Error>
where
V: MapAccess<'de>,
{
let mut vec =
Vec::<(Content, Content)>::with_capacity(
size_hint::cautious::<(Content, Content)>(visitor.size_hint()),
);
while let Some(kv) = tri!(visitor.next_entry()) {
vec.push(kv);
}
Ok(Content::Map(vec))
}
fn visit_enum<V>(self, _visitor: V) -> Result<Self::Value, V::Error>
where
V: EnumAccess<'de>,
{
Err(de::Error::custom(
"untagged and internally tagged enums do not support enum input",
))
}
}
/// This is the type of the map keys in an internally tagged enum.
///
@ -1065,7 +784,7 @@ mod content {
{
#[cold]
fn invalid_type(self, exp: &dyn Expected) -> E {
de::Error::invalid_type(self.content.unexpected(), exp)
de::Error::invalid_type(content_unexpected(&self.content), exp)
}
fn deserialize_integer<V>(self, visitor: V) -> Result<V::Value, E>
@ -1110,7 +829,8 @@ mod content {
V: Visitor<'de>,
E: de::Error,
{
let mut seq_visitor = SeqDeserializer::new(content.into_iter());
let mut seq_visitor =
SeqDeserializer::new(content.into_iter().map(ContentDeserializer::new));
let value = tri!(visitor.visit_seq(&mut seq_visitor));
tri!(seq_visitor.end());
Ok(value)
@ -1124,7 +844,11 @@ mod content {
V: Visitor<'de>,
E: de::Error,
{
let mut map_visitor = MapDeserializer::new(content.into_iter());
let mut map_visitor = MapDeserializer::new(
content
.into_iter()
.map(|(k, v)| (ContentDeserializer::new(k), ContentDeserializer::new(v))),
);
let value = tri!(visitor.visit_map(&mut map_visitor));
tri!(map_visitor.end());
Ok(value)
@ -1465,7 +1189,7 @@ mod content {
s @ Content::String(_) | s @ Content::Str(_) => (s, None),
other => {
return Err(de::Error::invalid_type(
other.unexpected(),
content_unexpected(&other),
&"string or map",
));
}
@ -1496,6 +1220,14 @@ mod content {
drop(self);
visitor.visit_unit()
}
fn __deserialize_content_v1<V>(self, visitor: V) -> Result<V::Value, Self::Error>
where
V: Visitor<'de, Value = Content<'de>>,
{
let _ = visitor;
Ok(self.content)
}
}
impl<'de, E> ContentDeserializer<'de, E> {
@ -1589,11 +1321,12 @@ mod content {
V: de::Visitor<'de>,
{
match self.value {
Some(Content::Seq(v)) => {
de::Deserializer::deserialize_any(SeqDeserializer::new(v.into_iter()), visitor)
}
Some(Content::Seq(v)) => de::Deserializer::deserialize_any(
SeqDeserializer::new(v.into_iter().map(ContentDeserializer::new)),
visitor,
),
Some(other) => Err(de::Error::invalid_type(
other.unexpected(),
content_unexpected(&other),
&"tuple variant",
)),
None => Err(de::Error::invalid_type(
@ -1613,13 +1346,19 @@ mod content {
{
match self.value {
Some(Content::Map(v)) => {
de::Deserializer::deserialize_any(MapDeserializer::new(v.into_iter()), visitor)
}
Some(Content::Seq(v)) => {
de::Deserializer::deserialize_any(SeqDeserializer::new(v.into_iter()), visitor)
de::Deserializer::deserialize_any(
MapDeserializer::new(v.into_iter().map(|(k, v)| {
(ContentDeserializer::new(k), ContentDeserializer::new(v))
})),
visitor,
)
}
Some(Content::Seq(v)) => de::Deserializer::deserialize_any(
SeqDeserializer::new(v.into_iter().map(ContentDeserializer::new)),
visitor,
),
Some(other) => Err(de::Error::invalid_type(
other.unexpected(),
content_unexpected(&other),
&"struct variant",
)),
None => Err(de::Error::invalid_type(
@ -1642,7 +1381,7 @@ mod content {
{
#[cold]
fn invalid_type(self, exp: &dyn Expected) -> E {
de::Error::invalid_type(self.content.unexpected(), exp)
de::Error::invalid_type(content_unexpected(self.content), exp)
}
fn deserialize_integer<V>(self, visitor: V) -> Result<V::Value, E>
@ -1690,7 +1429,7 @@ mod content {
V: Visitor<'de>,
E: de::Error,
{
let mut seq_visitor = SeqDeserializer::new(content.iter());
let mut seq_visitor = SeqDeserializer::new(content.iter().map(ContentRefDeserializer::new));
let value = tri!(visitor.visit_seq(&mut seq_visitor));
tri!(seq_visitor.end());
Ok(value)
@ -1704,10 +1443,16 @@ mod content {
V: Visitor<'de>,
E: de::Error,
{
fn content_ref_deserializer_pair<'a, 'de>(
fn content_ref_deserializer_pair<'a, 'de, E>(
(k, v): &'a (Content<'de>, Content<'de>),
) -> (&'a Content<'de>, &'a Content<'de>) {
(k, v)
) -> (
ContentRefDeserializer<'a, 'de, E>,
ContentRefDeserializer<'a, 'de, E>,
) {
(
ContentRefDeserializer::new(k),
ContentRefDeserializer::new(v),
)
}
let map = content.iter().map(content_ref_deserializer_pair);
@ -2043,7 +1788,7 @@ mod content {
ref s @ Content::String(_) | ref s @ Content::Str(_) => (s, None),
ref other => {
return Err(de::Error::invalid_type(
other.unexpected(),
content_unexpected(other),
&"string or map",
));
}
@ -2077,6 +1822,14 @@ mod content {
{
visitor.visit_unit()
}
fn __deserialize_content_v1<V>(self, visitor: V) -> Result<V::Value, Self::Error>
where
V: Visitor<'de, Value = Content<'de>>,
{
let _ = visitor;
Ok(self.content.clone())
}
}
impl<'a, 'de, E> ContentRefDeserializer<'a, 'de, E> {
@ -2182,7 +1935,7 @@ mod content {
// newtype_enum::tuple2
Some(Content::Seq(v)) => visit_content_seq_ref(v, visitor),
Some(other) => Err(de::Error::invalid_type(
other.unexpected(),
content_unexpected(other),
&"tuple variant",
)),
None => Err(de::Error::invalid_type(
@ -2209,7 +1962,7 @@ mod content {
// newtype_enum::empty_struct_from_seq
Some(Content::Seq(v)) => visit_content_seq_ref(v, visitor),
Some(other) => Err(de::Error::invalid_type(
other.unexpected(),
content_unexpected(other),
&"struct variant",
)),
None => Err(de::Error::invalid_type(
@ -2716,7 +2469,7 @@ fn flat_map_take_entry<'de>(
// and if the field is one recognized by the current data structure.
let is_recognized = match entry {
None => false,
Some((k, _v)) => k.as_str().map_or(false, |name| recognized.contains(&name)),
Some((k, _v)) => content_as_str(k).map_or(false, |name| recognized.contains(&name)),
};
if is_recognized {

View File

@ -19,6 +19,7 @@ fn main() {
println!("cargo:rustc-check-cfg=cfg(no_core_net)");
println!("cargo:rustc-check-cfg=cfg(no_core_num_saturating)");
println!("cargo:rustc-check-cfg=cfg(no_diagnostic_namespace)");
println!("cargo:rustc-check-cfg=cfg(no_serde_derive)");
println!("cargo:rustc-check-cfg=cfg(no_std_atomic)");
println!("cargo:rustc-check-cfg=cfg(no_std_atomic64)");
println!("cargo:rustc-check-cfg=cfg(no_target_has_atomic)");
@ -48,6 +49,11 @@ fn main() {
}
}
// Current minimum supported version of serde_derive crate is Rust 1.61.
if minor < 61 {
println!("cargo:rustc-cfg=no_serde_derive");
}
// Support for core::ffi::CStr and alloc::ffi::CString stabilized in Rust 1.64.
// https://blog.rust-lang.org/2022/09/22/Rust-1.64.0.html#c-compatible-ffi-types-in-core-and-alloc
if minor < 64 {

View File

@ -1255,6 +1255,16 @@ pub trait Deserializer<'de>: Sized {
fn is_human_readable(&self) -> bool {
true
}
// Not public API.
#[cfg(all(not(no_serde_derive), any(feature = "std", feature = "alloc")))]
#[doc(hidden)]
fn __deserialize_content_v1<V>(self, visitor: V) -> Result<V::Value, Self::Error>
where
V: Visitor<'de, Value = crate::private::Content<'de>>,
{
self.deserialize_any(visitor)
}
}
////////////////////////////////////////////////////////////////////////////////

View File

@ -261,6 +261,15 @@ pub use crate::ser::{Serialize, Serializer};
#[doc(hidden)]
pub use lib::from_utf8_lossy;
// Used by generated code. Not public API.
#[cfg(all(not(no_serde_derive), any(feature = "std", feature = "alloc")))]
#[doc(hidden)]
#[path = "private.rs"]
pub mod __private;
#[cfg(all(not(no_serde_derive), any(feature = "std", feature = "alloc")))]
use self::__private as private;
#[path = "de/seed.rs"]
mod seed;

265
serde_core/src/private.rs Normal file
View File

@ -0,0 +1,265 @@
use crate::de::{
self, size_hint, Deserialize, Deserializer, EnumAccess, MapAccess, SeqAccess, Visitor,
};
use crate::lib::*;
// Used from generated code to buffer the contents of the Deserializer when
// deserializing untagged enums and internally tagged enums.
//
// Not public API. Use serde-value instead.
//
// Obsoleted by format-specific buffer types (https://github.com/serde-rs/serde/pull/2912).
#[derive(Clone)]
#[doc(hidden)]
pub enum Content<'de> {
Bool(bool),
U8(u8),
U16(u16),
U32(u32),
U64(u64),
I8(i8),
I16(i16),
I32(i32),
I64(i64),
F32(f32),
F64(f64),
Char(char),
String(String),
Str(&'de str),
ByteBuf(Vec<u8>),
Bytes(&'de [u8]),
None,
Some(Box<Content<'de>>),
Unit,
Newtype(Box<Content<'de>>),
Seq(Vec<Content<'de>>),
Map(Vec<(Content<'de>, Content<'de>)>),
}
impl<'de> Deserialize<'de> for Content<'de> {
fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
where
D: Deserializer<'de>,
{
// Untagged and internally tagged enums are only supported in
// self-describing formats.
let visitor = ContentVisitor { value: PhantomData };
deserializer.__deserialize_content_v1(visitor)
}
}
#[doc(hidden)]
pub struct ContentVisitor<'de> {
value: PhantomData<Content<'de>>,
}
impl<'de> ContentVisitor<'de> {
pub fn new() -> Self {
ContentVisitor { value: PhantomData }
}
}
impl<'de> Visitor<'de> for ContentVisitor<'de> {
type Value = Content<'de>;
fn expecting(&self, fmt: &mut fmt::Formatter) -> fmt::Result {
fmt.write_str("any value")
}
fn visit_bool<F>(self, value: bool) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::Bool(value))
}
fn visit_i8<F>(self, value: i8) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::I8(value))
}
fn visit_i16<F>(self, value: i16) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::I16(value))
}
fn visit_i32<F>(self, value: i32) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::I32(value))
}
fn visit_i64<F>(self, value: i64) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::I64(value))
}
fn visit_u8<F>(self, value: u8) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::U8(value))
}
fn visit_u16<F>(self, value: u16) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::U16(value))
}
fn visit_u32<F>(self, value: u32) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::U32(value))
}
fn visit_u64<F>(self, value: u64) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::U64(value))
}
fn visit_f32<F>(self, value: f32) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::F32(value))
}
fn visit_f64<F>(self, value: f64) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::F64(value))
}
fn visit_char<F>(self, value: char) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::Char(value))
}
fn visit_str<F>(self, value: &str) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::String(value.into()))
}
fn visit_borrowed_str<F>(self, value: &'de str) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::Str(value))
}
fn visit_string<F>(self, value: String) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::String(value))
}
fn visit_bytes<F>(self, value: &[u8]) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::ByteBuf(value.into()))
}
fn visit_borrowed_bytes<F>(self, value: &'de [u8]) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::Bytes(value))
}
fn visit_byte_buf<F>(self, value: Vec<u8>) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::ByteBuf(value))
}
fn visit_unit<F>(self) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::Unit)
}
fn visit_none<F>(self) -> Result<Self::Value, F>
where
F: de::Error,
{
Ok(Content::None)
}
fn visit_some<D>(self, deserializer: D) -> Result<Self::Value, D::Error>
where
D: Deserializer<'de>,
{
let v = tri!(Deserialize::deserialize(deserializer));
Ok(Content::Some(Box::new(v)))
}
fn visit_newtype_struct<D>(self, deserializer: D) -> Result<Self::Value, D::Error>
where
D: Deserializer<'de>,
{
let v = tri!(Deserialize::deserialize(deserializer));
Ok(Content::Newtype(Box::new(v)))
}
fn visit_seq<V>(self, mut visitor: V) -> Result<Self::Value, V::Error>
where
V: SeqAccess<'de>,
{
let mut vec =
Vec::<Content>::with_capacity(size_hint::cautious::<Content>(visitor.size_hint()));
while let Some(e) = tri!(visitor.next_element()) {
vec.push(e);
}
Ok(Content::Seq(vec))
}
fn visit_map<V>(self, mut visitor: V) -> Result<Self::Value, V::Error>
where
V: MapAccess<'de>,
{
let mut vec = Vec::<(Content, Content)>::with_capacity(size_hint::cautious::<(
Content,
Content,
)>(visitor.size_hint()));
while let Some(kv) = tri!(visitor.next_entry()) {
vec.push(kv);
}
Ok(Content::Map(vec))
}
fn visit_enum<V>(self, _visitor: V) -> Result<Self::Value, V::Error>
where
V: EnumAccess<'de>,
{
Err(de::Error::custom(
"untagged and internally tagged enums do not support enum input",
))
}
}

View File

@ -2662,7 +2662,7 @@ fn deserialize_map(
if let _serde::__private::Some(_serde::__private::Some((__key, _))) =
__collect.into_iter().filter(_serde::__private::Option::is_some).next()
{
if let _serde::__private::Some(__key) = __key.as_str() {
if let _serde::__private::Some(__key) = _serde::__private::de::content_as_str(&__key) {
return _serde::__private::Err(
_serde::de::Error::custom(format_args!("unknown field `{}`", &__key)));
} else {