Provide access to a wrapper that validates that bytes being read are valid UTF-8

.github/workflows/rust.yml

@@ -14,7 +14,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
     - uses: actions/checkout@v4
-    - uses: dtolnay/rust-toolchain@1.56.0
+    - uses: dtolnay/rust-toolchain@1.79.0
     - run: cargo check
 
   minimal-versions:

Cargo.toml

@@ -10,7 +10,7 @@ repository = "https://github.com/tafia/quick-xml"
 keywords = ["xml", "serde", "parser", "writer", "html"]
 categories = ["asynchronous", "encoding", "parsing", "parser-implementations"]
 license = "MIT"
-rust-version = "1.56"
+rust-version = "1.79"
 # We exclude tests & examples & benches to reduce the size of a package.
 # Unfortunately, this is source of warnings in latest cargo when packaging:
 # > warning: ignoring {context} `{name}` as `{path}` is not included in the published package
@@ -62,6 +62,11 @@ name = "macrobenches"
 harness = false
 path = "benches/macrobenches.rs"
 
+[[bench]]
+name = "encoding"
+harness = false
+path = "benches/encoding.rs"
+
 [features]
 default = []
 
@@ -213,11 +218,6 @@ name = "async-tokio"
 required-features = ["async-tokio"]
 path = "tests/async-tokio.rs"
 
-[[test]]
-name = "encodings"
-required-features = ["encoding"]
-path = "tests/encodings.rs"
-
 [[test]]
 name = "html"
 required-features = ["escape-html"]

Changelog.md

@@ -32,6 +32,13 @@
   - `Attribute::decode_and_unescape_value_with()`
 
   Deprecated functions now behaves the same as newly added.
+-[#947]: Add new constructors to `Reader` and `NsReader` that perform automatic streaming UTF-8
+  validation on the underlying input. Validation failures are raised as errors when the `Reader`
+  is used. These APIs are currently considered "experimental".
+  - `Reader::from_reader_validating()`
+  - `Reader::from_file_validating()`
+  - `NsReader::from_reader_validating()`
+  - `NsReader::from_file_validating()`
 
 ### Bug Fixes
 
@@ -46,11 +53,13 @@
   accepts `XmlVersion` parameter to apply correct EOL normalization rules.
 - [#944]: `read_text()` now returns `BytesText` which allows you to get the content with
   properly normalized EOLs. To get the previous behavior use `.read_text().decode()?`.
+- [#947]: Bumped MSRV from 1.59 (Feb 2022) to 1.79 (June 2024)
 
 [#371]: https://github.com/tafia/quick-xml/issues/371
 [#914]: https://github.com/tafia/quick-xml/pull/914
 [#938]: https://github.com/tafia/quick-xml/pull/938
 [#944]: https://github.com/tafia/quick-xml/pull/944
+[#947]: https://github.com/tafia/quick-xml/pull/947
 
 
 ## 0.39.2 -- 2026-02-20

README.md

@@ -4,7 +4,7 @@
 [![Crate](https://img.shields.io/crates/v/quick-xml.svg)](https://crates.io/crates/quick-xml)
 [![docs.rs](https://docs.rs/quick-xml/badge.svg)](https://docs.rs/quick-xml)
 [![codecov](https://img.shields.io/codecov/c/github/tafia/quick-xml)](https://codecov.io/gh/tafia/quick-xml)
-[![MSRV](https://img.shields.io/badge/rustc-1.56.0+-ab6000.svg)](https://blog.rust-lang.org/2021/10/21/Rust-1.56.0.html)
+[![MSRV](https://img.shields.io/badge/rustc-1.79.0+-ab6000.svg)](https://blog.rust-lang.org/2024/06/13/Rust-1.79.0/)
 
 High performance xml pull reader/writer.
 

benches/encoding.rs

@@ -0,0 +1,57 @@
+use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Throughput};
+use quick_xml::encoding::Utf8ValidatingReader;
+use std::io::{BufReader, Read};
+
+static SAMPLE: &[u8] = include_bytes!("../tests/documents/sample_rss.xml");
+
+/// Read the entire input through the reader using a fixed-size buffer,
+/// returning the total number of bytes read.
+fn drain_reader(reader: &mut impl Read, buf: &mut [u8]) -> usize {
+    let mut total = 0;
+    loop {
+        match reader.read(buf) {
+            Ok(0) => break,
+            Ok(n) => total += n,
+            Err(e) => panic!("unexpected error: {e}"),
+        }
+    }
+    total
+}
+
+fn bench_utf8_validation(c: &mut Criterion) {
+    let mut group = c.benchmark_group("utf8_read");
+    group.throughput(Throughput::Bytes(SAMPLE.len() as u64));
+
+    for buf_size in [64, 1024, 8192] {
+        group.bench_with_input(
+            BenchmarkId::new("BufReader_only", buf_size),
+            &buf_size,
+            |b, &buf_size| {
+                b.iter(|| {
+                    let mut reader = BufReader::new(SAMPLE);
+                    let mut buf = vec![0u8; buf_size];
+                    let n = drain_reader(&mut reader, &mut buf);
+                    assert_eq!(n, SAMPLE.len());
+                });
+            },
+        );
+
+        group.bench_with_input(
+            BenchmarkId::new("Utf8ValidatingReader", buf_size),
+            &buf_size,
+            |b, &buf_size| {
+                b.iter(|| {
+                    let mut reader = Utf8ValidatingReader::new(BufReader::new(SAMPLE));
+                    let mut buf = vec![0u8; buf_size];
+                    let n = drain_reader(&mut reader, &mut buf);
+                    assert_eq!(n, SAMPLE.len());
+                });
+            },
+        );
+    }
+
+    group.finish();
+}
+
+criterion_group!(benches, bench_utf8_validation);
+criterion_main!(benches);

benches/macrobenches.rs

@@ -1,7 +1,3 @@
-// std::hint::black_box stable since 1.66, but our MSRV = 1.56.
-// criterion::black_box is deprecated in since criterion 0.7.
-// Running benchmarks assumed on current Rust version, so this should be fine
-#![allow(clippy::incompatible_msrv)]
 use criterion::{self, criterion_group, criterion_main, Criterion, Throughput};
 use quick_xml::events::Event;
 use quick_xml::reader::{NsReader, Reader};

benches/microbenches.rs

@@ -1,7 +1,3 @@
-// std::hint::black_box stable since 1.66, but our MSRV = 1.56.
-// criterion::black_box is deprecated in since criterion 0.7.
-// Running benchmarks assumed on current Rust version, so this should be fine
-#![allow(clippy::incompatible_msrv)]
 use criterion::{self, criterion_group, criterion_main, Criterion};
 use pretty_assertions::assert_eq;
 use quick_xml::escape::{escape, unescape};