Merge pull request #759 from Mingun/const

Make `const` as many functions as possible

Author: Mingun

Changelog.md

@@ -22,10 +22,37 @@
 ### Misc Changes
 
 - [#650]: Change the type of `Event::PI` to a new dedicated `BytesPI` type.
+- [#759]: Make `const` as much functions as possible:
+  - `resolve_html5_entity()`
+  - `resolve_predefined_entity()`
+  - `resolve_xml_entity()`
+  - `Attr::key()`
+  - `Attr::value()`
+  - `Attributes::html()`
+  - `Attributes::new()`
+  - `BytesDecl::from_start()`
+  - `Decoder::encoding()`
+  - `Deserializer::get_ref()`
+  - `IoReader::get_ref()`
+  - `LocalName::into_inner()`
+  - `Namespace::into_inner()`
+  - `NsReader::config()`
+  - `NsReader::prefixes()`
+  - `Prefix::into_inner()`
+  - `QName::into_inner()`
+  - `Reader::buffer_position()`
+  - `Reader::config()`
+  - `Reader::decoder()`
+  - `Reader::error_position()`
+  - `Reader::get_ref()`
+  - `SliceReader::get_ref()`
+  - `Writer::get_ref()`
+  - `Writer::new()`
 
 [#650]: https://github.com/tafia/quick-xml/issues/650
 [#755]: https://github.com/tafia/quick-xml/pull/755
 [#758]: https://github.com/tafia/quick-xml/pull/758
+[#759]: https://github.com/tafia/quick-xml/pull/759
 
 
 ## 0.32.0 -- 2024-06-10

src/de/mod.rs

@@ -2155,7 +2155,7 @@ impl<'i, R: XmlRead<'i>, E: EntityResolver> XmlReader<'i, R, E> {
     }
 
     /// Returns `true` if all events was consumed
-    fn is_empty(&self) -> bool {
+    const fn is_empty(&self) -> bool {
         matches!(self.lookahead, Ok(PayloadEvent::Eof))
     }
 
@@ -2166,7 +2166,7 @@ impl<'i, R: XmlRead<'i>, E: EntityResolver> XmlReader<'i, R, E> {
     }
 
     #[inline(always)]
-    fn need_trim_end(&self) -> bool {
+    const fn need_trim_end(&self) -> bool {
         // If next event is a text or CDATA, we should not trim trailing spaces
         !matches!(
             self.lookahead,
@@ -2464,7 +2464,7 @@ where
     /// assert_eq!(reader.error_position(), 28);
     /// assert_eq!(reader.buffer_position(), 41);
     /// ```
-    pub fn get_ref(&self) -> &R {
+    pub const fn get_ref(&self) -> &R {
         &self.reader.reader
     }
 
@@ -3127,7 +3127,7 @@ impl<R: BufRead> IoReader<R> {
     /// assert_eq!(reader.error_position(), 28);
     /// assert_eq!(reader.buffer_position(), 41);
     /// ```
-    pub fn get_ref(&self) -> &Reader<R> {
+    pub const fn get_ref(&self) -> &Reader<R> {
         &self.reader
     }
 }
@@ -3194,7 +3194,7 @@ impl<'de> SliceReader<'de> {
     /// assert_eq!(reader.error_position(), 28);
     /// assert_eq!(reader.buffer_position(), 41);
     /// ```
-    pub fn get_ref(&self) -> &Reader<&'de [u8]> {
+    pub const fn get_ref(&self) -> &Reader<&'de [u8]> {
         &self.reader
     }
 }

src/de/simple_type.rs

@@ -551,7 +551,7 @@ impl<'de, 'a> SimpleTypeDeserializer<'de, 'a> {
 
     /// Constructor for tests
     #[inline]
-    fn new(content: CowRef<'de, 'a, [u8]>, escaped: bool, decoder: Decoder) -> Self {
+    const fn new(content: CowRef<'de, 'a, [u8]>, escaped: bool, decoder: Decoder) -> Self {
         Self {
             content,
             escaped,

src/encoding.rs

@@ -62,7 +62,7 @@ impl Decoder {
     ///
     /// [`decode`]: Self::decode
     #[cfg(feature = "encoding")]
-    pub fn encoding(&self) -> &'static Encoding {
+    pub const fn encoding(&self) -> &'static Encoding {
         self.encoding
     }
 

src/escape.rs

@@ -284,7 +284,7 @@ where
 /// Behaves like [`resolve_xml_entity`] if feature is not enabled and as
 /// [`resolve_html5_entity`] if enabled.
 #[inline]
-pub fn resolve_predefined_entity(entity: &str) -> Option<&'static str> {
+pub const fn resolve_predefined_entity(entity: &str) -> Option<&'static str> {
     #[cfg(not(feature = "escape-html"))]
     {
         resolve_xml_entity(entity)
@@ -314,7 +314,7 @@ pub fn resolve_predefined_entity(entity: &str) -> Option<&'static str> {
 /// ```
 ///
 /// [specification]: https://www.w3.org/TR/xml11/#sec-predefined-ent
-pub fn resolve_xml_entity(entity: &str) -> Option<&'static str> {
+pub const fn resolve_xml_entity(entity: &str) -> Option<&'static str> {
     // match over strings are not allowed in const functions
     let s = match entity.as_bytes() {
         b"lt" => "<",
@@ -328,7 +328,7 @@ pub fn resolve_xml_entity(entity: &str) -> Option<&'static str> {
 }
 
 /// Resolves all HTML5 entities. For complete list see <https://dev.w3.org/html5/html-author/charref>.
-pub fn resolve_html5_entity(entity: &str) -> Option<&'static str> {
+pub const fn resolve_html5_entity(entity: &str) -> Option<&'static str> {
     // imported from https://dev.w3.org/html5/html-author/charref
     // match over strings are not allowed in const functions
     //TODO: automate up-to-dating using https://html.spec.whatwg.org/entities.json

src/events/attributes.rs

@@ -5,8 +5,8 @@
 use crate::errors::Result as XmlResult;
 use crate::escape::{escape, resolve_predefined_entity, unescape_with};
 use crate::name::QName;
-use crate::reader::{is_whitespace, Reader};
-use crate::utils::{write_byte_string, write_cow_string, Bytes};
+use crate::reader::Reader;
+use crate::utils::{is_whitespace, write_byte_string, write_cow_string, Bytes};
 use std::fmt::{self, Debug, Display, Formatter};
 use std::iter::FusedIterator;
 use std::{borrow::Cow, ops::Range};
@@ -195,20 +195,20 @@ pub struct Attributes<'a> {
 impl<'a> Attributes<'a> {
     /// Internal constructor, used by `BytesStart`. Supplies data in reader's encoding
     #[inline]
-    pub(crate) fn wrap(buf: &'a [u8], pos: usize, html: bool) -> Self {
+    pub(crate) const fn wrap(buf: &'a [u8], pos: usize, html: bool) -> Self {
         Self {
             bytes: buf,
             state: IterState::new(pos, html),
         }
     }
 
     /// Creates a new attribute iterator from a buffer.
-    pub fn new(buf: &'a str, pos: usize) -> Self {
+    pub const fn new(buf: &'a str, pos: usize) -> Self {
         Self::wrap(buf.as_bytes(), pos, false)
     }
 
     /// Creates a new attribute iterator from a buffer, allowing HTML attribute syntax.
-    pub fn html(buf: &'a str, pos: usize) -> Self {
+    pub const fn html(buf: &'a str, pos: usize) -> Self {
         Self::wrap(buf.as_bytes(), pos, true)
     }
 
@@ -416,7 +416,7 @@ impl<T> Attr<T> {
 impl<'a> Attr<&'a [u8]> {
     /// Returns the key value
     #[inline]
-    pub fn key(&self) -> QName<'a> {
+    pub const fn key(&self) -> QName<'a> {
         QName(match self {
             Attr::DoubleQ(key, _) => key,
             Attr::SingleQ(key, _) => key,
@@ -429,7 +429,7 @@ impl<'a> Attr<&'a [u8]> {
     ///
     /// [HTML specification]: https://www.w3.org/TR/2012/WD-html-markup-20120329/syntax.html#syntax-attr-empty
     #[inline]
-    pub fn value(&self) -> &'a [u8] {
+    pub const fn value(&self) -> &'a [u8] {
         match self {
             Attr::DoubleQ(_, value) => value,
             Attr::SingleQ(_, value) => value,
@@ -518,7 +518,7 @@ pub(crate) struct IterState {
 }
 
 impl IterState {
-    pub fn new(offset: usize, html: bool) -> Self {
+    pub const fn new(offset: usize, html: bool) -> Self {
         Self {
             state: State::Next(offset),
             html,

src/events/mod.rs

@@ -41,6 +41,7 @@ pub mod attributes;
 use encoding_rs::Encoding;
 use std::borrow::Cow;
 use std::fmt::{self, Debug, Formatter};
+use std::mem::replace;
 use std::ops::Deref;
 use std::str::from_utf8;
 
@@ -50,12 +51,10 @@ use crate::escape::{
     escape, minimal_escape, partial_escape, resolve_predefined_entity, unescape_with,
 };
 use crate::name::{LocalName, QName};
-use crate::reader::{is_whitespace, name_len};
-use crate::utils::write_cow_string;
 #[cfg(feature = "serialize")]
 use crate::utils::CowRef;
+use crate::utils::{name_len, trim_xml_end, trim_xml_start, write_cow_string};
 use attributes::{Attribute, Attributes};
-use std::mem::replace;
 
 /// Opening tag data (`Event::Start`), with optional attributes.
 ///
@@ -78,7 +77,7 @@ pub struct BytesStart<'a> {
 impl<'a> BytesStart<'a> {
     /// Internal constructor, used by `Reader`. Supplies data in reader's encoding
     #[inline]
-    pub(crate) fn wrap(content: &'a [u8], name_len: usize) -> Self {
+    pub(crate) const fn wrap(content: &'a [u8], name_len: usize) -> Self {
         BytesStart {
             buf: Cow::Borrowed(content),
             name_len,
@@ -406,7 +405,7 @@ impl<'a> BytesDecl<'a> {
     }
 
     /// Creates a `BytesDecl` from a `BytesStart`
-    pub fn from_start(start: BytesStart<'a>) -> Self {
+    pub const fn from_start(start: BytesStart<'a>) -> Self {
         Self { content: start }
     }
 
@@ -621,7 +620,7 @@ pub struct BytesEnd<'a> {
 impl<'a> BytesEnd<'a> {
     /// Internal constructor, used by `Reader`. Supplies data in reader's encoding
     #[inline]
-    pub(crate) fn wrap(name: Cow<'a, [u8]>) -> Self {
+    pub(crate) const fn wrap(name: Cow<'a, [u8]>) -> Self {
         BytesEnd { name }
     }
 
@@ -1019,7 +1018,7 @@ pub struct BytesPI<'a> {
 impl<'a> BytesPI<'a> {
     /// Creates a new `BytesPI` from a byte sequence in the specified encoding.
     #[inline]
-    pub(crate) fn wrap(content: &'a [u8], target_len: usize) -> Self {
+    pub(crate) const fn wrap(content: &'a [u8], target_len: usize) -> Self {
         Self {
             content: BytesStart::wrap(content, target_len),
         }
@@ -1257,38 +1256,6 @@ fn str_cow_to_bytes<'a, C: Into<Cow<'a, str>>>(content: C) -> Cow<'a, [u8]> {
     }
 }
 
-/// Returns a byte slice with leading XML whitespace bytes removed.
-///
-/// 'Whitespace' refers to the definition used by [`is_whitespace`].
-const fn trim_xml_start(mut bytes: &[u8]) -> &[u8] {
-    // Note: A pattern matching based approach (instead of indexing) allows
-    // making the function const.
-    while let [first, rest @ ..] = bytes {
-        if is_whitespace(*first) {
-            bytes = rest;
-        } else {
-            break;
-        }
-    }
-    bytes
-}
-
-/// Returns a byte slice with trailing XML whitespace bytes removed.
-///
-/// 'Whitespace' refers to the definition used by [`is_whitespace`].
-const fn trim_xml_end(mut bytes: &[u8]) -> &[u8] {
-    // Note: A pattern matching based approach (instead of indexing) allows
-    // making the function const.
-    while let [rest @ .., last] = bytes {
-        if is_whitespace(*last) {
-            bytes = rest;
-        } else {
-            break;
-        }
-    }
-    bytes
-}
-
 fn trim_cow<'a, F>(value: Cow<'a, [u8]>, trim: F) -> Cow<'a, [u8]>
 where
     F: FnOnce(&[u8]) -> &[u8],

src/name.rs

@@ -20,7 +20,7 @@ pub struct QName<'a>(pub &'a [u8]);
 impl<'a> QName<'a> {
     /// Converts this name to an internal slice representation.
     #[inline(always)]
-    pub fn into_inner(self) -> &'a [u8] {
+    pub const fn into_inner(self) -> &'a [u8] {
         self.0
     }
 
@@ -137,7 +137,7 @@ pub struct LocalName<'a>(&'a [u8]);
 impl<'a> LocalName<'a> {
     /// Converts this name to an internal slice representation.
     #[inline(always)]
-    pub fn into_inner(self) -> &'a [u8] {
+    pub const fn into_inner(self) -> &'a [u8] {
         self.0
     }
 }
@@ -187,7 +187,7 @@ pub struct Prefix<'a>(&'a [u8]);
 impl<'a> Prefix<'a> {
     /// Extracts internal slice
     #[inline(always)]
-    pub fn into_inner(self) -> &'a [u8] {
+    pub const fn into_inner(self) -> &'a [u8] {
         self.0
     }
 }
@@ -252,7 +252,7 @@ impl<'a> Namespace<'a> {
     /// [non-normalized]: https://www.w3.org/TR/xml11/#AVNormalize
     /// [IRI reference]: https://datatracker.ietf.org/doc/html/rfc3987
     #[inline(always)]
-    pub fn into_inner(self) -> &'a [u8] {
+    pub const fn into_inner(self) -> &'a [u8] {
         self.0
     }
     //TODO: implement value normalization and use it when comparing namespaces
@@ -618,7 +618,7 @@ impl NamespaceResolver {
     }
 
     #[inline]
-    pub fn iter(&self) -> PrefixIter {
+    pub const fn iter(&self) -> PrefixIter {
         PrefixIter {
             resolver: self,
             // We initialize the cursor to 2 to skip the two default namespaces xml: and xmlns:

src/reader/async_tokio.rs

@@ -9,9 +9,9 @@ use crate::events::Event;
 use crate::name::{QName, ResolveResult};
 use crate::reader::buffered_reader::impl_buffered_source;
 use crate::reader::{
-    is_whitespace, BangType, ElementParser, NsReader, ParseState, Parser, PiParser, ReadTextResult,
-    Reader, Span,
+    BangType, ElementParser, NsReader, ParseState, Parser, PiParser, ReadTextResult, Reader, Span,
 };
+use crate::utils::is_whitespace;
 
 /// A struct for read XML asynchronously from an [`AsyncBufRead`].
 ///

src/reader/buffered_reader.rs

@@ -8,7 +8,8 @@ use std::path::Path;
 use crate::errors::{Error, Result};
 use crate::events::Event;
 use crate::name::QName;
-use crate::reader::{is_whitespace, BangType, Parser, ReadTextResult, Reader, Span, XmlSource};
+use crate::reader::{BangType, Parser, ReadTextResult, Reader, Span, XmlSource};
+use crate::utils::is_whitespace;
 
 macro_rules! impl_buffered_source {
     ($($lf:lifetime, $reader:tt, $async:ident, $await:ident)?) => {