Remove InfallibleTokenizer (#102)

* Remove InfallibleTokenizer

InfallibleTokenizer and Tokenizer::infallible were added as shortcuts
for when the underlying input cannot fail, to make iteration slightly
more ergonomic.

In recent Rust versions it's possible to run `for Ok(token) in ...`,
which makes this useless.

* remove impl

* fix

* fix doc

* fix benchmarks

* link pr

CHANGELOG.md

@@ -0,0 +1,14 @@
+# 0.7.0
+
+- Removal of `Tokenizer.infallible()`. Use `for Ok(token) in Tokenizer::new()` instead. [PR 102](https://github.com/untitaker/html5gum/pull/102)
+- Add more convenience functions to `tree-builder` feature, equivalent to `html5ever::driver`. [PR 101](https://github.com/untitaker/html5gum/pull/101)
+
+# 0.6.1
+
+- Fix a bug where html5gum would interpret tags inside of `<script>`. [PR 98](https://github.com/untitaker/html5gum/pull/98)
+- Restructured the crate slightly, though there _should_ not be any breaking changes. [PR 99](https://github.com/untitaker/html5gum/pull/99)
+- Added a way to integrate with `scraper` crate and the `html5ever` tree builder, see `examples/scraper.rs`.
+
+# Before 0.6.1
+
+Who knows...

Cargo.toml

@@ -19,7 +19,8 @@ test-generator = "0.3.0"
 serde_bytes = "0.11.5"
 glob = "0.3.0"
 libtest-mimic = "0.8.1"
-iai = "0.1.1"
+# https://github.com/bheisler/iai/issues/34
+iai = { git = "https://github.com/sigaloid/iai", rev = "6c83e942" }
 # required for examples/scraper.rs
 scraper = "0.20.0"
 argh = "0.1.12"

README.md

@@ -12,7 +12,7 @@ use html5gum::{Tokenizer, Token};
 let html = "<title   >hello world</title>";
 let mut new_html = String::new();
 
-for token in Tokenizer::new(html).infallible() {
+for Ok(token) in Tokenizer::new(html) {
     match token {
         Token::StartTag(tag) => {
             write!(new_html, "<{}>", String::from_utf8_lossy(&tag.name)).unwrap();

benches/patterns.rs

@@ -4,7 +4,7 @@ use html5gum::Tokenizer;
 
 fn pattern(pattern: &str, i: usize) {
     let s: String = black_box((0..i).map(|_| pattern).collect());
-    for _ in Tokenizer::new(&s).infallible() {}
+    for Ok(_) in Tokenizer::new(&s) {}
 }
 
 macro_rules! pattern_tests {

src/emitters/callback.rs

@@ -29,9 +29,8 @@
 //! });
 //!
 //! let input = r#"<h1><span class=hello>Hello</span> world!</h1>"#;
-//! let text_fragments = Tokenizer::new_with_emitter(input, emitter)
-//!     .infallible()
-//!     .collect::<Vec<_>>();
+//! let Ok(text_fragments) = Tokenizer::new_with_emitter(input, emitter)
+//!     .collect::<Result<Vec<_>, _>>();
 //!
 //! assert_eq!(text_fragments, vec![b"Hello".to_vec()]);
 //! ```

src/lib.rs

@@ -61,4 +61,4 @@ pub use error::Error;
 pub use htmlstring::HtmlString;
 pub use reader::{IoReader, Readable, Reader, StringReader};
 pub use state::State;
-pub use tokenizer::{InfallibleTokenizer, Tokenizer};
+pub use tokenizer::Tokenizer;

src/reader.rs

@@ -126,7 +126,7 @@ impl<'a, R: 'a + Reader> Readable<'a> for R {
 /// let html = "<title   >hello world</title>";
 /// let mut new_html = String::new();
 ///
-/// for token in Tokenizer::new(html).infallible() {
+/// for Ok(token) in Tokenizer::new(html) {
 ///     match token {
 ///         Token::StartTag(tag) => {
 ///             write!(new_html, "<{}>", String::from_utf8_lossy(&tag.name)).unwrap();

src/tokenizer.rs

@@ -123,30 +123,3 @@ impl<R: Reader, E: Emitter> Iterator for Tokenizer<R, E> {
         }
     }
 }
-
-/// A kind of tokenizer that directly yields tokens when used as an iterator, so `Token` instead of
-/// `Result<Token, _>`.
-///
-/// This is the return value of [`Tokenizer::infallible`].
-#[derive(Debug)]
-pub struct InfallibleTokenizer<R: Reader<Error = Infallible>, E: Emitter>(Tokenizer<R, E>);
-
-impl<R: Reader<Error = Infallible>, E: Emitter> Tokenizer<R, E> {
-    /// Statically assert that this iterator is infallible.
-    ///
-    /// Call this to get rid of error handling when parsing HTML from strings.
-    pub fn infallible(self) -> InfallibleTokenizer<R, E> {
-        InfallibleTokenizer(self)
-    }
-}
-
-impl<R: Reader<Error = Infallible>, E: Emitter> Iterator for InfallibleTokenizer<R, E> {
-    type Item = E::Token;
-
-    fn next(&mut self) -> Option<Self::Item> {
-        match self.0.next()? {
-            Ok(token) => Some(token),
-            Err(e) => match e {},
-        }
-    }
-}