Merge pull request #238 from Kijewski/pr-is_ascii_whitespace

parser: use built-in whitespace trimming

.github/workflows/rust.yml

@@ -97,7 +97,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@master
         with:
-          toolchain: "1.71.1"
+          toolchain: "1.80.0"
       - run: cargo check --lib -p rinja --all-features
 
   Audit:

rinja/Cargo.toml

@@ -11,7 +11,7 @@ license = "MIT OR Apache-2.0"
 workspace = ".."
 readme = "../README.md"
 edition = "2021"
-rust-version = "1.71"
+rust-version = "1.80"
 
 [package.metadata.docs.rs]
 features = ["full"]

rinja_actix/Cargo.toml

@@ -13,7 +13,7 @@ repository = "https://github.com/rinja-rs/rinja"
 license = "MIT OR Apache-2.0"
 readme = "README.md"
 edition = "2021"
-rust-version = "1.71"
+rust-version = "1.80"
 
 [package.metadata.docs.rs]
 all-features = true

rinja_axum/Cargo.toml

@@ -5,7 +5,7 @@ members = ["."]
 name = "rinja_axum"
 version = "0.3.5"
 edition = "2021"
-rust-version = "1.71"
+rust-version = "1.80"
 description = "Axum integration for Rinja templates"
 keywords = ["markup", "template", "jinja2", "html", "axum"]
 categories = ["template-engine"]

rinja_derive/Cargo.toml

@@ -8,7 +8,7 @@ license = "MIT OR Apache-2.0"
 workspace = ".."
 readme = "README.md"
 edition = "2021"
-rust-version = "1.71"
+rust-version = "1.80"
 
 [package.metadata.docs.rs]
 all-features = true

rinja_derive_standalone/Cargo.toml

@@ -10,7 +10,7 @@ repository = "https://github.com/rinja-rs/rinja"
 license = "MIT OR Apache-2.0"
 readme = "README.md"
 edition = "2021"
-rust-version = "1.71"
+rust-version = "1.80"
 publish = false
 
 [package.metadata.docs.rs]

rinja_parser/Cargo.toml

@@ -11,7 +11,7 @@ license = "MIT OR Apache-2.0"
 workspace = ".."
 readme = "README.md"
 edition = "2021"
-rust-version = "1.71"
+rust-version = "1.80"
 
 [package.metadata.docs.rs]
 all-features = true

rinja_parser/src/expr.rs

@@ -8,11 +8,10 @@ use winnow::combinator::{
     terminated,
 };
 use winnow::error::{ErrorKind, ParserError as _};
-use winnow::token::{take_till0, take_till1};
 
 use crate::{
     CharLit, ErrorContext, Level, Num, ParseResult, PathOrIdentifier, StrLit, WithSpan, char_lit,
-    filter, identifier, keyword, not_ws, num_lit, path_or_identifier, str_lit, ws,
+    filter, identifier, keyword, num_lit, path_or_identifier, skip_ws0, skip_ws1, str_lit, ws,
 };
 
 macro_rules! expr_prec_layer {
@@ -181,7 +180,7 @@ impl<'a> Expr<'a> {
 
     fn concat(i: &'a str, level: Level) -> ParseResult<'a, WithSpan<'a, Self>> {
         fn concat_expr(i: &str, level: Level) -> ParseResult<'_, Option<WithSpan<'_, Expr<'_>>>> {
-            let ws1 = |i| opt(take_till1(not_ws)).parse_next(i);
+            let ws1 = |i| opt(skip_ws1).parse_next(i);
             let (j, data) = opt((ws1, '~', ws1, |i| Expr::muldivmod(i, level))).parse_next(i)?;
             if let Some((t1, _, t2, expr)) = data {
                 if t1.is_none() || t2.is_none() {
@@ -591,8 +590,6 @@ impl<'a> Suffix<'a> {
     }
 
     fn r#try(i: &'a str) -> ParseResult<'a, Self> {
-        preceded(take_till0(not_ws), '?')
-            .map(|_| Self::Try)
-            .parse_next(i)
+        preceded(skip_ws0, '?').map(|_| Self::Try).parse_next(i)
     }
 }

rinja_parser/src/lib.rs

@@ -15,7 +15,7 @@ use winnow::ascii::escaped;
 use winnow::combinator::{alt, cut_err, delimited, fail, not, opt, peek, preceded, repeat};
 use winnow::error::{ErrorKind, FromExternalError};
 use winnow::stream::AsChar;
-use winnow::token::{any, one_of, take_till0, take_till1, take_while};
+use winnow::token::{any, one_of, take_till1, take_while};
 
 pub mod expr;
 pub use expr::{Expr, Filter};
@@ -274,18 +274,25 @@ impl<'a> From<ErrorContext<'a>> for winnow::error::ErrMode<ErrorContext<'a>> {
     }
 }
 
-fn is_ws(c: char) -> bool {
-    matches!(c, ' ' | '\t' | '\r' | '\n')
+#[inline]
+fn skip_ws0(i: &str) -> ParseResult<'_, ()> {
+    Ok((i.trim_ascii_start(), ()))
 }
 
-fn not_ws(c: char) -> bool {
-    !is_ws(c)
+#[inline]
+fn skip_ws1(i: &str) -> ParseResult<'_, ()> {
+    let j = i.trim_ascii_start();
+    if i.len() != j.len() {
+        Ok((j, ()))
+    } else {
+        fail.parse_next(i)
+    }
 }
 
 fn ws<'a, O>(
     inner: impl Parser<&'a str, O, ErrorContext<'a>>,
 ) -> impl Parser<&'a str, O, ErrorContext<'a>> {
-    delimited(take_till0(not_ws), inner, take_till0(not_ws))
+    delimited(skip_ws0, inner, skip_ws0)
 }
 
 /// Skips input until `end` was found, but does not consume it.
@@ -892,7 +899,7 @@ fn filter<'a>(
     i: &'a str,
     level: &mut Level,
 ) -> ParseResult<'a, (&'a str, Option<Vec<WithSpan<'a, Expr<'a>>>>)> {
-    let (j, _) = take_till0(not_ws).parse_next(i)?;
+    let (j, _) = skip_ws0.parse_next(i)?;
     let had_spaces = i.len() != j.len();
     let (j, _) = ('|', not('|')).parse_next(j)?;
 

rinja_parser/src/node.rs

@@ -5,12 +5,12 @@ use winnow::Parser;
 use winnow::combinator::{
     alt, cut_err, delimited, eof, fail, not, opt, peek, preceded, repeat, separated1, terminated,
 };
-use winnow::token::{any, tag, take_till0};
+use winnow::token::{any, tag};
 
 use crate::memchr_splitter::{Splitter1, Splitter2, Splitter3};
 use crate::{
-    ErrorContext, Expr, Filter, ParseResult, State, Target, WithSpan, filter, identifier, is_ws,
-    keyword, not_ws, skip_till, str_lit_without_prefix, ws,
+    ErrorContext, Expr, Filter, ParseResult, State, Target, WithSpan, filter, identifier, keyword,
+    skip_till, skip_ws0, str_lit_without_prefix, ws,
 };
 
 #[derive(Debug, PartialEq)]
@@ -88,10 +88,7 @@ impl<'a> Node<'a> {
         let start = i;
         let (i, tag) = preceded(
             |i| s.tag_block_start(i),
-            peek(preceded(
-                (opt(Whitespace::parse), take_till0(not_ws)),
-                identifier,
-            )),
+            peek(preceded((opt(Whitespace::parse), skip_ws0), identifier)),
         )
         .parse_next(i)?;
 
@@ -1076,9 +1073,9 @@ impl<'a> Lit<'a> {
     }
 
     pub(crate) fn split_ws_parts(s: &'a str) -> Self {
-        let trimmed_start = s.trim_start_matches(is_ws);
+        let trimmed_start = s.trim_ascii_start();
         let len_start = s.len() - trimmed_start.len();
-        let trimmed = trimmed_start.trim_end_matches(is_ws);
+        let trimmed = trimmed_start.trim_ascii_end();
         Self {
             lws: &s[..len_start],
             val: trimmed,

rinja_rocket/Cargo.toml

@@ -13,7 +13,7 @@ repository = "https://github.com/rinja-rs/rinja"
 license = "MIT OR Apache-2.0"
 readme = "README.md"
 edition = "2021"
-rust-version = "1.71"
+rust-version = "1.80"
 
 [package.metadata.docs.rs]
 all-features = true

rinja_warp/Cargo.toml

@@ -13,7 +13,7 @@ repository = "https://github.com/rinja-rs/rinja"
 license = "MIT OR Apache-2.0"
 readme = "README.md"
 edition = "2021"
-rust-version = "1.71"
+rust-version = "1.80"
 
 [package.metadata.docs.rs]
 all-features = true

testing/Cargo.toml

@@ -4,7 +4,7 @@ version = "0.3.5"
 authors = ["rinja-rs developers"]
 workspace = ".."
 edition = "2021"
-rust-version = "1.71"
+rust-version = "1.80"
 publish = false
 
 [features]