Rework custom syntax

- Implement a dedicated optimized searcher for the default syntax. - Use the `aho-corasick` crate for custom syntax. - Put custom syntax behind a feature.
rossmacarthur · Jan 17, 2025 · 5054b80 · 5054b80
1 parent 69cf095
commit 5054b80
Show file tree

Hide file tree

Showing 26 changed files with 507 additions and 785 deletions.
diff --git a/.github/workflows/build.yaml b/.github/workflows/build.yaml
@@ -55,7 +55,7 @@ jobs:
           name: cargo-hack
 
       - name: Test
-        run: cargo test --workspace
+        run: cargo test --workspace --all-features
 
       - name: Test (feature powerset)
         run: cargo hack test --package upon --feature-powerset --depth 2 --lib --tests
@@ -71,12 +71,16 @@ jobs:
       - uses: actions/checkout@v4
 
       - uses: dtolnay/[email protected]
-      - name: Check (no filters)
+      - name: Check (no filters, no syntax)
         run: cargo check --no-default-features --features serde,unicode
 
+      - uses: dtolnay/[email protected]
+      - name: Check (no filters)
+        run: cargo check --no-default-features --features serde,syntax,unicode
+
       - uses: dtolnay/[email protected]
       - name: Test
-        run: cargo test
+        run: cargo test --all-features
         env:
           RUSTFLAGS: --deny warnings
 

diff --git a/.vscode/settings.json b/.vscode/settings.json
@@ -1,4 +1,6 @@
 {
+    "rust-analyzer.check.features": "all",
+    "rust-analyzer.cargo.features": "all",
     "rust-analyzer.server.extraEnv": {
         "RUSTFLAGS": "--cfg internal_debug"
     },

diff --git a/Cargo.lock b/Cargo.lock
diff --git a/Cargo.toml b/Cargo.toml
@@ -16,6 +16,7 @@ include = ["src/**/*", "LICENSE-*", "README.md"]
 rustdoc-args = ["--cfg", "docsrs"]
 
 [dependencies]
+aho-corasick = { version = "1.1.2", optional = true }
 serde = { version = "1.0.137", optional = true }
 unicode-ident = { version = "1.0.5", optional = true }
 unicode-width = { version = "0.1.9", optional = true }
@@ -35,6 +36,10 @@ filters = []
 # the context using `Value`'s '`From` impls.
 serde = ["dep:serde"]
 
+# Enables support for configuring custom delimiters in templates and pulls in
+# the `aho-corasick` crate.
+syntax = ["dep:aho-corasick"]
+
 # Allows unicode identifiers in templates and enables improved error
 # formatting.
 unicode = ["dep:unicode-ident", "dep:unicode-width"]

diff --git a/README.md b/README.md
@@ -130,6 +130,10 @@ The following crate features are available.
   [`render_from(..)`][render_from] to render templates and
   construct the context using [`Value`][value]’s `From` impls.
 
+- **`syntax`** *(disabled by default)* — Enables support for configuring
+  custom delimiters in templates (see `Engine::with_syntax`) and pulls in
+  the `aho-corasick` crate.
+
 - **`unicode`** *(enabled by default)* — Enables unicode support and pulls
   in the [`unicode-ident`][unicode-ident] and
   [`unicode-width`][unicode-width] crates. If disabled then unicode

diff --git a/SYNTAX.md b/SYNTAX.md
@@ -23,8 +23,8 @@ the binary or provided at runtime (e.g. read from a file). A template
 contains [**expressions**](#expressions) for rendering values and
 [**blocks**](#blocks) for controlling logic. These require you to use
 specific syntax delimiters in the template. Because `upon` allows you to
-configure these delimiters, this document will only refer to the
-**default** configuration.
+configure these delimiters, this document will only refer to the **default**
+configuration.
 
 ## Expressions
 

diff --git a/benches/Cargo.toml b/benches/Cargo.toml
@@ -7,13 +7,13 @@ publish = false
 [dependencies]
 handlebars = "4.3.7"
 liquid = "0.26.4"
-minijinja = "1.0.5"
+minijinja = { version = "1.0.5", features = ["custom_syntax"] }
 rand = "0.8.5"
 serde = { version = "1.0.137", features = ["derive"] }
 serde_json = "1.0.103"
 tera = "1.19.0"
 tinytemplate = "1.2.1"
-upon = { path = ".." }
+upon = { path = "..", features = ["syntax"] }
 
 [dev-dependencies]
 criterion = { version = "0.5.1", features = ["html_reports"] }

diff --git a/benches/benchdata/syntax/minijinja.html b/benches/benchdata/syntax/minijinja.html
@@ -0,0 +1,23 @@
+<!DOCTYPE html>
+<html>
+<head>
+    <title>{ title }</title>
+</head>
+<body>
+    <table>
+        <tr>
+            <th>Name</th>
+            <th>Age</th>
+        </tr>
+      <%- for user in users -%>
+        <% if not user.is_disabled %>
+            <# This is a comment #>
+            <tr>
+                <td>{ user.name }</td>
+                <td>{ user.age }</td>
+            </tr>
+        <% endif %>
+      <%- endfor -%>
+    </table>
+</body>
+</html>
diff --git a/benches/benchdata/syntax/upon.html b/benches/benchdata/syntax/upon.html
@@ -0,0 +1,23 @@
+<!DOCTYPE html>
+<html>
+<head>
+    <title>{ title }</title>
+</head>
+<body>
+    <table>
+        <tr>
+            <th>Name</th>
+            <th>Age</th>
+        </tr>
+      <%- for user in users -%>
+        <% if not user.is_disabled %>
+            <# This is a comment #>
+            <tr>
+                <td>{ user.name }</td>
+                <td>{ user.age }</td>
+            </tr>
+        <% endif %>
+      <%- endfor -%>
+    </table>
+</body>
+</html>
diff --git a/benches/benches/engines.rs b/benches/benches/engines.rs
@@ -11,6 +11,7 @@ criterion_group! {
     benches,
     bench_init,
     bench_compile,
+    bench_syntax,
     bench_render,
     bench_filters,
 }
@@ -57,6 +58,25 @@ pub fn bench_compile(c: &mut Criterion) {
     bench!(Upon, "../benchdata/basic/upon.html");
 }
 
+/// Benchmarks the time taken to compile a template with custom syntax.
+pub fn bench_syntax(c: &mut Criterion) {
+    let mut g = c.benchmark_group("syntax");
+
+    macro_rules! bench {
+        ($E:ty, $source:literal) => {{
+            g.bench_function(<$E as Engine>::name(), |b| {
+                let source = repeat(include_str!($source), 50);
+                let mut engine =
+                    <$E as Engine>::with_syntax(("{", "}"), ("<%", "%>"), ("<#", "#>"));
+                b.iter(|| engine.add_template("bench", &source));
+            });
+        }};
+    }
+
+    bench!(Minijinja, "../benchdata/syntax/minijinja.html");
+    bench!(Upon, "../benchdata/syntax/upon.html");
+}
+
 /// Benchmarks the time taken to render a template as a string.
 pub fn bench_render(c: &mut Criterion) {
     let mut g = c.benchmark_group("render");

diff --git a/benches/src/lib.rs b/benches/src/lib.rs
@@ -5,10 +5,19 @@ mod tests;
 use std::collections::HashMap;
 
 /// Abstraction for a template engine.
-pub trait Engine<'a> {
+pub trait Engine<'a>: Sized {
     fn name() -> &'static str;
     fn new() -> Self;
-    fn add_filters(&mut self);
+    fn with_syntax(
+        _expr: (&'static str, &'static str),
+        _block: (&'static str, &'static str),
+        _comment: (&'static str, &'static str),
+    ) -> Self {
+        unimplemented!()
+    }
+    fn add_filters(&mut self) {
+        unimplemented!()
+    }
     fn add_template(&mut self, name: &'static str, source: &'a str);
     fn render<S>(&self, name: &'static str, ctx: &S) -> String
     where
@@ -83,9 +92,6 @@ impl<'engine> Engine<'engine> for Liquid {
         }
     }
 
-    #[inline]
-    fn add_filters(&mut self) {}
-
     #[inline]
     fn add_template(&mut self, name: &'static str, source: &'engine str) {
         let template = self.parser.parse(source).unwrap();
@@ -123,6 +129,24 @@ impl<'engine> Engine<'engine> for Minijinja<'engine> {
     }
 
     #[inline]
+    fn with_syntax(
+        (variable_start, variable_end): (&'static str, &'static str),
+        (block_start, block_end): (&'static str, &'static str),
+        (comment_start, comment_end): (&'static str, &'static str),
+    ) -> Self {
+        let mut env = minijinja::Environment::new();
+        env.set_syntax(minijinja::Syntax {
+            block_start: block_start.into(),
+            block_end: block_end.into(),
+            variable_start: variable_start.into(),
+            variable_end: variable_end.into(),
+            comment_start: comment_start.into(),
+            comment_end: comment_end.into(),
+        })
+        .unwrap();
+        env
+    }
+
     fn add_filters(&mut self) {}
 
     #[inline]
@@ -156,7 +180,6 @@ impl<'engine> Engine<'engine> for Tera {
         tera::Tera::default()
     }
 
-    #[inline]
     fn add_filters(&mut self) {}
 
     #[inline]
@@ -195,9 +218,6 @@ impl<'engine> Engine<'engine> for TinyTemplate<'engine> {
         tt
     }
 
-    #[inline]
-    fn add_filters(&mut self) {}
-
     #[inline]
     fn add_template(&mut self, name: &'static str, source: &'engine str) {
         self.add_template(name, source).unwrap();
@@ -229,6 +249,21 @@ impl<'engine> Engine<'engine> for upon::Engine<'engine> {
         upon::Engine::new()
     }
 
+    #[inline]
+    fn with_syntax(
+        (begin_expr, end_expr): (&'static str, &'static str),
+        (begin_block, end_block): (&'static str, &'static str),
+        (begin_comment, end_comment): (&'static str, &'static str),
+    ) -> Self {
+        upon::Engine::with_syntax(
+            upon::Syntax::builder()
+                .expr(begin_expr, end_expr)
+                .block(begin_block, end_block)
+                .comment(begin_comment, end_comment)
+                .build(),
+        )
+    }
+
     #[inline]
     fn add_filters(&mut self) {
         self.add_filter("lower", str::to_lowercase);

diff --git a/benches/src/testdata/syntax_minijinja.golden b/benches/src/testdata/syntax_minijinja.golden
@@ -0,0 +1,25 @@
+<!DOCTYPE html>
+<html>
+<head>
+    <title>My awesome webpage!</title>
+</head>
+<body>
+    <table>
+        <tr>
+            <th>Name</th>
+            <th>Age</th>
+        </tr>
+
+            <tr>
+                <td>Nancy Wheeler</td>
+                <td>17</td>
+            </tr>
+
+
+            <tr>
+                <td>Steve Harrington</td>
+                <td>18</td>
+            </tr>
+        </table>
+</body>
+</html>
diff --git a/benches/src/testdata/syntax_upon.golden b/benches/src/testdata/syntax_upon.golden
@@ -0,0 +1,25 @@
+<!DOCTYPE html>
+<html>
+<head>
+    <title>My awesome webpage!</title>
+</head>
+<body>
+    <table>
+        <tr>
+            <th>Name</th>
+            <th>Age</th>
+        </tr>
+
+            <tr>
+                <td>Nancy Wheeler</td>
+                <td>17</td>
+            </tr>
+
+
+            <tr>
+                <td>Steve Harrington</td>
+                <td>18</td>
+            </tr>
+        </table>
+</body>
+</html>