Don’t use _all_ HTML entities in tests

Building matchers for all HTML entities can be a bit much for limited
hosts, e.g. micro VMs or limited Docker containers. This removes most of
the HTML entities; it only keeps entites that are all lowercase and
whose names start with one of `[aglqt]`.

This also fixes #14 (tests fail to build in Docker), so it removes the
existing work-around.

This resets all benchmarks.

Dockerfile

@@ -20,12 +20,6 @@ RUN curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs \
 
 ENV PATH=/root/.cargo/bin:$PATH
 
-# Prevent missing files (https://github.com/danielparks/matchgen/issues/14).
-# This won’t affect iai output since it puts stuff in matchgen_tests/target. Of
-# course, it might be affected by the bug, but at least that would provide more
-# data about the problem.
-ENV CARGO_TARGET_DIR=/root/target
-
 COPY . /work
 WORKDIR /work
 RUN cargo build --benches --profile bench --all-features

matchgen_tests/Cargo.toml

@@ -28,10 +28,10 @@ paste = "1.0.11"
 bench = false
 
 [[bench]]
-name = "all_entity_decode_iai"
+name = "most_entity_decode_iai"
 harness = false
 required-features = ["iai"]
 
 [[bench]]
-name = "all_entity_decode"
+name = "most_entity_decode"
 harness = false

matchgen_tests/benches/all_entity_decode.rs → matchgen_tests/benches/most_entity_decode.rs

@@ -11,19 +11,19 @@ macro_rules! benchmark {
         let input = $input;
         $group.throughput(Throughput::Bytes(input.len().try_into().unwrap()));
         $group.bench_with_input(
-            BenchmarkId::new("all_entity_decode", $test_name),
+            BenchmarkId::new("most_entity_decode", $test_name),
             input,
             |b, input| {
                 b.iter(|| {
                     let mut iter = input.iter();
-                    all_entity_decode(&mut iter)
+                    most_entity_decode(&mut iter)
                 })
             },
         );
         $group.bench_with_input(
-            BenchmarkId::new("all_entity_decode_slice", $test_name),
+            BenchmarkId::new("most_entity_decode_slice", $test_name),
             input,
-            |b, input| b.iter(|| all_entity_decode_slice(input)),
+            |b, input| b.iter(|| most_entity_decode_slice(input)),
         );
     }};
 }

matchgen_tests/benches/all_entity_decode_iai.rs → matchgen_tests/benches/most_entity_decode_iai.rs

@@ -9,11 +9,11 @@ macro_rules! iai_benchmarks {
                 fn [<$name _iter>]() -> Option<&'static str> {
                     let input = black_box($input);
                     let mut iter = input.iter();
-                    black_box(all_entity_decode(&mut iter))
+                    black_box(most_entity_decode(&mut iter))
                 }
 
                 fn [<$name _slice>]() -> (Option<&'static str>, &'static [u8]) {
-                    black_box(all_entity_decode_slice(
+                    black_box(most_entity_decode_slice(
                         black_box($input.as_slice())
                     ))
                 }

matchgen_tests/build.rs

@@ -83,12 +83,12 @@ fn main() -> Result<(), Box<dyn Error>> {
         .render(&mut out)?;
     writeln!(out)?;
 
-    writeln!(out, "/// Decode all HTML entities.")?;
-    let input = fs::read("html-entities.json")?;
+    writeln!(out, "/// Decode most HTML entities.")?;
+    let input = fs::read("most-html-entities.json")?;
     let input: serde_json::Map<String, serde_json::Value> =
         serde_json::from_slice(&input)?;
     let mut matcher =
-        TreeMatcher::new("pub fn all_entity_decode", "&'static str");
+        TreeMatcher::new("pub fn most_entity_decode", "&'static str");
     matcher
         .disable_clippy(true)
         .set_input_type(Input::Iterator)
@@ -101,8 +101,8 @@ fn main() -> Result<(), Box<dyn Error>> {
     matcher.render(&mut out)?;
     writeln!(out)?;
 
-    writeln!(out, "/// Decode all HTML entities.")?;
-    matcher.fn_name = "pub fn all_entity_decode_slice".to_string();
+    writeln!(out, "/// Decode most HTML entities.")?;
+    matcher.fn_name = "pub fn most_entity_decode_slice".to_string();
     matcher.set_input_type(Input::Slice);
     matcher.render(&mut out)?;
     writeln!(out)?;