Minify JS/CSS comments; faster removal of boolean attr values; sorted generated JSON objects

README.md

@@ -296,11 +296,9 @@ Any entities in attribute values are decoded, and then the shortest representati
 - Single quoted, with any `'` encoded.
 - Unquoted, with `"`/`'` first character (if applicable), `>` last character (if applicable), and any whitespace encoded.
 
-Some attributes have their whitespace (after any decoding) trimmed and collapsed:
+`class` attributes have their whitespace (after any decoding) trimmed and collapsed.
 
-- `class`
-
-[Boolean attributes](./gen/boolean_attrs.json) will have their values removed.
+[Boolean attribute](./gen/boolean_attrs.json) values are removed.
 
 `type` attributes on `script` tags with a value equaling a [JavaScript MIME type](https://mimesniff.spec.whatwg.org/#javascript-mime-type) are removed.  
 `type` attributes on `style` tags are removed.  
@@ -309,11 +307,16 @@ If an attribute value is empty after any processing, it is completely removed (i
 
 Spaces are removed between attributes if possible.
 
+# Script and style
+
+Insignificant whitespace is trimmed and collapsed inside `<script>` with JS code and `<style>`.
+
+JS and CSS comments are removed inside `<script>` and `<style>`.
+
 ### Other
 
 - Comments are removed.
 - Entities are decoded if valid (see relevant parsing section).
-- Whitespace is trimmed and collapsed inside `<script>` with JS code and `<style>`.
 
 ### Ignored
 

gen/boolean_attrs.json

@@ -1,4 +1,13 @@
 {
+  "allowfullscreen": [
+    "iframe"
+  ],
+  "allowtransparency": [
+    "iframe"
+  ],
+  "async": [
+    "script"
+  ],
   "autofocus": [
     "button",
     "input",
@@ -6,6 +15,21 @@
     "select",
     "textarea"
   ],
+  "autoplay": [
+    "media"
+  ],
+  "checked": [
+    "input"
+  ],
+  "controls": [
+    "media"
+  ],
+  "default": [
+    "track"
+  ],
+  "defer": [
+    "script"
+  ],
   "disabled": [
     "button",
     "fieldset",
@@ -16,32 +40,36 @@
     "select",
     "textarea"
   ],
+  "disablepictureinpicture": [
+    "video"
+  ],
   "formnovalidate": [
     "button",
     "input"
   ],
+  "loop": [
+    "media"
+  ],
+  "multiple": [
+    "input",
+    "select"
+  ],
+  "muted": [
+    "media"
+  ],
+  "nomodule": [
+    "script"
+  ],
+  "novalidate": [
+    "form"
+  ],
   "open": [
     "details",
     "dialog"
   ],
-  "novalidate": [
-    "form"
-  ],
-  "allowfullscreen": [
-    "iframe"
-  ],
-  "allowtransparency": [
-    "iframe"
-  ],
-  "seamless": [
-    "iframe"
-  ],
-  "checked": [
-    "input"
-  ],
-  "multiple": [
-    "input",
-    "select"
+  "playsinline": [
+    "media",
+    "video"
   ],
   "readonly": [
     "input",
@@ -52,44 +80,16 @@
     "select",
     "textarea"
   ],
-  "autoplay": [
-    "media"
-  ],
-  "controls": [
-    "media"
-  ],
-  "loop": [
-    "media"
-  ],
-  "muted": [
-    "media"
-  ],
-  "playsinline": [
-    "media",
-    "video"
-  ],
   "reversed": [
     "ol"
   ],
-  "selected": [
-    "option"
-  ],
-  "async": [
-    "script"
-  ],
-  "defer": [
-    "script"
-  ],
-  "nomodule": [
-    "script"
-  ],
   "scoped": [
     "style"
   ],
-  "default": [
-    "track"
+  "seamless": [
+    "iframe"
   ],
-  "disablepictureinpicture": [
-    "video"
+  "selected": [
+    "option"
   ]
 }

gen/build/dom.js

@@ -63,8 +63,9 @@ const processReactTypeDeclarations = async (source) => {
     node.forEachChild(c => void unvisited.push(c));
   }
 
+  // Sort output JSON object by property so diffs are clearer.
   await fs.writeFile(BOOLEAN_ATTRS_PATH, JSON.stringify(
-    Object.fromEntries(booleanAttributes.entries()),
+    Object.fromEntries([...booleanAttributes.entries()].sort((a, b) => a[0].localeCompare(b[0]))),
     null,
     2,
   ));

src/unit/attr/mod.rs

@@ -3,7 +3,7 @@ use phf::{phf_set, Set};
 use crate::err::ProcessingResult;
 use crate::proc::{Processor, ProcessorRange};
 use crate::spec::codepoint::is_control;
-use crate::unit::attr::value::{DelimiterType, process_attr_value, ProcessedAttrValue};
+use crate::unit::attr::value::{DelimiterType, process_attr_value, ProcessedAttrValue, skip_attr_value};
 
 mod value;
 
@@ -49,15 +49,14 @@ pub fn process_attr(proc: &mut Processor, element: ProcessorRange) -> Processing
     let (typ, value) = if !has_value {
         (AttrType::NoValue, None)
     } else {
-        // TODO Don't process if going to erase anyway.
-        let val = process_attr_value(proc, should_collapse_and_trim_value_ws)?;
         if is_boolean {
-            proc.erase_written(after_name);
+            skip_attr_value(proc)?;
             (AttrType::NoValue, None)
         } else {
-            match val {
+            match process_attr_value(proc, should_collapse_and_trim_value_ws)? {
                 ProcessedAttrValue { value: None, .. } => {
-                    // Value is empty, which is equivalent to no value, so discard `=` and any quotes.
+                    // Value is empty, which is equivalent to no value, so discard `=`.
+                    debug_assert_eq!(proc.written_count(after_name), 1);
                     proc.erase_written(after_name);
                     (AttrType::NoValue, None)
                 }

src/unit/attr/value.rs

@@ -133,6 +133,13 @@ impl Metrics {
     }
 }
 
+pub fn skip_attr_value(proc: &mut Processor) -> ProcessingResult<()> {
+    let src_delimiter = chain!(proc.match_pred(is_attr_quote).require_with_reason("attribute value delimiter quote")?.discard().char());
+    chain!(proc.match_while_not_char(src_delimiter).discard());
+    chain!(proc.match_char(src_delimiter).require_with_reason("attribute value delimiter quote")?.discard());
+    Ok(())
+}
+
 pub struct ProcessedAttrValue {
     pub delimiter: DelimiterType,
     pub value: Option<ProcessorRange>,

src/unit/script/js.rs

@@ -8,19 +8,19 @@ fn is_string_delimiter(c: u8) -> bool {
 
 fn parse_comment_single(proc: &mut Processor) -> ProcessingResult<()> {
     if cfg!(debug_assertions) {
-        chain!(proc.match_seq(b"//").expect().keep());
+        chain!(proc.match_seq(b"//").expect().discard());
     } else {
-        proc.accept_amount_expect(2);
+        proc.skip_amount_expect(2);
     };
 
     // Comment can end at closing </script>.
     // TODO Optimise
-    while !chain!(proc.match_line_terminator().keep().matched()) {
+    while !chain!(proc.match_line_terminator().discard().matched()) {
         if chain!(proc.match_seq(b"</script>").matched()) {
             break;
         }
 
-        proc.accept()?;
+        proc.skip()?;
     }
 
     Ok(())
@@ -28,19 +28,19 @@ fn parse_comment_single(proc: &mut Processor) -> ProcessingResult<()> {
 
 fn parse_comment_multi(proc: &mut Processor) -> ProcessingResult<()> {
     if cfg!(debug_assertions) {
-        chain!(proc.match_seq(b"/*").expect().keep());
+        chain!(proc.match_seq(b"/*").expect().discard());
     } else {
-        proc.accept_amount_expect(2);
+        proc.skip_amount_expect(2);
     };
 
     // Comment can end at closing </script>.
     // TODO Optimise
-    while !chain!(proc.match_seq(b"*/").keep().matched()) {
+    while !chain!(proc.match_seq(b"*/").discard().matched()) {
         if chain!(proc.match_seq(b"</script>").matched()) {
             break;
         }
 
-        proc.accept()?;
+        proc.skip()?;
     };
 
     Ok(())

src/unit/style.rs

@@ -11,14 +11,14 @@ fn is_string_delimiter(c: u8) -> bool {
 
 fn parse_comment(proc: &mut Processor) -> ProcessingResult<()> {
     if cfg!(debug_assertions) {
-        chain!(proc.match_seq(b"/*").expect().keep());
+        chain!(proc.match_seq(b"/*").expect().discard());
     } else {
-        proc.accept_amount_expect(2);
+        proc.skip_amount_expect(2);
     };
 
     // Unlike script tags, style comments do NOT end at closing tag.
-    while !chain!(proc.match_seq(b"*/").keep().matched()) {
-        proc.accept()?;
+    while !chain!(proc.match_seq(b"*/").discard().matched()) {
+        proc.skip()?;
     };
 
     Ok(())