From 37115fef7ae64324daef28e204b48a870817a392 Mon Sep 17 00:00:00 2001
From: Wilson Lin <code@wilsonl.in>
Date: Tue, 14 Jan 2020 21:10:49 +1100
Subject: [PATCH] Allow more whitespace w.r.t. attributes; do not format bench
 test pages

---
 README.md            | 11 -----------
 bench/fetch.js       | 38 ++++++++++++--------------------------
 bench/package.json   |  1 -
 src/err.rs           |  4 ----
 src/unit/attr/mod.rs |  4 +++-
 src/unit/tag.rs      |  7 +------
 6 files changed, 16 insertions(+), 49 deletions(-)

diff --git a/README.md b/README.md
index c0d30bb..2cc03a8 100644
--- a/README.md
+++ b/README.md
@@ -351,17 +351,6 @@ Numeric character references that do not reference a valid [Unicode Scalar Value
 Backticks (`` ` ``) are not valid quote marks and not interpreted as such.
 However, backticks are valid attribute value quotes in Internet Explorer.
 
-It is an error if there is:
-
-- whitespace between `=` and an attribute name/value; and/or
-- no whitespace before an attribute.
-
-For example:
-
-```html
-<div id = "a"class="abc"></div>
-```
-
 Special handling of some attributes require case sensitive names and values. For example, `CLASS` won't be recognised as an attribute to minify, and `type="Text/JavaScript"` on a `<script>` will cause the element to be parsed as a [data block](https://html.spec.whatwg.org/dev/scripting.html#data-block) instead of JavaScript code.
 
 ### Script and style
diff --git a/bench/fetch.js b/bench/fetch.js
index 0bb1147..62ade6f 100644
--- a/bench/fetch.js
+++ b/bench/fetch.js
@@ -1,7 +1,6 @@
 const {promises: fs} = require('fs');
 const request = require('request-promise-native');
 const path = require('path');
-const prettier = require('prettier');
 
 const tests = {
   "Amazon": "https://www.amazon.com/",
@@ -38,31 +37,18 @@ const fetchTest = async (name, url) => {
 
   // Format after fetching as formatting is synchronous and can take so long that connections get dropped by server due to inactivity.
   for (const [name, html] of await Promise.all(Object.entries(tests).map(([name, url]) => fetchTest(name, url)))) {
-    // Apply some fixes to HTML to allow strict formatter to work.
-    const formatted = prettier.format(
-      html
-        // Fix missing semicolon after entity in Amazon.
-        .replace(/&#x200b/g, '&#x200b;')
-        // Fix consecutive malformed entities in Amazon.
-        .replace(/&& window.ue_sbl/g, '&amp&amp window.ue_sbl')
-        .replace(/&&pf_rd_p/g, '&amp&amppf_rd_p')
-        // Fix early termination of conditional comment in Amazon.
-        .replace('--></style>\n<![endif]-->', '</style>\n<![endif]-->')
-        // Fix closing of void tag in Amazon.
-        .replace(/><\/hr>/g, '/>')
-        // Fix extra '</div>' in BBC.
-        .replace('</a></span></small></div></div></div></footer>', '</a></span></small></div></div></footer>')
-        // Fix consecutive malformed entities in Google.
-        .replace(/&&google.aft/g, '&amp&ampgoogle.aft')
-        // Fix parser failing to parse unquoted attribute value starting with forward slash in Stack Overflow.
-        .replace('action=/search', 'action="/search"')
-        // Fix broken attribute value in Stack Overflow.
-        .replace('height=151"', 'height="151"')
-      ,
-      {parser: 'html'},
-    );
-    console.log(`Formatted ${name}`);
-    await fs.writeFile(path.join(__dirname, 'tests', `${name}.html`), formatted);
+    // Apply some fixes to HTML.
+    const fixed = html
+      // Fix early termination of conditional comment in Amazon.
+      .replace('--></style>\n<![endif]-->', '</style>\n<![endif]-->')
+      // Fix closing of void tag in Amazon.
+      .replace(/><\/hr>/g, '/>')
+      // Fix extra '</div>' in BBC.
+      .replace('</a></span></small></div></div></div></footer>', '</a></span></small></div></div></footer>')
+      // Fix broken attribute value in Stack Overflow.
+      .replace('height=151"', 'height="151"')
+    ;
+    await fs.writeFile(path.join(__dirname, 'tests', `${name}.html`), fixed);
   }
 })()
   .catch(console.error);
diff --git a/bench/package.json b/bench/package.json
index 07de493..d5de7a9 100644
--- a/bench/package.json
+++ b/bench/package.json
@@ -8,7 +8,6 @@
     "hyperbuild": "file:../nodejs",
     "minimize": "2.2.0",
     "mkdirp": "^0.5.1",
-    "prettier": "^1.19.1",
     "request": "^2.88.0",
     "request-promise-native": "^1.0.8"
   },
diff --git a/src/err.rs b/src/err.rs
index 5a8f6c9..cbc0c5e 100644
--- a/src/err.rs
+++ b/src/err.rs
@@ -2,7 +2,6 @@
 #[derive(Debug)]
 pub enum ErrorType {
     ClosingTagMismatch,
-    NoSpaceBeforeAttr,
     MatchNotFound(&'static [u8]),
     NotFound(&'static str),
     ExpectedChar(u8),
@@ -15,9 +14,6 @@ impl ErrorType {
             ErrorType::ClosingTagMismatch => {
                 format!("Closing tag name does not match opening tag.")
             }
-            ErrorType::NoSpaceBeforeAttr => {
-                format!("Space required before attribute.")
-            }
             ErrorType::MatchNotFound(seq) => {
                 format!("Expected `{}`.", unsafe { std::str::from_utf8_unchecked(seq) })
             }
diff --git a/src/unit/attr/mod.rs b/src/unit/attr/mod.rs
index 36b5896..2f7071f 100644
--- a/src/unit/attr/mod.rs
+++ b/src/unit/attr/mod.rs
@@ -2,7 +2,7 @@ use phf::{phf_set, Set};
 
 use crate::err::ProcessingResult;
 use crate::proc::{Processor, ProcessorRange};
-use crate::spec::codepoint::is_control;
+use crate::spec::codepoint::{is_control, is_whitespace};
 use crate::unit::attr::value::{DelimiterType, process_attr_value, ProcessedAttrValue, skip_attr_value};
 
 mod value;
@@ -44,11 +44,13 @@ pub fn process_attr(proc: &mut Processor, element: ProcessorRange) -> Processing
     let after_name = proc.checkpoint();
 
     let should_collapse_and_trim_value_ws = COLLAPSIBLE_AND_TRIMMABLE_ATTRS.contains(&proc[name]);
+    let ws_accepted = chain!(proc.match_while_pred(is_whitespace).discard().matched());
     let has_value = chain!(proc.match_char(b'=').keep().matched());
 
     let (typ, value) = if !has_value {
         (AttrType::NoValue, None)
     } else {
+        let ws_accepted = chain!(proc.match_while_pred(is_whitespace).discard().matched());
         if is_boolean {
             skip_attr_value(proc)?;
             (AttrType::NoValue, None)
diff --git a/src/unit/tag.rs b/src/unit/tag.rs
index 0e96703..1ebb343 100644
--- a/src/unit/tag.rs
+++ b/src/unit/tag.rs
@@ -108,17 +108,12 @@ pub fn process_tag(proc: &mut Processor, prev_sibling_closing_tag: Option<Proces
             break;
         }
 
-        // This needs to be enforced as otherwise there would be difficulty in determining what is the end of a tag/attribute name/attribute value.
-        if !ws_accepted {
-            return Err(ErrorType::NoSpaceBeforeAttr);
-        }
-
         // Mark attribute start in case we want to erase it completely.
         let attr_checkpoint = proc.checkpoint();
         let mut erase_attr = false;
 
         // Write space after tag name or unquoted/valueless attribute.
-        // Don't write after unquoted.
+        // Don't write after quoted.
         match last_attr_type {
             Some(AttrType::Unquoted) | Some(AttrType::NoValue) | None => proc.write(b' '),
             _ => {}