From 6ab687c73795b899b3ab0e8b8e9a12c5377553db Mon Sep 17 00:00:00 2001
From: Kirsten Laskoski <kirsten.laskoski@gmail.com>
Date: Sun, 28 Apr 2024 13:45:37 -0400
Subject: [PATCH] Handle whitespace inside closing tags

Fixes #65 by adding a skip whitespace before and after the name of a
closing tag. Also adds a test to verify that this functions properly,
based on the example in #65.

Note that the [HTML spec] does not allow for whitespace between the `</`
and the tag name, although this is accepted by these changes to maintain
symmetry with the parsing of the start of a tag which started allowing
whitespace between `<` and the tag name in c532c72.

[HTML spec]: https://html.spec.whatwg.org/multipage/syntax.html#end-tags
---
 src/parser/base.rs | 22 ++++++++++++++--------
 src/tests.rs       | 33 ++++++++++++++++++++++++++++++++-
 2 files changed, 46 insertions(+), 9 deletions(-)

diff --git a/src/parser/base.rs b/src/parser/base.rs
index 7b8c498..7b9512e 100644
--- a/src/parser/base.rs
+++ b/src/parser/base.rs
@@ -217,17 +217,23 @@ impl<'a> Parser<'a> {
 
     fn read_end(&mut self) {
         self.stream.advance();
+        self.skip_whitespaces();
 
-        let closing_tag_name = self.read_to(b'>');
-        
-        self.stream.expect_and_skip_cond(b'>');
+        if let Some(closing_tag_name) = self.read_ident() {
+            self.skip_whitespaces();
+            self.stream.expect_and_skip_cond(b'>');
 
-        let closing_tag_matches_parent = self.stack.last()
-            .and_then(|last_handle| last_handle.get(self))
-            .and_then(|last_item| last_item.as_tag())
-            .map_or(false, |last_tag| last_tag.name() == closing_tag_name);
+            let closing_tag_matches_parent = self
+                .stack
+                .last()
+                .and_then(|last_handle| last_handle.get(self))
+                .and_then(|last_item| last_item.as_tag())
+                .map_or(false, |last_tag| last_tag.name() == closing_tag_name);
 
-        if !closing_tag_matches_parent {
+            if !closing_tag_matches_parent {
+                return;
+            }
+        } else {
             return;
         }
 
diff --git a/src/tests.rs b/src/tests.rs
index 046861a..ecf2fe5 100644
--- a/src/tests.rs
+++ b/src/tests.rs
@@ -1,4 +1,4 @@
-use crate::{parse, parse_owned, Bytes};
+use crate::{parse, parse_owned, Bytes, VDom};
 use crate::{parser::*, HTMLTag, Node};
 
 fn force_as_tag<'a, 'b>(actual: &'a Node<'b>) -> &'a HTMLTag<'b> {
@@ -778,3 +778,34 @@ fn tag_raw_abrupt_stop() {
     let from_raw = first_tag.raw().try_as_utf8_str().unwrap();
     assert_eq!(from_raw, "<p>abcd</p");
 }
+
+#[test]
+fn line_breaks_inside_tags() {
+    /// Asserts that the given node is a div with the given text as its only child.
+    fn assert_div_with_text(node: &NodeHandle, dom: &VDom, text: &[u8]) {
+        let tag = node.get(dom.parser()).unwrap().as_tag().unwrap();
+        assert_eq!(&tag._name, "div");
+        assert_eq!(tag._children.len(), 1);
+        assert_eq!(
+            tag._children[0]
+                .get(dom.parser())
+                .unwrap()
+                .as_raw()
+                .unwrap()
+                .as_bytes(),
+            text
+        );
+    }
+
+    let input = "<div>div 1</div><\ndiv\n>div 2</\ndiv\n><div>div 3</div><div>div 4</div>";
+    let dom = parse(input, Default::default()).unwrap();
+    let children = dom.children();
+
+    // each of the 4 divs should be a top level child
+    assert_eq!(children.len(), 4);
+
+    assert_div_with_text(&children[0], &dom, b"div 1");
+    assert_div_with_text(&children[1], &dom, b"div 2");
+    assert_div_with_text(&children[2], &dom, b"div 3");
+    assert_div_with_text(&children[3], &dom, b"div 4");
+}