Fix parsing of comments in parse/1 (#141)

philss · web-flow · commit 515daa038346 · 2024-03-23T13:34:04.000-03:00
Closes #121
diff --git a/native/html5ever_nif/src/flat_dom.rs b/native/html5ever_nif/src/flat_dom.rs
@@ -583,7 +583,9 @@ pub fn flat_sink_to_rec_term<'a>(
                 NodeData::Text { contents } => {
                     term = StrTendrilWrapper(contents).encode(env);
                 }
-                NodeData::Comment { .. } => continue,
+                NodeData::Comment { contents } => {
+                    term = (atoms::comment(), StrTendrilWrapper(contents)).encode(env);
+                }
                 _ => unimplemented!(""),
             }
 
diff --git a/test/html5ever_test.exs b/test/html5ever_test.exs
@@ -8,9 +8,17 @@ defmodule Html5everTest do
   end
 
   test "parse basic html" do
-    html = "<html><head></head><body></body></html>"
+    html = "<html><head></head><body><h1>Hello</h1><!-- my comment --></body></html>"
 
-    assert Html5ever.parse(html) == {:ok, [{"html", [], [{"head", [], []}, {"body", [], []}]}]}
+    assert Html5ever.parse(html) ==
+             {:ok,
+              [
+                {"html", [],
+                 [
+                   {"head", [], []},
+                   {"body", [], [{"h1", [], ["Hello"]}, {:comment, " my comment "}]}
+                 ]}
+              ]}
   end
 
   test "does not parse with not valid UTF8 binary" do

Original file line number	Diff line number	Diff line change
`@@ -583,7 +583,9 @@ pub fn flat_sink_to_rec_term<'a>(`
`583`	`583`	`NodeData::Text { contents } => {`
`584`	`584`	`term = StrTendrilWrapper(contents).encode(env);`
`585`	`585`	`}`
`586`		`- NodeData::Comment { .. } => continue,`
	`586`	`+ NodeData::Comment { contents } => {`
	`587`	`+ term = (atoms::comment(), StrTendrilWrapper(contents)).encode(env);`
	`588`	`+ }`
`587`	`589`	`_ => unimplemented!(""),`
`588`	`590`	`}`
`589`	`591`