Stricter URL parsing

magnetar_mmm_parser/src/lib.rs

@@ -991,7 +991,10 @@ impl Context {
         let (input, url_span) = recognize(tuple((protocol, url_chars(tag(">"), true))))(input)?;
         let (input, _) = tag(">")(input)?;
 
-        Ok((input, Token::UrlRaw(Cow::from(url_span.into_fragment()))))
+        Ok((
+            input,
+            Token::UrlNoEmbed(Cow::from(url_span.into_fragment())),
+        ))
     }
 
     fn link<'a>(&self, input: Span<'a>) -> IResult<Span<'a>, Token<'a>> {
@@ -1120,7 +1123,12 @@ fn protocol(input: Span) -> IResult<Span, Span> {
 
 #[inline]
 fn url_chars_base(input: Span) -> IResult<Span, Span> {
-    recognize(alt((alpha1, recognize(one_of(".,_/:%#$&?!~=+-()[]@")))))(input)
+    recognize(alt((
+        alpha1,
+        recognize(tuple((tag("["), many_till(url_chars_base, tag("]"))))),
+        recognize(tuple((tag("("), many_till(url_chars_base, tag(")"))))),
+        recognize(one_of(".,_/:%#$&?!~=+-@")),
+    )))(input)
 }
 
 #[inline]
@@ -1128,26 +1136,10 @@ fn url_chars<'a, T: 'a>(
     terminator: impl Fn(Span<'a>) -> IResult<Span<'a>, T> + 'a,
     spaces: bool,
 ) -> impl FnMut(Span<'a>) -> IResult<Span<'a>, Span<'a>> + 'a {
-    let terminating = move |input| {
-        tuple((
-            &terminator,
-            alt((
-                space1,
-                line_ending,
-                eof,
-                recognize(one_of("([<'\"")),
-                recognize(tuple((
-                    alt((alpha1, recognize(one_of("*")))),
-                    alt((space1, line_ending, eof)),
-                ))),
-            )),
-        ))(input)
-    };
-
     let chars = tuple((
         not(tuple((space1, eof))),
         not(tuple((space1, tag("\"")))),
-        not(tuple((opt(space1), terminating))),
+        not(tuple((opt(space1), terminator))),
         alt((url_chars_base, if spaces { space1 } else { fail })),
     ));
 
@@ -1167,49 +1159,48 @@ mod test {
 
     #[test]
     fn parse_url_chars() {
-        let test1 = "https://en.wikipedia.org/wiki/Sandbox_(computer_security))";
         assert_eq!(
+            url_chars(tag(")"), true)(Span::new(
+                "https://en.wikipedia.org/wiki/Sandbox_(computer_security))"
+            ))
+            .unwrap()
+            .1
+            .into_fragment(),
+            "https://en.wikipedia.org/wiki/Sandbox_(computer_security)"
+        );
+
+        assert_eq!(
+            url_chars(tag(")"), true)(Span::new(
+                "https://en.wikipedia.org/wiki/Sandbox_(computer_security)))"
+            ))
+            .unwrap()
+            .1
+            .into_fragment(),
             "https://en.wikipedia.org/wiki/Sandbox_(computer_security)",
-            url_chars(tag(")"), true)(Span::new(test1))
-                .unwrap()
-                .1
-                .into_fragment()
         );
 
-        let test2 = "https://en.wikipedia.org/wiki/Sandbox_(computer_security)))";
         assert_eq!(
-            "https://en.wikipedia.org/wiki/Sandbox_(computer_security))",
+            url_chars(tag(")"), true)(Span::new("https://cs.wikipedia.org/wiki/Among_Us  "))
-            url_chars(tag(")"), true)(Span::new(test2))
                 .unwrap()
                 .1
-                .into_fragment()
+                .into_fragment(),
-        );
-
-        let test3 = "https://en.wikipedia.org/wiki/(";
-        assert_eq!(
-            test3,
-            url_chars(tag(")"), true)(Span::new(test3))
-                .unwrap()
-                .1
-                .into_fragment()
-        );
-
-        let test4 = "https://cs.wikipedia.org/wiki/Among_Us  ";
-        assert_eq!(
             "https://cs.wikipedia.org/wiki/Among_Us",
-            url_chars(tag(")"), true)(Span::new(test4))
-                .unwrap()
-                .1
-                .into_fragment()
         );
 
-        let test5 = "https://cs.wikipedia.org/wiki/Among Us  )";
         assert_eq!(
-            "https://cs.wikipedia.org/wiki/Among Us",
+            url_chars(tag(")"), true)(Span::new("https://cs.wikipedia.org/wiki/Among Us  )"))
-            url_chars(tag(")"), true)(Span::new(test5))
                 .unwrap()
                 .1
-                .into_fragment()
+                .into_fragment(),
+            "https://cs.wikipedia.org/wiki/Among Us"
+        );
+
+        assert_eq!(
+            url_chars(tag(")"), false)(Span::new("https://en.wikipedia.org/wiki/Among Us  )"))
+                .unwrap()
+                .1
+                .into_fragment(),
+            "https://en.wikipedia.org/wiki/Among"
         );
     }
 
@@ -1381,6 +1372,20 @@ text</center>"#
             ])
         );
 
+        assert_eq!(
+            parse_full("<https://example.com>"),
+            Token::UrlNoEmbed("https://example.com".into())
+        );
+
+        // Adjacent links okay
+        assert_eq!(
+            parse_full("<https://example.com/><https://awawa.gay/>"),
+            Token::Sequence(vec![
+                Token::UrlNoEmbed("https://example.com/".into()),
+                Token::UrlNoEmbed("https://awawa.gay/".into())
+            ])
+        );
+
         assert_eq!(
             parse_full("Link test: ?[label](https://awawa.gay)"),
             Token::Sequence(vec![
@@ -1393,6 +1398,32 @@ text</center>"#
             ])
         );
 
+        assert_eq!(
+            parse_full("Link test: ?[label](https://awawa.gay)test"),
+            Token::Sequence(vec![
+                Token::PlainText("Link test: ".into()),
+                Token::Link {
+                    label: Box::new(Token::PlainText("label".into())),
+                    href: "https://awawa.gay".into(),
+                    embed: false
+                },
+                Token::PlainText("test".into())
+            ])
+        );
+
+        assert_eq!(
+            parse_full("Link test: (?[label](https://awawa.gay))"),
+            Token::Sequence(vec![
+                Token::PlainText("Link test: (".into()),
+                Token::Link {
+                    label: Box::new(Token::PlainText("label".into())),
+                    href: "https://awawa.gay".into(),
+                    embed: false
+                },
+                Token::PlainText(")".into())
+            ])
+        );
+
         assert_eq!(
             parse_full("Link test: ?[label](https://awawa.gay"), // Missing closing bracket
             Token::Sequence(vec![