From cb24837ff08cbb9a8ad5691ca2d8bce9f418d2e6 Mon Sep 17 00:00:00 2001
From: jutty <j@jutty.dev>
Date: Mon, 5 Jan 2026 00:35:38 -0300
Subject: [PATCH] Fix unbalanced anchor tags

---
 src/syntax/content/parser.rs                 | 61 ++++++++++++++------
 src/syntax/content/parser/context.rs         |  2 +-
 src/syntax/content/parser/context/anchor.rs  | 39 +++++++++++--
 src/syntax/content/parser/token/linebreak.rs |  4 +-
 src/syntax/content/parser/token/literal.rs   |  4 +-
 src/syntax/content/parser/token/paragraph.rs |  4 ++
 6 files changed, 88 insertions(+), 26 deletions(-)

diff --git a/src/syntax/content/parser.rs b/src/syntax/content/parser.rs
index 55fbf68..29e23cd 100644
--- a/src/syntax/content/parser.rs
+++ b/src/syntax/content/parser.rs
@@ -51,9 +51,7 @@ fn lex(text: &str, map: LexMap, config: &Config) -> Vec<Token> {
                     tokens.push(Token::Header(header));
                     continue;
                 } else if Paragraph::probe(lexeme) {
-                    log!(
-                        "Probed {lexeme:#?} from Block::None -> Block::Paragraph"
-                    );
+                    log!("Probed block context None -> Paragraph: {lexeme:?}");
                     state.context.block = Block::Paragraph;
                     tokens.push(Token::Paragraph(Paragraph::new(true)));
                 }
@@ -68,12 +66,8 @@ fn lex(text: &str, map: LexMap, config: &Config) -> Vec<Token> {
                 continue;
             },
             Block::Paragraph => {
-                if lexeme.text() == "\n"
-                    && matches!(state.context.inline, Inline::None)
-                {
-                    log!(
-                        "Probed {lexeme:#?} from Block::Paragraph -> Block::None"
-                    );
+                if Paragraph::probe_end(lexeme) {
+                    log!("Probed block context Paragraph -> None: {lexeme:?}");
                     tokens.push(Token::Paragraph(Paragraph::new(false)));
                     state.context.block = Block::None;
                 }
@@ -357,8 +351,7 @@ mod tests {
             read_noconfig("\n|SomeAnchor|\n"),
             concat!(
                 "\n",
-                r#"<p><a href="/node/SomeAnchor">SomeAnchor</a></p>"#,
-                "\n"
+                r#"<p><a href="/node/SomeAnchor">SomeAnchor</a></p>"#
             ),
         );
     }
@@ -368,10 +361,9 @@ mod tests {
         assert_eq!(
             read_noconfig("|SomeAnchor|\n|SomeOtherAnchor|\n"),
             concat!(
-                r#"<p><a href="/node/SomeAnchor">SomeAnchor</a></p>"#,
+                r#"<p><a href="/node/SomeAnchor">SomeAnchor</a>"#,
                 "\n",
-                r#"<p><a href="/node/SomeOtherAnchor">SomeOtherAnchor</a></p>"#,
-                "\n"
+                r#"<a href="/node/SomeOtherAnchor">SomeOtherAnchor</a></p>"#
             )
         );
     }
@@ -384,18 +376,53 @@ mod tests {
                 r#"<p><a href="/node/SomeAnchor">SomeAnchor</a></p>"#,
                 "\n",
                 "\n",
-                r#"<p><a href="/node/SomeOtherAnchor">SomeOtherAnchor</a></p>"#,
-                "\n",
+                r#"<p><a href="/node/SomeOtherAnchor">SomeOtherAnchor</a></p>"#
             ),
         );
     }
 
+    #[test]
+    fn homepage_footer() {
+        assert_eq!(
+            read_noconfig(
+                "made by jutty|https://jutty.dev • acknowledgments|Acknowledgments • |source code|https://codeberg.org/jutty/en"
+            ),
+            r#"<p>made by <a href="https://jutty.dev">jutty</a> • <a href="/node/Acknowledgments">acknowledgments</a> • <a href="https://codeberg.org/jutty/en">source code</a></p>"#
+        );
+    }
+
+    #[test]
+    fn trailing_anchor() {
+        assert_eq!(
+            read_noconfig("see acks|acks"),
+            r#"<p>see <a href="/node/acks">acks</a></p>"#
+        );
+    }
+
+    #[test]
+    fn trailing_anchor_with_newline() {
+        assert_eq!(
+            read_noconfig("\nsee acks|acks\n"),
+            concat!("\n", r#"<p>see <a href="/node/acks">acks</a></p>"#)
+        );
+    }
+
+    #[test]
+    fn trailing_oblique() {
+        assert_eq!(read_noconfig("see _acks_"), "<p>see <em>acks</em></p>");
+    }
+
+    #[test]
+    fn trailing_oblique_with_newline() {
+        assert_eq!(read_noconfig("see _acks_\n"), "<p>see <em>acks</em></p>");
+    }
+
     #[test]
     fn pre() {
         let payload = "D0qdJ184f3q1okbYu3Xm1d93jj6jy615";
         assert_eq!(
             read_noconfig(&format!("`\n{payload}\n`\n")),
-            format!("<pre>\n{payload}\n</pre>\n"),
+            format!("<pre>\n{payload}\n</pre>"),
         );
     }
 
diff --git a/src/syntax/content/parser/context.rs b/src/syntax/content/parser/context.rs
index 542e51c..a69d66c 100644
--- a/src/syntax/content/parser/context.rs
+++ b/src/syntax/content/parser/context.rs
@@ -1,6 +1,6 @@
 use crate::syntax::content::parser::{
-    token::{Token, paragraph::Paragraph, preformat::PreFormat},
     State,
+    token::{Token, paragraph::Paragraph, preformat::PreFormat},
 };
 
 pub mod anchor;
diff --git a/src/syntax/content/parser/context/anchor.rs b/src/syntax/content/parser/context/anchor.rs
index 2fecad3..d35265d 100644
--- a/src/syntax/content/parser/context/anchor.rs
+++ b/src/syntax/content/parser/context/anchor.rs
@@ -57,6 +57,7 @@ pub fn parse(
         );
 
         // Conditions in this decision tree should match the destination end
+        // or some intermediary state necessary to finding it
         if lexeme.match_as_char('s')
             && lexeme.is_next_boundary()
             && !lexeme.match_next_as_char('|')
@@ -80,7 +81,7 @@ pub fn parse(
             state.context.inline = Inline::None;
             return true;
         } else if lexeme.match_as_char('|') && !candidate.balanced {
-            log!("Found a pipe, but no boundary: Destination likely follows");
+            log!("State: Found a pipe, but no boundary: destination follows");
             candidate.balanced = true;
             return true;
         } else if lexeme.match_as_char('|') {
@@ -96,12 +97,13 @@ pub fn parse(
             tokens.push(Token::Anchor(candidate.clone()));
             state.context.inline = Inline::None;
             return false;
-        } else if lexeme.is_whitespace() {
-            log!("End: Whitespace");
+        } else if lexeme.is_next_whitespace() {
+            log!("End: next is whitespace");
+            buffer.destination.push_str(&lexeme.text());
             candidate.destination = Some(buffer.destination.clone());
             tokens.push(Token::Anchor(candidate.clone()));
             state.context.inline = Inline::None;
-            return false;
+            return true;
 
         // This else branch is the 'no end found yet' state and will keep
         // pushing lexemes into the buffer until an end is found above
@@ -137,3 +139,32 @@ pub fn parse(
     state.context.inline = Inline::None;
     false
 }
+
+#[cfg(test)]
+mod tests {
+    use crate::{syntax::content::parser, types::Graph};
+
+    fn read_noconfig(input: &str) -> String {
+        parser::read(input, &Graph::new(None).meta.config)
+    }
+
+    #[test]
+    fn indifferent_trailing_pipe() {
+        assert_eq!(read_noconfig("|a|a|"), read_noconfig("a|a|"));
+    }
+
+    #[test]
+    fn indifferent_leading_pipe() {
+        assert_eq!(read_noconfig("|a|a|"), read_noconfig("|a|a"));
+    }
+
+    #[test]
+    fn indifferent_multiline_trailing_pipe() {
+        assert_eq!(read_noconfig("|a|a|\nn"), read_noconfig("a|a|\nn"));
+    }
+
+    #[test]
+    fn indifferent_multiline_leading_pipe() {
+        assert_eq!(read_noconfig("|a|a|\nn"), read_noconfig("|a|a\nn"));
+    }
+}
diff --git a/src/syntax/content/parser/token/linebreak.rs b/src/syntax/content/parser/token/linebreak.rs
index 2341384..e9209f8 100644
--- a/src/syntax/content/parser/token/linebreak.rs
+++ b/src/syntax/content/parser/token/linebreak.rs
@@ -2,12 +2,12 @@ use crate::{
     syntax::content::{Parseable, parser::lexeme::Lexeme},
 };
 
-#[derive(Debug, Clone, Eq, PartialEq)]
+#[derive(Default, Debug, Clone, Eq, PartialEq)]
 pub struct LineBreak {}
 
 impl Parseable for LineBreak {
     fn probe(lexeme: &Lexeme) -> bool {
-        lexeme.text() == "\n"
+        lexeme.text() == "\n" && !lexeme.last()
     }
 
     fn lex(_lexeme: &Lexeme) -> LineBreak {
diff --git a/src/syntax/content/parser/token/literal.rs b/src/syntax/content/parser/token/literal.rs
index d311633..85f12ab 100644
--- a/src/syntax/content/parser/token/literal.rs
+++ b/src/syntax/content/parser/token/literal.rs
@@ -6,8 +6,8 @@ pub struct Literal {
 }
 
 impl Parseable for Literal {
-    fn probe(_lexeme: &Lexeme) -> bool {
-        true
+    fn probe(lexeme: &Lexeme) -> bool {
+        !(lexeme.last() && lexeme.is_whitespace())
     }
 
     fn lex(lexeme: &Lexeme) -> Literal {
diff --git a/src/syntax/content/parser/token/paragraph.rs b/src/syntax/content/parser/token/paragraph.rs
index b0def3d..bf46b28 100644
--- a/src/syntax/content/parser/token/paragraph.rs
+++ b/src/syntax/content/parser/token/paragraph.rs
@@ -9,6 +9,10 @@ impl Paragraph {
     pub fn new(open: bool) -> Paragraph {
         Paragraph { open: Some(open) }
     }
+
+    pub fn probe_end(lexeme: &Lexeme) -> bool {
+        lexeme.match_as_char('\n') && lexeme.match_next_as_char('\n')
+    }
 }
 
 impl Parseable for Paragraph {