Use syntect to properly process links

Cargo.toml

@@ -22,7 +22,7 @@ rusqlite_migration = { version = "1", default-features = false }
 rust-argon2 = "2.0.0"
 sha2 = "0"
 serde = { version = "1", features = ["derive"] }
-syntect = { version = "5", default-features = false, features = ["html", "plist-load", "regex-fancy"] }
+syntect = { version = "5", default-features = false, features = ["html", "plist-load", "regex-fancy", "yaml-load"] }
 thiserror = "2"
 time = { version = "0.3", features = ["macros", "serde"] }
 tokio = { version = "1", features = ["full"] }
@@ -34,7 +34,6 @@ url = "2"
 zstd = "0.13"
 parking_lot = "0.12.1"
 http = "1.1.0"
-regex = { version = "1.11.1", default-features = false, features = ["std"] }
 
 [dev-dependencies]
 reqwest = { version = "0", default-features = false, features = ["cookies", "json"] }

assets/LinkHighlight.sublime-syntax

@@ -0,0 +1,33 @@
+%YAML 1.2
+---
+# http://www.sublimetext.com/docs/3/syntax.html
+name: LinkHighlight
+file_extensions:
+  - link_highlight
+scope: text.highlighting
+contexts:
+  main:
+    - match: (<)((?:https?|ftp)://.*?)(>)
+      scope: meta.link.inet.markdown
+      captures:
+        1: punctuation.definition.link.begin.markdown
+        2: markup.underline.link.markdown
+        3: punctuation.definition.link.end.markdown
+    - match: (((https|http|ftp)://)|www\.)[\w-]+(\.[\w-]+)+
+      scope: markup.underline.link.markdown
+      push: # After a valid domain, zero or more non-space non-< characters may follow
+        - match: (?=[?!.,:*_~]*[\s<]) # Trailing punctuation (specifically, ?, !, ., ,, :, *, _, and ~) will not be considered part of the autolink, though they may be included in the interior of the link
+          pop: true
+        - match: (?={{html_entity}}[?!.,:*_~]*[\s<]) # If an autolink ends in a semicolon (;), we check to see if it appears to resemble an entity reference; if the preceding text is & followed by one or more alphanumeric characters. If so, it is excluded from the autolink
+          pop: true
+        - match: \( # When an autolink ends in ), we scan the entire autolink for the total number of parentheses. If there is a greater number of closing parentheses than opening ones, we don’t consider the last character part of the autolink, in order to facilitate including an autolink inside a parenthesis
+          push:
+            - meta_scope: markup.underline.link.markdown
+            - match: (?=[?!.,:*_~]*[\s<])
+              pop: true
+            - match: \)
+              pop: true
+        - match: (?=\)[?!.,:*_~]*[\s<])
+          pop: true
+        - match: '[^?!.,:*_~\s<&()]+|\S'
+          scope: markup.underline.link.markdown

src/highlight.rs

@@ -1,14 +1,12 @@
 use crate::db::read::Entry;
 use crate::errors::Error;
-use regex::{Captures, Regex};
 use sha2::{Digest, Sha256};
 use std::cmp::Ordering;
-use std::collections::HashMap;
 use std::io::Cursor;
 use std::sync::LazyLock;
 use syntect::highlighting::ThemeSet;
 use syntect::html::{css_for_theme_with_class_style, line_tokens_to_classed_spans, ClassStyle};
-use syntect::parsing::{ParseState, ScopeStack, SyntaxReference, SyntaxSet};
+use syntect::parsing::{ParseState, ScopeStack, SyntaxDefinition, SyntaxReference, SyntaxSet};
 use syntect::util::LinesWithEndings;
 
 const HIGHLIGHT_LINE_LENGTH_CUTOFF: usize = 2048;
@@ -34,6 +32,11 @@ pub static DATA: LazyLock<Data> = LazyLock::new(|| {
     let dark = Css::new("dark", &DARK_CSS);
     let syntax_set: SyntaxSet =
         syntect::dumps::from_binary(include_bytes!("../assets/newlines.packdump"));
+    let link_highlighting = SyntaxDefinition::load_from_str(
+        include_str!("../assets/LinkHighlight.sublime-syntax"), false, None).expect("loading link style");
+    let mut builder = syntax_set.into_builder();
+    builder.add(link_highlighting);
+    let syntax_set = builder.build();
     let mut syntaxes = syntax_set.syntaxes().to_vec();
     syntaxes.sort_by(|a, b| a.name.partial_cmp(&b.name).unwrap_or(Ordering::Less));
 
@@ -46,14 +49,6 @@ pub static DATA: LazyLock<Data> = LazyLock::new(|| {
     }
 });
 
-static LINK_REGEX: LazyLock<Regex> = LazyLock::new(|| {
-    Regex::new(r#"https?://[^ "><'\t\n`]+"#).unwrap()
-});
-
-static LINK_REGEX_2: LazyLock<Regex> = LazyLock::new(|| {
-    Regex::new(r#"LINKWASTEBIN\d+END"#).unwrap()
-});
-
 /// Combines CSS content with a filename containing the hash of the content.
 pub struct Css<'a> {
     pub name: String,
@@ -82,12 +77,13 @@ impl<'a> Css<'a> {
 }
 
 fn highlight(source: &str, ext: &str) -> Result<String, Error> {
-    let syntax_ref = DATA
-        .syntax_set
-        .find_syntax_by_extension(ext)
+    let syntax_ref = (ext != "txt").then(||
+        DATA
+            .syntax_set
+            .find_syntax_by_extension(ext)).flatten()
         .unwrap_or_else(|| {
             DATA.syntax_set
-                .find_syntax_by_extension("txt")
+                .find_syntax_by_extension("link_highlight")
                 .expect("finding txt syntax")
         });
 
@@ -96,18 +92,10 @@ fn highlight(source: &str, ext: &str) -> Result<String, Error> {
     let mut scope_stack = ScopeStack::new();
 
     for (mut line_number, line) in LinesWithEndings::from(source).enumerate() {
-        let mut links = HashMap::new();
-        let (mut formatted, delta) = if line.len() > HIGHLIGHT_LINE_LENGTH_CUTOFF {
+        // let mut links = HashMap::new();
+        let (formatted, delta) = if line.len() > HIGHLIGHT_LINE_LENGTH_CUTOFF {
             (line.to_string(), 0)
         } else {
-            // Add placeholder for link elements.
-            let line = LINK_REGEX.replace_all(&line, |x: &Captures| {
-                let num = links.len();
-                let placeholder = format!("LINKWASTEBIN{num}END");
-                links.insert(placeholder.clone(), x.get(0).unwrap().as_str().to_owned());
-                placeholder
-            }).to_string();
-
             let parsed = parse_state.parse_line(&line, &DATA.syntax_set)?;
             line_tokens_to_classed_spans(
                 &line,
@@ -128,13 +116,6 @@ fn highlight(source: &str, ext: &str) -> Result<String, Error> {
             html.push_str(&"<span>".repeat(delta.abs().try_into()?));
         }
 
-        // Process link element placeholders.
-        formatted = LINK_REGEX_2.replace_all(&formatted, |x: &Captures| {
-            let id = x.get(0).unwrap().as_str();
-            let link = links.get(id).map(|x| &**x).unwrap_or(id);
-            format!("<a href='{link}'>{link}</a>")
-        }).to_string();
-
         // Strip stray newlines that cause vertically stretched lines.
         for c in formatted.chars().filter(|c| *c != '\n') {
             html.push(c);

templates/base.html

@@ -23,6 +23,7 @@
       </header>
       <main>
         {% block content %}{% endblock %}
+		{% block content_post %}{% endblock %}
       </main>
     </div>
   </body>

templates/paste.html

@@ -53,3 +53,26 @@
     <li><a href="{{ base_path.join(id) }}?fmt=raw"><button>raw</button></a></li>
     <li><a href="{{ base_path.join(id) }}?fmt=qr"><button>qr</button></a></li>
 {% endblock %}
+
+{% block content_post %}
+<script>
+  while (true) {
+    const o = document.querySelector("span.link");
+    if (!o) {
+      break;
+    }
+    let l = o.nextSibling;
+    let link = o.innerHTML;
+    while(l != null && l.tagName == "SPAN" && l.classList.contains("link")) {
+      let prevL = l;
+      l = l.nextSibling;
+      link += prevL.innerText;
+      prevL.remove();
+    }
+    o.setAttribute("href", link);
+    o.innerText = link;
+	o.className = "";
+    o.outerHTML = o.outerHTML.replace(/^<span(.*)span>$/, "<a$1a>");
+  }
+</script>
+{% endblock %}