Merge pull request 'Post ap_url as valid IRI' (#947) from post-slug into main

Reviewed-on: https://git.joinplu.me/Plume/Plume/pulls/947

plume-common/src/utils.rs

@@ -27,6 +27,59 @@ pub fn make_actor_id(name: &str) -> String {
         .collect()
 }
 
+/**
+ * Percent-encode characters which are not allowed in IRI path segments.
+ *
+ * Intended to be used for generating Post ap_url.
+ */
+pub fn iri_percent_encode_seg(segment: &str) -> String {
+    segment.chars().map(iri_percent_encode_seg_char).collect()
+}
+
+pub fn iri_percent_encode_seg_char(c: char) -> String {
+    if c.is_alphanumeric() {
+        c.to_string()
+    } else {
+        match c {
+            '-'
+            | '.'
+            | '_'
+            | '~'
+            | '\u{A0}'..='\u{D7FF}'
+            | '\u{20000}'..='\u{2FFFD}'
+            | '\u{30000}'..='\u{3FFFD}'
+            | '\u{40000}'..='\u{4FFFD}'
+            | '\u{50000}'..='\u{5FFFD}'
+            | '\u{60000}'..='\u{6FFFD}'
+            | '\u{70000}'..='\u{7FFFD}'
+            | '\u{80000}'..='\u{8FFFD}'
+            | '\u{90000}'..='\u{9FFFD}'
+            | '\u{A0000}'..='\u{AFFFD}'
+            | '\u{B0000}'..='\u{BFFFD}'
+            | '\u{C0000}'..='\u{CFFFD}'
+            | '\u{D0000}'..='\u{DFFFD}'
+            | '\u{E0000}'..='\u{EFFFD}'
+            | '!'
+            | '$'
+            | '&'
+            | '\''
+            | '('
+            | ')'
+            | '*'
+            | '+'
+            | ','
+            | ';'
+            | '='
+            | ':'
+            | '@' => c.to_string(),
+            _ => {
+                let s = c.to_string();
+                Uri::percent_encode(&s).to_string()
+            }
+        }
+    }
+}
+
 /**
 * Redirects to the login page with a given message.
 *
@@ -476,6 +529,20 @@ mod tests {
         }
     }
 
+    #[test]
+    fn test_iri_percent_encode_seg() {
+        assert_eq!(
+            &iri_percent_encode_seg("including whitespace"),
+            "including%20whitespace"
+        );
+        assert_eq!(&iri_percent_encode_seg("%20"), "%2520");
+        assert_eq!(&iri_percent_encode_seg("é"), "é");
+        assert_eq!(
+            &iri_percent_encode_seg("空白入り 日本語"),
+            "空白入り%20日本語"
+        );
+    }
+
     #[test]
     fn test_inline() {
         assert_eq!(

plume-models/src/posts.rs

@@ -17,7 +17,7 @@ use plume_common::{
         inbox::{AsActor, AsObject, FromId},
         Hashtag, Id, IntoId, Licensed, Source, PUBLIC_VISIBILITY,
     },
-    utils::md_to_html,
+    utils::{iri_percent_encode_seg, md_to_html},
 };
 use riker::actors::{Publish, Tell};
 use std::collections::{HashMap, HashSet};
@@ -249,7 +249,12 @@ impl Post {
     }
 
     pub fn ap_url(blog: Blog, slug: &str) -> String {
-        ap_url(&format!("{}/~/{}/{}/", CONFIG.base_url, blog.fqn, slug))
+        ap_url(&format!(
+            "{}/~/{}/{}/",
+            CONFIG.base_url,
+            blog.fqn,
+            iri_percent_encode_seg(slug)
+        ))
     }
 
     // It's better to calc slug in insert and update