Fix npub mention bugs, fix slowness when parsing large posts

Switch the post parser to use the same code as the content parser. This
was causing many issues, including performance issues.

Changelog-Fixed: Fix lag when creating large posts
Changelog-Fixed: Fix npub mentions failing to parse in some cases
Changelog-Added: Add r tag when mentioning a url
Changelog-Removed: Remove old @ and & hex key mentions

damusTests/HashtagTests.swift

@@ -39,15 +39,37 @@ final class HashtagTests: XCTestCase {
     }
     
     func testHashtagWithEmoji() {
-        let parsed = parse_mentions(content: "some hashtag #bitcoin☕️ cool", tags: []).blocks
-        
+        let content = "some hashtag #bitcoin☕️ cool"
+        let parsed = parse_mentions(content: content, tags: []).blocks
+        let post_blocks = parse_post_blocks(content: content)
+
         XCTAssertNotNil(parsed)
         XCTAssertEqual(parsed.count, 3)
         XCTAssertEqual(parsed[0].is_text, "some hashtag ")
         XCTAssertEqual(parsed[1].is_hashtag, "bitcoin☕️")
         XCTAssertEqual(parsed[2].is_text, " cool")
+
+        XCTAssertEqual(post_blocks.count, 3)
+        XCTAssertEqual(post_blocks[0].is_text, "some hashtag ")
+        XCTAssertEqual(post_blocks[1].is_hashtag, "bitcoin☕️")
+        XCTAssertEqual(post_blocks[2].is_text, " cool")
     }
-    
+
+    func testPowHashtag() {
+        let content = "pow! #ぽわ〜"
+        let parsed = parse_mentions(content: content, tags: []).blocks
+        let post_blocks = parse_post_blocks(content: content)
+
+        XCTAssertNotNil(parsed)
+        XCTAssertEqual(parsed.count, 2)
+        XCTAssertEqual(parsed[0].is_text, "pow! ")
+        XCTAssertEqual(parsed[1].is_hashtag, "ぽわ〜")
+
+        XCTAssertEqual(post_blocks.count, 2)
+        XCTAssertEqual(post_blocks[0].is_text, "pow! ")
+        XCTAssertEqual(post_blocks[1].is_hashtag, "ぽわ〜")
+    }
+
     func testHashtagWithAccents() {
         let parsed = parse_mentions(content: "hello from #türkiye", tags: []).blocks