Fix the tag span generation for tags with nonascii characters (#2700)

* Update mention and tag regexes from mastodon

* Normalize nonascii tag names the same way that mastodon does

app/src/test/java/com/keylesspalace/tusky/SpanUtilsTest.kt

@@ -38,6 +38,7 @@ class SpanUtilsTest {
                 return listOf(
                     "@mention",
                     "#tag",
+                    "#tåg",
                     "https://thr.ee/meh?foo=bar&wat=@at#hmm",
                     "http://thr.ee/meh?foo=bar&wat=@at#hmm"
                 )

app/src/test/java/com/keylesspalace/tusky/util/LinkHelperTest.kt

@@ -86,6 +86,17 @@ class LinkHelperTest {
         }
     }
 
+    @Test
+    fun whenCheckingTags_tagNameIsNormalized() {
+        val mutator = "aeiou".toList().zip("åÉîøÜ".toList()).toMap()
+        for (tag in tags) {
+            val mutatedTagName = String(tag.name.map { mutator[it] ?: it }.toCharArray())
+            val tagName = getTagName("#$mutatedTagName", tags)
+            Assert.assertNotNull(tagName)
+            Assert.assertNotNull(tags.firstOrNull { it.name == tagName })
+        }
+    }
+
     @Test
     fun hashedUrlSpans_withNoMatchingTag_areNotModified() {
         for (tag in tags) {