From c4600411f73138489e3c3870de1693a297f4fe46 Mon Sep 17 00:00:00 2001
From: multiple creatures <dev@multiple-creature.party>
Date: Sat, 20 Jul 2019 23:19:26 -0500
Subject: [PATCH] fix anchor tagger filename matching

---
 app/lib/sanitize_config.rb | 34 ++++++++++++++++------------------
 1 file changed, 16 insertions(+), 18 deletions(-)

diff --git a/app/lib/sanitize_config.rb b/app/lib/sanitize_config.rb
index cd2299e8b..697d88fd3 100644
--- a/app/lib/sanitize_config.rb
+++ b/app/lib/sanitize_config.rb
@@ -43,30 +43,28 @@ class Sanitize
       text = node.text.strip
       return if href == text
 
-      # try to detect filenames
-      href_filename = href.rpartition('/')[-1]
-      url_filename = text.rpartition('/')[-1]
-
-      unless href_filename.blank?
-        if url_filename == href_filename
-          node.inner_html = "\xf0\x9f\x93\x8e #{node.inner_html}"
-          return
-        end
-
-        # many fedi servers obfuscate media filenames
-        ext = url_filename.rpartition('.')[-1]
-        if ext.downcase.in?(MEDIA_EXTENSIONS) && ext == href_filename.rpartition('.')[-1]
-          node.inner_html = "\xf0\x9f\x93\x8e #{node.inner_html}"
-          return
-        end
-      end
-
       # strip ellipse & replace keyword search obscuring
       text = text.sub(/ *(?:\u2026|\.\.\.)\Z/, '').gsub(/ dot /i, '.').gsub(/[\u200b-\u200d\ufeff\u200e\u200f]/, '')
 
       # href now matches text without obscuring?
       return if href == text
 
+      # try to detect filenames
+      href_filename = '/'.in?(href) ? href.rpartition('/')[2] : nil
+      unless href_filename.blank?
+        if text == href_filename
+          node.inner_html = "\xf0\x9f\x93\x8e #{node.inner_html}"
+          return
+        end
+
+        # many fedi servers obfuscate media filenames
+        ext = text.rpartition('.')[-1]
+        if ext.downcase.in?(MEDIA_EXTENSIONS) && ext == href_filename.rpartition('.')[2]
+          node.inner_html = "\xf0\x9f\x93\x8e #{node.inner_html}"
+          return
+        end
+      end
+
       # grab first url from link text
       first_url = text.scan(/[\w\-]+\.[\w\-]+(?:\.[\w\-]+)*\S*/).first