From c3c4f5a2a70b737500105206286f54696d643175 Mon Sep 17 00:00:00 2001
From: evazion <noizave@gmail.com>
Date: Fri, 14 Jan 2022 22:42:19 -0600
Subject: [PATCH] Fix #4957: Autotag non-web_source.

Autotag non-web_source on posts that have a non-http:// or https:// URL.
Add a fix script to backfill old posts.

Syntactically invalid URLs are still considered web sources. For
example, `https://google,com` technically isn't a valid URL, but it's
not considered a non-web source.
---
 app/models/post.rb                         |  6 +++++-
 script/fixes/089_add_non-web_source_tag.rb | 25 ++++++++++++++++++++++
 test/unit/post_test.rb                     | 24 +++++++++++++++++++++
 3 files changed, 54 insertions(+), 1 deletion(-)
 create mode 100755 script/fixes/089_add_non-web_source_tag.rb

diff --git a/app/models/post.rb b/app/models/post.rb
index 3350988d0..ede050c04 100644
--- a/app/models/post.rb
+++ b/app/models/post.rb
@@ -404,7 +404,7 @@ class Post < ApplicationRecord
     end
 
     def add_automatic_tags(tags)
-      tags -= %w[incredibly_absurdres absurdres highres lowres flash video ugoira animated_gif animated_png exif_rotation non-repeating_animation]
+      tags -= %w[incredibly_absurdres absurdres highres lowres flash video ugoira animated_gif animated_png exif_rotation non-repeating_animation non-web_source]
 
       if tags.size >= 30
         tags -= ["tagme"]
@@ -443,6 +443,10 @@ class Post < ApplicationRecord
         tags << "ugoira"
       end
 
+      if source.present? && source !~ %r{\Ahttps?://}i
+        tags << "non-web_source"
+      end
+
       # Allow only Flash files to be manually tagged as `animated`; GIFs, PNGs, videos, and ugoiras are automatically tagged.
       tags -= ["animated"] unless is_flash?
       tags << "animated" if media_asset.is_animated?
diff --git a/script/fixes/089_add_non-web_source_tag.rb b/script/fixes/089_add_non-web_source_tag.rb
new file mode 100755
index 000000000..d0b46bcf0
--- /dev/null
+++ b/script/fixes/089_add_non-web_source_tag.rb
@@ -0,0 +1,25 @@
+#!/usr/bin/env ruby
+
+require_relative "base"
+
+CurrentUser.scoped(User.system) do
+  Post.system_tag_match("-source:none -source:http://* -source:https://* -non-web_source").find_each do |post|
+    puts "post ##{post.id}: adding non-web_source"
+    post.save!
+  end
+
+  Post.system_tag_match("source:none non-web_source").find_each do |post|
+    puts "post ##{post.id}: removing non-web_source"
+    post.save!
+  end
+
+  Post.system_tag_match("source:http://* non-web_source").find_each do |post|
+    puts "post ##{post.id}: removing non-web_source"
+    post.save!
+  end
+
+  Post.system_tag_match("source:https://* non-web_source").find_each do |post|
+    puts "post ##{post.id}: removing non-web_source"
+    post.save!
+  end
+end
diff --git a/test/unit/post_test.rb b/test/unit/post_test.rb
index cd5b2c8af..a34ae6637 100644
--- a/test/unit/post_test.rb
+++ b/test/unit/post_test.rb
@@ -1046,6 +1046,30 @@ class PostTest < ActiveSupport::TestCase
         end
       end
 
+      context "a post with a non-web source" do
+        should "automatically add the non-web_source tag" do
+          @post.update!(source: "this was once revealed to me in a dream")
+          @post.save!
+          assert_equal("non-web_source tag1 tag2", @post.tag_string)
+        end
+      end
+
+      context "a post with a blank source" do
+        should "remove the non-web_source tag" do
+          @post.update!(source: "", tag_string: "non-web_source")
+          @post.save!
+          assert_equal("tagme", @post.tag_string)
+        end
+      end
+
+      context "a post with a https:// source" do
+        should "remove the non-web_source tag" do
+          @post.update!(source: "https://www.google.com", tag_string: "non-web_source")
+          @post.save!
+          assert_equal("tagme", @post.tag_string)
+        end
+      end
+
       should "have an array representation of its tags" do
         post = FactoryBot.create(:post)
         post.reload