get_blog_id_from_url() has wrong return

[lonnylot]

If the blog does not exist we sometimes return 0 and sometimes return false. Per the comment we should always be returning an int.

[ryan]

This function needs a lot of help. Its cache is never invalidated and escaped data is used to create cache keys. We need to do something like we did for #20950.

[ryan]

No to mention it is pretty much the same as domain_exists() and get_id_from_blogname().

[ryan]

Down the rabbit hole

[ryan]

• Register blog-id-cache group as global
• Introduce clean_blog_cache() so we can run it independently of refresh_blog_details() which assumes the blog still exists and get_blog_details() can be called.
• Don't db escape cache keys
• Return 0 for all failures in get_blog_id_from_url()
• clean_blog_cache() after dropping tables in wpmu_delete_blog() to make sure the cache is for real cleaned.

[ryan]

Unit tests

[ryan]

​[UT1060] and [22092]