That's not the case anymore. They built an application level filter. For example, Wikipedia block was due to http header reading. https wasn't also a solution because of hand-shaking part. Even TOR can be blocked this way but it's reachable through obfs4proxy.
Besides that, DNS queries to all servers are blocked if you try to get a blocked site's IP. No hijacking attempt here, as far as I can see.
I routed only Wikipedia through TOR with a browser extension. With this setup not all DNS queries goes through proxy. So I had to take advantage of CoreDNS' dns-over-tls support.
Because people who knows basic networking and also read English are less than a thousandth of the country, their method works very well enough.
“DNS queries to all servers are blocked if you try to get a blocked site's IP”
Wait, so if I was in Turkey, a request for wikipedia.org to my ISP DNS server would not only have failed, but caused all my other queries to fail as well? For how long?
I think I couldn't explain well.
When we use ISP's DNS, we get a local IP address and a web-page at that address, explaining who blocked the site and why.
But if I use Google's or Cloudflare's public DNS, the query won't get an answer, falls to timeout.
A funny method to circumvent the block was to visit wikimedia. As long as your connection was kept alive, you could keep surfing wikipedia like it wasn't blocked at all!
This worked well in domestic telecommunications providers. Universities implemented the block themselves, and you were unable to use the workaround stated in those universities.
Besides that, DNS queries to all servers are blocked if you try to get a blocked site's IP. No hijacking attempt here, as far as I can see.
I routed only Wikipedia through TOR with a browser extension. With this setup not all DNS queries goes through proxy. So I had to take advantage of CoreDNS' dns-over-tls support.
Because people who knows basic networking and also read English are less than a thousandth of the country, their method works very well enough.