Files
once-campfire/test/models/opengraph/fetch_test.rb
T
Marcello Costagliola 414ff3376e Put the whole unfurl, lookups included, under one deadline
The deadline in Opengraph::Fetch started after the lookup of the pasted
host, and an unfurl looks up more hosts outside any fetch: the canonical
URL's, and the image's, once to check its content type and again to
validate it. Each of those waits as long as the resolver takes to give
up. Move the deadline up to UnfurlLinksController, around everything the
link leads to; Opengraph::Fetch keeps its per-operation timeouts and
makes no retries.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Bj8KnxpTf9sj2Ysa8aLAVa
2026-10-06 14:03:26 +02:00

119 lines
4.7 KiB
Ruby

require "test_helper"
require "restricted_http/private_network_guard"
class Opengraph::FetchTest < ActiveSupport::TestCase
setup do
@fetch = Opengraph::Fetch.new
@url = URI.parse("https://www.example.com")
end
test "#fetch_document fetches valid HTML" do
WebMock.stub_request(:get, "https://www.example.com/")
.to_return(status: 200, body: "<body>ok<body>", headers: { content_type: "text/html" })
assert_equal "<body>ok<body>", @fetch.fetch_document(@url)
end
test "#fetch_document discards other content types" do
WebMock.stub_request(:get, "https://www.example.com/")
.to_return(status: 200, body: "I'm not HTML!", headers: { content_type: "text/plain" })
assert_nil @fetch.fetch_document(@url)
end
test "#fetch_document follows redirects" do
WebMock.stub_request(:get, "https://www.example.com/")
.to_return(status: 302, headers: { location: "https://www.other.com/" })
WebMock.stub_request(:get, "https://www.other.com/")
.to_return(status: 200, body: "<body>ok<body>", headers: { content_type: "text/html" })
assert_equal "<body>ok<body>", @fetch.fetch_document(@url)
end
test "#fetch_document does not follow redirects to private networks" do
WebMock.stub_request(:get, "https://www.example.com/")
.to_return(status: 302, headers: { location: "https://www.other.com/" })
WebMock.stub_request(:get, "https://www.other.com/")
.to_return(status: 200, body: "<body>ok<body>", headers: { content_type: "text/html" })
Resolv.stubs(:getaddresses).with("www.other.com").returns([ "127.0.0.1" ])
assert_raises RestrictedHTTP::Violation do
@fetch.fetch_document(@url, ip: "1.2.3.4")
end
end
test "#fetch_document resolves hostnames once to avoid DNS rebinding" do
# Allow but interrupt a real connection to demonstrate that we connect
# to a resolved IP, not a hostname to re-resolve.
WebMock.disable_net_connect! allow: [ @url.host ]
Resolv.stubs(:getaddresses).with(@url.host).returns([ "1.2.3.4" ], [ "127.0.0.1" ])
TCPSocket.expects(:open).with { |*args, **| args.first == @url.host }.never
TCPSocket.expects(:open).with { |*args, **| args.first == "1.2.3.4" && args[1] == 443 }.throws(:dns_not_rebound)
assert_throws :dns_not_rebound do
@fetch.fetch_document(@url)
end
end
test "#fetch_document resolves redirect location hostnames once to avoid DNS rebinding" do
# Stub the initial URL to redirect to a DNS-rebound location
WebMock.stub_request(:get, "https://www.other.com/")
.to_return(status: 302, headers: { location: @url.to_s })
# Allow but interrupt a real connection to demonstrate that we connect
# to a resolved IP, not a hostname to re-resolve.
WebMock.disable_net_connect! allow: [ @url.host ]
Resolv.stubs(:getaddresses).with(@url.host).returns([ "1.2.3.4" ], [ "127.0.0.1" ])
TCPSocket.expects(:open).with { |*args, **| args.first == @url.host }.never
TCPSocket.expects(:open).with { |*args, **| args.first == "1.2.3.4" && args[1] == 443 }.throws(:dns_not_rebound)
assert_throws :dns_not_rebound do
@fetch.fetch_document(URI.parse("https://www.other.com/"), ip: "1.2.3.4")
end
end
test "#fetch_document is empty following redirects that never finish" do
WebMock.stub_request(:get, "https://www.example.com/")
.to_return(status: 302, headers: { location: "https://www.example.com/" })
assert_raises Opengraph::Fetch::TooManyRedirectsError do
@fetch.fetch_document(@url)
end
end
test "#fetch_document ignores large responses" do
WebMock.stub_request(:get, "https://www.example.com/")
.to_return(status: 200, body: "too large", headers: { content_length: 1.gigabyte, content_type: "text/html" })
assert_nil @fetch.fetch_document(@url)
end
test "#fetch_document ignores large responses that were missing their content length" do
WebMock.stub_request(:get, "https://www.example.com/")
.to_return(status: 200, body: large_body_content, headers: { content_type: "text/html" })
assert_nil @fetch.fetch_document(@url)
end
test "#fetch_document ignores large responses that were lying about their content length" do
WebMock.stub_request(:get, "https://www.example.com/")
.to_return(status: 200, body: large_body_content, headers: { content_length: 1.megabyte, content_type: "text/html" })
assert_nil @fetch.fetch_document(@url)
end
test "fetch content type" do
WebMock.stub_request(:head, "https://example.com/image.png").to_return(status: 200, headers: { content_type: "image/png" })
url = URI.parse("https://example.com/image.png")
assert_equal "image/png", @fetch.fetch_content_type(url)
end
private
def large_body_content
"x" * (Opengraph::Fetch::MAX_BODY_SIZE + 1)
end
end