Merge branch 'feat/rich-media-head' into 'develop'
RichMedia: Do a HEAD request to check content type/length See merge request pleroma/pleroma!2995
This commit is contained in:
commit
4d543fcb75
4 changed files with 99 additions and 1 deletions
|
@ -87,6 +87,50 @@ defmodule Pleroma.Web.RichMedia.Helpers do
|
|||
def rich_media_get(url) do
|
||||
headers = [{"user-agent", Pleroma.Application.user_agent() <> "; Bot"}]
|
||||
|
||||
Pleroma.HTTP.get(url, headers, @options)
|
||||
head_check =
|
||||
case Pleroma.HTTP.head(url, headers, @options) do
|
||||
# If the HEAD request didn't reach the server for whatever reason,
|
||||
# we assume the GET that comes right after won't either
|
||||
{:error, _} = e ->
|
||||
e
|
||||
|
||||
{:ok, %Tesla.Env{status: 200, headers: headers}} ->
|
||||
with :ok <- check_content_type(headers),
|
||||
:ok <- check_content_length(headers),
|
||||
do: :ok
|
||||
|
||||
_ ->
|
||||
:ok
|
||||
end
|
||||
|
||||
with :ok <- head_check, do: Pleroma.HTTP.get(url, headers, @options)
|
||||
end
|
||||
|
||||
defp check_content_type(headers) do
|
||||
case List.keyfind(headers, "content-type", 0) do
|
||||
{_, content_type} ->
|
||||
case Plug.Conn.Utils.media_type(content_type) do
|
||||
{:ok, "text", "html", _} -> :ok
|
||||
_ -> {:error, {:content_type, content_type}}
|
||||
end
|
||||
|
||||
_ ->
|
||||
:ok
|
||||
end
|
||||
end
|
||||
|
||||
@max_body @options[:max_body]
|
||||
defp check_content_length(headers) do
|
||||
case List.keyfind(headers, "content-length", 0) do
|
||||
{_, maybe_content_length} ->
|
||||
case Integer.parse(maybe_content_length) do
|
||||
{content_length, ""} when content_length <= @max_body -> :ok
|
||||
{_, ""} -> {:error, :body_too_large}
|
||||
_ -> :ok
|
||||
end
|
||||
|
||||
_ ->
|
||||
:ok
|
||||
end
|
||||
end
|
||||
end
|
||||
|
|
|
@ -36,6 +36,14 @@ defmodule Pleroma.Web.RichMedia.Parser do
|
|||
{:ok, _data} = res ->
|
||||
res
|
||||
|
||||
{:error, :body_too_large} = e ->
|
||||
e
|
||||
|
||||
{:error, {:content_type, _}} ->
|
||||
e
|
||||
|
||||
# The TTL is not set for the errors above, since they are unlikely to change
|
||||
# with time
|
||||
{:error, _} = e ->
|
||||
ttl = Pleroma.Config.get([:rich_media, :failure_backoff], 60_000)
|
||||
Cachex.expire(:rich_media_cache, url, ttl)
|
||||
|
|
|
@ -1262,4 +1262,21 @@ defmodule HttpRequestMock do
|
|||
inspect(headers)
|
||||
}"}
|
||||
end
|
||||
|
||||
# Most of the rich media mocks are missing HEAD requests, so we just return 404.
|
||||
@rich_media_mocks [
|
||||
"https://example.com/ogp",
|
||||
"https://example.com/ogp-missing-data",
|
||||
"https://example.com/twitter-card"
|
||||
]
|
||||
def head(url, _query, _body, _headers) when url in @rich_media_mocks do
|
||||
{:ok, %Tesla.Env{status: 404, body: ""}}
|
||||
end
|
||||
|
||||
def head(url, query, body, headers) do
|
||||
{:error,
|
||||
"Mock response not implemented for HEAD #{inspect(url)}, #{query}, #{inspect(body)}, #{
|
||||
inspect(headers)
|
||||
}"}
|
||||
end
|
||||
end
|
||||
|
|
|
@ -56,6 +56,27 @@ defmodule Pleroma.Web.RichMedia.ParserTest do
|
|||
|
||||
%{method: :get, url: "http://example.com/error"} ->
|
||||
{:error, :overload}
|
||||
|
||||
%{
|
||||
method: :head,
|
||||
url: "http://example.com/huge-page"
|
||||
} ->
|
||||
%Tesla.Env{
|
||||
status: 200,
|
||||
headers: [{"content-length", "2000001"}, {"content-type", "text/html"}]
|
||||
}
|
||||
|
||||
%{
|
||||
method: :head,
|
||||
url: "http://example.com/pdf-file"
|
||||
} ->
|
||||
%Tesla.Env{
|
||||
status: 200,
|
||||
headers: [{"content-length", "1000000"}, {"content-type", "application/pdf"}]
|
||||
}
|
||||
|
||||
%{method: :head} ->
|
||||
%Tesla.Env{status: 404, body: "", headers: []}
|
||||
end)
|
||||
|
||||
:ok
|
||||
|
@ -144,4 +165,12 @@ defmodule Pleroma.Web.RichMedia.ParserTest do
|
|||
test "returns error if getting page was not successful" do
|
||||
assert {:error, :overload} = Parser.parse("http://example.com/error")
|
||||
end
|
||||
|
||||
test "does a HEAD request to check if the body is too large" do
|
||||
assert {:error, :body_too_large} = Parser.parse("http://example.com/huge-page")
|
||||
end
|
||||
|
||||
test "does a HEAD request to check if the body is html" do
|
||||
assert {:error, {:content_type, _}} = Parser.parse("http://example.com/pdf-file")
|
||||
end
|
||||
end
|
||||
|
|
Loading…
Reference in a new issue