help request: How to use proxy-pass plugin with regex_uri attributes when the url is encoded

[yujinchoi-94]

Description

Hi, I have to call an upstream server with encoded string.

For example,
I want to make a request like this.

curl -X GET "{apisix-server-address}/rest/alimi/alimianswer/1234/N/%ED%85%8C%EC%8A%A4%ED%8A%B8"

(%ED%85%8C%EC%8A%A4%ED%8A%B8 is encoded string)

I've found out that the apisix uses decoded URI coming from nginx through this
So, i've tried using use_real_request_uri_unsafe option, but it doesn't work with regex_uri.

Is there any way to call the upstream server with encoded string?

---

Here's my configuration.

- name: proxy-rewrite
  enable: true
	config:
		regex_uri:
			- /rest/alimi/alimianswer/(.+)/(.+)/(.+)
			- /cv/v1/emergency/answer?emerNtceSeq=$1&confirmYn=$2&answerCtnt=$3
    use_real_request_uri_unsafe: true

When I used the above configuration, the log result were like this

10.205.164.68 - - [12/Jun/2023:04:42:21 +0000] .... "GET /rest/alimi/alimianswer/1234/N/%ED%85%8C%EC%8A%A4%ED%8A%B8 HTTP/1.1" 400 808 0.077 "-" "curl/7.79.1" - - - "http://{upstream-server}"

What I want is to call the upstream server with below url

http://{upstream-server}/cv/v1/emergency/answer?emerNtceSeq=1234&confirmYn=N&answerCtnt=%ED%85%8C%EC%8A%A4%ED%8A%B8

Environment

• APISIX version (run apisix version):
  • /usr/local/openresty//luajit/bin/luajit /usr/local/apisix/apisix/cli/apisix.lua version 3.2.0
• Operating system (run uname -a):
  • Linux dev-apisix-6d55c4cfd5-nqw9x 5.4.226-129.415.amzn2.x86_64 #1 SMP Fri Dec 9 12:54:21 UTC 2022 x86_64 GNU/Linux
• OpenResty / Nginx version (run openresty -V or nginx -V):
  • openresty -V nginx version: openresty/1.21.4.1
  • nginx -V nginx version: openresty/1.21.4.1
• etcd version, if relevant (run curl http://127.0.0.1:9090/v1/server_info):
• APISIX Dashboard version, if relevant:
• Plugin runner version, for issues related to plugin runners:
• LuaRocks version, for installation issues (run luarocks --version):

[Revolyssup]

@yujinchoi-94 You dont need to use use_real_request_uri_unsafe here since you are passing the uri in regex_uri. You can remove it and it should work

[yujinchoi-94]

@Revolyssup
When i removed use_real_request_uri_unsafe attributes, encoded string is changed like below log.
Encoded string(%ED%85%8C%EC%8A%A4%ED%8A%B8) becomes\xED\x85\x8C\xEC\x8A\xA4\xED\x8A\xB8

Log
10.205.164.68 - - [12/Jun/2023:06:47:11 +0000] ... "GET /rest/alimi/alimianswer/1234/N/%ED%85%8C%EC%8A%A4%ED%8A%B8 HTTP/1.1" 400 808 0.330 "-" "curl/7.79.1" - - - "{upstream-server}/cv/v1/emergency/answer?emerNtceSeq=1234&confirmYn=N&answerCtnt=\xED\x85\x8C\xEC\x8A\xA4\xED\x8A\xB8"

[Revolyssup]

ncoded string is changed like belo

Yep you're right. What you're trying to do is currently not supported because the support for allowing encoded string is only available currently with use_real_request_uri_unsafe and this field only supports proxying to real request.
Basically the issue is that two different concerns real_request and unsafe_uri are cobbed together in one field.
What your use case requires is just unsafe_uri along with a regex field.
Or at least regex_uri should internally have the unsafe field.
I will create a fix for this.

[Sn0rt]

/assign @Sn0rt pls take a look