[regext] Re: RDAP responses for invalid domain names

[Jasdip Singh <[e-mail]>]

Hi Paweł,

Interesting observations. :)

"Is 404 the right behaviour?”

Does not seem. 404 (Not Found) should be for a non-extant resource if the query is deemed well-formed syntax- and semantics-wise. Examples 1 and 3 should return 400 (Bad Request) assuming these are considered syntax errors per section 5.4 of RFC 9083. As for example 2, if it is deemed a syntax error by that same section, then 400; otherwise, 422 (Unprocessable Content) since mixing  jp characters with .de doesn’t make sense semantically. Example 2 seems tricky. :)

Looks like it should be 400 if syntax error in query value, and 422 if semantic error. That said, 400 could cover 422 scenarios but not vice-versa, AFAIU.

Thanks,
Jasdip

From: Pawel Kowalik <kowalik=[e-mail]>
Date: Friday, April 10, 2026 at 10:20 AM
To: [e-mail] <[e-mail]>
Subject: [regext] RDAP responses for invalid domain names

Hi,

rfc7480 Section 5.3 specifies 404 as an answer for an non existing
object (for example domain name).

Section 5.4 specifies 400 code for malformed queries.

How about queries which are not malformed on the RDAP protocol level but
invalid objects (generally or registry-specific)?

Example1: xn–fdkp0gtc.com - wrongly encoded IDN domain

Example2: ロバート.de, where de does not support jp characters

Example3: 1234567890123456789012345678901234567890123456789012345678901234567890.com
- label too long

My observation is that RDAP servers would respond with 404 in all of
those cases, but the standard would not specify it clearly whether it is
correct or not.

rfc7482 only suggests 400 as a response code, if UTF-8 decoding of query
strings would fail, which is in line with rfc7480 Section 5.4, but does
not cover the other cases.

Did I miss something? Is 404 the right behaviour?

Kind regards,
Pawel