BTW, could you add a note to the Rejected Alternatives section that removing the implicit fallback to OAuthBearerUnsecuredLoginCallbackHandler is out of scope for this KIP? That would make it clear to readers that the fallback behavior is unchanged.
Best, Chia-Ping On Thu, Nov 12, 2026 05:24 PM, Chia-Ping Tsai <[email protected]> wrote: > Hi Ming-Yen, > > Could you add the exact log messages to the KIP, i.e. the new INFO message > and the updated WARN message? Otherwise, LGTM. > > Best, > Chia-Ping > > On Fri, Nov 06, 2026 11:56 PM, Ming-Yen Chung <[email protected]> > wrote: > >> Hi Chia-Ping, >> >> chia_02: Sure, I added it to the Motivation section. >> chia_03: I updated the KIP so the new handler will log it at INFO level. >> >> Best, >> Ming-Yen >> >> Chia-Ping Tsai <[email protected]> 於 2026年11月5日週四 下午11:02寫道: >> >> > Hi Ming-Yen, >> > >> > chia_02: >> > Would you mind highlighting in the Motivation section that the unsecured >> > classes are not production-ready? I think readers need to be aware of >> that. >> > >> > chia_03: >> > The new OAuthBearerTokenlessLoginCallbackHandler will be released as >> > production-ready, so we need to think about how to avoid misuse. For >> > example, we could add a log when users configure this handler, such as >> > "Login succeeded without a token". >> > >> > Best, >> > Chia-Ping >> > >> > On Mon, Nov 02, 2026 10:26 PM, Ming-Yen Chung <[email protected]> >> > wrote: >> > >> > > Hi Chia-Ping, >> > > >> > > Thanks for the review. >> > > >> > > chia_00: You're right, I was overthinking it. I removed the check from >> > the >> > > KIP, so the new handler will ignore all JAAS options. >> > > >> > > chia_01: That makes sense. I updated the KIP so the log will be WARN >> and >> > > suggest OAuthBearerTokenlessLoginCallbackHandler. >> > > >> > > Best, >> > > Ming-Yen >> > > >> > > Chia-Ping Tsai via dev <[email protected]> 於 2026年11月2日週一 >> 上午10:43寫道: >> > > >> > > > Hi Ming-Yen >> > > > >> > > > chia_00: >> > > > Why does the new handler need to throw an exception when it sees the >> > > > unsecuredLogin* configs? Those belong to another handler, so I don’t >> > get >> > > > the point of checking them here. >> > > > >> > > > chia_01: >> > > > Should OAuthBearerUnsecuredLoginCallbackHandler switch to use warn >> > level >> > > > instead of info level? >> > > > ```java >> > > > log.info("Token not provided since n ... >> > > > ``` >> > > > >> > > > Best, >> > > > Chia-Ping >> > > > >> > > > On 2026-11-02 02:21 UTC Ming-Yen Chung wrote: >> > > > > Hi all, >> > > > > >> > > > > I would like to start a discussion on >> > > > > KIP-1390: Add a tokenless login callback handler for >> SASL/OAUTHBEARER >> > > > > <https://cwiki.apache.org/confluence/x/qAKWGw> >> > > > > >> > > > > A broker that only validates OAuth tokens does not need its own >> > token, >> > > > but >> > > > > today it relies on the internal OAuthBearerUnsecuredLoginCallb >> > > ackHandler >> > > > to >> > > > > log in without one. This KIP proposes a new >> > > > > OAuthBearerTokenlessLoginCallbackHandler to support tokenless >> login >> > > > > officially. >> > > > > >> > > > > Best, >> > > > > Ming-Yen >> > > > > >> > > > >> > > >> > >> >
