allthingssecurity opened a new pull request, #30248:
URL: https://github.com/apache/camel/pull/30248

   # Description
   
   [CAMEL-27042](https://issues.apache.org/jira/browse/CAMEL-27042)
   
   Found while checking the calls of the consumer against Paho 1.2.5 for the 
Paho DST world of #30138 (CAMEL-26958). Reproduced here with a plain unit test.
   
   The consumer subscribes on start (`doStart`), after an automatic reconnect 
of the client (`resubscribe`, CAMEL-24511), and after its own reconnect when 
the broker did not keep the session (`reconnect`). It never looked at the 
result of `MqttClient.subscribe`. In MQTT 5 a broker that refuses a 
subscription still answers with a SUBACK, with a reason code of 0x80 or higher: 
0x80 Unspecified error, 0x83 Implementation specific error, 0x87 Not 
authorized, 0x8F Topic Filter invalid, 0x97 Quota exceeded, 0x9E, 0xA1, 0xA2 
(MQTT 5.0, 3.9.3; 2.4 says a reason code of 0x80 or greater indicates failure). 
Paho 1.2.5 completes the token of the SUBSCRIBE with those reason codes and no 
exception (javap: `ClientState.notifyReceivedAck` raises an exception only for 
a failed PUBREC, `Token.markComplete` copies the reason codes of the SUBACK to 
the token, and `MqttClient.subscribe` returns the token once it is complete). 
The Paho MQTT 3 client does throw (`MqttClient.subscribe` throws `MqttException(
 128)` for a granted QoS of 0x80), so camel-paho gets this check from Paho, and 
camel-paho-mqtt5 did not have it.
   
   So a route whose subscription is refused, for example by an ACL of the 
broker, started and stayed Started without ever receiving a message, and after 
a reconnect the resubscribe of CAMEL-24511 counted a refused subscription as 
done.
   
   Change: the consumer subscribes through one method that throws 
`MqttException` with the reason code when the SUBACK has a reason code of 0x80 
or higher (the message comes from Paho's catalog, for example `Not 
authorized.`), used in the three places. So a refused subscription is handled 
as a subscribe that throws:
   
   - on start the route does not start, and the own client is disconnected and 
closed (as before for a failed subscribe);
   - after an automatic reconnect, the route is restarted for an own client, 
and an error is logged for a client given with the `client` option 
(CAMEL-24511);
   - after the consumer's own reconnect (a failed message with 
`manualAcksEnabled` on a kept session, CAMEL-25793), when the broker did not 
keep the session, the route is restarted too. That connect created a session 
the broker keeps, so a later reconnect would find it and not subscribe, and the 
consumer would stay connected without its subscription. This also applies when 
that subscribe fails because the connection is lost; before, the failure went 
to the exception handler and the next reconnect found the session and did not 
subscribe.
   
   A granted QoS lower than the requested one (0x00 or 0x01 for a QoS 2 
request) is not a failure and is left as is. A token without reason codes is 
taken as granted.
   
   Tests, new `PahoMqtt5SubscribeRefusedTest` (a mocked `MqttClient` whose 
`subscribe` returns a token with the reason codes, as Paho does):
   
   - `routeDoesNotStartWhenTheBrokerRefusesTheSubscription` (0x87): on main the 
route started.
   - `refusedSubscriptionOnReconnectStopsTheRoute`: the subscription is granted 
on start and refused after `connectComplete(reconnect=true)`; the test waits 
until the restart closed the client twice and subscribed three times, then 
checks Stopped. On main the route stayed Started.
   - `refusedSubscriptionAfterTheReconnectOfAFailedMessageStopsTheRoute`: 
`manualAcksEnabled`, `cleanStart=false`, `sessionExpiryInterval=60`, a message 
that fails, and a reconnect that finds no session and gets a refused 
subscription; on main the route stayed Started.
   - `routeStartsWhenTheBrokerGrantsTheSubscription` (reason code 2): the 
control.
   
   Without the change the three refused cases fail; with it all four pass, and 
the camel-paho-mqtt5 unit tests pass (79 run, 0 failures, 0 errors, 0 skipped). 
The ITs need Docker and were not run locally (CI runs them). I built offline 
with the versions cached on my machine (Caffeine 3.3.0, commons-lang3 3.21.0, 
Guava 33.7.2); Paho 1.2.5 is the version of main.
   
   #30138 is merged, and the fake of its world answers only the token methods 
it models and fails a seed on any other one, so this PR carries a test-only 
change to it (`FakeMqtt`): it answers `getReasonCodes` (the granted QoS; the 
fake broker refuses no subscription), and it also models `disconnectForcibly`, 
which the CAMEL-27041 fix calls (with the two corrections from the review of 
#30138: a forced disconnect during an automatic reconnect reschedules the 
cycle, and an older connect does not clear the connecting flag of a newer one). 
Without it every seed of the world's default set fails once the consumer reads 
the reason codes (seed 1: "not modelled: IMqttToken.getReasonCodes"). With it, 
on main the world passes 1000 of 1000 seeds and the camel-paho-mqtt5 unit tests 
pass (1079 run including the seeds, 0 failures).
   
   The three camel-paho-mqtt5 fixes of this batch (CAMEL-27040, CAMEL-27041, 
CAMEL-27042) all change `PahoMqtt5Consumer`, in different methods. They merge 
cleanly in either order, and merged together on main (with the fake change of 
the CAMEL-27042 PR) the camel-paho-mqtt5 unit tests and the world pass (1089 
run including 1000 seeds, 0 failures).
   
   Upgrade guide: a 4.24 entry, "camel-paho-mqtt5 - a subscription the broker 
refuses fails the consumer", as a route that started before now fails to start 
when the broker refuses its subscription.
   
   # Target
   
   - [x] I checked that the commit is targeting the correct branch (Camel 4 
uses the `main` branch)
   
   # Tracking
   - [x] If this is a large change, bug fix, or code improvement, I checked 
there is a [JIRA issue](https://issues.apache.org/jira/browse/CAMEL) filed for 
the change (usually before you start working on it): 
[CAMEL-27042](https://issues.apache.org/jira/browse/CAMEL-27042).
   
   # Apache Camel coding standards and style
   
   - [x] I checked that each commit in the pull request has a meaningful 
subject line and body.
   - [ ] I have run `mvn clean install -DskipTests` locally from root folder 
and I have committed all auto-generated changes.
     (I built camel-paho-mqtt5 and the core modules with the formatter and 
import-sort plugins and ran the camel-paho-mqtt5 unit tests. I did not run the 
full root build.)
   
   # AI-assisted contributions
   
   - [x] If this PR includes AI-generated code, commits have proper 
co-authorship attribution (e.g., `Co-authored-by` trailers) and the PR 
description identifies the AI tool used.
     This PR was prepared with Claude Code (Claude Opus 5.5). The commit 
carries a `Co-Authored-By` trailer.
   
   _Claude Code on behalf of allthingssecurity_
   
   🤖 Generated with [Claude Code](https://claude.com/claude-code)
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to