Fix up to a 1s delay in first produce to new topic

[milindl]

Currently, when creating a new rkt via the public api, rd_kafka_topic_new(), causes us to immediately make a leader query if the topic is unknown.

On the other hand, using the internal rd_kafka_topic_new0() does not do this. producev and produceva both call rd_kafka_topic_new0 in case topic is provided by name. So this delays produce until scan_all is run, which runs on a 1s timer.

Definitely affects JS client and I think it affects python client and dotnet client as well.

[confluent-cla-assistant]

🎉 All Contributor License Agreements have been signed. Ready to merge.
_{Please push an empty commit if you would like to re-run the checks to verify CLA status for all contributors.}

[Copilot]

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

Comments suppressed due to low confidence (1)

tests/0055-producer_latency.c:496

• [nitpick] The variable 'wait_time_mss' appears ambiguous; consider renaming it to 'wait_times_ms' for clarity.

int wait_time_mss[] = {500, 1200, 2500, 3700, 4900};

[Copilot]

Copilot reviewed 4 out of 4 changed files in this pull request and generated no comments.

Comments suppressed due to low confidence (1)

tests/0055-producer_latency.c:500

• [nitpick] The variable name 'wait_time_mss' appears to contain a typo; consider renaming it to 'wait_time_ms' for clarity.

int wait_time_mss[] = {500, 1200, 2500, 3700, 4900};

[emasab]

Thanks for the fix Milind, sent some improvements

[emasab]

let's repeat this 50 times instead of changing the sleep. Every time with a different topic name. this way we can also evaluate the number of metadata requests done

[emasab]

Thanks Milind! Approving