Files
ProcessControl/conv/04-part-IV.md
T

2.5 KiB
Raw Blame History

Conversation Log — Part IV — Rewind · Steps 28–34 (turns 62)

⬅ Previous: Part III · Index · Next: Part V ➡

All parts: Setup & Step 0 · Part I · Part II · Part III · Part IV · Part V · Part VI · Part VII · Parts VIII & IX


[62] Student

so after test, after a ctrl+d, the console in is still working, but nothing is output anymore, i'd say that the subscription to topic which is interupt, explaining why nothing is output because consoleout don't receive anything from rot13 and rot13 don't receive anything from consolein

[62] Tutor — actions taken

  • Read 04-dynamic-in-jvm/DynamicQueueRot13.java lines 37–52 (runConsoleIn) and 120–140 (main) to check the mechanism behind the observation.
  • Earlier in Part IV: corrected who publishes TOPIC (it is main(), lines 122–133, not ConsoleIn) and what the dynamic version changes (ROT13 takes straight from consoleIn.outbox; rot13.inbox now carries control messages only).

[62] Tutor — reply

Corrected the mechanism: nothing is interrupted — runConsoleIn returns on EOF, so the producer is gone while both subscriber threads are still blocked in take() on sources that will never produce again. main's Thread.currentThread().join() is what keeps the JVM alive. Also answered the second half: with no mediator, only the three capabilities are left in the data path, plus a startup-only dependency on the configurator publishing TOPIC. Then set Steps 31–34 as the next reading.

Progress record

  • Current step: 30 done → 31–34 (Part IV)
  • Last observed result: dynamic in-JVM version runs; after Ctrl+D the producer thread has returned and the two subscribers wait forever
  • Unresolved problem: none
  • Next action: student reads Steps 31–34 and answers the milestone question (who knows whom, where the topology lives, what drives the interaction)

⬅ Previous: Part III · Index · Next: Part V ➡

All parts: Setup & Step 0 · Part I · Part II · Part III · Part IV · Part V · Part VI · Part VII · Parts VIII & IX