Destroy Rialto pipeline in Stop() instead of Configure()

[anshephe]

AampRialtoPlayer::Stop() now calls m_pipeline.reset() right after stop(), releasing the pipeline immediately rather than leaving it alive until the next Configure() overwrites it. ShouldRecreatePipeline() uses !m_pipeline instead of the now-redundant m_pipelineStopped flag.

AampRialtoMediaSource::respondAbandonedRequest() treats a null pipeline as an expected post-Stop() condition (dropped with an INFO log) instead of an error, since a stale needData can legitimately arrive after Stop() destroys the pipeline.

Updated AampRialtoPlayerTestCases.cpp: added Stop_DestroysPipeline, and fixed tests that asserted on m_mockPipelinePtr after Stop() (now a dangling pointer) to instead verify no crash on late needData.

[Copilot]

Pull request overview

This PR updates the Direct-Rialto playback teardown semantics so AampRialtoPlayer::Stop() immediately destroys the Rialto pipeline (instead of deferring destruction until the next Configure()), and adjusts post-Stop handling to treat late/stale needData notifications as an expected condition.

Changes:

• Destroy m_pipeline in AampRialtoPlayer::Stop() and simplify pipeline recreation logic to key off !m_pipeline (removing m_pipelineStopped).
• Treat a null pipeline in AampRialtoMediaSource::respondAbandonedRequest() as an expected post-Stop state (log INFO and drop).
• Update/add L1 tests to reflect immediate pipeline teardown and avoid dereferencing a now-dangling mock pipeline pointer after Stop().

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
test/utests/tests/AampRialtoPlayerTests/AampRialtoPlayerTestCases.cpp	Adds/updates Stop-related tests for immediate pipeline destruction and safe handling of late needData.
direct-rialto/AampRialtoPlayer.h	Removes the m_pipelineStopped flag now made redundant by immediate pipeline reset.
direct-rialto/AampRialtoPlayer.cpp	Resets m_pipeline in Stop() and updates ShouldRecreatePipeline() to recreate when !m_pipeline.
direct-rialto/AampRialtoMediaSource.h	Documents new behavior when pipeline is null during abandoned request response.
direct-rialto/AampRialtoMediaSource.cpp	Logs INFO (not ERR) and drops abandoned requests when the pipeline is already gone.

Suppressed comments (2)

test/utests/tests/AampRialtoPlayerTests/AampRialtoPlayerTestCases.cpp:1213

• This test no longer verifies that EOS flags are reset (it now only checks that a stale needData call after Stop() does not crash). Rename the test so it accurately reflects what is being validated.

	// Stop() destroys the pipeline (m_mockPipelinePtr is now a dangling
	// pointer - no EXPECT_CALL can be set against it), so it must not
	// resurrect the EOS state and re-fire haveData(EOS, ...); the stale
	// needData arriving after Stop() has nothing left to respond to and
	// must be dropped safely rather than crashing.

test/utests/tests/AampRialtoPlayerTests/AampRialtoPlayerTestCases.cpp:1217

• Same as above: please document why this sleep_for() is required for the test’s oracle (negative assertion / allow async processing). If it’s not required, consider removing it to reduce flakiness.

	std::this_thread::sleep_for(std::chrono::milliseconds(20));