Skip to content

close-range: remove test_close_range_threshold (RSDEV-13516) - #15522

Merged
Nir-Az merged 1 commit into
realsenseai:developmentfrom
ymodlin:RSDEV-13516-skip-close-range-threshold
Aug 9, 2026
Merged

Nir-Az merged 1 commit into
realsenseai:developmentfrom
ymodlin:RSDEV-13516-skip-close-range-threshold

Conversation

@ymodlin

@ymodlin ymodlin commented Aug 9, 2026 •

Copy link
Copy Markdown
Contributor

Removes test_close_range_threshold (D555/D585, unit-tests/live/d500/pytest-close-range.py) and its now-unused THRESHOLD_DEFAULT constant. Tracked by RSDEV-13516.

Why

The test asserted that the threshold embedded-filter option reports the spec default 550.0:

check.equal( embedded_filter.get_option( rs.option.threshold ), THRESHOLD_DEFAULT )  # 550.0

FW reports a different, per-unit value instead:

Linux   [D555-343622300739]:  FAILURE: check 491.0 == 550.0
Windows [D555-344522301005]:  FAILURE: check 487.0 == 550.0

This blocked both the Windows and Linux legs of LRS_libci_pipeline nightly (#17548, #17552, #17553) — it was the only failing test in an otherwise green pipeline.

Evidence it is not flaky, and not an SDK regression

  • Both rigs retried after a full device power-cycle (UniFi port down/up, device re-enumerated) and read back the identical value. Deterministic.
  • Two different D555 units report different-but-close values (487.0 / 491.0) — consistent with a calibration-derived per-unit default rather than a fixed constant.
  • On the DDS path the SDK relays the FW value verbatim; there is no host-side conversion involved.
  • The earlier "passing" nightlies (#17549–#17551) skipped the whole file — those cameras carried FW 7.58.38429.9662, below the MIN_FW_BY_DEVICE gate (7.58.39807.10573). There is no prior passing baseline for this assertion; it is day-1 exposure following RSDEV-12321 ("D555/D585 – MinZ Control (DDS/USB)"), not a regression.

Open question (owned by RSDEV-13516)

Is 550.0 a universal spec default (FW bug), or is the default calibration-derived (in which case the assertion was wrong to begin with)? Until that is decided the expected value cannot be re-baselined, so the test is removed rather than re-pointed at 487/491.

Scope note

The whole test is removed, which also drops the _set_get_round_trip( threshold, 600.0 ) coverage that currently passes — set/get on the option itself is working. When RSDEV-13516 is resolved, the test should be reintroduced with the correct expectation for the default value (or with the default-value assertion dropped and only the round-trip kept). A comment recording this is on the ticket.

THRESHOLD_DEFAULT had no other reference in the repo, so it is removed with the test.

Testing

  • python3 -m py_compile clean.
  • No hardware run — the change only deletes a test.

@sysrsbuild-gh-agentic sysrsbuild-gh-agentic left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Auto-generated by rs-agentic bot

Enhancement: Both log.warning and pytest.skip label this as "FW bug", but the PR description notes the root cause is still open — it may be a wrong constant in the test rather than a FW defect. Consider a neutral wording, e.g.:

log.warning( "Skipping test_close_range_threshold: threshold default mismatch, see RSDEV-13516" )
pytest.skip( "threshold default mismatch tracked by RSDEV-13516" )

This avoids prematurely attributing blame before RSDEV-13516 is resolved.

Signed-off-by: Yogev Modlin <yogev.modlin@realsenseai.com>
@ymodlin
ymodlin force-pushed the RSDEV-13516-skip-close-range-threshold branch from 48964eb to 4b6c12b Compare August 9, 2026 06:27
@ymodlin ymodlin changed the title close-range: skip test_close_range_threshold pending RSDEV-13516 close-range: remove test_close_range_threshold (RSDEV-13516) Aug 9, 2026
@ymodlin
ymodlin marked this pull request as ready for review August 9, 2026 06:30
@ymodlin
ymodlin requested a review from Nir-Az August 9, 2026 06:30

@Nir-Az Nir-Az left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@Nir-Az
Nir-Az merged commit e1bda42 into realsenseai:development Aug 9, 2026
35 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants