CNA-BUG-206: Software_XnaLineCoverage passes when a scene differs from the XNA reference: run-oracle-corpus-diff-software.sh fails only on render errors

CNA snapshot 009d40f5  ·  Known Issues › Current bugs  ·  source links pinned to 009d40f5

✓

Evidence basis: source-verified at the pinned commit; executed for this entry (the Evidence section names exactly what was run); tests exist (not executed for this page). Claims on this page were checked by reading the CNA source at commit 009d40f5; unless a sentence says otherwise, nothing here was built or executed. Nothing on this page was executed unless the Evidence section says so.

The CTest described as the exact Software regression for the two aliased-line scenes runs a script that prints DIFF for a pixel mismatch and still exits 0; only a missing reference or a failed render makes it fail.

Identifier
CNA-BUG-206
Category
Bug
Subsystem
Testing & evidence
Status
Open
Verified against
CNA 009d40f5 (009d40f5dd085c4e674d3479675fac84b12b3e0a)
Severity
Low (a triage suggestion, not a project priority)
Evidence basis
Reproduced: executed for this entry (the Evidence section names exactly what was run)
Tests touching this area
Yes: see Current tests
Affected contract
CTest Software_XnaLineCoverage (modules/renderers/software/examples/CMakeLists.txt)

Expected behaviour

The registration comment calls the two measured line scenes an exact Software regression (SOFTWARE-344). The EasyGL twin, EasyGL_XnaLineCoverage, fails on any mismatch through run-oracle-corpus-diff-easygl.sh.

Actual behaviour at TARGET

run-oracle-corpus-diff-software.sh counts DIFFERENT and RENDER_FAILED separately and exits 1 only when RENDER_FAILED is non-zero. A mismatch, and equally an xna-diff.py crash such as a missing Pillow module, is printed as DIFF and the test passes. cna_register_renderer_test adds no failure regular expression for the Software renderer.

Source locations

Evidence

Checked by reading at 009d40f5; not executed.

Focused reproduction

# Illustrative; not run. In a SOFTWARE test tree, corrupt one reference image, then:
ctest -R '^Software_XnaLineCoverage$' --output-on-failure
# Expected: failure. Actual by reading: the scene prints DIFF and the test passes.

Current tests

Only the script itself; nothing tests the script's exit contract.

Regression test

Make the script exit non-zero when DIFFERENT is non-zero (as the EasyGL script does), and add a self-test with a deliberately different reference.

Blast radius

The Software renderer's line-rasterisation regression guard; no CI workflow runs it at present (see the Known Issues entry on oracle comparisons).

Workaround

No workaround is known.

The same subject is explained at several altitudes. These are the neighbouring pages at each one.

Internals
SOFTWARE renderer internals
Deep dives
Oracles: the XNA corpus · Evidence tiers of the native modern GPU renderers
Known issues
Bug index