Restore: #1362 regression test dropped by #1366 auto-merge (script + fixtures + CTest registration)

[colourbill-ctrl]

What

Restores the #1362 regression test that #1366 was supposed to add. Although #1366 is marked merged (merge commit 21da415), its auto-merge silently dropped all four artifacts of the test when resolving the Testing/CMakeLists.txt conflict against master:

• .github/scripts/iccdev-issue-1362-v5dspobs-xyztype-regression.sh
• .github/ci/test-data/v5dspobs-lcddisplay.icc (fixture)
• .github/ci/test-data/v5dspobs-cat8lab.icc (fixture)
• the iccdev.issue-1362-v5dspobs-xyztype-regression iccdev_add_script_test() registration

Master currently has the #1364 producer fix (CIccTagXYZ) and the #1356/#1361 tests, but no #1362 test — so the rXYZ/gXYZ/bXYZ XYZType fix has no regression guard. This branch (fix-rev-1362, prepared by @xsscx) re-adds exactly those four pieces.

Why this matters

The test runs iccV5DspObsToV4Dsp on the committed V5 display + observer inputs and inspects the on-disk type signature of each matrix-column tag, failing if any is still 'sf32' (s15Fixed16ArrayType, 73663332) instead of 'XYZ ' (XYZType, 58595a20). Without it, a regression of #1362 would land unnoticed.

Notes

• The #1362 iccdev_add_script_test() block is restored with its closing ) (paren balance verified); the dropped paren in the earlier Test: regression for iccV5DspObsToV4Dsp rXYZ/gXYZ/bXYZ tag type (#1362) #1366 merge was the root of the Parse error. Function missing ending ")" configure failure.
• In-repo branch (script + CI test data + CMake), per the convention that .github/scripts / CodeQL / CI changes live on an upstream branch.
• Closes the gap left by Test: regression for iccV5DspObsToV4Dsp rXYZ/gXYZ/bXYZ tag type (#1362) #1366. Ref iccV5DspObsToV4Dsp writes rXYZ/gXYZ/bXYZ matrix-column tags as s15Fixed16ArrayType ('sf32') instead of XYZType ('XYZ ') #1362, Sub-Issues from #1355 #1357.

🤖 Generated with Claude Code

[xsscx]

2026-06-15 22:38:35 UTC

[xsscx]

TODO @xsscx

2026-06-15 22:44:01 UTC
Note - Removed the Fail Label the Bot added. Need to Research. I watched the PR open, and ci-pr-action is running and does not indicate any failure, nor do any sibling workflows.

[colourbill-ctrl]

Merged — #1362 regression test restored on master

Confirming the close-out for the archive.

Why this PR existed: #1366 is marked merged (21da415), but its auto-merge silently dropped all four #1362 test artifacts when resolving the Testing/CMakeLists.txt conflict against master — the regression script, both .icc fixtures, and the CTest registration. So the #1364 producer fix (rXYZ/gXYZ/bXYZ → CIccTagXYZ / 'XYZ ') had shipped to master without a regression guard. (Full timeline on #1366.)

What landed here (00f04e1):

• .github/scripts/iccdev-issue-1362-v5dspobs-xyztype-regression.sh
• .github/ci/test-data/v5dspobs-lcddisplay.icc, .github/ci/test-data/v5dspobs-cat8lab.icc
• the iccdev.issue-1362-v5dspobs-xyztype-regression iccdev_add_script_test() registration — with its closing ) (paren balance verified)

Verified post-merge on master: all four pieces present, exactly one #1362 registration (no duplicate), CMakeLists.txt parens balanced, 30 checks green / 0 fail. The test runs iccV5DspObsToV4Dsp on the committed inputs and fails if any matrix-column tag is still 'sf32' (73663332) instead of 'XYZ ' (58595a20).

The #1362 thread is now fully closed: fix #1364 + test #1368. Thanks @xsscx for the fix-rev-1362 restore branch.