#1459 Segment A (governance): register XRCM / ICC / none / LOGO / ccox manufacturer signatures

[colourbill-ctrl]

Summary

Split off from #1459 (S3 creator/manufacturer validation). The non-governance code fix is handled in #1470 / #1471. This issue tracks the part that cannot be fixed in code: signatures that appear in ICC-published and widely-used profiles but are absent from the public Manufacturer Signatures registry (registry.color.org/manufacturer-signatures), and therefore require an ICC registry submission or policy decision.

Spec basis: ICC.1:2022-05 §7.2.17 — the header manufacturer/creator fields shall match a signature in the device-manufacturer section of the ICC signature registry, or be zero. Until these are registered, iccPawgReport's S3 correctly continues to WARN on them.

Evidence

Found by validating a 95-profile corpus (ICC-published + vendor profiles) with the #1470-fixed iccPawgReport. Distinct header signatures that are neither zero nor in the registry snapshot (2026-06-20):

Signature	Hex	Occurrences	Seen in	Disposition
XRCM	0x5852434D	16	CGATS21_CRPC2..7, GRACoL2013*, SWOP2013C3_CRPC5, APTEC_PC1x_*	Submit — creator of the entire ICC CRPC/GRACoL/SWOP/APTEC family; confirm owner (X-Rite production?)
ICC	0x49434320	23	sRGB_D65_colorimetric/MAT, iccDEV regression fixtures	Register/clarify the consortium's own signature (registry lists SICC/iccd, not 'ICC ')
none	0x6E6F6E65	25	AdobeRGB1998, ISO22028-2_ROMM-RGB, sRGB_v4_ICC_preference_displayclass	Policy — sanction 'none' as a placeholder (≡ zero) or recommend producers emit zero?
LOGO	0x4C4F474F	2	MCPPiPF5000Glossy	Investigate/submit (Logo/GretagMacbeth tooling?)
ccox	0x63636F78	1	corpus	Investigate/submit (likely CHROMiX; registry has CMiX 0x434D6958)

Related CMM-field observations (separate from manufacturer/creator; the IccProfLib GetCmmSigName gaps are handled in a separate library PR):

• APPL (0x4150504C, uppercase) used in the CMM field of SNAP2007, MCPPiPF5000Glossy — this is the platform signature, not Apple's registered CMM appl (0x6170706C).
• appl (0x6170706C) appears once as a manufacturer — it is in the CMM registry but not the Manufacturer registry.

Requested ICC actions

1. Register XRCM (confirm owner) so the ICC's own published CRPC/GRACoL/SWOP/APTEC profiles validate cleanly.
2. Register or canonically document the 'ICC ' consortium signature.
3. Decide policy on the 'none' placeholder.
4. Confirm/register LOGO and ccox (or provide correct mappings).

Meanwhile (already in #1470)

S3 keeps WARNing on all of the above (spec-correct). For 'none' and 'ICC ' the WARN wording is softened to identify them as known conventions rather than suspect data. No signature is silently accepted.

🤖 Generated with Claude Code

[github-actions]

Thanks for opening this issue.

The triage bot added initial labels based on the template and issue text. Maintainers may adjust labels as they review scope, reproduction details, CTest/regression needs, and security impact.

If you have questions, contact a listed Maintainer.

[ChrisCoxArt]

Nope, ccox is me. I wrote it into a few pieces of test code and haven't gotten around to registering it.

[colourbill-ctrl]

Yes, I know ccox is you @ChrisCoxArt :)
This issue was from a validation sweep I OCD'd into recently. The governance is to get Phil et al to register the unknown signatures into the registry site online. I split this off from the true bugs with validation for signatures.

[ChrisCoxArt]

My testing also found missing creators for Monaco and Linocolor (both of which got bought out before the registry took off).
Here's about 700-ish profiles from before the dawn of time (and some recent and standards) that I use for testing: https://www.dropbox.com/scl/fi/y0qznz1l3y1o8rd12alvu/Ccox_CollectedProfiles.7z?rlkey=lt00qzidavimac2043vjvdwj2&st=r01o99rb&dl=0
Warning - many of those won't parse, there were a LOT of mistakes made in the early days of profiles.