[OhadMeir]

Tracked on [RSDEV-12267]

[Copilot]

Pull request overview

This PR adds Windows support for Dual-RGB devices whose color frames are delivered via the depth sensor by enabling correct metadata parsing and ensuring the Windows metadata registry configuration accounts for the extra media pin.

Changes:

• Allow multiple metadata parsers to be registered for the same metadata key by chaining parsers and selecting the first one that successfully validates against a frame’s metadata layout.
• Add Dual-RGB color-metadata registration on the D500 depth sensor so color frames routed through it expose the expected RGB metadata fields.
• Update Windows metadata enablement logic (PowerShell + helper heuristic) to account for a 4th media pin on Dual-RGB SKUs.

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
src/sensor.cpp	Chains metadata parsers on duplicate registrations instead of ignoring later registrations.
src/metadata-parser.h	Adds md_layout_selecting_parser to try multiple parsers; suppresses md-type mismatch debug spam.
src/ds/d500/d500-dual-rgb.h	Declares register_color_metadata() for Dual-RGB devices.
src/ds/d500/d500-dual-rgb.cpp	Registers RGB-layout metadata parsers on the depth sensor and reuses ds_color_common metadata registration.
scripts/realsense_metadata_win10.ps1	Adds MetadataBufferSizeInKB3 handling for the additional Dual-RGB media pin.
common/metadata-helper.cpp	Updates media-pin count heuristic to return 4 for Dual-RGB PIDs on MI_00.

[sysrsbuild-gh-agentic]

[rs-agentic bot] Code review of latest push (commit 16a3288)

Reviewed the changes in common/metadata-helper.cpp, scripts/realsense_metadata_win10.ps1, and src/metadata-parser.h. Inline comments added below. Additional note:

🐛 BUG — operator== not updated to include vid (common/metadata-helper.cpp lines 48-54)

The vid field was added to device_id in this PR, but operator== still only compares pid, guid, mi, and sn. Two devices from different vendors (e.g. VID_8086 vs VID_38E5) sharing the same PID, MI, and SN would be incorrectly treated as equal in foreach_device_path. Please add equal(a.vid, b.vid) && to the comparison.

⚠️ Missing tests
No unit tests cover the new md_layout_selecting_parser or the updated number_of_mediapins(vid, pid, mi) logic. The unit-tests/ framework supports targeted test files — consider adding a test for these two behaviours.

[Nir-Az]

LGTM

[sysrsbuild-gh-agentic]

[rs-agentic bot] Code review of latest push (commit 3bf1b6e)

Reviewed the changes in common/metadata-helper.cpp, scripts/realsense_metadata_win10.ps1, and src/metadata-parser.h.

• ✅ Cross-layer #include <src/usb/usb-types.h> removed; VID_REALSENSE_CAMERA now defined locally as a string constant.
• ✅ std::stoul usage dropped; explanatory comment added for the intentional 4-pin behavior across all MIs on VID_38E5.
• ✅ Code style fixed (space after //, American English "behavior").

One item from the previous review body is still not addressed — see inline comment below.

[sysrsbuild-gh-agentic]

[rs-agentic bot] Code review of latest push (commit 6804a8e)

✅ operator== updated to include vid — the previous bug is fixed. See inline comments.

[Nir-Az]

LGTM

[Nir-Az]

LGTM

[sysrsbuild-gh-agentic]

[rs-agentic bot] Code review of latest push (commit 274952f)

Reviewed changes in common/metadata-helper.cpp, src/metadata-parser.h, and scripts/realsense_metadata_win10.ps1.

• ✅ Extra spaces in operator== (equal(a.vid, b.vid ) / double-space before &&) are cleaned up — all five comparisons now use consistent formatting.
• ✅ src/metadata-parser.h: md_always_enabled_param_parser::is_attribute_valid — the LOG_DEBUG spam for type-mismatch has been removed and replaced with the single explanatory comment // Some mismatched metadata types are expected, e.g dual-RGB color metadata on a depth sensor. as requested by @Nir-Az.
• ✅ scripts/realsense_metadata_win10.ps1: $prefix is now computed once with a length guard (if ($devInstance.Length -ge 27)) before the multi-pin list lookup, preventing a potential Substring exception on short device instance strings.

No new issues found in this commit. All previously raised bot comments are now resolved.