[sareluzi]

Summary

Removes the Intel prefix from camera visible names exposed via RS2_CAMERA_INFO_NAME. E.g. Intel RealSense D455 -> RealSense D455. Rest of the name string is unchanged.

Changes

Name source (single source of truth):

• src/ds/d400/d400-private.h rs400_sku_names and src/ds/d500/d500-private.h rs500_sku_names - dropped Intel from every entry (D4xx/D5xx, F416, Recovery, IMU, USB2, DS5U, Tracking variants).
• src/media/ros/ros_reader.cpp - legacy TM2 playback name.

Prefix consumers updated (read the name, depend on the literal prefix):

• src/gl/camera-shader.cpp - 3D camera mesh selection (starts_with).
• common/device-model.cpp - preset-file lookup regex.
• common/on-chip-calib.cpp - D415 emitter special-case.
• wrappers/python/pyrs_device.cpp and tools/dds/dds-adapter/rs-dds-adapter.cpp - prefix strip (length 16 -> 10).

Tests / fixtures / examples:

• unit-tests/dds/{d455,d435i,d405}.py, unit-tests/live/calib/calibrations_common.py, unit-tests/infra-tests/test_iq_snapshot_dedup.py, wrappers/python/examples/d500_triggered_calibration.py, tools/realsense-viewer/presets/sample.preset.
• unit-tests/py/rspy/devices.py already handles both Intel RealSense and RealSense prefixes (kept).

Out of scope (not the camera visible name)

OS USB descriptors, metadata field names, GPU vendor checks, L500/L515 (Unity) matching, C# assembly/window-title branding, docs.

Notes

• Safety camera (D585S) covered; its Safety Camera sensor name never had an Intel prefix.
• Builds verified locally (realsense-viewer).

[Copilot]

Pull request overview

This PR updates librealsense device naming by removing the "Intel " prefix from RS2_CAMERA_INFO_NAME (e.g., "Intel RealSense D455" → "RealSense D455") and adjusts downstream consumers (viewer logic, DDS adapter, Python bindings, ROS playback, and tests/fixtures) that depend on the literal name prefix.

Changes:

• Updated D400/D500 SKU name maps to emit "RealSense ..." visible names.
• Updated multiple prefix consumers (regex/prefix stripping/mesh selection/legacy ROS reader) to use the new prefix.
• Updated unit tests, examples, and preset fixtures to match the new visible device names.

Reviewed changes

Copilot reviewed 15 out of 15 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
wrappers/python/pyrs_device.cpp	Updates Python device __repr__ prefix stripping to the new "RealSense " prefix
wrappers/python/examples/d500_triggered_calibration.py	Updates D585S safety-camera name check to "RealSense D585S"
unit-tests/live/calib/calibrations_common.py	Updates D415 name check used for emitter behavior
unit-tests/infra-tests/test_iq_snapshot_dedup.py	Updates mocked pipeline device names used for snapshot dedup tests
unit-tests/dds/d455.py	Updates DDS fixture device name
unit-tests/dds/d435i.py	Updates DDS fixture device name
unit-tests/dds/d405.py	Updates DDS fixture device name
tools/realsense-viewer/presets/sample.preset	Updates sample preset device name
tools/dds/dds-adapter/rs-dds-adapter.cpp	Updates ROS2 node-name prefix stripping to new "RealSense " prefix
src/media/ros/ros_reader.cpp	Updates legacy TM2 playback name to "RealSense TM2"
src/gl/camera-shader.cpp	Updates viewer camera-mesh selection string matching to new "RealSense ..." names
src/ds/d500/d500-private.h	Removes "Intel " prefix from D500 SKU name map entries
src/ds/d400/d400-private.h	Removes "Intel " prefix from D400 SKU name map entries
common/on-chip-calib.cpp	Updates D415 special-case name string
common/device-model.cpp	Updates preset-file lookup regex to parse "RealSense ..." names

Comments suppressed due to low confidence (1)

tools/dds/dds-adapter/rs-dds-adapter.cpp:49

• build_node_name() strips only the new "RealSense " prefix. If a device name still starts with "Intel RealSense " (e.g., playback from older recordings), node names/topic-roots will change unexpectedly (e.g. Intel_RealSense_D455_<sn>). Consider stripping both prefixes to keep node names stable across device sources.

    constexpr char const * DEVICE_NAME_PREFIX = "RealSense ";
    constexpr size_t DEVICE_NAME_PREFIX_CCH = 10;
    // We don't need the prefix in the path
    std::string model_name = rs2_camera_name;
    if( model_name.length() > DEVICE_NAME_PREFIX_CCH
        && 0 == strncmp( model_name.data(), DEVICE_NAME_PREFIX, DEVICE_NAME_PREFIX_CCH ) )

[Nir-Az]

Overall lookd good, I triggered a nightly CI to see all cases runs,

[Nir-Az]

Found as breaking change for ROS, we need to align it too

  realsense2_camera/src/realsense_node_factory.cpp:453
  std::string device_name(_device.get_info(RS2_CAMERA_INFO_NAME));
  ...
  if (device_name == "Intel RealSense D555")   // exact full-name match
  {
      pid = RS555_PID;                          // DDS PID workaround
  }
  else
  {
      pid = std::stoi(pid_str, 0, 16);          // D555 falls here after rename
  }

Opened a matching PR here for your review, should be merged together:
realsenseai/realsense-ros#3524

[Nir-Az]

LGTM

[sareluzi]

Pushed review fixes (commit 9c48f4c):

Addressed

• DDS adapter (@OhadMeir): build_node_name now strips both Intel RealSense and RealSense prefixes → topic-roots stable across sources.
• Substring matching (@OhadMeir): calibrations_common.py and d500_triggered_calibration.py switched from == to in. Also applied the same to common/on-chip-calib.cpp (D415 emitter check) which had the identical equality pattern.

Dismissed (with rationale in threads)

• sample.preset / existing user presets (@OhadMeir): preset lookup keys on the model token (<model> *.preset), unchanged by the rename; the JSON name field isn't validated on load — existing presets still resolve.
• Copilot backward-compat ×3 (camera-shader / device-model regex / pyrs_device __repr__): per @Nir-Az, viewer/display/cosmetic; live devices always get the new name. Left as-is.

Cross-repo coordination

• ROS break (@Nir-Az): librealsense-ros#3524 must merge together with this PR. Confirmed the only full-name exact-match in this repo's tree are the ones fixed above; the ROS one lives in the external repo and is covered by that PR.

Local build (realsense-viewer) green after the changes.

[Nir-Az]

@sareluzi I understand from @OhadMeir that on DDS camera name is sent by FW, meaning same change needs to occur on D555 DDS FW (and also D555 + D400 USB FW is we want the descriptors to change)

[sareluzi]

@Nir-Az - i will follow-up on D500 names change as well.
D400 PR already opened.

[Nir-Az]

Noticed another miss

librealsense/src/gl/camera-shader.cpp

Line 166 in 09937b5

if (starts_with(dev_name, "Intel RealSense D415")) index = 0;

I suggest changing all line to find instead of start with and drop Intel

[Nir-Az]

fix needed in camera shader

[sareluzi]

@Nir-Az done (commit b1ab176) — camera_renderer::process_frame now selects the mesh by model token using find instead of starts_with, so it's prefix-agnostic (matches D415/D435/D436/D45 regardless of Intel RealSense / RealSense / playback names):

std::string dev_name = dev.get_info(RS2_CAMERA_INFO_NAME);
if (dev_name.find("D415") != std::string::npos) index = 0;
if (dev_name.find("D435") != std::string::npos ||
    dev_name.find("D436") != std::string::npos) index = 1;
if (dev_name.find("D45")  != std::string::npos) index = 2;

Local viewer build green.

[aikido-pr-checks]

camera_renderer::process_frame mixes device-name parsing, index selection, and full GL rendering logic; split responsibilities into helpers to reduce cognitive complexity

Details

✨ AI Reasoning
​1. What is this code trying to accomplish? It identifies a camera model index from the device name and then performs GL rendering of a selected mesh. 2. Does this harm maintainability? Combining parsing logic and full rendering flow in one function increases cognitive load. 3. Is this appropriate? The method orchestrates diverse responsibilities (device name parsing, index selection, shader setup, fixed-function rendering), so it's harder to reason about at a glance. 4. Would fixing meaningfully improve the codebase? Yes — separating name-matching and rendering paths would localize concerns and simplify testing. 5. Is this a common pattern? Renderers often separate resource selection and draw calls; here they're combined. 6. Fixable in PR scope? The PR introduced a small change inside the function; refactoring into helper functions is reasonable but beyond this tiny change. Conclusion: the modified function remains a medium-length, multi-responsibility function that is more complex than desirable.

🔧 How do I fix it?
Break down long functions into smaller helper functions. Aim for functions under 60 lines with fewer than 10 local variables.

_{Reply @AikidoSec feedback: [FEEDBACK] to get better review comments in the future.
Reply @AikidoSec ignore: [REASON] to ignore this issue.
More info}

[sareluzi]

Makes sense but not related to this PR. Can be handled seperately.

[sareluzi]

Rebased on latest development (resolved the d500-private.h SKU-map conflict — kept all new D500 SKUs, dropped Intel from the 5 pre-existing entries). PR is mergeable again. The camera-shader find-by-model-token fix is included (6b7ebc5). @Nir-Az re-review please.

[sareluzi]

RetriggerGated

[Nir-Az]

Verified on ROS