Changelog for package rosbot_utils

1.3.0 (2026-09-24)

• Bump firmware to v2.2.0-jazzy and drop micro-ROS

• Enter the STM32 bootloader over CBUS, without usbreset CBUS1 drives the MCU reset through an inverting stage (high holds it in reset), the opposite of what the code assumed. That is why the CBUS pulse never seemed to reach the bootloader and usbreset looked unavoidable; a port reset is also what leaves this FTDI unresponsive until a physical replug. Entering via CBUS with BOOT0 held keeps the FTDI on the bus; usbreset remains only as the second-attempt fallback, and a failed open now flushes pyftdi's device cache instead of resetting the port. ROSbot XL, 2026-09-23: 6 flashes (one through the read-protection recovery), 0 port resets, 0 drops, ~38 s instead of ~70 s per flash.

• Clear STM32 write protection only after a failed flash Clearing it up front cost a second usbreset on every USB flash, and each usbreset can drop the FTDI off the bus until a physical replug. ROSbot XL, 2026-09-23: old flow 2 resets per flash, FTDI lost in ~5 of 10 flashes; new flow 1 reset, 0 of 5.

• Bump firmware to v2.1.0-jazzy

• Auto-detect ALSA device and re-enable battery_alert (#195)

  * Auto-detect ALSA device and re-enable battery_alert The node played to ALSA [default]{.title-ref} (card 0), which is the speaker on none of the SBCs: the Jetson's card 0 is the APE with no playback PCM, the NUC's is the internal HD-Audio codec, the RPi5's is HDMI. The speaker PCB sits behind a USB sound card everywhere, so the device is now resolved from /proc/asound as plughw:CARD=<id> (stable across boots, unlike the card index) and detection is lazy so a card plugged in after boot works.

  * Auto-detect ALSA device and re-enable battery_alert The node played to ALSA [default]{.title-ref} (card 0), which is the speaker on none of the SBCs: the Jetson's card 0 is the APE with no playback PCM, the NUC's is the internal HD-Audio codec, the RPi5's is HDMI. The speaker PCB sits behind a USB sound card everywhere, so the device is now resolved from /proc/asound as plughw:CARD=<id> (stable across boots, unlike the card index) and detection is lazy so a card plugged in after boot works.

• Contributors: Rafal Gorecki, rafal-gorecki

1.2.4 (2026-09-01)

• Bump dependencies
• Contributors: rafal-gorecki

1.2.3 (2026-08-26)

• Add ROSbot 3 pro config + move camera_mount to custom yaml component (#194)

  * Add ROSbot 3 pro configuration (OAK-D Pro + RPLIDAR S2) basic stays OAK-D Lite + RPLIDAR C1; pro is the ROSbot 3 PRO sensor set, mirroring the existing basic/telepresence/autonomy split for rosbot_xl.

  * Move camera_mount to a custom yaml component for ROSbot XL Was built into rosbot_xl_macro.urdf.xacro behind an internal include_camera_mount toggle; now telepresence.yaml/autonomy.yaml add it as a [type: custom]{.title-ref} component (like man01_bracket), with camera_mount_angle_1/2 read straight off that yaml entry via the new husarion_components_description [component]{.title-ref} passthrough (bumped pin: f059bb6 -> 5f783f8).

  • Use canonical component names instead of SKU codes in manipulation configs
  • Bump firmware

  * Fix test_spawn_robot_configuration_choices to cover both models configuration is now a shared arg (rosbot: basic/pro/custom, rosbot_xl: full set) since the ROSbot 3 pro config landed, but the test still only diffed spawn_robot.yaml's choices against rosbot_xl/*.yaml, so adding [pro]{.title-ref} (which lives under config/rosbot/) tripped it as drift.

• Contributors: Rafal Gorecki

1.2.2 (2026-08-25)

• Fix flash_firmware NACK on read-protected STM32 MCUs
• Contributors: rafal-gorecki

1.2.1 (2026-08-14)

1.2.0 (2026-08-12)

• Restore vendored stb_image.h and the controllers pin rationale (#189) clang-format rewrote 6.3k lines of upstream stb_image v2.30 before the third_party exclusion landed, so the file could no longer be resynced against upstream.

File truncated at 100 lines see the full file