51230 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
Ramon Roche	8d446794ca	docs(skills): say PRs may be squashed at maintainer discretion (#28871) ... The pr and commit skills claimed every PR is squash-merged, which contradicts CONTRIBUTING.md: merge strategy is case by case, with both squash and rebase merges enabled. The wrong claim led to fix-up commits being kept on branches on the assumption they would be squashed away. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com>	2026-09-27 16:27:20 -06:00
PX4 Build BotandCrowdin Bot	4ce66eb466	docs(i18n): PX4 guide translations (Crowdin) - uk (#28866) ... Co-authored-by: Crowdin Bot <support+bot@crowdin.com>	2026-09-27 13:34:48 +10:00
PX4 Build BotandCrowdin Bot	56d5e8661a	docs(i18n): PX4 guide translations (Crowdin) - zh-CN (#28867) ... Co-authored-by: Crowdin Bot <support+bot@crowdin.com>	2026-09-27 13:34:19 +10:00
Ramon Roche	939b420488	ci(codecov): upload unit test coverage (#28869) ... * ci(codecov): upload unit test coverage and limit SITL coverage to main Coverage on main came from a single standard_vtol SITL flight, unit tests were never uploaded, and the Coverage build never hit ccache because ccache 4.9.1 (Ubuntu 24.04) rejects -fprofile-update=atomic. The flag only keeps hit counts exact under thread races; line coverage is unaffected and lcov already ignores the resulting negative counts. Unit tests now build with coverage and upload as the unittests flag on every PR and push to main. standard_vtol builds as Coverage on main only and uploads as sitl-vtol, carried forward on PRs. codecov.yml sets the default branch, informational project/patch statuses and ignores. Per-build-type ccache keys stop the Coverage and RelWithDebInfo jobs from overwriting each other's cache. Fixes #28862 Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com> * ci(codecov): run unit test coverage as its own job The Coverage build shifts EKF replay float output, so running it in the Unit Tests job made the change indication step auto-commit new baselines to the PR branch. Unit Tests goes back to the plain build and a separate job produces the unittests upload. disable_search keeps the Codecov uploader to the lcov report instead of running gcov over the build tree. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com> --------- Signed-off-by: Ramon Roche <mrpollo@gmail.com>	2026-09-26 19:11:53 -07:00
Ramon Roche	bf9b1610ca	fix(ros2): pin BuildKit for larger dev container SBOMs ... Publishing px4-dev-ros2-gazebo fails because its SPDX attestation exceeds BuildKit 0.32's 40 MiB limit. Pin the builder to v0.33.0 (80 MiB), matching the fix already applied to build_deb_package.yml in 61907962c9. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com>	2026-09-26 18:51:08 -07:00
3 people	e35171c3f9	fix(logger): make logger off close the active log file (#28636) ... * fix(logger): make off stop active file logging Handle logger on and off as explicit requests so logger off closes an active file log without disabling future automatic activations. Continuous logging modes fall back to arm-to-disarm sessions until reboot after their active log is stopped. Assisted-by: Codex:gpt-5 Signed-off-by: jc-works <jc-works@users.noreply.github.com> * fix(logger): keep the configured log mode on manual stop Falling back to LogMode::while_armed by overwriting _log_mode destroyed the configured mode for the rest of the boot, and nothing reported it: neither logger status nor LoggerStatus carries the mode. Only boot_until_shutdown needs the fallback at all, because its early return would otherwise keep logging disabled once the boot log is stopped. Express that with a dedicated flag so _log_mode stays the configured value, and say so on the console when it happens. arm_until_shutdown needs no fallback. After the forced stop _prev_file_log_start_state is false, so full_log_continues is false and the desired state falls back to plain arming. The next arming starts a log and sets the flag again, so the following disarm keeps logging to shutdown as configured instead of ending the log early. Assisted-by: Claude:claude-opus-5 Signed-off-by: Julian Oes <julian@oes.ch> * fix(logger): restore SDLOG_MODE 2 logging behaviour after rearming After logger off stops a boot-to-shutdown log, suspend automatic logging only until the next arming. Restore the configured lifecycle when the new log starts so it continues across subsequent disarms until shutdown. Update the logger message and documentation accordingly. Assisted-by: Codex:gpt-5 Signed-off-by: jc-works <jc-works@users.noreply.github.com> * fix(logger): make repeated off requests idempotent Regression testing found that repeating logger off after the active log had closed cleared the manual stop latch. When the arming or AUX logging condition remained active, automatic logging started a new log. Preserve the manual stop latch across repeated off requests so logging remains stopped until the automatic logging condition deactivates. The latch is also cleared explicitly by logger on. Assisted-by: Codex:gpt-5 Signed-off-by: jc-works <jc-works@users.noreply.github.com> --------- Signed-off-by: jc-works <jc-works@users.noreply.github.com> Signed-off-by: Julian Oes <julian@oes.ch> Co-authored-by: jc-works <jc-works@users.noreply.github.com> Co-authored-by: Julian Oes <julian@oes.ch>	2026-09-27 14:27:50 +13:00
Ramon Roche	786893436c	ci(macos): save ccache only on main ... The macOS build runs on GitHub-hosted runners, so its ccache lives in the repository's 10 GB Actions cache. Every run saved a new entry, about 300 a week and 80% from pull requests, which kept the cache full and evicting. Pull requests already restore from main's newest entry, so only main needs to save. Pull requests still print the final ccache stats. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com>	2026-09-26 18:13:56 -07:00
Beniamino Pozzan	fba1622d0c	docs(uxrce_dds): bump agent version from 3.0.1 to 3.0.2 ... Flow control now is available in 3.0.2 and only need cherry-pick for 2.4.2 and 2.4.3 Signed-off-by: Beniamino Pozzan <beniamino.pozzan@gmail.com>	2026-09-26 18:13:37 -07:00
Ramon Roche	39e82389e2	docs(dev_setup): drop the stale JDK sudo note from the macOS guide (#28861) ... Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com>	2026-09-27 09:18:00 +10:00
LtxxandBeniamino Pozzan	d2de298122	fix(uxrce_dds): align SITL namespace prefix (#28338) ... * fix(uxrce_dds): align SITL namespace prefix Use the same uav_ namespace prefix for SITL multi-instance startup and UXRCE_DDS_NS_IDX fallback. This keeps simulation and hardware topic names consistent without changing the existing hardware default. Assisted-by: OpenAI Codex: GPT-5 * docs(uxrce): update uxrce_dds docs * docs(release): update `main` release page --------- Signed-off-by: Beniamino Pozzan <beniamino.pozzan@gmail.com> Co-authored-by: Beniamino Pozzan <beniamino.pozzan@gmail.com>	2026-09-27 00:12:55 +01:00
Ramon Roche	5ecc673e23	ci(macos): drop the Homebrew download cache and double ccache ... The Homebrew download cache restored from any older key and saved the result under the new one, so it only ever grew: each entry is now about 2.9 GB, three of them fill most of the 10 GB repository cache, and GitHub evicts the ccache entries to make room. Restoring and saving it cost about 72 s per run, more than the Homebrew install it saves. The 200M ccache is full after px4_sitl, so px4_fmu-v5_default only hits about 81%. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com>	2026-09-26 15:34:46 -07:00
Ramon Roche	a8b0e91ebe	feat(setup): install macOS Gazebo from a locked conda-forge environment ... Homebrew rebuilds and drops the gz bottles whenever homebrew-core bumps protobuf or abseil, which breaks the macOS Gazebo install until the pins catch up. conda-forge keeps every package it has published, so Tools/setup/macos/pixi.toml and its lock install the same gz, protobuf, abseil, OpenCV and GStreamer binaries until the lock is updated on purpose. OpenCV is the headless build, and GStreamer brings the good, bad and ugly plugin sets the gz camera plugin's pipeline uses. macos.sh installs the environment with pixi and loads it from the venv activation through a small script, since pixi shell-hook would export the installing shell's PATH. The build and the gz launch do not depend on that activation: CMake searches the environment like it pins the .venv Python, and px4-rc.gzsim sources gz_env.sh before it looks for gz, which loads the environment when it is not already active and puts its lib dir ahead of Homebrew's in the dyld fallback path. conda-forge's gz-transport13 does not pull in cppzmq, so the manifest lists it. The osrf/simulation pin, the protobuf pin and the weekly bot that refreshed them are removed. The homebrew-core commit moves into homebrew-pins.txt unchanged. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com>	2026-09-26 15:12:00 -07:00
Ramon Roche	de9c69316e	docs(simulation): replace jMAVSim with SIH and Hawkeye ... The jMAVSim pages become redirects to SIH, examples and HITL options move to SIH, Gazebo or Gazebo Classic, and SIH on hardware gets Hawkeye as its 3D display-only view through a MAVLink bridge to its UDP port. The release notes flag the removal in the upgrade guide. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com>	2026-09-26 14:28:26 -07:00
Ramon Roche	a23449cbe5	feat(simulation)!: remove jMAVSim ... PX4 SIH with the Hawkeye visualizer replaces jMAVSim, in SITL and as the display-only viewer for SIH running on hardware. Drop the jMAVSim submodule, its make targets, px4-rc.jmavsim, the 10017_jmavsim_iris airframe and the VS Code debug helpers, stop installing Java and ant in the macOS and Arch setup scripts, and point the remaining mentions at SIH and Hawkeye. BREAKING CHANGE: make px4_sitl jmavsim and SYS_AUTOSTART 10017 no longer exist. Use make px4_sitl_sih sihsim_quadx with Hawkeye instead. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com>	2026-09-26 14:28:26 -07:00
Ramon Roche	a4b722d400	ci(packaging): seed the container build caches from main (#28859) ... * ci(packaging): seed the container build caches from main The workflow only ran on pull requests, tags and manual dispatch, so nothing ever saved a cache under main. Pull requests can only read their own ref and main, so every first run on a PR built all images cold (30 min vs 13 min warm). Running on main and weekly gives PRs a cache to fall back to; the weekly run also keeps it inside the runner cache's 10-day expiry. The ROS ccache fallback also skipped the image name, so sih and gazebo could restore each other's cache. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com> * ci(packaging): clarify why main is rebuilt weekly Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com> --------- Signed-off-by: Ramon Roche <mrpollo@gmail.com>	2026-09-26 14:27:05 -07:00
Ramon Roche	55a577ffd0	ci(docs): skip site artifact upload on pull requests ... Only T4: Deploy reads px4_docs_build, and it runs on push and workflow_dispatch. On pull requests the upload spent ~93 s and ~1 GB of artifact storage per run for nothing. The site is still built, so the check keeps failing on broken docs. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Ramon Roche <mrpollo@gmail.com>	2026-09-26 13:43:41 -07:00
Jacob Dahl	7c8d6daca9	docs(agents): move the shared driver code rule to drivers.instructions.md (#28853) ... It only matters when working under src/drivers/, which the scoped file already covers, and AGENTS.md is loaded for every task. The section was also not wrapped like the rest of AGENTS.md. Assisted-by: Claude:claude-opus-5-5	2026-09-25 23:41:55 -06:00
Jacob Dahl	2d277ec00b	docs(agents): say where code shared between drivers goes (#28852) ... Agents put helpers used by several drivers of one type next to those drivers in src/drivers/<type>/, where the decoders they sit beside already live in src/lib/. Assisted-by: Claude:claude-opus-5-5	2026-09-25 23:37:02 -06:00
PX4BuildBot	5fb15552bb	docs: auto-sync metadata [skip ci] ... Co-Authored-By: PX4 BuildBot <bot@px4.io>	2026-09-26 02:47:37 +00:00
Andrea Bernasconi	b38dcb1f6c	fix(navigator): use "Return" instead of "RTL" in user-facing text (#28543)	2026-09-26 12:40:56 +10:00
Jacob Dahl	bd7d86e141	fix(uavcan): keep GNSS jamming and spoofing with RelPosHeading (#28847) ... The Fix2 ecef_position_velocity block also carries the node's noise, jamming and spoofing state, but it was skipped whenever a RelPosHeading was pending, so jamming and spoofing read unknown for as long as a dual-antenna heading was valid. Assisted-by: Claude:claude-opus-5-5	2026-09-25 16:41:30 -06:00
Jacob Dahl	4e91ae1c42	ci(flash_analysis): rename size table columns to Δ and Total (#28844) ... Shorter headers keep the five-column table narrow, and "used" read as the amount used by the change rather than the target's total. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Jacob Dahl <dahl.jakejacob@gmail.com>	2026-09-25 15:59:06 -06:00
Jacob Dahl	4dbd2e069a	fix(simulation): cast gz airspeed temperature to double (#28843) ... The macOS build has been failing since #28842: clang reports the implicit float to double conversion under -Wdouble-promotion, which the build treats as an error. GCC does not warn for argument conversions, so the Linux builds passed. Assisted-by: Claude:claude-opus-5-5	2026-09-25 12:05:22 -06:00
PX4BuildBot	579f7d161e	docs: auto-sync metadata [skip ci] ... Co-Authored-By: PX4 BuildBot <bot@px4.io>	2026-09-25 17:37:12 +00:00
yannickpoffetandMichael Fritsche	78c5eafa85	feat(new-mode): Add new parking mode (#28469) ... * feature(new-mode): Add new parking mode This mode is an extension of the manual mode. It is to be used to manoevre platforms precisely in tight spaces, such as when parking rovers or when docking a boat. It is designed to be used with a joystick and allows for precise control of the vehicle's position and orientation. * feat(rover): implement manual parking modes and add docs * fix(rover): move from percent based to normalized max throttle in manual parking mode * fix(rover): improve parking mode description in docs --------- Co-authored-by: Michael Fritsche <frimi999@gmail.com>	2026-09-25 13:29:19 -04:00
Balduin	1cd0440b68	fix(simulation): set temp in gz airspeed plugin (#28842) ... The plugin already calculates a temperature which is used to adjust air density influence on differential pressure. But it does not give the airspeed message the temperature, so the differential pressure temperature ends up being 0.	2026-09-25 09:32:07 -06:00
PX4BuildBot	0b3e166d4c	docs: auto-sync metadata [skip ci] ... Co-Authored-By: PX4 BuildBot <bot@px4.io>	2026-09-25 09:53:14 +00:00
Daniel Buleandră	ba140de1e4	feat(drivers/pca9685_pwm_out): merge PCA9685_SCHD_HZ into PCA9685_PWM_FREQ (#28705) ... PCA9685_PWM_FREQ and PCA9685_SCHD_HZ controlled the same physical PWM rate through two independent parameters that had to be kept in sync manually by the user, and neither was exposed in the ground station's Actuators tab. Merge both into PCA9685_PWM_FREQ, which now drives both the chip's PWM frequency and the I2C update rate, and expose it as a single config field on the actuator output group. PCA9685_SCHD_HZ is removed outright, with no automatic migration: a saved PCA9685_PWM_FREQ carries over for free since the name is unchanged, and PCA9685_SCHD_HZ was intentionally left orphaned rather than migrated, so that inspecting main.cpp or PCA9685.cpp is enough to understand the current behavior of the driver. Verified on real Skynode S hardware: the parameter reaches the Actuators tab in the ground station, and the PWM frequency measured with a logic analyzer on the PCA9685 output pin matches the configured value. Assisted-by: Claude:claude-sonnet-5 Signed-off-by: danielbuleandra <daniel.buleandra@auterion.com>	2026-09-25 11:47:54 +02:00
Jacob Dahl	253479bbeb	feat(ci): report flash and RAM use against capacity, analyze fmu-v2 (#28838) ... * feat(ci): report flash and RAM use against capacity, analyze fmu-v2 The size comment showed only the delta, which does not say how close a target is to its limit. Capacity comes from the board's own linker script, because the flash-analysis build links against an oversized flash region so that a change which overflows still gets a report. fmu-v2 adds an F4 next to the F7 and H7 targets, and with 1 MB of flash it is the FMU that overflows first. Assisted-by: Claude:claude-opus-5-5 * fix(ci): shorten the size comment's usage columns Five columns with capacities spelled out made the table too wide to read at a glance. Assisted-by: Claude:claude-opus-5-5	2026-09-24 17:54:56 -06:00
Saibernard	d2921f3567	docs(releases): add main notes for the AlphaFilter API change, Gazebo GNSS failure injection and the macOS test suite (#28548) ... Assisted-by: Claude:claude-fable-5 Signed-off-by: Saibernard Yogendran <bernie97@seas.upenn.edu>	2026-09-25 08:52:16 +10:00
Silvan Fuhrer	e66fcae998	feat: add Marco to maintainers list (#28831)	2026-09-24 12:42:32 -06:00
Jacob Dahl	816a234175	docs(dronecan): drop EKF2_RNG_A_HMAX from rangefinder setup (#28836) ... EKF2_RNG_A_HMAX is the height below which conditional range aid uses the rangefinder as the height source, not the maximum range the EKF uses; that comes from the sensor's reported max distance (UAVCAN_RNG_MAX). The DIST and smartmicro pages set it to the sensor range (30, 50, 175 m), above the parameter's 10 m max, so QGroundControl only accepts it with a force-save. The right value depends on the vehicle's baro ground effect rather than the sensor, so the default applies. Assisted-by: Claude:claude-opus-5-5	2026-09-24 11:17:24 -06:00
PX4BuildBot	86716507be	docs: auto-sync metadata [skip ci] ... Co-Authored-By: PX4 BuildBot <bot@px4.io>	2026-09-24 10:17:06 +00:00
Silvan	1bcefdd86c	feat(docs): improve IC engine docs ... - add dedicated page for ICE config and explanation - update diagrams and screenshots Signed-off-by: Silvan <silvan@auterion.com>	2026-09-24 12:10:34 +02:00
Silvan	2c5c54a3d5	chore(ice): improve format and description of ICEControl.msg ... Signed-off-by: Silvan <silvan@auterion.com>	2026-09-24 12:10:34 +02:00
Silvan	6a63433195	chore(ice): change enum for user_request into bool ... Signed-off-by: Silvan <silvan@auterion.com>	2026-09-24 12:10:34 +02:00
Silvan	8d53fa0bd6	fix(ice_controller): zero-initialize the published engine status ... Only four fields of internal_combustion_engine_status are assigned before publishing, so the remaining ones carried indeterminate stack contents. The topic is forwarded to the ground station as MAVLink EFI_STATUS, which therefore reported garbage engine speed, temperatures, pressures and fuel figures on setups without a DroneCAN ECU filling them in. Assisted-by: Claude:claude-opus-5[1m] Signed-off-by: Silvan <silvan@auterion.com>	2026-09-24 12:10:34 +02:00
Hamish Willee	c428709beb	docs(docs) typo (#28828)	2026-09-24 17:01:29 +10:00
Hamish Willee	e370d98280	fix(mavlink): allow p1/p2 for standalone NAV_VTOL_TAKEOFF command (#28732) ... The command-side param mask for MAV_CMD_NAV_VTOL_TAKEOFF (0x7C) allowed param3 (unused) but forbade non-zero param1 and param2, even though Navigator's standalone-command handler reads param1 as the post-transition loiter height and param2 as the "use specified transition heading" flag (navigator_main.cpp). A GCS sending non-default values for either, per the documented MAV_CMD semantics, was denied at the MAVLink boundary with MAV_RESULT_DENIED before ever reaching Commander or Navigator. Assisted-by: Claude:claude-sonnet-5 Signed-off-by: Hamish Willee <hamishwillee@gmail.com>	2026-09-24 14:27:20 +10:00
Hamish Willee	860aca9864	docs(docs): fix broken links and missing uORB TOC entries (#28827) ... * docs(msg): add missing uORB message pages to SUMMARY The msg_docs block in SUMMARY.md is not regenerated by the metadata sync, so seven message pages were unreachable from the sidebar (PageNotInTOC). Replace the block with the fragment written by generate_msg_docs.py, which also moves VehicleControlMode to the unversioned list. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Hamish Willee <hamishwillee@gmail.com> * docs: fix broken and permanently redirected external links CUAV moved its docs to a /en/<product>/ layout, so the Nora, C-RTK 9Ps and C-RTK2 links returned 404. Other vendor links (InvenSense, Bosch, ARK, CORVON) permanently redirect; point them at the final URL. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Hamish Willee <hamishwillee@gmail.com> * docs: fix remaining dead external links RaccoonLab's store product pages, the CubeNode pin descriptions, wiki.ubuntu.com/LTS and the Bitbucket NuttX repos no longer resolve. RaccoonLab now only has docs pages for these devices, so link those. ubuntu.sh and macos.sh already install kconfig-frontends, so the NuttX porting guide no longer tells Ubuntu users to build it from source; the old steps are kept in a comment. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Hamish Willee <hamishwillee@gmail.com> * docs: ignore unresolvable external links for 6 months These links fail the external link check for reasons a docs edit can't fix: bot blocking (raccoonlab.co, gitlab.arm.com), expired vendor certificates (svehicle.cn, jmarple.ai), and the ARKV6X-RT order link, which is waiting on a replacement URL from ARK. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Hamish Willee <hamishwillee@gmail.com> --------- Signed-off-by: Hamish Willee <hamishwillee@gmail.com>	2026-09-24 09:55:45 +10:00
PX4BuildBot	72acb3c3ec	docs: auto-sync metadata [skip ci] ... Co-Authored-By: PX4 BuildBot <bot@px4.io>	2026-09-23 21:51:09 +00:00
3 people	6bf2d5a91b	docs(param): remove stale RC preserve text from SYS_AUTOCONFIG (#28824) ... The SYS_AUTOCONFIG description still claimed RC* params are preserved, but #26798 removed that behavior. Drop the outdated sentence. Assisted-by: Cursor:composer-2.5-fast Signed-off-by: Luka Dragar <luka.dragar@ssrd.io> Co-authored-by: Luka Dragar <luka.dragar@ssrd.io> Co-authored-by: Cursor <cursoragent@cursor.com>	2026-09-23 15:44:29 -06:00
Jacob Dahl	a7646e2d6d	fix(drivers/imu/st/lsm6dsv): reset before telling the 16X from the 32X (#28823) ... CTRL8 bit2 identifies the 32X only at its reset value, and probe() read it before any reset. After a warm restart from firmware that cleared the bit (drivers without 32X support write the 16X CTRL8 encoding), a 32X was detected as a 16X and its accel ran at ±32 g while scaled as ±16 g. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Jacob Dahl <dahl.jakejacob@gmail.com>	2026-09-23 15:43:39 -06:00
Jacob Dahl	bb4e77b00c	fix(boards/ark_fpv): drive SPIX_SYNC only on the IIM-42653 revision (#28825) ... SPIX_SYNC is the 32 kHz CLKIN on IMU pin 9. On FMUM 1 that pin is the LSM6DSV32X INT2 push-pull output, so the MCU drove against it: the IMU die ran about 10 degC hotter and accel read about 0.19 g off on one axis. Assisted-by: Claude:claude-opus-5-5	2026-09-23 15:13:04 -06:00
Claudio Chies	2521e78a7f	fix(commander): allow RC calibration in SIH mode (#28821)	2026-09-23 17:39:03 +02:00
PX4BuildBot	c2526065e4	docs: auto-sync metadata [skip ci] ... Co-Authored-By: PX4 BuildBot <bot@px4.io>	2026-09-23 08:08:22 +00:00
bresch	d38eb8f5b3	fix(ekf2): update GSF composite yaw when predicting ... The composite yaw and variance were only recomputed when GNSS velocity was fused, freezing them when the speed accuracy was too poor while the models kept tracking the gyro, which risks an emergency yaw reset to a stale heading. Also reset the estimator after 60 s without fusion. Assisted-by: Claude:claude-opus-5 Signed-off-by: bresch <brescianimathieu@gmail.com>	2026-09-23 10:01:51 +02:00
Hamish Willee	e6d55568ae	perf(docs): skip rendering off-screen parameters in parameter reference (#28814) ... parameter_reference.md renders ~3100 parameters as ~111k elements, so the browser spends seconds on style recalculation and layout before the page responds (about 1.2 s of long tasks on a fast desktop and 6.8 s at 4x CPU throttling), and sidebar navigation to it takes 0.8-3.7 s. Wrap each parameter group and each parameter in a div, and give them content-visibility: auto so the browser skips off-screen content. The per-parameter level keeps deep links into large groups such as Actuator Outputs (701 parameters) fast. Group headings stay outside the wrappers so the page outline still tracks them. Exclude the generated page from Prettier: with the wrappers it needs over 4 GB of heap, which would break the docs metadata auto-sync. Assisted-by: Claude:claude-opus-5-5 Signed-off-by: Hamish Willee <hamishwillee@gmail.com>	2026-09-23 17:45:47 +10:00
mahima-yoga	68e9cbf23d	fix(commander): don't report "compass uncalibrated" while calibrating	2026-09-23 09:32:37 +02:00
PX4BuildBot	5fd9398462	docs: auto-sync metadata [skip ci] ... Co-Authored-By: PX4 BuildBot <bot@px4.io>	2026-09-23 05:21:00 +00:00