Research window: Past 24 hours (2026-09-02 11:18 ~ 2026-09-03 11:18 Beijing Time) Sources: RT-Thread official news API, GitHub (org: RT-Thread / main repo commits+PR / packages index / topic:rt-thread ecosystem repos), Google News aggregation (zh-CN / en-US, via proxy), etc. (see Appendix: Source Verification Table for details)

Index

• I. Kernel and Version Updates
  • Main repo merged [lwp]: move terminal/ out of global CPPPATH (09-03)
  • In progress: CAN framework structural refactoring PR pending review, led by community developer (09-02)
• II. Official and Community Events
  • Official news API sees no new entries for the fourteenth consecutive day, elexcon 2026 countdown at 6 days (09-03)
• III. Ecosystem and Vendor Adaptation
  • canopennode-rtt feature branch remains active, no new commits in mainline window (09-02)
• IV. Industry Applications
  • Ecosystem reference: consumer drone H120D firmware reverse-engineering documentation updated (09-02)
• V. Summary and Trend Observations
• Appendix: Source Verification Table

1. Kernel and Version Updates

1.1 Main repo merges [lwp]: moving terminal/ out of the global CPPPATH

Date: 2026-09-03 (merged 09:20 Beijing time)　Source: PR #11776 (merge commit 54e5164, submitted by TTLLDD)

components/lwp/SConscript previously added the ./terminal/ directory to the global CPPPATH; this commit removes it. A repo-wide check shows that all unqualified includes of terminal subsystem headers (terminal.h, tty_config.h, bsd_porting.h, etc.) come from source files inside components/lwp/terminal/, which the compiler’s quoted-form lookup already satisfies, so there is no need to leak an internal directory into the global include path. This is the minimal demonstration landing of the convention agreed in the November 2025 issue #10980 “Standardize the use of CPPPATH” — components no longer leak internal directories into the global include path, a small step forward in build-system/include hygiene. It is the only merge in the main repo during the window; the weekday-morning merge cadence matches recent weeks, with no major features.

1.2 In progress: community developer submits structural refactor PR for the CAN framework

Date: 2026-09-02 (created 13:21 Beijing time, open, CI clean, not draft)　Source: PR #11779

wdfk-prog (the author of the CANopenNode port package that was promoted to the official index yesterday) submitted a refactor[CAN] PR: a structural refactor of the Generic CAN Framework and the STM32 bxCAN backend (2 commits, 8 files, +2902/-1308), redefining the framework’s responsibility boundaries — Generic CAN handles hardware resource management, synchronization, ownership, and lifecycle, while TX buffering/retry/scheduling and protocol QoS policies are left to the upper protocol stack or the application. The proposal traces back to two threads: the 8-26 feature request #11759 (CANopen CiA 305 LSS Activate Bit Timing runtime baud-rate switching requires TX Full Flush and Hardware Abort capabilities; now closed) and the 8-28 framework refactor discussion #11767 (unifying TX Engine, Mailbox Ownership, and Lifecycle; this PR was opened 3 minutes after the discussion closed). Both issues were closed within the window and funneled into this PR, which carries the labels BSP: STM32 and component: drivers/can. The PR elevates the low-level capability gaps exposed by the CANopenNode port into a framework-level refactor proposal; it awaits maintainer review, and if merged it will be the largest adjustment to the Generic CAN Framework’s responsibility boundaries in recent years.

2. Official and Community Activities

2.1 Official news API sees no new items for the fourteenth consecutive day; 6 days until elexcon 2026

Date: 2026-09-03　Source: RT-Thread official news API

The official news API (rt_org category) still holds 5 old entries, the latest still the 2026-06-17 Abu Dhabi investment forum report — no new items for the fourteenth consecutive day. Multiple supplementary gnews checks (RT-Thread when:3d in English and Chinese, RT-Thread 睿赛德 when:3d, 睿擎派 OR 睿擎 when:3d): English results remain Rotten Tomatoes word-splitting noise and unrelated tech posts, and all Chinese queries returned empty. The next external window is elexcon 2026 in Shenzhen on September 9-11 (RuiSaide exhibiting; the 8-30 preview has been logged), 6 days away; the continued pre-event quiet period is the expected cadence.

3. Ecosystem and Vendor Adaptation

3.1 canopennode-rtt feature branches remain active, no new commits on mainline during the window

Date: 2026-09-02 (repo pushed 22:30 Beijing time)　Source: github.com/wdfk-prog/canopennode-rtt

One of the topic:rt-thread repos pushed during the window. The latest commit on the default branch master is still the 09-02 11:02 Beijing “co_rtt_mainline_deadline_min changed to extern to work with mainline scheduling” (logged yesterday); pushes during the window landed on the two feature branches A04 and ds402 — continued evolution of branches for different CANopen profiles (DS402 motion control, A04 application), with no new content on mainline. The same day, the repo’s author proposed the CAN framework refactor PR in the RT-Thread main repo (see 1.2), a two-way push from “application layer (CANopenNode package) → framework layer (Generic CAN refactor)”, making CAN the most active ecosystem thrust in the community right now.

4. Industry Applications

4.1 Ecosystem reference: consumer drone H120D firmware reverse-engineering docs updated

Date: 2026-09-02 (pushed 13:34 Beijing time)　Source: H4ck3rTR/h120d-protocol

The topic:rt-thread ecosystem repo h120d-protocol was updated during the window (README revision, consolidating the April Hackaday submission). The project reverse-engineers and documents the WiFi communication protocol of the consumer drone Holy Stone H120D and the RT-Thread internals in its firmware, and provides control tools, with code mostly in C++. There were no substantive code updates during the window — this is documentation maintenance; the project’s existence itself corroborates RT-Thread’s deployment in consumer drone firmware (a claim by the reverse-engineering project itself, not officially verified), logged as a low-activity ecosystem reference.

5. Summary and Trend Observations

• Kernel side: a single merge during the window — [lwp] moving terminal/ out of the global CPPPATH (09-03 09:20), continuing the landing cadence of the build/include hygiene convention (issue #10980), with no major feature merges.
• In progress: the CAN framework structural refactor PR (#11779, +2902/-1308) was submitted by community developer wdfk-prog, CI clean and not draft, awaiting maintainer review; the late-August CAN discussion (#11767) and feature request (#11759) have both been closed and funneled into this PR, giving the refactor direction community-side consensus.
• Official side: the official news API has seen no new items for 14 consecutive days; the next external window is elexcon 2026 in Shenzhen on September 9-11, 6 days away.
• Ecosystem side: the two-track CAN push (application layer expanding profile branches such as DS402 after CANopenNode’s promotion + framework-layer refactor proposal) is currently the clearest community thrust; the package index saw no new merges during the window, and the Tencent IoT SDK config PR #2034 remains open.
• Three-day trend (09-01 ~ 09-03): kernel robustness batch of 4 PRs, CANopenNode event-driven scheduling, RTC utest and the two packages Fudan Micro FM33/CANopenNode entering the index (09-01) → Tencent IoT SDK config PR, CAN framework refactor PR proposed, h120d docs updated (09-02) → lwp CPPPATH hygiene merged, CAN refactor PR awaiting review, 14th day of official silence (09-03). The main thrust has shifted from “package index expansion” to “deep CAN ecosystem push + build-system standardization”.

Appendix: Source Verification Table