refactor(bldc_haptics): standard OTA / coredump / haptics modules + discovery by finger563 · Pull Request #768 · esp-cpp/espp

finger563 · 2026-09-04T17:57:54Z

Modernize the bldc_haptics example: standard OTA / coredump / haptics modules + discovery

Brings the bldc_haptics example fully in line with the other framed-USB examples
and wires it into dispatcher capability discovery (#767).

1. Migrate to `espp::Dispatcher` + advertise

The example drove a raw stream_frame::StreamParser with a manual
frame.module == kModule filter. It now routes through an espp::Dispatcher and
advertises itself for capability discovery, so the browser Device Hub lists it.

2. Split OTA / core dump onto their own modules

Instead of multiplexing an OTA subset + a crash-report command into the haptics
protocol (module 2), the example now runs the standard protocols on their own
dispatcher modules:

OTA → module 0 — a handler speaking the espp ota_stream protocol, reusing
the existing espp::Ota. The plain ota_console.html updates this device.
Core dump → module 4 — an espp::CoreDump + espp::CoreDumpService (full
download/erase). The dump is no longer erased at boot (the console downloads +
erases it); the boot summary log stays. coredump_console.html works against it.
Haptics stays module 2 — commands + telemetry only.

All three are registered + advertised (OTA / Core Dump / BLDC Haptics), so the hub
discovers a device with three modules and links to each dedicated console.

3. Drop the inline OTA / crash UI from the haptics console

webapp/index.html loses its Firmware-update panel + crash fetch (and the
associated protocol types / plumbing) — the hub + dedicated consoles cover it.
PROTOCOL.md / README updated accordingly.

Verified

Firmware builds clean, manager-off, on IDF v6.0.1 (esp32s3).
Haptics web console passes node --check with no dangling references.

🤖 Generated with Claude Code

The bldc_haptics example was the one framed-USB example still driving a raw stream_frame::StreamParser (with a manual `frame.module == kModule` filter) instead of an espp::Dispatcher, so it could not advertise itself to the browser Device Hub. Migrate it and advertise. - Replace the raw parser + module filter with an espp::Dispatcher: register the haptics protocol on module 2 (proto::kModule) with a handler that gates on !is_reply() (unchanged semantics), and feed()/reset() the dispatcher from the USB RX worker. The example's OTA subset + haptics commands all ride module 2's message types, so this stays a single-module protocol. - Advertise it: register_module(..., ModuleInfo{"BLDC Haptics", "haptics_console.html", ...}), set_device_info(usb_cfg.product), and serve_discovery() routed through the existing tx_mutex-guarded usb_send. - Add the dispatcher component to the example's EXTRA_COMPONENT_DIRS and to the main component's REQUIRES. Also complete the dispatcher README module-id table (add 2 = BLDC haptics, 6 = MCP266). This finishes the set: every framed-USB example (ota, coredump, can_bridge, mcp266, haptics) now advertises over the reserved discovery module 0xFF. Verified: builds clean, manager-off, on IDF v6.0.1 (esp32s3). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Copilot

Warning

Copilot couldn't run its full agentic review because it didn't start before the timeout. Make sure your repository has a runner available, or add a copilot-code-review.yml file specifying one with the runs-on attribute. See the docs for more details.

Pull request overview

Migrates the bldc_haptics framed-USB example from a raw StreamParser + manual module filtering to espp::Dispatcher, and enables capability discovery advertising so it can be listed by the browser Device Hub.

Changes:

Replace StreamParser parsing/filtering with espp::Dispatcher routing on module 2.
Advertise the module and serve discovery responses via the existing USB TX path.
Wire the new component dependency in CMake and update dispatcher module-id documentation.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

File	Description
components/dispatcher/README.md	Adds missing module-id entries for BLDC haptics (2) and MCP266 (6).
components/bldc_haptics/example/main/bldc_haptics_example.cpp	Switches to `Dispatcher`, registers/advertises module 2, and routes discovery replies through USB send.
components/bldc_haptics/example/main/CMakeLists.txt	Adds `dispatcher` to the example main component requirements.
components/bldc_haptics/example/CMakeLists.txt	Adds `dispatcher` to `EXTRA_COMPONENT_DIRS` for the example build.

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

+  // same tx_mutex-guarded usb_send as every other frame.
+  dispatcher.serve_discovery(
+      [&](std::span<const uint8_t> f) { usb_send(std::vector<uint8_t>(f.begin(), f.end())); });


github-actions · 2026-09-04T18:21:42Z

✅Static analysis result - no issues found! ✅

Bring the haptics example in line with the other framed-USB examples: instead of multiplexing an OTA subset + a crash-report command into the haptics protocol (module 2), run the STANDARD protocols on their own dispatcher modules so the device is discovered as three modules and the plain ota / coredump web consoles work against it directly. - OTA -> module 0: a handler speaking the espp ota_stream protocol (Begin/Data/ Finish/Abort -> make_ok/make_error), reusing the existing espp::Ota. Removed the Ota* message types from haptics_usb_protocol.hpp + the cases from the haptics handler. - Core dump -> module 4: an espp::CoreDump + espp::CoreDumpService (full download/ erase protocol). The dump is no longer erased at boot (the coredump console downloads + erases it); the boot-time summary log stays. Removed GetCrash/Crash from the haptics protocol + the raw esp_core_dump usage. - Haptics stays module 2 (commands + telemetry only). - Register + advertise all three (OTA / Core Dump / BLDC Haptics); add the coredump component to the example deps. Web console cleanup (drop the now-moved OTA + crash UI) follows in a separate commit. Verified: builds clean, manager-off, IDF v6.0.1 (esp32s3). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Now that OTA is dispatcher module 0 and core dump is module 4 (not multiplexed into the haptics protocol), remove their UI from the haptics console: the device hub discovers all three modules and links to the standard ota_console / coredump_console, which speak those modules directly. - webapp/index.html: remove the Firmware-update (OTA) panel + upload flow, the GET_CRASH fetch, the OTA/crash message types + progress plumbing, and the now- unused constants/CSS. Haptics controls + telemetry are unchanged. - PROTOCOL.md: module 2 now documents haptics only; note OTA=module 0 / coredump=module 4 as separate standard protocols. - README.md: point firmware update / crash inspection at the ota / coredump consoles (and the device hub) instead of an inline panel. Web app passes node --check; no dangling references remain. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

…e checks The haptics console refactor left a call to updateUploadUI() (removed along with the OTA upload panel) inside setConnectedUI(). Since setConnectedUI runs early in connect(), it threw a ReferenceError before the RX pump and the init handshake started -- the page connected (dial showed "waiting for telemetry") but never sent GET_INFO / SET_STREAMING, so no telemetry ever arrived. Drop the stale call. Also harden the ota / coredump / haptics WebUSB consoles: - default device filter is now VID-only (any espp device, VID 0x1209) instead of VID+PID, so a multi-module device is offered in every console; "show all devices" uses {filters:[{}]} (ota_console's invalid acceptAllDevices fixed). - on connect each console queries the reserved dispatcher discovery module (0xFF ListModules, reusing Dispatcher::describe's TLV) and warns if its own module is not advertised, then continues anyway; firmware without the discovery module simply never replies and is tolerated silently. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

The OTA cases that used this std::error_code moved to handle_ota_frame (which has its own), leaving handle_frame's `ec` unused -- cppcheck's unusedVariable failed the static-analysis gate. Remove it. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Copilot

🔵 Needs a closer look

The cross-layer firmware, protocol, discovery, and console changes warrant final human review despite only minor findings.

Review details

Suppressed comments (3)

components/bldc_haptics/example/PROTOCOL.md:47

The protocol no longer has two unsolicited frame kinds after OTA_PROGRESS was removed; only telemetry remains. Please make this singular so implementers are not told to expect a second undocumented event.

sending the next. Two device-to-host frame kinds may arrive *unsolicited* and
must be tolerated at any time:

- `TELEMETRY (0x93)` — when streaming is enabled.

components/bldc_haptics/example/webapp/index.html:513

The adjacent comment still says the default matches this example's exact VID/PID, but the changed filter now matches every PID under the espp VID; the checkbox broadens that further to all USB devices. Update the comment to match the selector behavior.

          : { filters: [{ vendorId: DEFAULT_VID }] };  // any espp device (VID 0x1209), any PID

components/coredump/web/coredump_console.html:617

The adjacent comment still describes an exact example VID/PID default and says the checkbox only adds other espp devices. The new selector already includes every espp PID by default, while the checkbox includes all USB vendors, so the documentation is now misleading.

        : { filters: [{ vendorId: DEFAULT_VID }] };  // any espp device (VID 0x1209), any PID

Files reviewed: 10/10 changed files
Comments generated: 2
Review effort level: Balanced

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Addresses the review note on the discovery reply path: usb_send took a `const std::vector<uint8_t>&`, so the serve_discovery and CoreDumpService reply callbacks (which receive a std::span) had to copy their bytes into a fresh vector on every reply. usb_send now takes `std::span<const uint8_t>` and calls write_vendor (which already accepts a span) directly: - the vector-building callers (proto::build / ota_stream make_*) still work via the implicit vector->span conversion, no copy; - the discovery and coredump callbacks forward their span with no allocation. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Copilot AI lite review requested due to automatic review settings September 4, 2026 17:57

Copilot AI reviewed Sep 4, 2026

View reviewed changes

finger563 and others added 2 commits September 4, 2026 13:29

finger563 changed the title ~~feat(bldc_haptics): advertise for dispatcher capability discovery~~ refactor(bldc_haptics): standard OTA / coredump / haptics modules + discovery Sep 4, 2026

Copilot started reviewing on behalf of finger563 September 4, 2026 18:40 View session

finger563 and others added 2 commits September 4, 2026 16:14

finger563 requested a balanced review from Copilot September 4, 2026 21:41

Copilot started reviewing on behalf of finger563 September 4, 2026 21:56 View session

Copilot AI reviewed Sep 4, 2026

View reviewed changes

Comment thread components/bldc_haptics/example/main/bldc_haptics_example.cpp Outdated

Comment thread components/bldc_haptics/example/main/haptics_usb_protocol.hpp Outdated

finger563 and others added 3 commits September 4, 2026 17:00

Fix comment formatting in haptics_usb_protocol.hpp

403e3a2

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Clarify coredump web console functionality

109fca9

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Navigation Menu

Uh oh!

refactor(bldc_haptics): standard OTA / coredump / haptics modules + discovery - #768

refactor(bldc_haptics): standard OTA / coredump / haptics modules + discovery#768
finger563 wants to merge 8 commits into
mainfrom
feat/haptics-advertise

finger563 commented Sep 4, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

github-actions Bot commented Sep 4, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

finger563 commented Sep 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Modernize the bldc_haptics example: standard OTA / coredump / haptics modules + discovery

1. Migrate to espp::Dispatcher + advertise

2. Split OTA / core dump onto their own modules

3. Drop the inline OTA / crash UI from the haptics console

Verified

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

github-actions Bot commented Sep 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

✅Static analysis result - no issues found! ✅

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🔵 Needs a closer look

Suppressed comments (3)

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

finger563 commented Sep 4, 2026 •

edited

Loading

1. Migrate to `espp::Dispatcher` + advertise

github-actions Bot commented Sep 4, 2026 •

edited

Loading