Repository navigation
Conversation
1、Add Sonoff valve subdriver lazy loading. 2、Modify comment language and file name
|
I have tested this with a physical Sonoff SWV2C device. It pairs successfully in the SmartThings app, and both valve controls work normally. |
That's the easiest part, but does it stay on for longer than 10 minutes or does it automatically turn itself off after 10 minutes? What happens after - say - 15 minutes? The device is completely useless when it turns itself off after 10 minutes. Users don't want to turn it on every 10 minutes if they want to irrigate for an hour. Discussion: It's a bit more complicated (link to (outdated) code and technical documentation): |
|
Another issue: before a channel is turned on, the opposite channel must be turned off. According to the official manual only one channel should be open at any time. |
|
I tested this with a physical Sonoff SWV2C device. For the 10-minute behavior: yes, the device has an internal default auto-close/manual watering duration. In my test, if a valve is opened normally, the device closes it automatically after the configured duration. The default value appears to be 10 minutes. Users can change the irrigation/manual watering duration through the device's own configuration path outside of this basic valve control behavior. For the dual-channel behavior: I also tested the mutual exclusion behavior on the physical device. The two channels cannot stay open at the same time. If channel 1 is open and then channel 2 is opened, channel 1 is automatically closed and channel 2 opens. The device firmware reports the valve state automatically when a valve is operated, and the SmartThings app state updates correctly from those reports. If preferred, I can update the driver to explicitly send an Off command to the opposite endpoint before opening the requested channel, so the SmartThings state stays aligned with the device behavior. |
|
Thanks for the feedback. I’d prefer to keep this PR focused on SWV2C. SWV2C is a different hardware product from SWV1C. SWV1C is a single-channel valve, while SWV2C is a dual-channel valve with two endpoints and mutual-exclusion behavior between the two channels. There is already a separate PR for SWV1C, and that PR is focused on the single-channel device. Since SWV2C requires different handling for the second endpoint and child device, I think it is clearer and safer to review SWV2C in this dedicated PR instead of combining both hardware products into one submission. This also keeps the scope of each PR aligned with one hardware product and one WWST certification flow. |
| device:try_update_metadata({ provisioning_state = "PROVISIONED" }) | ||
| if device.network_type == st_device.NETWORK_TYPE_ZIGBEE then | ||
| device:emit_event(capabilities.valve.valve.closed()) | ||
| device:emit_event(capabilities.battery.battery(100)) |
There was a problem hiding this comment.
what's the source for the 100% battery.
Battery events should be created as the response for the battery report that comes from device.
There was a problem hiding this comment.
We should also not have the valve events emitted without the device telling us. If the device is actually open it would not be accurately reflected in the app for the user.
There was a problem hiding this comment.
@yezi289 Please change the code according to provided remarks. Thank You.
| [capabilities.valve.ID] = { | ||
| [capabilities.valve.commands.open.NAME] = valve_open_handler, | ||
| [capabilities.valve.commands.close.NAME] = valve_close_handler, | ||
| } |
There was a problem hiding this comment.
This should be handled by the framework.
Please check without that handler.
There was a problem hiding this comment.
@yezi289 Did You check how Your device works without those handlers ?(default handling should work)
| local function valve_open_handler(driver, device, command) | ||
| device:send(OnOff.server.commands.On(device)) | ||
| device:send(OnOff.attributes.OnOff:read(device)) | ||
| end | ||
|
|
||
| --- valve.close | ||
| local function valve_close_handler(driver, device, command) | ||
| device:send(OnOff.server.commands.Off(device)) | ||
| device:send(OnOff.attributes.OnOff:read(device)) | ||
| end |
There was a problem hiding this comment.
This should be handled by the framework.
Please check without those handlers.
There was a problem hiding this comment.
@yezi289 Please confirm if Your device reports new state itself after the command (without the need to read the attirbute here)
|
Profile category check: ✅ Passed - all profiles have a category defined. |
|
Duplicate profile check: Passed - no duplicate profiles detected. |
|
Invitation URL: |
|
zigbee-valve_coverage.xml
Minimum allowed coverage is Generated by 🐒 cobertura-action against 92b8fbe |
Test Results 73 files 544 suites 0s ⏱️ For more details on these failures, see this check. Results for commit 92b8fbe. ♻️ This comment has been updated with latest results. |
|
I tested the updated Edge driver (commit a4565e0) with a physical SONOFF SWV2C on a SmartThings v3 Hub. Both valve channels can be controlled from the SmartThings app, and the app reflects the device-reported valve states. Switching from channel 1 to channel 2 closes channel 1 and opens channel 2 as expected. No device firmware changes were made. |
|
Thank you for the review. I pushed e5e09ae to remove a duplicate health_check assignment reported by Luacheck; this does not change driver behavior. I also checked the Driver Tests failures on the previous commit. The GitHub Actions test results reported 3,294/3,294 passing, and I ran all 44 zigbee-valve tests locally against both the 0.57 and 0.58 Lua libraries, with 44/44 passing on each. The external Jenkins statuses for 57, 58, and dev do not include a log or a link to the failing test cases, so I cannot yet determine whether those failures involve this driver. Could you please share the failing test names/logs, or rerun those checks on this commit? I will address any failures attributable to this PR. Thank you. |
|
Quick update: I merged the current upstream main into this PR. Driver Tests for 57, 58, 59, 60, 61, and dev now all pass. This suggests the earlier 57/58/dev failures were related to the outdated base branch, so the failure logs I requested are no longer needed. This update did not change the SWV2C driver logic or device firmware. Thank you, and I would appreciate a re-review when convenient. |
|
Conflicts will need to be fixed prior to merging. Note the single failing tests is a known intermittent failure in matter test framework. |
Check all that apply
Type of Change
Checklist
Description of Change
Summary of Completed Tests