LED_BUILTIN is NOT defined for all ESP32 boards - some do not have a built-in LED! Instead, Status LED now defaults to Pin 13, as opposed to LED_BUILTIN. Also, added a new method, homeSpan.getStatusPin(), to return the pin number used for the Status LED, whether or not it remains the default (13) or is changed by user with homeSpan.setStatusPin(pin). This method is now used in the DEV_Identify.h file for each example, instead of using LED_BUILTIN (which otherwise won't compile for boards without a built-in LED)
WiFi connection now occurs asynchronously so that if WiFi is lost, or cannot be connected at start-up, the device still operates intead of blocking while it tries to connect. The new logic should allow for a wider range of chips to connect.
Program now halts if more than 41 Accessories are defined. This allows for a bridge to have 40 Accessories. HAP limit is 150, but there is not enough memory in ESP32 to robustly handle more (testing up to 50 LightBulb Services worked, but was sluggish).
1) Used `const char *` instead of `char *` where appropriate, including the need to create a dummy blank string for us in certain places.
2) Set initialization of WiFiClient to 0 instead of NULL, since WiFiClient is not a pointer (probably don't need to set it to anything since WiFiClient overrides the boolean operator anyway).
3) Cleaned up some of the messaging and logic when WiFi tries to connect so that users know to wait a bit.
4) Only remaining warning messages are for casting SpanService to (void *), which I think i unavailable (and is not forbidden).
To Do: Go through examples and check for warnings (will likely need to convert `char *` to `const char *` in many places.
* if Accessory ID is NOT specified, the default is aid=1+aid of last Accessory, with aid of very first Accessory always set to 1.
* aid of first Accessory must ALWAYS be set to 1 - if user over-rides with another value, error will be thrown
* validation now includes checking for duplicate aids, as well as ensuring the first Accessory is always aid=1
If not specified, "aid" defaults to an index number representing the order in which the Accessories were created. Values are stored as uint32_t (which is more limited than HAP requirements of uint64_t, but makes the code easier).
TO DO: Must add validation to ensure duplicate aid values are not used.
button(int pin, boolean isLong) changed to button(int pin, int pressType), where pressType can be SpanButton::LONG, SpanButton::SINGLE, or SpanButton::DOUBLE. Updated Example 16 and confirmed everything works as expected.
To do: Review all prior examples and update SpanButton when needed. Also need to update Zephyr Vent Hood code!
Added ServiceLabelIndex Characteristic, which seems to be required as noted in HAP specifications. However, linking to a ServiceLabel Service does not appear to be required, so there is no need to build any Service-Linking logic into homeSpan - will wait for use-case to be identified.
To do: Add support for double-click to SpanButton, then include in Example 16.
Modified SpanCharacteristic::sprintfAttributes to streamline logic and add special handling for the ProgrammableSwitchEvent Characteristic as required by HAP: the value returned for the database or any read request must be set to null (i.e. "value":null). The only time the real value should be returned is when the device sends a EV/Notify message (when the button is pressed). Verified that the example works as expected!
TO DO: Add functionality to allow for Service Namespace and Index label so that you can have multiple programmable pushbuttons in one service - this requires logic for HAP LINKED SERVICE functionality.
Added validation to ensure all Accessories have a AccessoryInformation Service and all Accessories have a HAPProtocolInformation Service, unless the device is configured as a bridge, in which case only the first Accessory requires a HAPProtocolInformation Service.
Checks to ensure that: (A) Characteristic is allowable (i.e. either required or optional) for the defined Service, and (B) Characteristic has not already been instantiated for the defined Service.
The 'A' command now resets any existing WiFi functionality so that it can start an Access Point without needing to first reboot.
To Do: Add a method to cancel out of apConfigure() and reboot by pressing control button.
To Do: Create time-out for commandMode()
Changing code so that it is possible to run HomeSpan without configuring WiFi. HomeSpan can already run without being paired, but requires WiFi at start-up. No reason to force WiFi to be used if device has buttons to operate the appliance withouth the need for WiFi or a connection to HomeKit. Work in progress...
Zephyr3, the latest version, was moved into its own repo and renamed ZephyrVentHood.
Deleted Zephyr1, Zephyr2, and Zephyr3, which are no longer needed.
Flattened folder structure for Examples and placed them all in top-level directory
Started adding lists to the first few Services in Service.h. This is a longer-term project to enable HomeSpan to validate a configuration and check that all Services have all required Characteristics and that any optional Characteristics defined are indeed permitted.
Full initialization now depends on "isInitialized" flag instead of monitoring Wifi connection, to protect against a loss of WiFi causing a re-initialization, instead of just a wifi reconnect with initWifi() alone.
SpanButton now instantiates PushButton instead of keeping track of its own triggering logic. This means that SpanButton longPress() is triggered by holding down the button for longTime WITHOUT the need to release. This is a desired change.
Next: Must update Example 15 to state that SpanButtons trigger a long press withouth needing to be released.
To do: Consider adding a "repeating mode" for longPress in which the SpanButton re-triggers every longTime while button continues to be pressed. Best way to implement is probably by creating a PushButton::extend() method.
Also moved StatusCode definitions from Settings.h to HAPConstants.h. There was no reason to have the user return a StatusCode of OK or UNABLE from update() since the only other choice was BUSY and HomeKit does not seem to distinguish errors. Either it's okay or not okay, which can more easily be represented by a simply return of TRUE or FALSE from update().
The user now needs to know nothing about StatusCodes. All Examples were modified to change StatusCode update() to boolean update() and return(StatusCode:OK) to return (true).
Much simpler.
Convention is that every *.cpp references, in this order:
1) Any required core librries
2) Its own *.h (i.e. Foo.cpp -> Foo.h)
3) All other *.h files needed by the cpp, UNLESS those *.h files are already in its own *.h file. Must include all *.h files that are not in its own *.h file even if it is others that are referenced.
This change allows the user to set the log-level with homeSpan.setLogLevel() instead of needing to change in Settings.h. Next up: Update Example that introduces VERBOSE to instead use setLogLevel(). Consider adding an interactive 'L' command to change this during run-time.
Also, added "#pragma once" to every *.h file to prevent duplication of definitions. TO DO: Review all #include to ensure sanity across all files.
These optional methods can be "chained" when instantiating a new SpanService to configure the Service as Primary or Hidden (or both). This alleviates the need to ServiceType to be propogated through the Service constructor.
Next Up: Eliminate ServiceType construct and update Examples to use setPrimary instead.
Created separate init() methods for PushButton and Blinker. Created HomeSpan:setControlPin and HomeSpan:setStatusPin that can be called from main setup() to modify the defaults. All default settings being migrated to Settings.h
Next Up: Add method to break out of WiFi connection loop by using push button in case WiFi SSID/PWD has changed but there is no desire to unpair controllers. To Do: Move blinking status light setting to named macros.
Setup Code now saved after serialConfig() or apConfig().
Next up: Create 'U' command to unpair controller. Change function of control button so it unpairs (short press) or resets WiFi (long press).
To do: Revisit statusLED blinking paterns to make sure they are clear and consistent - save them as macros in settings.h
Gregg