STM32H7B3I - TouchGFX Application Framework: MVP and the ModelListener

RE: STM32H7B3I - TouchGFX Application Framework: MVP and the ModelListener

Jan Cumps — Fri, 28 Aug 2020 22:33:21 GMT

For my own reference later (I need these conversions): From Ascii / Unicode / UTF8 tricks:

[embed:dc8ab71f-3b98-42d9-b0f6-e21e02a0f8e2:3ef06e6b-9b7d-4651-b12d-db8be7f5e991:type=c_cpp&text=++touchgfx%3A%3AUnicode%3A%3Asnprintf%28currentConnectedWifi.password%2C+64%2C+%22%25s%22%2C+password%29%3B%0A++touchgfx%3A%3AUnicode%3A%3Asnprintf%28currentConnectedWifi.wifi.ssid%2C+32%2C+%22%25s%22%2C+wifiAccessPoints%5Bid%5D.ssid%29%3B%0A++touchgfx%3A%3AUnicode%3A%3Asnprintf%28currentConnectedWifi.wifi.encryption%2C+30%2C+%22%25s%22%2C+wifiAccessPoints%5Bid%5D.encryption%29%3B%0A++currentConnectedWifi.wifi.signalStrength+%3D+wifiAccessPoints%5Bid%5D.signalStrength%3B%0A++%0A++touchgfx%3A%3AUnicode%3A%3AtoUTF8%28password%2C+password_ascii%2C+64%29%3B%0A++touchgfx%3A%3AUnicode%3A%3AtoUTF8%28wifiAccessPoints%5Bid%5D.ssid%2C+ssid_ascii%2C+32%29%3B%0A++touchgfx%3A%3AUnicode%3A%3AtoUTF8%28wifiAccessPoints%5Bid%5D.encryption%2C+encryption_ascii%2C+30%29%3B]

RE: STM32H7B3I - TouchGFX Application Framework: MVP and the ModelListener

Jan Cumps — Sun, 23 Aug 2020 17:17:05 GMT

demo of the interface in its current state:

[View:https://www.youtube.com/watch?v=PjsQ1yuo5TA:1110:699]

RE: STM32H7B3I - TouchGFX Application Framework: MVP and the ModelListener

Jan Cumps — Sat, 22 Aug 2020 19:09:16 GMT

There is a reason to design this mechanism between the single Model class and all screens.

It's a 1-to-n relation, and you don't want to put all those "if screen 1 then", "if screen 2 then" forks in the model.

It has one duty: be the mediator between the LCD and the backend. It should not have to deal with screen layout, flow, actions.

The other way doesn't need the same abstraction, because there's only one Model. And the screens are allowed to know the model.

That's why the ModelListener deals with any message coming from the Model, going to he screens.

But when the screens want to ask the Model something, there's no compelling reason for abstraction. The single Model has the duty to deal with Backend requests. And the screens should be able to just invoke those functions.