martinberlin
commented
2 years ago @cgreening test is failing:
In file included from src/main.cpp:16:
lib/Epub/Renderer/EpdiyRenderer.h:6:10: fatal error: EpdFrameBufferRenderer.h: No such file or directoryI think it’s because it’s using the Lilygo EPD47 environment. Is posible to set the environment dynamically?
on the other hand I get what you commented about preprocessor if conditionals. The question is if it’s worth the whole refactoring. To be sincere using EPDiy as a GFX library is the wrong concept because they are much more efficient and capable GFX components. But on the other hand like this you are saving a big refactoring and apart of M5 all the other epapers use EPDiy as a base component so I think although is not the cleanest alternative it’s one that can solve the problem and let you test to the end if M5 proves to work well. Then maybe at some point we can do a bigger refactoring. And for that we can draw a nice class dependence chart and try to figure out what it will be the best way to reorganize all the mess :)
cgreening
No worries I will never merge something if you mark it as "draft" :) I'm thinking in the direction that is not the double SPI config (Something you could also check disabling the SDCard and using SPIFFS)
But that epdiy is somehow instantiated and taking all DMA channels available before the M5 SPI config initializes. Only a theory but if the two SPI's are clearly differentiated and using different pins they should not conflict.