jevans3142 / adc-bar-lx Goto Github PK

SD card IO is currently done inline with refreshing the screen. IO could be performed in the background to improve snappyness

No proper setup to ensure these arrive in sensible configs if no SD card settings loaded?

Control of building front floods and other FOH DMX devices: would require a major rewrite to accomodate second control panel, possibly second DMX output

This is because the UART is triggering a data event every 120 bytes which is the default - during a break event the last set of bytes get lost.

See:
espressif/esp-idf#4537

Believed to be when output routine can't take the mutex in time from the input code when the two threads are trying to both read and write the incoming DMX buffer at the same time. Solution will be to double buffer or similar.

When unplugging an external lighting desk the values should not hold but instead be reset to zero for the HTP merge

draw_byte, which is called by the text drawing functions, defaults to using 'WHITE' draw mode regardless of what parameter was passed to the calling function.

When using the up and down buttons in menu screens, holding the up and down button should speed up entry to reduce button hammering

Across all files. Sort out using signed vs. unsigned. Remove unneeded data checks >0 when using unsigned.

Total screen refreshes as implemented in screen_refresh() are very visually apparent. Blanking the display during the refresh hasn't made it better, neither has altering the I2C speed.

Possible fixes

• Implement a partial refresh for only a defined area of the screen (wish to do anyway for live data etc)
• We're creating and deleting the I2C command link many, many times per refresh. Why did I write it like this? Surely this is stupid...

• Makes it easier to edit and review on computer
• More resilient to changes - can use default values for anything not present. See similar storage module in sm-desk-video-switcher project