introduce new_with_rmt_driver() to be able to use multiple mem blocks for TxRmtDriver to mitigate flickering

[weber-software]

I had the issue that my output was flickering like describe over here: #33

Using two mem blocks solved the flickering entirely for me (and i think for many others).

[aligator]

I tried it and on my first tests this improved it a lot.
Thanks for this finding.

I stil got veeeery few flickerings, but in dev mode. I expect it to be even better in production mode.

My test just used one extra thread with a webserver set up using std threads.
i did not change the CPU core. (Just the default stack size)
Led code was directly in the main thread.

Before this change the leds freaked out with this configuration.

[cat-in-136]

@weber-software Thank you for great survey and PR.

This fix also corresponds to the contents of the following URL, so it is a good workaround.

• https://docs.espressif.com/projects/esp-faq/en/latest/software-framework/peripherals/rmt.html#when-esp32-rmt-controls-the-ws2812-light-strip-with-wi-fi-or-bluetooth-enabled-there-are-some-data-frame-exceptions-how-to-solve-the-issue
• http://esp32.io/viewtopic.php?t=19298

But, the fundamental fix should be to be able to change mem_block_num. For example, there is a report on the latter forum post that says the issue was solved by setting mem_block_num = 8.

[weber-software]

@cat-in-136 i agree that it would be a good idea to make the mem_block_num configurable. In which way would you prefere?

• adding mem_block_num as a parameter to new
• adding another method new_with_mem_block_num
• adding a builder
• a different solution that i don't see

[cat-in-136]

@weber-software It would be appropriate to create a builder for Ws2812Esp32RmtDriver. I havenot investigated yet what to do for smartled and embeddedgraphics: passing driver as argument or create a builder for each?

(The existing new() function should be retained. These legacy functions are still useful for convenience.)

[weber-software]

@cat-in-136 i scribbled around a bit and came to the conclusion that it could be useful to have a function that accepts a TxRmtDriver. That way we could keep complexity out of this library. What do you think?

[cat-in-136]

@weber-software Could you please my review comment on lib_smart_leds.rs? Thank you.

[cat-in-136]

Thanks a lot!