Community appreciation

[2bndy5]

This PR fixes a couple changes I made in #691 that seemed to upset community members concerning backward compatibility.

1. open pipes 0 & 1 in begin()noticed in write(...) returns false on sucess unless stopListening() is called first. #734
2. enableAckPayloads() also enables dynamic payloads for pipe 1 automatically. Reddit and Arduino forum users have gotten quite upset about this one (to which my reaction has been "RTFD" or "pay more attention to examples' comments").

Concerning WiringPi build

Also, I'm disabling the WiringPi CI build to discontinue "Linux build CI failed" notifications that I keep getting. I've successfully built WiringPi locally on my RPi2 (and subsequently RF24 built successfully) using upstream WiringPi repo... I think that libcrypt.so.1 is not available for cross-compilation on x86 Ubuntu, but it seems to ship with armhf-based OSs (like Raspberry Pi OS). Pretty sure that libcrypt is part of the libc6 std libs, but I may be wrong. Lastly, I've changed the Linux build workflow to use the upstream WiringPi fork per @CoRfr recommendation (but it is disabled anyway).

[Avamander]

@2bndy5

a) I wouldn't revert 1. and 2. if it can cause hassle for those that have already updated their sketches. We should've maybe said better that behavior changes, but sketches should contain the version of library they were tested with.

b) stopListening being required before write would maybe be under write's docs? I will update the release notes to contain info about the change as well.

c) Is there a way enableAckPayload could make sure the ACK feature is enabled and usable without wasting a lot of performance?

[2bndy5]

Both points shouldn't cause any hassle with either old or new sketches.

Point 1 actually aligns with what the datasheet says about PoR condition (still recommend calling stopListening()...).

Point 2 would cure problems that some people are experiencing with sketches prior to v1.3.11; namely those that aren't aware the docs and examples indicate ACK payloads require dynamic payloads feature.

Is there a way enableAckPayload could make sure the ACK feature is enabled and usable without wasting a lot of performance?

The true problem is that using dynamic payloads (pre-req for ACK payloads) requires configuring 2 registers (DYNPD & FEATURE -- unfortunately the nRF24 slave device doesn't support configuring sequential registers with 1 SPI transaction), so I don't think performance can be improved. Additionally, when dynamic payloads are enabled, RX nodes without dynamic payloads enabled cannot receive payloads from TX nodes with dynamic payloads enabled, so it's probably best to not call enableDynamicPayloads() from within enableAckPayload(). TMRh20 and I brushed upon this topic when he disputed my notion to make dynamic payloads enabled by default, and the notion seemed to have consequences on RF24Network due to supporting counterfeits/clones.

stopListening being required before write would maybe be under write's docs?

Yes! Can't believe I didn't add that. Is that what you call "footgun"? Maybe also add a comment to examples about this? This requirement is nothing new, so adjusting any release notes seems redundant to me. Actually, there is a code snippet in write()'s docs that demonstrates proper usage of stopListening(), but maybe that's not enough. It's more likely that people don't know how to navigate the docs, so they don't bother reading them (yet another reason to update the doxygen docs to a more modern "theme" -- I have made progress on that).

[2bndy5]

@TMRh20 I think a new release is in order. I can retroactively adjust the release description with my current privileges if that makes it easier for you. For my next trick, I want to overload begin() (for Arduino platform only) to take a reference to a non-default SPI object, but an addition like that would be better categorized as v1.4.0 because it's a new feature (not just a patch/hotfix).

EDIT: Please see #737 first

[TMRh20]

I think a new release is in order. I can retroactively adjust the release description...

Sure, I added info about my change, so if you would add info for yours that would be great!

For my next trick, I want to overload begin()

Cool, will wait to see how it comes out

[2bndy5]

if you would add info for yours that would be great!

done. I get a notification about releases also, so I will continue to add info about my submitted changes in the future (if necessary).