Added error handling

[ufanders]

I found many avenues where this library could lock the system and patched them so we can always return.

• Fixed AS7261 hardware ID.
• Changed protos for all functions with Wire failure potential.
• Added fcn exits for Wire failures.
• Added local timeout for measurement fcn.
• Added retry limit.

[nseidle]

Thank you for this! Much better with your improvements. But do please explain how the retries line works. We realize it works just fine, but how? It has me and a few fellow engineers debating post increment variable tests.

[ufanders]

Hi @nseidle, thanks so much for merging I think this will save people a lot of work down the road. The retries line works because post-(in/de)crement operators execute after the variable is used in the conditional statement. pre-(in/de)crement operators execute before the variable is used. More here: https://www.geeksforgeeks.org/pre-increment-and-post-increment-in-c/

=]

[nseidle]

Thank you! We assumed this but that doesn't explain how we see x stopping at 255. uint_8 wraps at 255 so:

x = 254
(x++ > x) equates to "is 254 > 254?" No, so x becomes 255

x = 255
(x++ > x) -> is 255 > 255? No -> x=0

x = 0
(0++ > 0) -> is 0 > 0? No-> x=1
...

I know this doesn't reflect reality, but what am I missing?

[ufanders]

Durrr yeah I made the assumption the number of retries would never be set to such a high number, so it should break under that condition! 😅

[nseidle]

number of retries would never be set to such a high number

I am even more confused. The number of retries is not set as far as I can see. I assumed you were relying on a roll-over trick to cap the retries at 255. Am I mistaken? What was your intent with the retries line?

[ufanders]

I define retries as 3 here: https://github.com/sparkfun/SparkFun_AS726X_Arduino_Library/blob/master/src%2FAS726X.h#L172

Other than that, I dunno!

[nseidle]

Ah! Now I see you basically have a typo but got really lucky :)

if(retries++ > retries) return 0xFF;

should be

if(retries++ > RETRIES) return 0xFF;

We just got lucky that (retries++ > retries) fails at 255 and not infinity. I'll fix it.

[ufanders]

Oh boy, thanks for catching that! Now to make sure I'm not doing the same in my production code...

As an aside, I remember talking directly to you over email sometime when I was in college (2001-2005). I'm thrilled to see how SparkFun has become bedrock for the embedded community, and how responsive the team still is 20 years later.

=]

[nseidle]

Np!

Crazy! I'm pleased (and just as shocked as you are!) that we're still around. Thanks again for helping out the open source world. Let us know if anything else pops up.