Small bugs in the code about apa102-pi

[pcjacobse]

I just checked out sebastian's fork. You should also merge in his commits, it looks like they fix a couple of other issues :)

[tinue]

Thanks for your comments! This explains why the LEDs sometimes did not clear up when I hit Ctrl-C ;-)
Usually they did, but this obviously depends on when in the entire cycle Ctrl-C is being hit.

The new samples are not quite debugged yet, I suspect that there will be some more errors present. For example I don't check at all for overflow, so for example stuff like "currentCycle*numLEDs+currentStep" is bound to break after a few hours, and with a lot of LEDs.

As for the fork: I don't understand the change. What error does it prevent? The clearStrip method is only to clear the strip e.g. before exiting the application (the comment explains this). Why would one want to release and reallocate the memory for the strip buffer?

-> Anyway, I'll have a more detailed look at your comments on the weekend.

[pcjacobse]

The clearing of the memory help if you only want to set specific leds after clearing the strip. I guess its not really usefullI for exiting the application, but when you built out some of the samples and want to clear the strip midrun, and after that just set a couple of pixels, you dont have to set all other leds to black manually.

And the other commit he did was swap the green and blue byte, since those are in the wrong order in your code. For example your colorwheel starts on blue instead of green.

[tinue]

Coloring is a different issue, that I haven't tackled yet: There are different strips out on the market. Some are rgb, others are bgr, and there are possibly even more. It will get tricky if strips of different makes are coupled together. For example the strandtest would suddenly change color on the boundary of different types.
It looks like I have a different strip than you have, because on mine the colors are correct.

[tinue]

I think I fixed the reported issues, so I am closing it. Feel free to reopen if you disagree.