Fix wrong check on observers retaintion

[jiayihu]

I actually had this commits for a long while, but wanted to check to be sure it worked correctly, then forgot to ever open the PR.

• It adds the log feature to help with debugging
• It adds the create_notification method to create the packet which notifies observers about an update
• It corrects the ack check, ensuring that the ack corresponds to the last notification and not for some old one

[martindisch]

Good to hear from you again! I made some changes that I'd like you to review (you might have to git fetch && git reset --hard origin/master since I rebased) and have some questions too:

1. In https://github.com/martindisch/coap-lite/pull/14/commits/105dc977e99673a9a3ad9c6f019b3d2ff37d91e3 I added a short description to the new function you introduced so it's not the only publicly exposed functionality without documentation. Is what I wrote correct or should it be something else?
2. In https://github.com/martindisch/coap-lite/pull/14/commits/412b310c11b3e5d9294d8a9f18a9a89675aef73b I removed a #[macro_use] that Clippy told me was unnecessary. Could you test that this doesn't break anything, even when using the log feature?
3. Is there a reason why in the case where the log feature isn't activated the macro does let _ = $arg; in order to turn it into a no-op instead of actually doing nothing (e.g. {})?
4. Do you think it would be useful to have an explicit feature for log? Something like

   [features]
   with-log = ["log"]

   It's not technically necessary since having an optional dependency adds an implicit feature with the same name, but it would make it more obvious to potential users that the functionality exists. I don't mind either way, you decide.

[martindisch]

I'm ready to merge this as soon as you give my changes the green light. Just a small ping in case this slipped through the cracks, I'm in no rush at all :-)

[jiayihu]

Sorry, time flies so fast these days! 1 and 2 okay 3 The honest answer is that I don't know. Macro definitions are still a mystery to me 😄 The implementation was taken from https://github.com/smoltcp-rs/smoltcp/blob/master/src/macros.rs IIRC 4 I think it would make sense for better discoverability, but smoltcp doesn't do it. I guess we can assume one knows about the implications of optional = true.

Thanks for the patience, it really slipped through my mind :)

[martindisch]

Macro definitions are still a mystery to me

Same here ;-) Guess we'll leave it like that, it doesn't hurt and the compiler removes it anyway. Thank you.