[RFC] Improve rhizome_sync.c (follower of #117)

As noted by @lakeman in #117 annouce_cli_bundle add "... is not doing it's job when the rhizome store was empty; https://github.com/servalproject/serval-dna/blob/30aa1c16b85a8c8362403d92d9d62b698e461333/rhizome_sync.c#L255"

To make this work I changed rhizome_sync.c to work in my usage scenario. It now passes my own tests and also all of rhizomeprotocol.

Please give comments on the changed file: rhizome_sync.c.txt

I provide the whole file (not only diffs) because I think it is easier to understand.

I'm especially interested in:

Do the comments added really reflect the original intention of the commented functionality? In the course of understanding rhizome sync I had to deduce from the source code. I'm in no way an expert in rhizome sync.
Is the general scheduling mechanism chosen acceptable?
- The alarm for rhizome_sync_announce is now running all the time, only the interval is changed depending on neighbours available or not.
- Additional stati are added for announcements and broken payloads. This was necessary as updates on bundles did not lead to a new announcement, the same for broken/ missing payloads.
Are the renamed function names acceptable? To help me to better understand the functions I tried to give them names that reflect their function and/or associate them to the data structures the function is working on. I only renamed static functions, so no change in the interface.
Is rhizome_exists(&m->filehash) really sufficient to check for a missing or broken payload? See line 713 in rhizome_sync.c.
Does a comparison of two BARs to be done by bcmp(&bar1, &bar2, ..) or is there are better way to compare whether two BARs reference the same bundle? See line 144 in rhizome_sync.c.
What else do I have to make to make my variant of rhizome_sync.c acceptable to the serval project?

Happy new year Bobby

servalproject / serval-dna