For this very reason I also want to invest time in more robust testing. All that said, I do not see any more delays leading to the release at this moment. (2/2)
The new Skytable will land by the end of this month. Probably earlier.
Several major low level cluster bugs in the spec that were previously undetected have been detected and fixed. I’m currently working on pushing what will be the final spec for 0.9…(1/2)
Great progress again. I keep saying this but indeed - last few steps. I will be finishing up work on the control console, automated testing, pending misc. work & docs this weekend. Monday is the day perhaps?
Isn't it great when your:
1. test suite explodes
2. you compile the test suite target and summon gdb
3. gdb hits a SIGABRT
4. the trace points at a rust proc macro and provides absolutely no information whatsoever?
5. back to caveman style dbg!()
Happy debugging thursday!
And yes, we're actually very close this time.
There's a bunch of things remaining: (1) porting the remaining clustering implementation code from the testing code base to our code base (2) updating client drivers (3) compatibility options (4) documentation.
Yes, there were delays. Plenty of them.
But they delivered something better: a stack of new features and extensions for faster performance, easier scaling, simpler deployments, and powerful querying.
The new Skytable ships in just a couple of weeks :)
I am exploring options to automate documentation generation but since "public API" is not a straightforward construct for a database server, we'll likely have to find some silly way to do it ... for example, we previously used an external file to generate docs. We'll see.