fix #74 - ensure table are initialized before flushing them

[duritong]

fix #74

[nbarrientos]

Hehe, hacky fix. You're lucky that executing table inet foo {} when the table foo exists does not make nftables complain 😆

[duritong]

I didn't really see a way to conditionally check for the existence of the table

[traylenator]

I noticed this myself after adding that original patch - but could not think of a good way.
The other option is just to completly remove the flush from start up since it is redundant. We would rely on the flush in reload only.

Lets go with this unless you think removing the flush completly is a better idea.

[nbarrientos]

Perhaps a good time to create a release, there are two interesting fixes in master now.

[duritong]

Question: do you know why the author was rewritten when merged? @traylenator

[traylenator]

This commt 92e0fcb is written by you ?

[duritong]

I pushed & proposed to merge a4c3c76 but merged was 92e0fcb (also on master is this) - which is content-wise the same commit, except that the author got re-written, thus a different commit-sha.

So I was wondering why the author got re-written and whether you clicked anything special.

Background: When I now try to merge back upstream (this repo) I need either to merge or throw away my old commit, but this means I need to internally force push, which is also weird. Otherwise github started doing that which would be a really weird thing to start doing...

[traylenator]

whether you clicked anything special.

I clicked what ever was the default option when doing the merge assuming GH knows best.
I will take note next time what the options are and try and preserve.