Re: [nft PATCH 1/4] monitor: Fix printing of range elements in named sets

[Phil Sutter <phil@nwl.cc>]

On Wed, Jul 12, 2017 at 09:05:45PM +0200, Phil Sutter wrote:
> On Wed, Jul 12, 2017 at 06:30:47PM +0200, Arturo Borrero Gonzalez wrote:
[...]
> > Could this logic be merged into that function? My goal is to only
> > print from one code path.
> 
> Yes, that makes sense. I found it too cumbersome to squeeze the
> additional logic into netlink_events_setelem_cb(), hence why I went with
> a separate function. I'll give it another try. In doubt I'll move the
> printing logic into a separate function to be called from both places.
> 
> I'll prepare a v2 tomorrow, also merging the previous two patches as
> suggested.

Just a quick status update: It's a mess. ;)

There are so many different cases, I actually started drawing flow
diagrams (can't remember when I did that last time). In addition to what
we discussed already, I realized that via 'nft -f', I can make multiple
changes to even different sets within a single transaction - this
requires dealing with cached half-open ranges everywhere, not just in
NEWGEN callback. Another trap is 'nft flush set': The elements are
reported in reverse order. Anyway, I have something that seems to work
but needs quite some cleanup before I dare to publish it. :)

I should probably look into ways to write tests for this to get all the
cases covered.

Cheers, Phil