=== cpaelzer_ is now known as cpaelzer
[19:24] <minimal> caribou: I've been looking into the IPv6 default route prefix issue and think I see where the problem is occurring
=== stw__ is now known as stw
[22:02] <falcojr> minimal: whoops, sorry, just saw this now. I just put up a PR for it
[22:02] <falcojr> (if we're talking about the same issue)
[22:19] <minimal> falcojr: ah ok, yes it's the same issue. I had a quick look at your PR but need to test it to understand what exactly the change does ;-)
[22:20] <falcojr> basically, if there was a falsey value in the config, we'd strip it out. A prefix of 0 is falsey, so it'd get stripped out. Later we'd see no prefix and put in a default one
[22:20] <falcojr> the change says "0 is valid, don't strip it out"
[22:20] <minimal> ah "0" as falsey rather than numeric 0, right that makes sense
[22:21] <minimal> yes I'd figured out the prefix was being stripped and so the big if then elseif block was falled through to the bottom and assuming /64 for IPv6 and /24 for IPv4
[22:22] <minimal> I was in the middle of testing some other IPv4 & IPv6 routes to see if the problem only affected defaults or also others
[22:23] <falcojr> if you already have them written/running, it'd be good to run them against my PR
[22:23] <falcojr> I ran it against the current tests, but we're probably missing some use cases there
[22:24] <minimal> locally I've been creating a set of various static IPv4, static IPv6, static IPv4 & IPv6, SLAAC-related etc configs to test as I'm working on some other network renderering issues
[22:25] <minimal> so far only testing network config v2 but also want to check v1 as well
[22:26] <falcojr> that's good. v2 probably needs more love anyway as it's the less common version
[22:28] <minimal> yeah, I'm trying to figure out various DHCPv6 related stuff for eni that are NOT documented as supported, e.g. privext, request_prefix, dhcp, autoconf
[22:32] <minimal> falcojr: I settled in v2 in general quite some time ago as there was stuff I was using that I couldn't do via v1
[22:32] <falcojr> minimal: Do you remember what kind of stuff? Just curious
[22:33] <minimal> falcojr: hmm, I'd have to dig throgh notes/files to figure out what, might have been vlan related, or renaming interface related
[22:34] <falcojr> gotcha, not a big deal. Was just curious
[22:36] <minimal> falcojr: ah, maybe it was route related, with v1 you can't associate routes with particular interfaces from memory
[22:37] <falcojr> oooh...yeah, I tried unifying how v1 and v2 get represented under the hood and ran into that
[22:37] <minimal> falcojr: another v2 routes issue is something I'm trying to fix currently