16:17 <kentonv> ugh every time I get disconnected, XChat fails to re-join to #sandstorm since it tries to join before verifying my identity and so then I lose a day or two of logs before I notice and rejoin myself

16:19 <kentonv> abliss, yeah "unbound" is just (relatively-)unlimited CPU time. node-to-node messaging is not announced yet.

16:21 <kentonv> and yeah our "broad language support" really just meant showing people how you can transpile a lot of stuff to JS. TBH that actually works better than Wasm for most "managed" languages.

19:21 <JacobWeisz[m]> Ian, is what I'm talking about making sense?

19:24 <abliss> it makes sense to me.... but it's an inherently weird topic and I doubt there's any good answer in the real world. mounting the server's /usr/share/zoneinfo into the grain probably isn't a security hazard, but it does reduce reproducibility (an app might work on one server and not on another)

19:25 <JacobWeisz[m]> I suppose alternatively Sandstorm could package and include it. It'd still be much more current than relying on what people stuffed in an SPK.

19:26 <JacobWeisz[m]> I see those zoneinfo files grabbed in a surprisingly large number of packages.

19:27 <abliss> i think this may be one of those rare cases where the isolation model of sandstorm just crashes against the real world and it just doesn't quite fit.

19:28 <JacobWeisz[m]> But it probably doesn't cost much to fix. Apps should work fine when serves current tzdata instead of what they had when they were packaged, right?

19:29 <JacobWeisz[m]> s/serves/served

19:29 <abliss> i'd guess: probably, yes

19:30 <JacobWeisz[m]> Not an expert, but I suspect the format of those files is fairly set in stone if so many Linux platforms and packages depend on them.