01:16 <sprin> for capnp list serialization in c++, you need to know the list size before initializing the builder, correct?

01:18 <sprin> so if you are munging some list before capnp serialization, you need a temporary data structure (eg, vector), to hold the result before you serialize it

01:18 <sprin> in other words, if we need to remove some elements from a list, the data flow is:

01:19 <sprin> orig list -> temp list -> capnp list (via builder)

01:20 <dwrensha> in theory, if you're removing elements, you could just decrease the length of the list

01:20 <sprin> oh!

01:20 <dwrensha> you would want to make sure that you've zeroed out the removed elements

01:21 <dwrensha> I don't think the c++ implementation exposing an API for such list-length adjustment

01:22 <sprin> I don't see any such API in the docs at least

01:25 <dwrensha> so sounds like you'll need a temporary list

01:25 <dwrensha> orphans may come in handy. https://capnproto.org/cxx.html#orphans

01:26 <sprin> interesting! I will see if that works, thanks dwrensha

02:33 <zarvox> "I don't always write tests, but when I do, they're 100 lines long and involve multiple kinda-ridiculous hacks"

02:33 <zarvox> https://github.com/sandstorm-io/sandstorm/pull/1262

17:07 <secher> Heyas all. I'm trying to see if Etherpad on Sandstorm can support syntax highlighting. Any ideas? It seems that Etherpad proper does, but I can't figure out how to enable it in Sandstorm either via plugins or any other fashion.

17:11 <dwrensha> hi secher!

17:12 <dwrensha> Sandstorm's Etherphad SPK comes with a bunch of plugins

17:12 <dwrensha> to add a new one, you would need to convince kentonv to add it to the package

17:13 <dwrensha> you could open an issue here: https://github.com/kentonv/etherpad-lite

21:08 <|jemc|> hat's the best place to read more details about the planned features/mechanisms for the "Powerbox UI"?

21:10 <dwrensha> |jemc|: there are a bunch of interfaces defined here: https://github.com/sandstorm-io/sandstorm/blob/master/src/sandstorm/grain.capnp

21:11 <dwrensha> |jemc|: they are only partially implemented as yet

21:11 <|jemc|> dwrensha: perfect, that should keep me busy reading for a while, thanks :)

23:20 <dwrensha> eek, apparently appdemo links are semi-broken at the moment

23:20 <dwrensha> you get a "match failure"

23:20 <dwrensha> and then you can click through and successfully continue

23:20 <dwrensha> looking into it....

23:30 <dwrensha> it tries to create a new grain after the demo identity is created and logged in

23:30 <dwrensha> but before an account is created

23:31 <dwrensha> I guess this is one downside of our new two-stage login dance

23:31 <dwrensha> ... it's now tricker to specify a callback that should take place directly after login