16:06 <yomimono> drup: halp how do i functoria

16:06 <yomimono> (specifically, see the comments on https://github.com/mirage/mirage-www/pull/486 )

16:07 <yomimono> I strongly suspect there's a "your doin it rong" in there somewhere, but I'm not sure where or how to fix it

16:08 <Drup> I'm impressed

16:09 <Drup> I ... will have to investigate

16:09 <yomimono> ....that it ever worked?

16:09 <yomimono> In that case I'll make a functoria issue I guess?

16:09 <yomimono> I think the workaround of just not specifying --tls false will get the tests passing for mirage-www, from my local tests, but that's very unsatisfactory

16:10 <Drup> let me look at this calmly during the evening ^^'

16:11 <yomimono> sure. I'm not even sure how to write up an issue about it so I'll just leave it for now I guess :)

16:11 <yomimono> thanks for looking at it!

16:11 <Drup> I'm impressed because functoria is not even supposed to build a graph that doesn't respect the invariants

16:12 <yomimono> yeah -- I saw the checks and exceptions for cyclic graphs and >1 root and became impressed then that we were still triggering the ocamlgraph exception

16:12 <Drup> yomimono: note that "mirage describe --dot" doesn't fully evaluate the graph for nodes that are options that are not specified, that's on purpose

16:13 <Drup> if you want to fully evaluate, you need to give --eval

16:13 <yomimono> aaaaaaaaaaaaaaah

16:13 <yomimono> thanks

16:14 <Drup> (this way, you can see in the dot the parameters that you didn't yet specified, and you have the option to see "what it would be if I took the default" with the --eval option)