<gchristensen>
okay, a major milestone I feel comfortable messaging about here: the RAID array / mirror is clean and recovered after a one of the disks was added
<qyliss>
Profpatsch: have you tried mruby for your rubyMinimal use case?
<samueldr>
qyliss: I've been working with mruby for 1> week right now and it probably wouldn't work for our use case
<MichaelRaskin>
gchristensen: sounds like a bit of a relief!
<samueldr>
it's not compatible with the ruby ecosystem directly
<qyliss>
samueldr: But a rubyMinimal would?
<qyliss>
It disables basically the whole stdlib
<gchristensen>
MichaelRaskin: definitely ready to go for dinner and a beer :'D
<samueldr>
depending what they're doing with their minimal rubies :)
<qyliss>
well, for this to be in nixpkgs it has to be something that's generically useful
<samueldr>
it's definitely not a drop-in replacement for mattz's ruby though
<samueldr>
like, there's no require support without an mrbgem
<samueldr>
matz*
<samueldr>
(and funny enough, both are from matz)
<samueldr>
and I'm not convinced the simple package in nixpkgs, as it is, is useful for mruby, due to the way mrbgems are expected to be built with the interpreter
<samueldr>
most of the time I spent with mruby was figuring out a neat builder thing for it, so you can give it mrbgems
<samueldr>
adding `bundlerApp = bundlerApp.override({ ruby = rubyMinimal; });` to asciidoctor's callPackage is not enough to make it use rubyMinimal it seems
<gchristensen>
ermm that page looks to be saying we're not, but I'm pretty sure it is just because the statistics pipeline hasn't made it to prometheus yet
<gchristensen>
I'd like to see a PR marking everything in the very narrowest part of the stdenv bootstrapping phase marked as big-parallel, plus other large builds marked as big-parallel
<samueldr>
gchristensen: it already is big-parallel
<gchristensen>
nice
<gchristensen>
and then switch a certain % of machines to have big-parallel in mandatoryFeatures
<samueldr>
what could make it go for longer than 4h builds?
<lovesegfault>
aarch64 contention?
<samueldr>
I was thinking that it's building on a machine where a buttload of builds are going on
<gchristensen>
quite possibly
<samueldr>
which could starve the parallel builds of the kernel, and maybe even make bad iowait
<gchristensen>
but, forgive me, I'm done thinking about hydra for the next 8h :)
<samueldr>
yeah :)
<samueldr>
I rescheduled to build
<gchristensen>
back later -- possibly tomorrow morning, though at a minimum after a taco and beer.