16:59:48 <jcajka> #startmeeting Go SIG meeting
16:59:48 <zodbot> Meeting started Mon Aug  3 16:59:48 2020 UTC.
16:59:48 <zodbot> This meeting is logged and archived in a public location.
16:59:48 <zodbot> The chair is jcajka. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:59:48 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:59:48 <zodbot> The meeting name has been set to 'go_sig_meeting'
16:59:55 <jcajka> #topic Roll Call
17:01:06 <jcajka> #chair alexsaezm
17:01:06 <zodbot> Current chairs: alexsaezm jcajka
17:01:17 <jcajka> #chair Eighth_Doctor
17:01:17 <zodbot> Current chairs: Eighth_Doctor alexsaezm jcajka
17:06:22 <jcajka> there are again :) no tagged issues so moving on to the open floor
17:06:28 <jcajka> #topic Open Floor
17:07:48 <jcajka> from my side go1.15rc1 managed to get in to the mass-rebuild, from FTBFS I have checked at random non seem as golang bug/issue(either something that has been seen in the scratch rebuild or infra issues)
17:08:01 <Eighth_Doctor> so there's something odd going on with go 1.15
17:08:11 <jcajka> Eighth_Doctor: snapd?
17:08:15 <Eighth_Doctor> mborzecki noticed that it's ignoring the compiler flags we're passing in the snapd package
17:08:22 <Eighth_Doctor> jcajka: yep
17:09:53 <jcajka> Eighth_Doctor: go1.15 now supports internal linker for the pie mode, you either need to force linkmode=external(ldflag) or make sure that the snapd doesn't use any C code so internal linker generates static(linux) binary(AFAIK)
17:10:15 <jcajka> pie mode on x86_64 and aarch64 that is it
17:10:34 <Eighth_Doctor> ah
17:10:44 <Eighth_Doctor> I'll have to look at that sometime later
17:10:48 <Eighth_Doctor> but good to know
17:10:51 <Eighth_Doctor> (annoying, though...)
17:11:01 <Eighth_Doctor> I wish Go compiler behavior wasn't so unstable
17:12:18 <jcajka> Eighth_Doctor: IMHO it is more on the side that many expect static binaries in "wrong" way or wrong use cases(but AFAIK snapd shouldn't be the case),  let me know if that works(it should)
17:12:50 <Eighth_Doctor> yeah, I'll have to test it
17:13:05 <Eighth_Doctor> you're right that a lot use this the wrong way, but this is one of the rare cases where it makes sense
17:13:46 <Eighth_Doctor> jcajka: also, carlwgeorge is likely going to need some help with bringing caddy 2.x into Rawhide, so I'd appreciate it if you could be on hand to help :)
17:14:17 <Eighth_Doctor> upstream seems to be unwilling to commit to sa stable set of dependencies, so we're going to be forced to bundle them until they calm down
17:14:17 <jcajka> from what I have been checking, snapd is using C code outside of the stdlib(it will not build with CGO disabled), so I hope that it will work
17:15:54 <carlwgeorge> i mean, if we bundled i probably won't need much help
17:16:15 <Eighth_Doctor> carlwgeorge: I'm mostly concerned about go compiler changes, and the go mod stuff
17:16:42 <jcajka> carlwgeorge: hopefully :), what I see in the koji is some broken deps/api
17:16:47 <Eighth_Doctor> I expect that we'll eventually unbundle caddy, but I don't see us doing that right now with upstream literally swapping dependencies like mad in point releases
17:17:22 <carlwgeorge> it's a rather hostile upstream if i'm being honest
17:17:48 <jcajka> at worst will be most probably not using macros as there might be need for something like "-mod=vendor"
17:18:18 <carlwgeorge> jcajka: i'm not sure what eclipseo was doing with the most recent caddy builds, i haven't had time to dig in
17:19:13 <jcajka> deps in Go are unfortunately  kind of hostile on their own...(from distro perspective)
17:20:08 <jcajka> carlwgeorge: IIRC caddy haven't show up on my list during scratch rebuild as it has been already FTBFS in the rawhide
17:21:42 <jcajka> anyway fell free to ping me with any Go related FTBFS, I will try to help or at least point in the direction of the solution
17:24:18 <jcajka> any more topics?
17:25:13 <Eighth_Doctor> none from me
17:27:25 <jcajka> last call :)
17:30:36 <jcajka> #endmeeting