<beerman>
https://crux.nu/gitweb/?p=ports/opt.git;a=commitdiff;h=c7ed5f40d0c65439f6574a118a940850c488d02b <- isn't the release bump a little over the top here? also it might be needed to actually check if bash-completions is installed
<beerman>
the meson options file says its auto, so it should figure that fact out for itself but maybe it doesn't and we need to help it..
<jue>
beerman: the footprint changes, so a release bump seems appropriate IMO
<beerman>
ok, i mean doesn't really matter, but i just hope that the average CRUX user can solve these types of problems on their own and just pkgadd'ed
<beerman>
so they get to rebuild again with no further changes but that prt-get doesn't fail on the footprint
<jue>
auto-detect works fine for me, so no additional stuff is needed
<beerman>
and those that aren't able to do that will just sysup the next time and it will be ok ;)
<beerman>
ok
<jue>
well, a missing shouldn't happen, it's IMO a bug
<beerman>
it's not though
<beerman>
gtk4 is a good example
<jue>
for me it is a bug
<beerman>
if you have cups, it will produce a cups library, if you don't have cups, it will build something else
<beerman>
it used to be imo, and it's changing, we should change too to adjust for that
<beerman>
it should be a big fat warning, yes. but not fail right away
<beerman>
(i wish that gnome and gtk just dies, but that doesn't seem to happen anytime soon)
<jue>
well, for p11-kit the fix is really simple, I don't get it why you are against my commit, though
<beerman>
i am not against it
<beerman>
just wondering that you did it, tbh, and i wouldn't have bumped the release for that :)
<jue>
why I did it? Because the update failed, that's the reason
<beerman>
i support matts proposal to not make footprint mismatches a hard error, i expect it to get harder in the near future to produce reliable footprints
<beerman>
of course thats a fine reason, no question. all good
<beerman>
i think the topic this leads us too is more interesting. gtk4 is already a pain with cups/no cups
<beerman>
best i could do is edit the footprint manually and force a NEW with every build
<beerman>
which leads me to: maintaining is a lot of work, lets make it as easy as possible for us?
<jue>
I don't have gtk4 installed yet, so I cannot say anything about it, but editing of the footprint seems no very handsome
<beerman>
nope... we already started doing that with rust and i just picked it up for grafana because it will create these weird files with random hashes
<beerman>
go might be another candidate where this would make sense, but go, for me, takes the bullshit crown anyway.. buggy build system, i can't get it right in a container...
<beerman>
if i were given the chance, i would throw it in the same dumpster fire that i would dump gnome in..
<jue>
:)
<beerman>
no, for real. sometimes it produces a go binary, that will just segfault, other times (for no appaerent reason) it will generate no go binary at all
<beerman>
gcc-go doesn't help at all
<beerman>
it's just horrible. outside a container it behaves "fine".. so i need to pkgadd a built package from outside to build anything go... grafana, prometheus, pistol, glow, whatever else...
<beerman>
issues say its gcc 10.3 related and it shoudl be better with 10.4 or 11.x, but my 3.7 container doesn't behave differently :D