<flokli>
blackriversoftwa: currently building your PR once again
lovesegfault has joined #nixos-aarch64
<blackriversoftwa>
flokli: is it taking a long time?
<blackriversoftwa>
I have a warm cache for it but it didn't seem like that big to build
<flokli>
I built it
<blackriversoftwa>
flokli: any remaining outstanding issues? It does still leak build-platform things into the runtime closure but all the issues I encountered on a Pi before are cleared up now
<blackriversoftwa>
except the part where the closure is >1GB
<flokli>
should coreutils really just point to coreutils' bin/ directory?
ryantrinkle has joined #nixos-aarch64
<flokli>
otherwise, the paths all seem to exist
<blackriversoftwa>
flokli yes
<blackriversoftwa>
that was what I had to fix with this last push
<flokli>
where does it get used?
<blackriversoftwa>
in the PATH for the channel fetcher scripts in corepkgs
<flokli>
right
<flokli>
cool
<blackriversoftwa>
it's expected to point at the /bin directory because it just gets added to PATH without any preamble
<blackriversoftwa>
I tried it the other way and it complained that it couldn't find `mkdir`
<flokli>
merged :-)
<blackriversoftwa>
Yay
<blackriversoftwa>
flokli: thanks
lovesegfault has quit [Ping timeout: 268 seconds]
<blackriversoftwa>
flokli: I'm starting to get an idea of why we get that huge gcc output in the closure.
<clever>
sometimes, its simpler to just edit nixpkgs directly
<blackriversoftwa>
Well I just got done doing that to make current stable nix not import the wrong bash
<blackriversoftwa>
when cross-compiled
<blackriversoftwa>
so no reason I can't do it again :P
lovesegfault has joined #nixos-aarch64
<blackriversoftwa>
clever: aha I might know why?
<blackriversoftwa>
I think there is some logic to auto-split /lib directories out, but since this is a cross-compiler the libraries get put under /nix/store/<gcc>/<system>/lib
<blackriversoftwa>
so build system libraries get dumped to .lib but not x-compiled libraries, I think?
<blackriversoftwa>
clever: yeah the `.lib` output just has x86 libraries
<clever>
blackriversoftwa: i think you need 2 outputs then, for x86 and arm libraries
<clever>
so it doesnt depend on the x86 libs at arm runtime
<blackriversoftwa>
clever: yeah but then do you need to explicitly tell everything downstream about the new output so it goes in the rpath?
<blackriversoftwa>
maybe
<blackriversoftwa>
it is very late here so I'll look at it tomorrow
<{^_^}>
#81716 (by eamsden, 1 minute ago, open): gcc: Place target system libraries in a separate output
<blackriversoftwa>
flokli: I'm fairly convinced that the reason there are still build-system paths in the output closure for `nix` is because they are runtime dependencies of `gcc`, so fixing this might finally clean all of those up
cidkid has joined #nixos-aarch64
<cidkid>
I see the boot selection screen has finally been merged
<cidkid>
will attempt to boot on the 5T, but not much more work will be done for 5T for awhile
<cidkid>
as mainline is booting and my main focus is helping the dude developing it
<cidkid>
we only get simple graphics using simplefb
<cidkid>
for now
cidkid has quit [Remote host closed the connection]
jackdk has quit [Read error: Connection reset by peer]
jackdk has joined #nixos-aarch64
TheNumb has quit [Read error: Connection reset by peer]
TheNumb has joined #nixos-aarch64
dongcarl has quit [Read error: Connection reset by peer]
TheNumb has quit [Ping timeout: 256 seconds]
TheNumb has joined #nixos-aarch64
rajivr___ has quit [Ping timeout: 240 seconds]
NekomimiScience has quit [Ping timeout: 252 seconds]
NekomimiScience has joined #nixos-aarch64
rajivr___ has joined #nixos-aarch64
ryantrinkle has quit [Ping timeout: 258 seconds]
zupo has joined #nixos-aarch64
ryantrinkle has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<samueldr>
>> If we reach a funding target of more than 490k Euro, we can offer Nemo Mobile as an alternative operating system for subsequent installation
<samueldr>
I wonder if nemo was involved
<andi->
490k€.. good luck
<samueldr>
I see nothing about mainline though
<samueldr>
only that the pinephone and librem 5 are mainlined
<samueldr>
>> and if it'll ever arrive to the point of running a stable, fully featured "mainline" Linux kernel
<samueldr>
and that sentence full of doubt
<andi->
Yeah
<andi->
ALso being able to make a phone in <7m sounds like they are just using some ready made device with a slightly changed user space.
<samueldr>
I don't doubt that they'll release the sources, even though mediatek is reportedly against it, others do release the sources
<samueldr>
oh, it *is* that, that's for sure
<samueldr>
I'm sure you could figure it out through ODMs
<samueldr>
maybe on taobao or alibaba you can find the same exact phone shell
<samueldr>
but slightly unlikely
<samueldr>
what might be more likely, is on source release, find who the ODM is and then find matching devices
<samueldr>
I forgot again the name of the ODM that makes some xiaomi devices
<samueldr>
I figure it's like all the other things: can't reboot without being near a computer
<gchristensen>
yeah
<flokli>
andi-: samueldr: forget about mainline kernel. I saw a similar promise with the gemini pda "once hitting xxk, you'll be able to install Debian on it as well"
<samueldr>
flokli: the thought never crossed my mind
<flokli>
Turns out, still uses the android kernel, and there's some libhybris stuff
<flokli>
they rolled their own distro on top of that
<samueldr>
yep
<flokli>
or at least some packages
<flokli>
they're buggy and not pleasant at all
<samueldr>
when even google is not arsed to pick up behind the vendor (chromeos team) I figure the vendor is hopeless
<flokli>
Not to talk about the amount of hacks required to get the sp-flash-tool to flash that mediatek device
<flokli>
mediatek is very "eeeh, don't touch, it might be contagious"
<flokli>
at the end, I gave it away to adisbladis :-)
<samueldr>
ah, I thought he got one for himself, not that it was hazardous waste from someone else ;)
<adisbladis>
Hazardous waste indeed
<adisbladis>
I spent way to much time on that thing already :P
<blackriversoftwa>
flokli: I think I figured out why there is still cross-contamination, and it is a more general issue with x-compiling gcc. In fact there's already an open PR to fix it: https://github.com/NixOS/nixpkgs/pull/58606
<{^_^}>
#58606 (by illegalprime, 48 weeks ago, open): [WIP] gcc: put target-specific libs in lib output
<adisbladis>
flokli: Actually for the purpose of mobile-nixos I'm pretty grateful it relies on hybris
<adisbladis>
I managed to test a bunch of stuff using their android containers with nix-built hybris
zupo has joined #nixos-aarch64
lovesegfault has quit [Quit: WeeChat 2.7.1]
<flokli>
blackriversoftwa: yeah, I saw the highlight. Just didn't get into reading through aaaall of this PR
<blackriversoftwa>
flokli: I tried patching it as matthewbauer suggested and merging in master
<blackriversoftwa>
it doesn't build
<blackriversoftwa>
the actual code changes are minimal
<blackriversoftwa>
and the principal seems correct
lovesegfault has joined #nixos-aarch64
lovesegfault has quit [Client Quit]
<flokli>
I currently can't sink too much time into it, sorry :-/
<flokli>
it might get better in 2 weeks
<blackriversoftwa>
flokli: no it's no problem: but do you think I should ask to take over? the original PR hasn't been updated since april last year
<blackriversoftwa>
in terms of being pushed to
<flokli>
if you want to work on it, sure.
<flokli>
Just push a new PR, link it to the old one, and ask s.o. to close the original one
<blackriversoftwa>
OK
<flokli>
I hijacked PRs that were far less stale than this one ;-)
<blackriversoftwa>
flokli: OK I am still new to the nixpkgs community so I am trying to be polite. But I'll probably swipe this one
<flokli>
if sb decides to keep pursuing the original authors intent, I assume they'll have zero problems with so else finishing it up
<blackriversoftwa>
Do you have time to push for it to be merged if I do?
<flokli>
I usually keep the original Author as author in the git commit, and add myself as Co-Authored-By
<blackriversoftwa>
flokli: hrm that isn't something I've had to do before: how to?
<flokli>
and ask the original author for some feedback, if possible