<Sonarpulse> bgamari: ah yeah
<Sonarpulse> that's hit even in native builds
<Sonarpulse> no idea
<Sonarpulse> but I assume it has something to do with broken setup hook
<Sonarpulse> due to different setup hook logic
yegortimoshenko has joined joined #nixos-dev
<bgamari> Sonarpulse, yes, it's very odd
<bgamari> I'm trying to build GHC
<bgamari> and hitting it
<bgamari> Setting SOURCE_DATE_EPOCH to something later in preConfigure seems to work around it
orivej_ has joined joined #nixos-dev
orivej has quit [(Ping timeout: 272 seconds)]
orivej_ has quit [(Ping timeout: 248 seconds)]
orivej has joined joined #nixos-dev
mbrgm has quit [(Ping timeout: 264 seconds)]
mbrgm has joined joined #nixos-dev
nix-build has joined joined #nixos-dev
nix-build has quit [(Changing host)]
nix-build has joined joined #nixos-dev
{^_^} is now known as Guest63354
Guest63354 has quit [(Read error: Connection reset by peer)]
alunduil has joined joined #nixos-dev
<bgamari> Sonarpulse, why doesn't binutils-wrapper wrap nm?
alunduil has quit [(Quit: leaving)]
yegortimoshenko has quit [(Ping timeout: 264 seconds)]
phreedom has quit [(Remote host closed the connection)]
ma27 has joined joined #nixos-dev
ma27 has quit [(Ping timeout: 272 seconds)]
taktoa has joined joined #nixos-dev
taktoa has quit [(Ping timeout: 264 seconds)]
taktoa_ has joined joined #nixos-dev
phreedom has joined joined #nixos-dev
vcunat has joined joined #nixos-dev
taktoa_ is now known as taktoa
taktoa has quit [(Remote host closed the connection)]
pie_ has quit [(Ping timeout: 256 seconds)]
taktoa has joined joined #nixos-dev
ma27 has joined joined #nixos-dev
ma27 has quit [(Client Quit)]
ma27 has joined joined #nixos-dev
pie_ has joined joined #nixos-dev
ma27 has quit [(Ping timeout: 240 seconds)]
goibhniu has joined joined #nixos-dev
ma27 has joined joined #nixos-dev
ma27 has quit [(Ping timeout: 268 seconds)]
lezed1 has quit [(Quit: Connection closed for inactivity)]
laggy_wifi has joined joined #nixos-dev
__Sander__ has joined joined #nixos-dev
stqism has quit [(Ping timeout: 264 seconds)]
disasm has quit [(Ping timeout: 264 seconds)]
disasm has joined joined #nixos-dev
coup has joined joined #nixos-dev
coup is now known as Guest32460
pie_ has quit [(Ping timeout: 240 seconds)]
vcunat has quit [(Quit: Leaving.)]
orivej has quit [(Ping timeout: 264 seconds)]
pie_ has joined joined #nixos-dev
yegortimoshenko has joined joined #nixos-dev
Mic92 has quit [(Ping timeout: 255 seconds)]
contrapumpkin has quit [(Ping timeout: 265 seconds)]
<gchristensen> is a change in boost actually a change in the stdenv?
<gchristensen> see also https://github.com/NixOS/nixpkgs/pull/32832 -- I'm surprised this was tagged a stdenv rebuild
vcunat has joined joined #nixos-dev
pie_ has quit [(Ping timeout: 260 seconds)]
JosW has joined joined #nixos-dev
pie_ has joined joined #nixos-dev
orivej has joined joined #nixos-dev
contrapumpkin has joined joined #nixos-dev
yegortimoshenko has quit [(Read error: Connection reset by peer)]
<gchristensen> Sonarpulse: why didn't you merge a more recent master in to staging? I thought you were going to do it to fix the eval error
<vcunat> gchristensen: do you have nix on that/some RHEL6 derivative?
<vcunat> (trying to finish that glibc stuff)
<gchristensen> you bet I do!
<vcunat> when you nix-store --realize /nix/store/047hnqfir5fai40f7j9w8hzv1ycfw271-hello-2.10
<vcunat> what does this print?
<vcunat> /nix/store/047hnqfir5fai40f7j9w8hzv1ycfw271-hello-2.10/bin/hello
<gchristensen> $ /nix/store/047hnqfir5fai40f7j9w8hzv1ycfw271-hello-2.10/bin/hello
<gchristensen> FATAL: kernel too old
<gchristensen> 2.6.32-431.el6.x86_64
<vcunat> Ah, I missed a piece of code. Let me try to patch the check without disabling it completely.
<gchristensen> ok
taktoa has quit [(Ping timeout: 268 seconds)]
pauldub has joined joined #nixos-dev
ma27 has joined joined #nixos-dev
ma27 has quit [(Quit: WeeChat 1.9.1)]
ma27 has joined joined #nixos-dev
<vcunat> Hmm, probably. Building hello on Hydra so you can try it out: https://hydra.nixos.org/build/66357650
<vcunat> (rebuilding gcc is a bit uncomfortable)
<gchristensen> cool :)
<vcunat> The plan is to simply ignore the version check if (and only if) the kernel is exactly 2.6.32
<gchristensen> oh cool
laggy_wifi has quit [(Remote host closed the connection)]
yegortimoshenko has joined joined #nixos-dev
laggy_wifi has joined joined #nixos-dev
pie_ has quit [(Remote host closed the connection)]
pie_ has joined joined #nixos-dev
laggy_wifi has quit [(Remote host closed the connection)]
laggy_wifi has joined joined #nixos-dev
<Sonarpulse> bgamari: no reason, let's make it!
<Sonarpulse> gchristensen: I wanted to base my PR off the master commit, for a known good state for a reference point
<gchristensen> ack
<gchristensen> I was hoping staging would get the eval fix
<Sonarpulse> I could have merged more into staging too, all i needed to do was ensure the PR didn't contain too many commmits according to github
<Sonarpulse> but I figured staging was like only 2 master commits behind, so this was low impact
ma27 has quit [(Quit: WeeChat 1.9.1)]
ma27 has joined joined #nixos-dev
FRidh has joined joined #nixos-dev
ma27 has quit [(Ping timeout: 240 seconds)]
propumpkin has joined joined #nixos-dev
contrapumpkin has quit [(Ping timeout: 240 seconds)]
propumpkin is now known as contrapumpkin
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
FRidh has quit [(Client Quit)]
FRidh has joined joined #nixos-dev
Mic92 has joined joined #nixos-dev
FRidh has quit [(Quit: Konversation terminated!)]
__Sander__ has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
FRidh has quit [(Client Quit)]
FRidh has joined joined #nixos-dev
pie_ has quit [(Ping timeout: 263 seconds)]
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
pie_ has joined joined #nixos-dev
FRidh has quit [(Remote host closed the connection)]
FRidh has joined joined #nixos-dev
ma27 has joined joined #nixos-dev
pie_ has quit [(Read error: Connection reset by peer)]
pie_ has joined joined #nixos-dev
goibhniu has quit [(Ping timeout: 272 seconds)]
pie_ has quit [(Read error: Connection reset by peer)]
pie_ has joined joined #nixos-dev
<vcunat> uh, Hydra is so much slower than my machine in terms of a single build
<vcunat> gchristensen: you can finally test the hello /nix/store/4g4gmizl8cgwhwqsdrz4y70zjvg0rmil-hello-2.10
<yegortimoshenko> copumpkin: btw, all done (fetchurl user-agent), but now it doesn't report nixpkgs version. the only way i see out of it is impureEnvVars, but probably next time (i'll open an issue).
<copumpkin> that's probably fine
<copumpkin> it's better than nothing
<vcunat> uh
<gchristensen> =)
<Dezgeg> LD_PRELOAD ?
<Dezgeg> or LD_LIBRARY_PATH ?
<vcunat> yes, that's what I meant to ask
<vcunat> --print-interpreter and --print-rpath seem OK in my build
<gchristensen> fun ...
<Dezgeg> that always happen if you don't use nixpkgs ldd
goibhniu has joined joined #nixos-dev
<gchristensen> I don't have an LD_ env vars, and https://gist.github.com/grahamc/241174fd51361be2a476efb0c0f03f46
<Dezgeg> ehh
<Dezgeg> strace then? maybe it loads a ld.so.cache?
<vcunat> hmm, I can't explain this
<vcunat> $ env -i /nix/store/4g4gmizl8cgwhwqsdrz4y70zjvg0rmil-hello-2.10/bin/hello
<vcunat> Hello, world!
<vcunat> (I replaced my local paths for these two packages from binary cache, to be sure)
<gchristensen> though I think I know what is happening
<gchristensen> and we can't fix it
<vcunat> uh, I have no idea what's happening
<gchristensen> $ lsmod | grep "fake corporate security garbage" | wc -l
<gchristensen> 1
<vcunat> I can't see any problem until the error message
<Dezgeg> maybe there are multiple kernel version checks and you missed one?
<Dezgeg> (i.e. an additional one for dependent shared objects)
<vcunat> it doesn't seem like kernel check, if they're sensible with error messages
<Dezgeg> LD_DEBUG=all could help
<vcunat> (and I think I grepped the libc sources enough)
<vcunat> I suppose tomorrow I'll put CentOS 6 into some VM and try playing with it
<vcunat> hopefully it was just something local to the machine
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
<vcunat> hmm, the last file does exist - and the previous strace showed it succeeded with openat -> FD 3
FRidh has quit [(Client Quit)]
FRidh has joined joined #nixos-dev
<vcunat> hmm, libc.6.so is opened, 832 bytes are read, then it's closed, and then the whole process again
<vcunat> it's as if it didn't like the contents?
<gchristensen> say, vcunat, sha256sum /nix/store/igq6z6lvnc7ykikc5xljwsvsjfczy4s3-glibc-2.26-115/lib/libc-2.26.so /nix/store/4g4gmizl8cgwhwqsdrz4y70zjvg0rmil-hello-2.10/bin/hello
<Dezgeg> yes, I was thinking that it notices the .so says it's for linux >= 3.2.0 and ignores it
<vcunat> right, that seems possible
<vcunat> my strace shows some mmaps happening before closing
<Dezgeg> check 1681 /* Check .note.ABI-tag if present. */
<Dezgeg> in elf/dl-load.c
<vcunat> apparently it's rejected for some reason, and kernel seems likely
<vcunat> right, that will be it
<gchristensen> I love working with smart people ♥‿♥
goibhniu has quit [(Remote host closed the connection)]
<vcunat> ah, yes, I see it now. Using macro within a macro to set value of a global variable.
<vcunat> How come I've missed it.
<dtzWill> lol
<gchristensen> O.o
<dtzWill> ⊙_ʘ
<yegortimoshenko> do we need another glibc patch
<dtzWill> no surprise, can repro the behavior and the ABI-tag indicates requires 3.2.0
<yegortimoshenko> (or ld.so patch, rather)
<vcunat> yegortimoshenko: trying to persuade glibc-2.26 to work on RHEL 6 (and derivatives)
<dtzWill> (readelf -a translates the ABI-tag neatly, ending with: "Displaying notes found in: .note.ABI-tag
<dtzWill> Owner Data sizeDescription
<dtzWill> GNU 0x00000010NT_GNU_ABI_TAG (ABI version tag)
<dtzWill> OS: Linux, ABI: 3.2.0
<dtzWill> ")
<dtzWill> oh jeepers sorry meant to just paste that last line O:)
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
goibhniu has joined joined #nixos-dev
FRidh has quit [(Client Quit)]
FRidh has joined joined #nixos-dev
<yegortimoshenko> vcunat: have you tried patching sysdeps/unix/sysv/linux/configure.ac L61?
<vcunat> yegortimoshenko: my plan was to only allow 2.6.32 explicitly
<vcunat> to decrease the likelihood the checks won't fire with some other kernels that don't have so many backported features as RHEL ones
<gchristensen> rhel's kernels are such a nightmare
<Dezgeg> pronounced r-hell
<gchristensen> bingo
<vcunat> heh, my superior worked at Red Hat a year ago :-)
<yegortimoshenko> glibc is a real nightmare, to patch, at the very least. it's arcane.
<vcunat> I'll work it out somehow
<yegortimoshenko> that line that Dezgeg pointed to is supposed to just check for ABI tag. instead, it's...
<vcunat> I suppose I'll leverage dependency rewriting for testing, because recompiling gcc on each iteration is just painful
<vcunat> yes, this computes `osversion` from the ELF header and compares it
<Dezgeg> I guess instead of hello you can build something from the bootstrap stages
<vcunat> (probably)
<yegortimoshenko> sometimes i wish everyone would use musl libc
<vcunat> hmm, right, the first binary compiled with glibc :-)
peti has quit [(Quit: x-mas vacation)]
<vcunat> tomorrow I can finish this into something reasonable and working, probably
<gchristensen> wow, so cool :D
<vcunat> it's fortunate I'm on vacation
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
FRidh has quit [(Client Quit)]
FRidh has joined joined #nixos-dev
<yegortimoshenko> could someone review https://github.com/NixOS/nixpkgs/pull/31094 so that i can merge it? it's targeting staging
<yegortimoshenko> (IBus fix)
JosW has quit [(Quit: Konversation terminated!)]
<Dezgeg> fwiw, I don't think musl code is that great either
<Sonarpulse> https://hydra.nixos.org/jobset/nixpkgs/ericson2314-staging bgamari I just remembed the python failures
<Sonarpulse> with the timestamps
<Sonarpulse> have something to do with setup hooks being run in a slightly different order
<Sonarpulse> real nightmare to debug :(
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
<vcunat> right, we have to write a completely new libc, in rust I guess ;-)
<gchristensen> sgtm
<Sonarpulse> somebody was incrementally doing things with musl and rust
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
Guest32460 is now known as stqism
stqism is now known as Guest85011
Guest85011 is now known as stqism
stqism has quit [(Changing host)]
stqism has joined joined #nixos-dev
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
<bgamari> Sonarpulse, I just added SOURCE_DATE_EPOCH to workaround the issue
<Sonarpulse> bgamari: hmm
<Sonarpulse> well that's good
<bgamari> seems to work
<Sonarpulse> but I worry the larger ordering problem should be fixed
<bgamari> yes
<bgamari> indeed
<bgamari> I'm just in hack-and-slash mode at the moment
<Sonarpulse> sure, no worries
<bgamari> needed to get things running asap
<bgamari> Sonarpulse, for the record, I have a cut at reworking the ghc expression to properly cross-compile
<Sonarpulse> bgamari: oh i had that
<bgamari> oops
<Sonarpulse> cross-elegant in the obsidian nixpkgs repo
<bgamari> well, it's probably more correct than mine is
<Sonarpulse> sorry I didn't say!
<bgamari> quite alright
<Sonarpulse> GHC is weird cause it uses CC as CC_FOR_BUILD :D
<Sonarpulse> or something like that
<Sonarpulse> I mentioned it once
<Sonarpulse> we should fix that :D
<bgamari> yep
<bgamari> that is very true
<bgamari> I have cross-compiled GHC before
<bgamari> but didn't realize how odd it was until I tried to fix the nix expression
<Sonarpulse> :)
<Sonarpulse> bgamari: I think it's from the perspective of stage 2
<Sonarpulse> so in that sense its not totally wrong
<Sonarpulse> I forget if I said this, but I hope we can have a canadian cross CI job
<Sonarpulse> build != host != target, to ensure that all the variables are used correctly
<bgamari> yes
<bgamari> that would be nice
<Sonarpulse> (single stage canadian cross, no TH is fine)0
FRidh has quit [(Quit: Konversation terminated!)]
<Sonarpulse> so need a (G)HC (G)HC_FOR_BUILD and (G)HC_FOR_HOST
FRidh has joined joined #nixos-dev
<Sonarpulse> *_FOR_TARGET
ma27 has quit [(Ping timeout: 265 seconds)]
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
orivej has quit [(Ping timeout: 240 seconds)]
yegortimoshenko has quit [(Ping timeout: 264 seconds)]
<bgamari> Sonarpulse, right
<bgamari> opening a ghc ticket about this would be helpful
<Sonarpulse> bgamari: i'm about to head to airport now, but maybe I'll do it there :)
vcunat has quit [(Quit: Leaving.)]
<bgamari> Sonarpulse, thanks!
<bgamari> safe travels!
Sonarpulse has quit [(Ping timeout: 264 seconds)]
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
orivej has joined joined #nixos-dev
FRidh has quit [(Quit: Konversation terminated!)]
yegortimoshenko has joined joined #nixos-dev
FRidh has joined joined #nixos-dev
taktoa has joined joined #nixos-dev
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
taktoa has quit [(Read error: Connection reset by peer)]
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev
infinisil has quit [(Quit: ZNC 1.6.5 - http://znc.in)]
laggy_wifi has quit [(Read error: Connection reset by peer)]
infinisil has joined joined #nixos-dev
laggy_wifi has joined joined #nixos-dev
FRidh has quit [(Remote host closed the connection)]
FRidh has joined joined #nixos-dev
FRidh has quit [(Client Quit)]
FRidh has joined joined #nixos-dev
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined joined #nixos-dev