#vito with the latest changes to 5757
fluid hare
appreciated! I was able to reproduce the examples that Erik left in the PR last night, but not this morning.
weak flame
works here, but I'm on Linux, maybe this is a Mac thing?
what if you do go install ./cmd/dagger?
fluid hare
Can try that. Is there a current set of steps that works for you at least?
I was kind of thrashing a bit to build things and invoke things with the right env var at the right time etc
There used to be a README.md in the other Zentith PR that had steps to follow. Maybe that one is still valid.
weak flame
It should be yeah, the basic dev steps haven't changed, there are just different commands / examples to run
Can you run file ./bin/dagger in the dagger repo? Wonder if it built for the wrong platform or something
fluid hare
That's what I was guessing.
lemme look at the binary again
dagger-fork ➤ file bin/dagger git:zenith-functions
bin/dagger: Mach-O 64-bit executable arm64
that's after running ./hack/make engine:build
I'm not getting anything useful from go install ./cmd/dagger but maybe I need some Golang environmental stuff set up differently
I've use go install on other things in the past with success though, I think
Oh, wait. Yeah, now it's executing
The new binary, that is.
hmmm
weak flame
from go install?
could also try running ./hack/dev, I usually do that instead of ./hack/make engine:build
fluid hare
Yeah, that's why I just want a set of steps 😄
so I can copy what works for you.
I'll try that 👆 from a fresh checkout
weak flame
yeah totally. I just haven't had a chance to put it together yet, it only compiled as of last night 😛
I can start updating zenith/README.md now - seems like the easiest doc to find
fluid hare
# start from root of dagger/dagger clone
gh pr checkout 5757
# check `docker ps` and remove all running dagger engine containers
./hack/dev
# try to run `dagger`
# check `docker ps` and see the localhost/dagger-engine.dev:latest (you'll also see the other dagger engine that was used to build it)
# check `env` to see if you have the needed env vars. Mine don't get set, though they are displayed at end of `./hack/dev`
export PATH=$(pwd)/bin:$PATH
export _EXPERIMENTAL_DAGGER_CLI_BIN=$(pwd)/bin/dagger
export _EXPERIMENTAL_DAGGER_RUNNER_HOST=docker-container://dagger-engine.dev
cd core/integration/testdata/modules/go/basic/
dagger query --progress=plain <<EOF
{basic{getCustomObj(stringArg: "yo"){sayField}}}
EOF
Okay, got something that works for me 🙂
Might be some stuff to prune.
weak flame
nice! one thing I'll do is pop those env vars in an .envrc so you don't have to export them all the time
been wanting to do that for ages but it gums up ./hack/dev and ./hack/make (since those need to run against a stable engine). but that's an easy fix
eh maybe I'll put that in ./zenith/ and let that be the "playpen" of sorts
fluid hare
That would be sweet!
sullen juniper
I'm hitting the killed dagger thing now after pulling the latest commits. This is after running ./hack/dev
did you find the root cause of that on your end @fluid hare ?
weak flame
something something macOS signed binaries??? really a head scratcher
don't have a Mac around to test unfortunately
sullen juniper
maybe... the version I built earlier this afternoon the exact same way worked fine. Is there anything in the environment that would make it produce a weird binary? I unset all the dagger envs
i'll try in a fresh terminal. who knows
negative
weak flame
it was worth a shot 😂
anything in Console.app? or wherever macOS logs things these days
weak flame
lol `Termination Reason: CODESIGNING 2`
sullen juniper
Incident Identifier: 1590F72B-CED3-4927-8587-5720A71590F7
CrashReporter Key: 45922F14-3938-1988-EF37-E345BBE5357F
Hardware Model: MacBookPro18,3
Process: dagger [18190]
Path: /Users/USER/*/dagger
Identifier: dagger
Version: ???
Code Type: ARM-64 (Native)
Role: Unspecified
Parent Process: zsh [16598]
Coalition: dev.warp.Warp-Stable [206253]
Responsible Process: stable [8795]
Date/Time: 2023-09-13 18:01:03.5776 -0400
Launch Time: 2023-09-13 18:01:03.3028 -0400
OS Version: macOS 12.6 (21G115)
Release Type: User
Report Version: 104
Exception Type: EXC_BAD_ACCESS (SIGKILL (Code Signature Invalid))
Exception Subtype: UNKNOWN_0x32 at 0x0000000104f34000
Exception Codes: 0x0000000000000032, 0x0000000104f34000
VM Region Info: 0x104f34000 is in 0x104f34000-0x105d20000; bytes after start: 0 bytes before end: 14598143
REGION TYPE START - END [ VSIZE] PRT/MAX SHRMOD REGION DETAIL
UNUSED SPACE AT START
---> mapped file 104f34000-105d20000 [ 13.9M] r-x/rwx SM=COW ...t_id=1aad27c1
mapped file 105d20000-1068e0000 [ 11.8M] rw-/rw- SM=COW ...t_id=1aad27c1
Exception Note: EXC_CORPSE_NOTIFY
Termination Reason: CODESIGNING 2
Highlighted by Thread: 0
Backtrace not available
🤔
weak flame
maybe the cross-compiled version doesn't work, locally built one does?
does that mean our github released binaries also blow up?
sullen juniper
Don't really know what's going on. The version I was using all afternoon was built with ./hack/dev and worked fine