#Im running into a strange issue locally
1 messages ยท Page 1 of 1 (latest)
๐ค
interesting
๐ค
i do not see this
you also have this error in the trace: https://dagger.cloud/levs-test-org/traces/2a5b42d1f0ee570cedac9d17e0384080?span=e9efc53d2c69c7ca#d2fc916c144fde35
# github.com/dagger/dagger/modules/dirdiff
./dirdiff.go:25:13: not enough arguments in call to dag.Wolfi().Container
have (unknown type)
want ([]string, []*dagger.Container)
./dirdiff.go:25:20: undefined: dagger.WolfiContainerOpts
just to confirm and be sure, dagger version gives you 0.12.5? and echo "{version}" | dagger query gives the same?
hm, yeah, this is confusing, i really can't manage to confirm this at all, i've been working in 0.12.5 dagger/dagger all day long ๐ค no issues like this seen
Iโm just calling dagger functions - I did not change any code yet
Any tips on how to flush whatever seems to be broken here? Maybe to do with caching or volumes
Yeah, maybe you could try removing the engine container and trying again, that would flush everything
Ugh this is so annoying lol
Removing the engine completely gives me this
dagger functions
Full trace at https://dagger.cloud/levs-test-org/traces/163040deb5895032af3e6cbeeccf0943
โ connect 4.0s
โ initialize 1m34.3s
! input: moduleSource.withContextDirectory.asModule resolve: failed to create module: select: failed to update module dependencies: failed to initialize dependency modules: failed to initialize dependency module: select: failed to initialize module: failed to call module "php-sdk-dev" to get functions: call constructor: process "/src/sdk/php/dev/entrypoint.php" did not complete successfully: exit code: 127
โ resolving module ref 0.2s
โ installing module 1m34.2s
! input: moduleSource.withContextDirectory.asModule resolve: failed to create module: select: failed to update module dependencies: failed to initialize dependency modules: failed to initialize dependency module: select: failed to initialize module: failed to call module "php-sdk-dev" to get functions: call constructor: process "/src/sdk/php/dev/entrypoint.php" did not complete successfully: exit code: 127
โ ModuleSource.resolveFromCaller: ModuleSource! 24.2s
โ ModuleSource.asModule: Module! 1m9.9s
! failed to create module: select: failed to update module dependencies: failed to initialize dependency modules: failed to initialize dependency module: select: failed to initialize module: failed to call module "php-sdk-dev" to get functions: call constructor: process "/src/sdk/php/dev/entrypoint.php" did not complete successfully: exit code: 127
Error: input: moduleSource.withContextDirectory.asModule resolve: failed to create module: select: failed to update module dependencies: failed to initialize dependency modules: failed to initialize dependency module: select: failed to initialize module: failed to call module "php-sdk-dev" to get functions: call constructor: process "/src/sdk/php/dev/entrypoint.php" did not complete successfully: exit code: 127
Stderr:
': No such file or directory
But I am not doing anything just trying to run dagger functions off of main so it makes no sense why this is only breaking on my machine
yeah, okay, i'm seeing the same sort of issues (in your new trace)
just looking through your trace
yeah, is bizarre, it looks like dependencies for modules are just not getting loaded?
did this break at some point for you? i assume this used to work?
~~does 0.12.4 work? or any earlier version? ~~ nope, v0.12.5 is now the minimum version
do you have time before the community call tomorrow to try to work through this? i have to head off now, but could do then
oh wait, it looks like parameters aren't getting detected as optional from the wolfi module???
are you using dagger cloud by any chance? that would be weiiiiird if that had an effect on this
Yes Dagger Cloud.
Yeah would love some help - but just to be 1000% clear
I did not do anything other than:
- clone a fresh version of dagger/dagger
- run dagger functions
๐
This all worked before I had COVID, so maybe there is an infection on my computer :p
FWIW this does work just fine in my Linux server, and this is all happening on my Mac :/
dagger cloud cache as well? sorry, i brainfarted
if you go to ./modules/dirdiff, and do a dagger develop.
what does the generated internal/dagger/dagger.gen.go look like?
i'm trying to understand how this part of your trace is possible - https://dagger.cloud/levs-test-org/traces/163040deb5895032af3e6cbeeccf0943?span=555d078061a9f8f1#046d4d352bc0b0e6
somehow the modules seem like they're being served the wrong api ๐ค
Yeah I think I have both things in place
Running dagger logout to remove cloud cache did not make a difference
Heres the result of dagger develop inside of dirdiff
Cloud cache isn't controlled by dagger login/logout sadly, it's controlled by the _EXPERIMENTAL_DAGGER_CLOUD_TOKEN in the environment when the engine starts
Ah, I dont have this token set ๐
So does that mean my cloud cache just does not work at all lol
Yeah well at least it's not that influencing it ๐ค
If you do a dagger develop in the wolfi module...
What does the top-level dagger.gen.go there have in it? (Not the internal one)
Sorry remotely debugging over text chat while I'm mobile on the bus is really not ideal
Huh okay, there's no optional in there
Just to make 1000000000% sure - what's wolfi's main.go
// A Dagger Module to integrate with Wolfi Linux
//
// Wolfi is a container-native Linux distribution with an emphasis on security.
// https://wolfi.dev
package main
import "github.com/dagger/dagger/modules/wolfi/internal/dagger"
// A Wolfi Linux configuration
type Wolfi struct{}
// Build a Wolfi Linux container
func (w *Wolfi) Container(
// APK packages to install
// +optional
packages []string,
// Overlay images to merge on top of the base.
// See https://twitter.com/ibuildthecloud/status/1721306361999597884
// +optional
overlays []*dagger.Container,
) *dagger.Container {
ctr := dag.Apko().Wolfi(packages)
for _, overlay := range overlays {
ctr = ctr.WithDirectory("/", overlay.Rootfs())
}
return ctr
}
Cool great wonderful, how's that possible?
๐ฑ
Hmm waiiiiit
Do you have any funky newlines? As in, are there windows line endings in this repo?
That would indeed be odd, but you can usually tell by looking at vscode somewhere, or just dump the file using xxd and looking at that
That's broken some things like that before
But, huh it just looks like that optional gets entirely ignored
I have been seeing some weird shit with new lines lol
But I have no idea why, but also remember I am not opening this in vscode, its
- clone repo
- run some dagger commands
Oh wait this is so interesting
But, huh it just looks like that optional gets entirely ignored
Someone else was reporting an issue like this:
https://discord.com/channels/707636530424053791/1274092489137061999
sigh
yeah, this is https://github.com/dagger/dagger/pull/7121 again
turns out. i didn't fix it :)))))))))))))
there's a very helpful little note in the go ast parser:
// The Text field contains the comment text without carriage returns (\r) that
// may have been present in the source. Because a comment's end position is
// computed using len(Text), the position reported by [Comment.End] does not match the
// true source end position for comments containing carriage returns.
unhelpful, very annoying
need to rework the entire comment parsing logic to consider this i guess ๐
ah silly me for thinking that .End() would consistently get the token end position