#hello folks, when adding a new field to
lyric atlas
on trying to reproduce this issue locally, its reproducible if I run:
dagger call -m . check --targets=sdk/php
but works as expected when I run:
hack/with-dev dagger call -m . check --targets=sdk/php
@tough zealot @karmic wasp - oh, do we need to add another dagql.Func with a different View for WithExec for this change?
tough zealot
can you share a link/code-snippet to what you're trying to do?
when adding new features you should add them to all the views, as long as it's reasonable to do so imo
lyric atlas
and this is the failure i am trying to debug: https://github.com/dagger/dagger/actions/runs/10850936806/job/30113511601?pr=8427#step:5:1250
tough zealot
why is php trying to access that field? it seems like a bug in the codegen
cc @median dock any ideas? it seems like the codegen for the importable client is being used in the module maybe?
or are the tests not using the dev engine that we set?
no that doesn't appear to be it at least
2 weird things are happening
- why is php explicitly populating that field? it shouldn't be passing it if the user hasn't set it?
- why doesn't the engine see that field?
lyric atlas
not sure if you noticed this msg earlier: #1284143498718609429 message
doh. specifically this:
on trying to reproduce this issue locally, its reproducible if I run:
dagger call -m . check --targets=sdk/php
but works as expected when I run:
hack/with-dev dagger call -m . check --targets=sdk/php