I have a dagger pipeline where I am explicitly setting the ~/.docker/config.json (before starting the engine) so that the runner picks it up and sets it so that I can pull public docker images within my corp network. This has been working great until v0.9.0.
After updating to v0.9.0, the docker pull does timeout if it's trying to pull from public docker which leads me to think the proxy is not properly set within the container.
To help pinpoint the issue this seems to happen when I try to run the docker:dind service container as a sidecar. The only change I did was change the service container to a *dagger.Service. So this leads me to think that the regression is with the services V2 code. The error also happens in one of my unit tests, which is using dagger-in-dagger.
The pipeline still works fine with v0.8.8.