phlax

> Can you elaborate on how you hooked that shim in? very hackily - see this PR for details https://github.com/envoyproxy/envoy/pull/25359/files seems to have survived a `rules_python` upgrade, i didnt check...

interesting we just landed a workaround to inject namespacing init pys to our repo folders to avoid pypi namespace clashing - seems like we could avoid that workaround - ill...

this is still an issue

cc @mattklein123

we dont currently have the website served from any other URL, and inevitably "proxy" comes up quite a lot throughout the website could you ask the admin to allowlist the...

fwiw pulling and creating the image gives it a better chance of hitting the ground running, but not sufficiently that you could call it reliable

the problem can be seen on `main` where it attempts to regenerate the toolchains - https://github.com/envoyproxy/envoy-build-tools/actions?query=branch%3Amain i have added some CI, so that we test toolchain regeneration in PRs but...

the problem with this fix is that it will leave the windows build with ~1/10 chance of having its toolchains regenerated where i have needed to test downstream i have...

checking through my previous test attempts i noticed that when i had tried pinning the base image to the earlier one i had inadvertently pinned it to current i had...

strangely this seems to have resolved itself afaict if confirmed ill remove the pin in the envoy repo