mirror of https://github.com/pulumi/pulumi.git
101a72ed7f
Copilot templates require an access token to look up the conversation data and extract the code Here we try to download any template as usual, but then add the following logic to the response checking: - *If* a `403` is returned from what we identify as a pulumi-service backend (via `X-Pulumi-Request-ID` header, which has been present for the last 6 years) - *Then* we re-request w/ the auth token. If there is no token present then the login flow is automatically initiated. This does result in always raising a 403 on the server every time we request a copilot template. If we want to avoid these 403s we could check the domain against known pulumi backends in the CLI first, at the cost of intro'ing another code path. ## To test - Go through the Copilot flow, e.g "Create an s3 bucket please" - Click "Deploy with Pulumi" - Choose CLI, click next - Check out this branch - Copy most of the command but first: `cd ~/pulumi/pulumi/pkg` and then `go run ./cmd/pulumi new $ARGS_FROM_PULUMI_CLOUD_UI` It should succeed |
||
---|---|---|
.. | ||
go | ||
nodejs | ||
proto/go | ||
python | ||
.version | ||
README.md | ||
go.mod | ||
go.sum | ||
version.go |