Pass headers to download url

[mruz]

Is it possible to pass auth token in the headers to download_url? I need it to download asset from api.github.com:

curl -X POST registry.example.com/v1/api/modules/my-module/provider/1.0.0/upload \
    -H "Authorization: Bearer x-api-key:$TERRALIST_API_KEY" \
    -d '{
        "download_url": "https://api.github.com/repos/{OWNER}/{REPO}/releases/assets/{ASSET-ID}",
        "headers": {
           "Accept": "application/octet-stream",
           "Authorization:": "Bearer {TOKEN}",
           "X-GitHub-Api-Version": "2022-11-28"
        }
    }'

[valentindeaconu]

The short answer is no. Let me explain it: depending on how you're using Terralist there are 2 possibilities:

1. You're using the "proxy" mode, which means it will forward the URLs to the clients. In that case, you can use go-getter's git protocol and save the URLs in this format. Then, on your clients, you must ensure they have git installed and configured with credentials that can read from those sources.
2. You're using the S3 backend, meaning Terralist needs access to that resource. Again, you can use the go-getter's git protocol and configure Terralist to have git access to your resources.

However, in your case, you're trying to download assets from a particular release, and not the entire tag, and AFAIK this is not possible with go-getter, since this is not a git-specific feature, but a GitHub-specific.

I see there's a way for the go-getter package to forward some HTTP headers using the HTTP protocol, but unfortunately this is not currently available in Terralist.

[mruz]

In the latest release we have git installed, it works if I download the entire tag, but I have some modules that have symlinks to other modules and it doesn't work hashicorp/go-getter#60.

When I release a new version of modules that reference other modules, I upload an asset that follows symlinks, but to download it I have to use api.github.com and HTTP.

[mruz]

Instead of uploading the asset to github and downloading from the api, I uploaded it to s3 and downloaded from there.