Dimitri John Ledkov

Non-fetch APIs are popular as they are supported by very old openssl, new openssl, and modern forks of old openssl (boringssl, aws-lc, libressl). Hence they are not going away.

> I just saw that you guys are also building a gitlab-runner-helper and when taking a peak at the specifications it seems like you're using a non root user for...

> I think there was a misunderstanding. I'm not a premium customer sorry about that. I just saw that you already offer images for the use case that i'm building....

> we should figure out what the replaces semantics should actually be I guess we should check what happens with: - directories - regular files - device nodes files -...

Looking at apko code, writeHeader: https://github.com/chainguard-dev/apko/blob/41f5c557366ee4ad9b70ddbf39fc89cf13fd2810/pkg/tarfs/fs.go#L437 implements replaces logic, but Symlink https://github.com/chainguard-dev/apko/blob/41f5c557366ee4ad9b70ddbf39fc89cf13fd2810/pkg/tarfs/fs.go#L678 does not.

@iamrajiv yes they do => and the replaces logic higher up is not at all present there. And there is also not enough information passed through to this function. Thus...

@javacruft fun, but i am also confused about "amd64" and "arm64" cause these are not the right arch names.

https://gitlab.alpinelinux.org/alpine/infra/infra/-/issues/10862 So alpine says not to rely on etags, as they use nginx with time + size (basically etag is based on stat).

> > The introduction of the caching client for offline build usage uses etag for indexing - however: > > > > ``` > > $ curl -sI "https://alpinelinux.org/keys/alpine-devel%40lists.alpinelinux.org-616ae350.rsa.pub" |...

we also have a different regression fix for which was merged just now. Would want to get that out first, before investigating this.