As hinted at on the readme.
The idea would be that the CI server would use the Azure SDK to build ready-to-go web farm packages that could then be published somewhere for people to use (maybe add them to the GitHub release?).
That way if people don't need to add custom code to the RoleEntryPoint and they use one of the standard vmsizes we decide to build (I think XS, S and M all on the standard tier - no point in not having load balancing) then they have an easier time because they can simply upload a package in the portal or with PowerShell etc. rather than having to maintain a Visual Studio solution and potentially struggling with things like binding redirect mismatches.
As hinted at on the readme.
The idea would be that the CI server would use the Azure SDK to build ready-to-go web farm packages that could then be published somewhere for people to use (maybe add them to the GitHub release?).
That way if people don't need to add custom code to the RoleEntryPoint and they use one of the standard vmsizes we decide to build (I think XS, S and M all on the standard tier - no point in not having load balancing) then they have an easier time because they can simply upload a package in the portal or with PowerShell etc. rather than having to maintain a Visual Studio solution and potentially struggling with things like binding redirect mismatches.