CDF on Azure: File downloads with browsers should preserve the original file name

Due to differences between Google Cloud Storage and Azure Blob Storage Services we have somewhat different behavior for CDF on Google Cloud and CDF on Azure. The main difference is the set of characters supported in file/blob names by the two services.

The Azure Blob Storage API has much stricter rules on what characters are permitted in blob names compared to Google Cloud Storage, so we URL encode all file names to ensure that we do not come into conflict with the blob naming rules in Azure Blob Storage. This encoding process is performed automatically by the Azure Storage SDK which we have used to build the Files API on Azure.

More information from Microsoft about blob names: Naming and Referencing Containers, Blobs, and Metadata - Azure Storage

This can be contrasted with the Google service’s naming restrictions, which are basically “anything goes except newlines, directory traversals and “.well-known/acme-challenge” (which could be used to acquire a SSL certificate for Google’s cloud storage service using the ACME protocol, so naturally not allowed ) - About Cloud Storage objects  |  Google Cloud

As the blob name is visible to the browser when downloading, the file gets the percent encoded name when downloading unless the downloading client chooses a different local file name.