I'm trying to download a file from a browser using socket.io-stream. In a basic form, this is actually doable and there is a working example here.
However, that solution:
First streams the file contents to the browser using socket.io-stream.
Assemble the chunks in the client as a blob.
Create a hidden link to the blob location.
This forces the browser to contain the whole blob in memory before it can initiate the download. I'm working with really large blobs, so that is not advisable.
I would prefere to download the stream directly, instead of buffering it in memory in the browser.
Is that possible?
I know this is easy to do just with plain HTTP, but there are some reasons that make this simplest option not available in my case.