Multipart upload question

[Jim Nelson <jim yorba org>]

Hello,

I'm working with libsoup client (Shotwell) that does multipart uploads to various web services.  The problem I'm facing is how to upload a file that is quite large (> 400MB) without loading the entire contents of the file into memory.

I've looked over the API and I see various ways to hook into the HTTP request process, but I don't see an obvious way to tell Soup that, when it reaches a particular section of the multipart upload, to read from a file (or to call me via a callback/signal so I can provide it with buffers).

I suppose I could generate the multipart body by hand and start appending SoupBuffers after "wrote-headers" and "wrote-chunk", inserting the file data that way, but I'd like to avoid that.

Any suggestions or pointers would be most welcome.

-- Jim