A bunch of protocol libraries implement file transfer using a file API that reads/writes files from the disk. Unlike Telepathy, which transfers files to/from the client using sockets.
This makes implementing FT for these protocols a pain, where the only way it's possible is to implement the API is to spool the data from the socket to a temporary file. For large files this is both a huge slow-down, and a way to potentially run out of disk space. I was thinking the other day that it might be possible to abuse the new FileTransfer.URI property to work around this problem. For outgoing files, we could check the Size and ContentHash matched the URI and if that's the case, ignore the data coming via the socket. For incoming files, we could write to a temporary file and then move it into the location of URI once the client had closed its socket. It's pretty hacky, and might melt for some implementations. For clients that don't implement URI we could fall back to spooling. Thoughts? --danni -- Danielle Madeley Software Developer, Collabora Ltd. Melbourne, Australia www.collabora.co.uk _______________________________________________ telepathy mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/telepathy
