File transfert using In Band Bytestream ( jep-0047 )

[misc@…]

Jep-0096 ( file transfer ) mandate the usage of two file transfer method, socks 5 ( already in gajim ) and in band bytestream ( jep-0047 ), but the second is not implemented at the moment.

Ibb is useful when you cannot use socks5, ie restricting firewall. I also think it is much easier to implement than socks5.

[dkirov]

You are right. However IBB cannot replace OBB, it can only be used for

small files. AFAIK only psi has support for it. There is an ongoning talk about "File Transfer Preview", IBB will be very handy for that. xmpppy includes [http://trac.gajim.org/browser/trunk/src/common/xmpp/filetransfer.py plugin for ibb]. It needs little rework to use gajim sid and to send notification for transfer status - shouldn't be much work. Then we have to think of some criteria for choosing ibb as transfer method. I suggest files <= 8kb.

[anonymous]

Gajim needs to be able to fall back to this as a last resort.

If falling back it might be a good idea to warn the user before transferring large documents (> 100 KiB, perhaps), but the fall back is still needed. Some clients might have no connection to the public internet.

[js]

Server administrators would not be happy to see this and most of the XMPP community think that IBB file transfers should die. I'm voting for a WONTFIX.

[misc]

If a admin do not want to have users taking the bandwidth, I can configure the server for this, it would take me 5 minutes to filter this server side. So I think that's not a good reason to forbid it to people who want to use it.

For the moment, all my workmates and friends tell me that files transfers are unreliable, and always send me file by mail, and I think that IIB is one of the method that would perfectly work in such cases, especially for small files, as suggested by previous comment.

Limiting user choice