Comment 9 for bug 77133

New Infos: I managed to install a debug version of the mapping-daemon (by the way, how should the orignal profile_log method work? I changed it to write to a file).

The interesting thing is that in the case the export fails, you can see that right after the operation MAPPING_PROTOCOL_OP_CREATE_FILE the mapping daemon receives a request MAPPING_PROTOCOL_OP_REMOVE_FILE to delete the file again. This MAPPING_PROTOCOL_OP_REMOVE_FILE is missing when the export succeeds (i.e. nautilus is not running).

So I think that somehow nautilus monitors the burn directory and immediately sends out a remove operation to the mapping daemon. That maybe also the solution why it sometimes works - it seems to be a race condition if nautilus is running during the export. Most of the time the remove operation is "fast enough" to let the export fail, but sometimes nautilus is to slow to send out the remove operation.

But it still has to be investigated why this remove operation is done by nautilus.

Regards

Jens