Le 14 juin 2010 à 20:37, Jens Alfke a écrit :
Last I heard (admittedly, a few years ago) there wasn’t any good way to do
async/nonblocking filesystem operations. The standard system calls block; the
O_NONBLOCK mode for fcntl only works on network socket streams; and the AIO
API has (had?)
Hi guys,
i was just wondering, is there any way to determine when for example a network
cable is plugged off while copying data to a network share? The Finder has a
pretty long timeout...
Is there a way in the FS API or ANY other way to get notified, when a volume
isn't available for writing
On Jun 14, 2010, at 4:52 AM, Boris Prohaska wrote:
i was just wondering, is there any way to determine when for example a
network cable is plugged off while copying data to a network share? The
Finder has a pretty long timeout...
Is there a way in the FS API or ANY other way to get
Thats exactly the problem. IF the OS was aware of it, then there would be no
problem. There is also no setting for timeouts etc when performing blocking
operations onto the filesystem.
Maybe that is a OS X limitation??
Boris
On 14.06.2010, at 17:04, Ken Thomases wrote:
On Jun 14, 2010, at
On Jun 14, 2010, at 11:40 AM, Boris Prohaska wrote:
Thats exactly the problem. IF the OS was aware of it, then there would be no
problem. There is also no setting for timeouts etc when performing blocking
operations onto the filesystem.
Maybe that is a OS X limitation??
The way around it
On Jun 14, 2010, at 11:13 AM, Ken Thomases wrote:
The way around it is to not do blocking operations. Use non-blocking,
asynchronous I/O operations, and impose the timeout yourself.
Last I heard (admittedly, a few years ago) there wasn’t any good way to do
async/nonblocking filesystem
I don't know which kind of connection you want to observe, but
basically SCNetwork and SCNetworkReachability are your friends.
then control these (and get the right conf) using two callback handlers
use NSWorkspace to get your entry and your exit (if this case happens)
no you are in the weed there
On Mon, Jun 14, 2010 at 9:40 AM, Boris Prohaska borisproha...@gmx.at wrote:
Thats exactly the problem. IF the OS was aware of it, then there would be no
problem. There is also no setting for timeouts etc when performing blocking
operations onto the filesystem.
On Jun 14, 2010, at 1:37 PM, Jens Alfke wrote:
On Jun 14, 2010, at 11:13 AM, Ken Thomases wrote:
The way around it is to not do blocking operations. Use non-blocking,
asynchronous I/O operations, and impose the timeout yourself.
Last I heard (admittedly, a few years ago) there wasn’t