On Sun, Mar 04, 2007 at 11:02:20PM +0100, Tomas Groth wrote: > Hi, > > Just to end this thread properly, i just wanted to point out that mostly due > to > me failing to read documentation and specs properly, we won't need to switch > from the current curl_adapter, though it will be extended a bit, and some has > already been done. > > What still remains is primarily the download thread, which will most likely be > implemented as a seperate class. A quick design for it could be that it > receives a tu_file, and a mutex from the as_object that needs something > downloaded. The mutex is used to avoid the thread writing new data to the file > while the as_object is reading. If needed the as_object can pass a number (X) > and a callback function to the download class which will call the callback > when > there has been downloaded X bytes of the file.
WARNING! Let's don't let loader threads execute arbitrary code ! I wouldn't want to protect *everything* with mutexes. We definitely don't the advance loop to be in a race with loader threads. --strk; _______________________________________________ Gnash-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnash-dev
