On Mon, 2009-02-16 at 14:08 -0500, David Zeuthen wrote: > On Mon, 2009-02-16 at 20:02 +0100, Nicolò Chieffo wrote: > > I opened a bugreport some time ago > > http://bugzilla.gnome.org/show_bug.cgi?id=555715 and they answered > > that this change is not scheduled, even in glib 3. > > Thay also say it is not enough to invalidate the cache, because some > > app might have already asked for the special directory before, so they > > cannot free the pointer, becouse it could be in use. (bad C!!!!) > > I think it's an OK trade off to leak this memory since invalidating the > cache would rarely happen. > > > > Maybe we could have some call to flush the glib cache, and then we could > > > call that e.g. each time a file selector is displayed, should be "good > > > enough". > > That, and if gio is initialized (same place where all the extension > points are registered), we can set up a file monitor to flush it > whenever something changes....
I don't think we want all glib apps to monitor that file. There is a non-neglible cost in file watches. I think having a call to poll for updates makes much better sense. Then we can call that from nautilus when we know it changed, and from the file selector, say each time it is displayed. -- nautilus-list mailing list nautilus-list@gnome.org http://mail.gnome.org/mailman/listinfo/nautilus-list