<html> <head> <base href="https://bugzilla.gnome.org/" /> </head> <body> <div> <a class="bz_bug_link bz_status_REOPENED " title="REOPENED - File descriptor leak in gdk_wayland_selection_request_target()" href="https://bugzilla.gnome.org/show_bug.cgi?id=751414#c22">Comment # 22</a> on <a class="bz_bug_link bz_status_REOPENED " title="REOPENED - File descriptor leak in gdk_wayland_selection_request_target()" href="https://bugzilla.gnome.org/show_bug.cgi?id=751414">bug 751414</a> from <a href="page.cgi?id=describeuser.html&login=carlosg%40gnome.org" title="Carlos Garnacho <carlosg@gnome.org>"> Carlos Garnacho</a> <pre>(In reply to Michael Catanzaro from <a href="show_bug.cgi?id=751414#c21">comment #21</a>) > I confirm this bug is fixed when I apply these five patches. Great, thanks! Thanks for verifying! Will push now. > > (In reply to Carlos Garnacho from <a href="show_bug.cgi?id=751414#c16">comment #16</a>) > > Yeah, that'd be one way to work around the "1 request on the fly" > > limitation, is it too much of a change to?: > > > > * forward all mimetypes to the other process over internal IPC > > * In the other process, pick one and tell about it to the original process > > * call wl_data_offer_receive on that one, pass the fd. > > * In the other process, read() and close > > If that would work it would have been the first thing to try, but Chrome > expects the data to already be present before it deciding which type of data > to pick, unfortunately. I feared as much... Then there's no way around one-by-one requests and/or selection data caching ATM. > > To be clear, the problems I see are: > > * If GTK+ were to allow handling multiple data transfers at once in the > future, there would need to be policy either in GTK+ or in the compositor to > ensure that the drop target cannot initiate an arbitrary number of > simultaneous transfers. Agreed. > > * If GTK+ continues to not handle multiple data transfers at once, it's > arguably not implementing its end of the protocol properly. Since it works > with the reference Weston client, I would expect it to work with GTK+ as > well. But if there was upstream documentation to say "don't expect this to > work" (the simplest solution, possibly the best solution) then there would > be no problem. Oh, if it were the only unclear point in the DnD protocol. I found a few in the doing of the proposal for wayland DnD actions :/</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are on the CC list for the bug.</li> </ul> </body> </html>