mutter/src/wayland/meta-wayland-data-device.c

1006 lines
31 KiB
C
Raw Normal View History

/*
* Copyright © 2011 Kristian Høgsberg
*
* Permission to use, copy, modify, distribute, and sell this software and its
* documentation for any purpose is hereby granted without fee, provided that
* the above copyright notice appear in all copies and that both that copyright
* notice and this permission notice appear in supporting documentation, and
* that the name of the copyright holders not be used in advertising or
* publicity pertaining to distribution of the software without specific,
* written prior permission. The copyright holders make no representations
* about the suitability of this software for any purpose. It is provided "as
* is" without express or implied warranty.
*
* THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
* INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
* EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
* CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
* DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
* TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
* OF THIS SOFTWARE.
*/
/* The file is based on src/data-device.c from Weston */
#include "config.h"
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <stdio.h>
#include <glib.h>
#include "meta-wayland-data-device.h"
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
#include "meta-wayland-data-device-private.h"
#include "meta-wayland-seat.h"
#include "meta-wayland-pointer.h"
#include "meta-wayland-private.h"
#include "meta-dnd-actor-private.h"
struct _MetaWaylandDataOffer
{
struct wl_resource *resource;
MetaWaylandDataSource *source;
struct wl_listener source_destroy_listener;
};
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
typedef struct _MetaWaylandDataSourcePrivate
{
struct wl_array mime_types;
gboolean has_target;
} MetaWaylandDataSourcePrivate;
typedef struct _MetaWaylandDataSourceWayland
{
MetaWaylandDataSource parent;
struct wl_resource *resource;
} MetaWaylandDataSourceWayland;
G_DEFINE_TYPE_WITH_PRIVATE (MetaWaylandDataSource, meta_wayland_data_source,
G_TYPE_OBJECT);
G_DEFINE_TYPE (MetaWaylandDataSourceWayland, meta_wayland_data_source_wayland,
META_TYPE_WAYLAND_DATA_SOURCE);
static MetaWaylandDataSource *
meta_wayland_data_source_wayland_new (struct wl_resource *resource);
static void
drag_grab_data_source_destroyed (gpointer data, GObject *where_the_object_was);
2014-07-10 10:19:47 -04:00
static void
unbind_resource (struct wl_resource *resource)
{
wl_list_remove (wl_resource_get_link (resource));
}
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
static void
meta_wayland_data_source_target (MetaWaylandDataSource *source,
const char *mime_type)
{
META_WAYLAND_DATA_SOURCE_GET_CLASS (source)->target (source, mime_type);
}
void
meta_wayland_data_source_send (MetaWaylandDataSource *source,
const char *mime_type,
int fd)
{
META_WAYLAND_DATA_SOURCE_GET_CLASS (source)->send (source, mime_type, fd);
}
gboolean
meta_wayland_data_source_has_target (MetaWaylandDataSource *source)
{
MetaWaylandDataSourcePrivate *priv =
meta_wayland_data_source_get_instance_private (source);
return priv->has_target;
}
void
meta_wayland_data_source_set_has_target (MetaWaylandDataSource *source,
gboolean has_target)
{
MetaWaylandDataSourcePrivate *priv =
meta_wayland_data_source_get_instance_private (source);
priv->has_target = has_target;
}
struct wl_array *
meta_wayland_data_source_get_mime_types (const MetaWaylandDataSource *source)
{
MetaWaylandDataSourcePrivate *priv =
meta_wayland_data_source_get_instance_private (source);
return &priv->mime_types;
}
static void
meta_wayland_data_source_cancel (MetaWaylandDataSource *source)
{
META_WAYLAND_DATA_SOURCE_GET_CLASS (source)->cancel (source);
}
static void
data_offer_accept (struct wl_client *client,
struct wl_resource *resource,
guint32 serial,
const char *mime_type)
{
MetaWaylandDataOffer *offer = wl_resource_get_user_data (resource);
/* FIXME: Check that client is currently focused by the input
* device that is currently dragging this data source. Should
* this be a wl_data_device request? */
if (offer->source)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
meta_wayland_data_source_target (offer->source, mime_type);
meta_wayland_data_source_set_has_target (offer->source,
mime_type != NULL);
}
}
static void
data_offer_receive (struct wl_client *client, struct wl_resource *resource,
const char *mime_type, int32_t fd)
{
MetaWaylandDataOffer *offer = wl_resource_get_user_data (resource);
if (offer->source)
meta_wayland_data_source_send (offer->source, mime_type, fd);
else
close (fd);
}
static void
data_offer_destroy (struct wl_client *client, struct wl_resource *resource)
{
wl_resource_destroy (resource);
}
static const struct wl_data_offer_interface data_offer_interface = {
data_offer_accept,
data_offer_receive,
data_offer_destroy,
};
static void
destroy_data_offer (struct wl_resource *resource)
{
MetaWaylandDataOffer *offer = wl_resource_get_user_data (resource);
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
if (offer->source)
g_object_remove_weak_pointer (G_OBJECT (offer->source),
(gpointer *)&offer->source);
g_slice_free (MetaWaylandDataOffer, offer);
}
static struct wl_resource *
meta_wayland_data_source_send_offer (MetaWaylandDataSource *source,
struct wl_resource *target)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
MetaWaylandDataSourcePrivate *priv =
meta_wayland_data_source_get_instance_private (source);
MetaWaylandDataOffer *offer = g_slice_new0 (MetaWaylandDataOffer);
char **p;
offer->source = source;
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
g_object_add_weak_pointer (G_OBJECT (source), (gpointer *)&offer->source);
offer->resource = wl_resource_create (wl_resource_get_client (target),
&wl_data_offer_interface,
wl_resource_get_version (target), 0);
wl_resource_set_implementation (offer->resource,
&data_offer_interface,
offer,
destroy_data_offer);
wl_data_device_send_data_offer (target, offer->resource);
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
wl_array_for_each (p, &priv->mime_types)
wl_data_offer_send_offer (offer->resource, *p);
return offer->resource;
}
static void
data_source_offer (struct wl_client *client,
struct wl_resource *resource, const char *type)
{
MetaWaylandDataSource *source = wl_resource_get_user_data (resource);
if (!meta_wayland_data_source_add_mime_type (source, type))
wl_resource_post_no_memory (resource);
}
static void
data_source_destroy (struct wl_client *client, struct wl_resource *resource)
{
wl_resource_destroy (resource);
}
static struct wl_data_source_interface data_source_interface = {
data_source_offer,
data_source_destroy
};
struct _MetaWaylandDragGrab {
MetaWaylandPointerGrab generic;
MetaWaylandSeat *seat;
struct wl_client *drag_client;
MetaWaylandSurface *drag_focus;
struct wl_resource *drag_focus_data_device;
struct wl_listener drag_focus_listener;
MetaWaylandSurface *drag_surface;
struct wl_listener drag_icon_listener;
MetaWaylandDataSource *drag_data_source;
ClutterActor *feedback_actor;
MetaWaylandSurface *drag_origin;
struct wl_listener drag_origin_listener;
int drag_start_x, drag_start_y;
};
static void
destroy_drag_focus (struct wl_listener *listener, void *data)
{
MetaWaylandDragGrab *grab = wl_container_of (listener, grab, drag_focus_listener);
grab->drag_focus_data_device = NULL;
}
void
meta_wayland_drag_grab_set_focus (MetaWaylandDragGrab *drag_grab,
MetaWaylandSurface *surface)
{
MetaWaylandSeat *seat = drag_grab->seat;
struct wl_client *client;
struct wl_resource *data_device_resource, *offer = NULL;
if (drag_grab->drag_focus == surface)
return;
if (drag_grab->drag_focus)
{
meta_wayland_surface_drag_dest_focus_out (drag_grab->drag_focus);
drag_grab->drag_focus = NULL;
}
if (!surface)
return;
if (!drag_grab->drag_data_source &&
wl_resource_get_client (surface->resource) != drag_grab->drag_client)
return;
client = wl_resource_get_client (surface->resource);
data_device_resource = wl_resource_find_for_client (&seat->data_device.resource_list, client);
if (drag_grab->drag_data_source && data_device_resource)
offer = meta_wayland_data_source_send_offer (drag_grab->drag_data_source,
data_device_resource);
drag_grab->drag_focus = surface;
drag_grab->drag_focus_data_device = data_device_resource;
meta_wayland_surface_drag_dest_focus_in (drag_grab->drag_focus,
offer ? wl_resource_get_user_data (offer) : NULL);
}
MetaWaylandSurface *
meta_wayland_drag_grab_get_focus (MetaWaylandDragGrab *drag_grab)
{
return drag_grab->drag_focus;
}
static void
drag_grab_focus (MetaWaylandPointerGrab *grab,
MetaWaylandSurface *surface)
{
MetaWaylandDragGrab *drag_grab = (MetaWaylandDragGrab*) grab;
meta_wayland_drag_grab_set_focus (drag_grab, surface);
}
static void
drag_grab_motion (MetaWaylandPointerGrab *grab,
const ClutterEvent *event)
{
MetaWaylandDragGrab *drag_grab = (MetaWaylandDragGrab*) grab;
if (drag_grab->drag_focus)
meta_wayland_surface_drag_dest_motion (drag_grab->drag_focus, event);
if (drag_grab->drag_surface)
meta_feedback_actor_update (META_FEEDBACK_ACTOR (drag_grab->feedback_actor),
event);
}
static void
data_device_end_drag_grab (MetaWaylandDragGrab *drag_grab)
{
meta_wayland_drag_grab_set_focus (drag_grab, NULL);
if (drag_grab->drag_origin)
{
drag_grab->drag_origin = NULL;
wl_list_remove (&drag_grab->drag_origin_listener.link);
}
if (drag_grab->drag_surface)
{
drag_grab->drag_surface = NULL;
wl_list_remove (&drag_grab->drag_icon_listener.link);
}
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
if (drag_grab->drag_data_source)
g_object_weak_unref (G_OBJECT (drag_grab->drag_data_source),
drag_grab_data_source_destroyed,
drag_grab);
if (drag_grab->feedback_actor)
{
clutter_actor_remove_all_children (drag_grab->feedback_actor);
clutter_actor_destroy (drag_grab->feedback_actor);
}
drag_grab->seat->data_device.current_grab = NULL;
meta_wayland_pointer_end_grab (drag_grab->generic.pointer);
g_slice_free (MetaWaylandDragGrab, drag_grab);
}
static void
drag_grab_button (MetaWaylandPointerGrab *grab,
const ClutterEvent *event)
{
MetaWaylandDragGrab *drag_grab = (MetaWaylandDragGrab*) grab;
MetaWaylandSeat *seat = drag_grab->seat;
ClutterEventType event_type = clutter_event_type (event);
if (drag_grab->generic.pointer->grab_button == clutter_event_get_button (event) &&
event_type == CLUTTER_BUTTON_RELEASE)
{
gboolean success = FALSE;
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
if (meta_wayland_data_source_has_target (drag_grab->drag_data_source))
{
meta_wayland_surface_drag_dest_drop (drag_grab->drag_focus);
success = TRUE;
}
/* Finish drag and let actor self-destruct */
meta_dnd_actor_drag_finish (META_DND_ACTOR (drag_grab->feedback_actor),
success);
drag_grab->feedback_actor = NULL;
}
if (seat->pointer.button_count == 0 &&
event_type == CLUTTER_BUTTON_RELEASE)
data_device_end_drag_grab (drag_grab);
}
static const MetaWaylandPointerGrabInterface drag_grab_interface = {
drag_grab_focus,
drag_grab_motion,
drag_grab_button,
};
static void
destroy_data_device_origin (struct wl_listener *listener, void *data)
{
MetaWaylandDragGrab *drag_grab =
wl_container_of (listener, drag_grab, drag_origin_listener);
drag_grab->drag_origin = NULL;
meta_wayland_data_device_set_dnd_source (&drag_grab->seat->data_device, NULL);
data_device_end_drag_grab (drag_grab);
}
static void
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
drag_grab_data_source_destroyed (gpointer data, GObject *where_the_object_was)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
MetaWaylandDragGrab *drag_grab = data;
drag_grab->drag_data_source = NULL;
meta_wayland_data_device_set_dnd_source (&drag_grab->seat->data_device, NULL);
data_device_end_drag_grab (drag_grab);
}
static void
destroy_data_device_icon (struct wl_listener *listener, void *data)
{
MetaWaylandDragGrab *drag_grab =
wl_container_of (listener, drag_grab, drag_icon_listener);
drag_grab->drag_surface = NULL;
if (drag_grab->feedback_actor)
clutter_actor_remove_all_children (drag_grab->feedback_actor);
}
void
meta_wayland_data_device_start_drag (MetaWaylandDataDevice *data_device,
struct wl_client *client,
const MetaWaylandPointerGrabInterface *funcs,
MetaWaylandSurface *surface,
MetaWaylandDataSource *source,
MetaWaylandSurface *icon_surface)
{
MetaWaylandSeat *seat = wl_container_of (data_device, seat, data_device);
MetaWaylandDragGrab *drag_grab;
ClutterPoint pos, stage_pos;
data_device->current_grab = drag_grab = g_slice_new0 (MetaWaylandDragGrab);
drag_grab->generic.interface = funcs;
drag_grab->generic.pointer = &seat->pointer;
drag_grab->drag_client = client;
drag_grab->seat = seat;
drag_grab->drag_origin = surface;
drag_grab->drag_origin_listener.notify = destroy_data_device_origin;
wl_resource_add_destroy_listener (surface->resource,
&drag_grab->drag_origin_listener);
clutter_input_device_get_coords (seat->pointer.device, NULL, &pos);
clutter_actor_transform_stage_point (CLUTTER_ACTOR (meta_surface_actor_get_texture (surface->surface_actor)),
pos.x, pos.y, &stage_pos.x, &stage_pos.y);
drag_grab->drag_start_x = stage_pos.x;
drag_grab->drag_start_y = stage_pos.y;
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
g_object_weak_ref (G_OBJECT (source),
drag_grab_data_source_destroyed,
drag_grab);
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
drag_grab->drag_data_source = source;
meta_wayland_data_device_set_dnd_source (data_device,
drag_grab->drag_data_source);
if (icon_surface)
{
drag_grab->drag_surface = icon_surface;
drag_grab->drag_icon_listener.notify = destroy_data_device_icon;
wl_resource_add_destroy_listener (icon_surface->resource,
&drag_grab->drag_icon_listener);
drag_grab->feedback_actor = meta_dnd_actor_new (CLUTTER_ACTOR (drag_grab->drag_origin->surface_actor),
drag_grab->drag_start_x,
drag_grab->drag_start_y);
meta_feedback_actor_set_anchor (META_FEEDBACK_ACTOR (drag_grab->feedback_actor),
-drag_grab->drag_surface->offset_x,
-drag_grab->drag_surface->offset_y);
clutter_actor_add_child (drag_grab->feedback_actor,
CLUTTER_ACTOR (drag_grab->drag_surface->surface_actor));
meta_feedback_actor_set_position (META_FEEDBACK_ACTOR (drag_grab->feedback_actor),
pos.x, pos.y);
}
meta_wayland_pointer_start_grab (&seat->pointer, (MetaWaylandPointerGrab*) drag_grab);
}
void
meta_wayland_data_device_end_drag (MetaWaylandDataDevice *data_device)
{
if (data_device->current_grab)
data_device_end_drag_grab (data_device->current_grab);
}
static void
data_device_start_drag (struct wl_client *client,
struct wl_resource *resource,
struct wl_resource *source_resource,
struct wl_resource *origin_resource,
struct wl_resource *icon_resource, guint32 serial)
{
MetaWaylandDataDevice *data_device = wl_resource_get_user_data (resource);
MetaWaylandSeat *seat = wl_container_of (data_device, seat, data_device);
MetaWaylandSurface *surface = NULL, *icon_surface = NULL;
MetaWaylandDataSource *drag_source = NULL;
if (origin_resource)
surface = wl_resource_get_user_data (origin_resource);
if (!surface)
return;
if (seat->pointer.button_count == 0 ||
seat->pointer.grab_serial != serial ||
!seat->pointer.focus_surface ||
seat->pointer.focus_surface != surface)
return;
/* FIXME: Check that the data source type array isn't empty. */
if (data_device->current_grab ||
seat->pointer.grab != &seat->pointer.default_grab)
return;
if (icon_resource)
icon_surface = wl_resource_get_user_data (icon_resource);
if (source_resource)
drag_source = wl_resource_get_user_data (source_resource);
if (icon_resource &&
!meta_wayland_surface_assign_role (icon_surface,
META_TYPE_WAYLAND_SURFACE_ROLE_DND))
{
wl_resource_post_error (resource, WL_DATA_DEVICE_ERROR_ROLE,
"wl_surface@%d already has a different role",
wl_resource_get_id (icon_resource));
return;
}
meta_wayland_pointer_set_focus (&seat->pointer, NULL);
meta_wayland_data_device_start_drag (data_device, client,
&drag_grab_interface,
surface, drag_source, icon_surface);
}
static void
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
selection_data_source_destroyed (gpointer data, GObject *object_was_here)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
MetaWaylandDataDevice *data_device = data;
MetaWaylandSeat *seat = wl_container_of (data_device, seat, data_device);
struct wl_resource *data_device_resource;
struct wl_client *focus_client = NULL;
data_device->selection_data_source = NULL;
focus_client = meta_wayland_keyboard_get_focus_client (&seat->keyboard);
if (focus_client)
{
data_device_resource = wl_resource_find_for_client (&data_device->resource_list, focus_client);
if (data_device_resource)
wl_data_device_send_selection (data_device_resource, NULL);
}
}
static void
meta_wayland_source_send (MetaWaylandDataSource *source,
const gchar *mime_type,
gint fd)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
MetaWaylandDataSourceWayland *source_wayland =
META_WAYLAND_DATA_SOURCE_WAYLAND (source);
wl_data_source_send_send (source_wayland->resource, mime_type, fd);
close (fd);
}
static void
meta_wayland_source_target (MetaWaylandDataSource *source,
const gchar *mime_type)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
MetaWaylandDataSourceWayland *source_wayland =
META_WAYLAND_DATA_SOURCE_WAYLAND (source);
wl_data_source_send_target (source_wayland->resource, mime_type);
}
static void
meta_wayland_source_cancel (MetaWaylandDataSource *source)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
MetaWaylandDataSourceWayland *source_wayland =
META_WAYLAND_DATA_SOURCE_WAYLAND (source);
wl_data_source_send_cancelled (source_wayland->resource);
}
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
static void
meta_wayland_source_finalize (GObject *object)
{
G_OBJECT_CLASS (meta_wayland_data_source_parent_class)->finalize (object);
}
static void
meta_wayland_data_source_wayland_init (MetaWaylandDataSourceWayland *source_wayland)
{
}
static void
meta_wayland_data_source_wayland_class_init (MetaWaylandDataSourceWaylandClass *klass)
{
GObjectClass *object_class = G_OBJECT_CLASS (klass);
MetaWaylandDataSourceClass *data_source_class =
META_WAYLAND_DATA_SOURCE_CLASS (klass);
object_class->finalize = meta_wayland_source_finalize;
data_source_class->send = meta_wayland_source_send;
data_source_class->target = meta_wayland_source_target;
data_source_class->cancel = meta_wayland_source_cancel;
}
static void
meta_wayland_data_source_finalize (GObject *object)
{
MetaWaylandDataSource *source = META_WAYLAND_DATA_SOURCE (object);
MetaWaylandDataSourcePrivate *priv =
meta_wayland_data_source_get_instance_private (source);
char **pos;
wl_array_for_each (pos, &priv->mime_types)
g_free (*pos);
wl_array_release (&priv->mime_types);
G_OBJECT_CLASS (meta_wayland_data_source_parent_class)->finalize (object);
}
static void
meta_wayland_data_source_init (MetaWaylandDataSource *source)
{
MetaWaylandDataSourcePrivate *priv =
meta_wayland_data_source_get_instance_private (source);
wl_array_init (&priv->mime_types);
}
static void
meta_wayland_data_source_class_init (MetaWaylandDataSourceClass *klass)
{
GObjectClass *object_class = G_OBJECT_CLASS (klass);
object_class->finalize = meta_wayland_data_source_finalize;
}
static void
meta_wayland_drag_dest_focus_in (MetaWaylandDataDevice *data_device,
MetaWaylandSurface *surface,
MetaWaylandDataOffer *offer)
{
MetaWaylandDragGrab *grab = data_device->current_grab;
struct wl_display *display;
struct wl_client *client;
wl_fixed_t sx, sy;
if (!grab->drag_focus_data_device)
return;
client = wl_resource_get_client (surface->resource);
display = wl_client_get_display (client);
grab->drag_focus_listener.notify = destroy_drag_focus;
wl_resource_add_destroy_listener (grab->drag_focus_data_device,
&grab->drag_focus_listener);
meta_wayland_pointer_get_relative_coordinates (grab->generic.pointer,
surface, &sx, &sy);
wl_data_device_send_enter (grab->drag_focus_data_device,
wl_display_next_serial (display),
surface->resource, sx, sy, offer->resource);
}
static void
meta_wayland_drag_dest_focus_out (MetaWaylandDataDevice *data_device,
MetaWaylandSurface *surface)
{
MetaWaylandDragGrab *grab = data_device->current_grab;
if (grab->drag_focus_data_device)
wl_data_device_send_leave (grab->drag_focus_data_device);
wl_list_remove (&grab->drag_focus_listener.link);
grab->drag_focus_data_device = NULL;
}
static void
meta_wayland_drag_dest_motion (MetaWaylandDataDevice *data_device,
MetaWaylandSurface *surface,
const ClutterEvent *event)
{
MetaWaylandDragGrab *grab = data_device->current_grab;
wl_fixed_t sx, sy;
meta_wayland_pointer_get_relative_coordinates (grab->generic.pointer,
grab->drag_focus,
&sx, &sy);
wl_data_device_send_motion (grab->drag_focus_data_device,
clutter_event_get_time (event),
sx, sy);
}
static void
meta_wayland_drag_dest_drop (MetaWaylandDataDevice *data_device,
MetaWaylandSurface *surface)
{
MetaWaylandDragGrab *grab = data_device->current_grab;
wl_data_device_send_drop (grab->drag_focus_data_device);
}
static const MetaWaylandDragDestFuncs meta_wayland_drag_dest_funcs = {
meta_wayland_drag_dest_focus_in,
meta_wayland_drag_dest_focus_out,
meta_wayland_drag_dest_motion,
meta_wayland_drag_dest_drop
};
const MetaWaylandDragDestFuncs *
meta_wayland_data_device_get_drag_dest_funcs (void)
{
return &meta_wayland_drag_dest_funcs;
}
void
meta_wayland_data_device_set_dnd_source (MetaWaylandDataDevice *data_device,
MetaWaylandDataSource *source)
{
if (data_device->dnd_data_source == source)
return;
if (data_device->dnd_data_source)
g_object_remove_weak_pointer (G_OBJECT (data_device->dnd_data_source),
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
(gpointer *)&data_device->dnd_data_source);
data_device->dnd_data_source = source;
if (source)
g_object_add_weak_pointer (G_OBJECT (data_device->dnd_data_source),
(gpointer *)&data_device->dnd_data_source);
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
wl_signal_emit (&data_device->dnd_ownership_signal, source);
}
void
meta_wayland_data_device_set_selection (MetaWaylandDataDevice *data_device,
MetaWaylandDataSource *source,
guint32 serial)
{
MetaWaylandSeat *seat = wl_container_of (data_device, seat, data_device);
struct wl_resource *data_device_resource, *offer;
struct wl_client *focus_client;
if (data_device->selection_data_source &&
data_device->selection_serial - serial < UINT32_MAX / 2)
return;
if (data_device->selection_data_source)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
meta_wayland_data_source_cancel (data_device->selection_data_source);
g_object_weak_unref (G_OBJECT (data_device->selection_data_source),
selection_data_source_destroyed,
data_device);
data_device->selection_data_source = NULL;
}
data_device->selection_data_source = source;
data_device->selection_serial = serial;
focus_client = meta_wayland_keyboard_get_focus_client (&seat->keyboard);
if (focus_client)
{
data_device_resource = wl_resource_find_for_client (&data_device->resource_list, focus_client);
if (data_device_resource)
{
if (data_device->selection_data_source)
2014-04-20 11:35:53 -04:00
{
offer = meta_wayland_data_source_send_offer (data_device->selection_data_source, data_device_resource);
wl_data_device_send_selection (data_device_resource, offer);
2014-04-20 11:35:53 -04:00
}
else
{
wl_data_device_send_selection (data_device_resource, NULL);
2014-04-20 11:35:53 -04:00
}
}
}
if (source)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
g_object_weak_ref (G_OBJECT (source),
selection_data_source_destroyed,
data_device);
}
wl_signal_emit (&data_device->selection_ownership_signal, source);
}
static void
data_device_set_selection (struct wl_client *client,
struct wl_resource *resource,
struct wl_resource *source_resource,
guint32 serial)
{
MetaWaylandDataDevice *data_device = wl_resource_get_user_data (resource);
MetaWaylandDataSource *source;
if (source_resource)
source = wl_resource_get_user_data (source_resource);
else
source = NULL;
/* FIXME: Store serial and check against incoming serial here. */
meta_wayland_data_device_set_selection (data_device, source, serial);
}
static void
data_device_release(struct wl_client *client, struct wl_resource *resource)
{
wl_resource_destroy(resource);
}
static const struct wl_data_device_interface data_device_interface = {
data_device_start_drag,
data_device_set_selection,
data_device_release,
};
static void
destroy_data_source (struct wl_resource *resource)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
MetaWaylandDataSourceWayland *source = wl_resource_get_user_data (resource);
source->resource = NULL;
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
g_object_unref (source);
}
static void
create_data_source (struct wl_client *client,
struct wl_resource *resource, guint32 id)
{
struct wl_resource *source_resource;
source_resource = wl_resource_create (client, &wl_data_source_interface,
wl_resource_get_version (resource), id);
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
meta_wayland_data_source_wayland_new (source_resource);
}
static void
get_data_device (struct wl_client *client,
struct wl_resource *manager_resource,
guint32 id, struct wl_resource *seat_resource)
{
MetaWaylandSeat *seat = wl_resource_get_user_data (seat_resource);
2014-07-10 10:19:47 -04:00
struct wl_resource *cr;
cr = wl_resource_create (client, &wl_data_device_interface, wl_resource_get_version (manager_resource), id);
wl_resource_set_implementation (cr, &data_device_interface, &seat->data_device, unbind_resource);
wl_list_insert (&seat->data_device.resource_list, wl_resource_get_link (cr));
}
static const struct wl_data_device_manager_interface manager_interface = {
create_data_source,
get_data_device
};
static void
bind_manager (struct wl_client *client,
void *data, guint32 version, guint32 id)
{
struct wl_resource *resource;
resource = wl_resource_create (client, &wl_data_device_manager_interface, version, id);
wl_resource_set_implementation (resource, &manager_interface, NULL, NULL);
}
void
meta_wayland_data_device_manager_init (MetaWaylandCompositor *compositor)
{
if (wl_global_create (compositor->wayland_display,
&wl_data_device_manager_interface,
META_WL_DATA_DEVICE_MANAGER_VERSION,
NULL, bind_manager) == NULL)
g_error ("Could not create data_device");
}
void
meta_wayland_data_device_init (MetaWaylandDataDevice *data_device)
{
wl_list_init (&data_device->resource_list);
wl_signal_init (&data_device->selection_ownership_signal);
wl_signal_init (&data_device->dnd_ownership_signal);
}
void
meta_wayland_data_device_set_keyboard_focus (MetaWaylandDataDevice *data_device)
{
MetaWaylandSeat *seat = wl_container_of (data_device, seat, data_device);
struct wl_client *focus_client;
struct wl_resource *data_device_resource, *offer;
MetaWaylandDataSource *source;
focus_client = meta_wayland_keyboard_get_focus_client (&seat->keyboard);
if (!focus_client)
return;
data_device_resource = wl_resource_find_for_client (&data_device->resource_list, focus_client);
if (!data_device_resource)
return;
source = data_device->selection_data_source;
if (source)
{
offer = meta_wayland_data_source_send_offer (source, data_device_resource);
wl_data_device_send_selection (data_device_resource, offer);
}
else
wl_data_device_send_selection (data_device_resource, NULL);
}
gboolean
meta_wayland_data_device_is_dnd_surface (MetaWaylandDataDevice *data_device,
MetaWaylandSurface *surface)
{
return data_device->current_grab &&
data_device->current_grab->drag_surface == surface;
}
void
meta_wayland_data_device_update_dnd_surface (MetaWaylandDataDevice *data_device)
{
MetaWaylandDragGrab *drag_grab;
if (!data_device->current_grab)
return;
drag_grab = data_device->current_grab;
if (!drag_grab->feedback_actor || !drag_grab->drag_surface)
return;
meta_feedback_actor_set_anchor (META_FEEDBACK_ACTOR (drag_grab->feedback_actor),
-drag_grab->drag_surface->offset_x,
-drag_grab->drag_surface->offset_y);
}
gboolean
meta_wayland_data_source_has_mime_type (const MetaWaylandDataSource *source,
const gchar *mime_type)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
MetaWaylandDataSourcePrivate *priv =
meta_wayland_data_source_get_instance_private (source);
gchar **p;
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
wl_array_for_each (p, &priv->mime_types)
{
if (g_strcmp0 (mime_type, *p) == 0)
return TRUE;
}
return FALSE;
}
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
static MetaWaylandDataSource *
meta_wayland_data_source_wayland_new (struct wl_resource *resource)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
MetaWaylandDataSourceWayland *source_wayland =
g_object_new (META_TYPE_WAYLAND_DATA_SOURCE_WAYLAND, NULL);
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
source_wayland->resource = resource;
wl_resource_set_implementation (resource, &data_source_interface,
source_wayland, destroy_data_source);
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
return META_WAYLAND_DATA_SOURCE (source_wayland);
}
gboolean
meta_wayland_data_source_add_mime_type (MetaWaylandDataSource *source,
const gchar *mime_type)
{
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
MetaWaylandDataSourcePrivate *priv =
meta_wayland_data_source_get_instance_private (source);
gchar **pos;
wayland: Make MetaWaylandDataSource ownership protocol specific Firstly, this patch makes MetawaylandDataSource a GObject. This is in order to easier track its lifetime without adding destroy signals etc. It also makes the vfunc table GObject class functions instead while at it, as well as moves protocol specific part of the source into their own implementations. An important part of this patch is the change of ownership. Prior to this patch, MetaWaylandDataDevice would kind of own the source, but for Wayland sources it would remove it if the corresponding wl_resource was destroyed. For XWayland clients it would own it completely, and only remove it if the source was replaced. This patch changes so that the protocol implementation owns the source. For Wayland sources, the wl_resource owns the source, and the MetaWaylandDataDevice sets a weak reference (so in other words, no semantical changes really). For XWayland sources, the source is owned by the selection bridge, and not removed until replaced or if the client goes away. Given the changes in ownership, data offers may now properly track the lifetime of a source it represents. Prior to this patch, if an offer with an XWayland source would loose its source, it wouldn't get notified and have an invalid pointer it would potentally crash on. For Wayland sources, an offer would have a weak reference and clean itself up if the source went away. This patch changes so the behavior is consistent, meaning a weak reference is added to the source GObject so that the offer can behave correctly both for Wayland sources and XWayland sources. https://bugzilla.gnome.org/show_bug.cgi?id=750680
2015-06-17 22:25:11 -04:00
pos = wl_array_add (&priv->mime_types, sizeof (*pos));
if (pos)
{
*pos = g_strdup (mime_type);
return *pos != NULL;
}
return FALSE;
}