3 - scanner: wl_* prefix removal: split it out into a namespace part so
4 we can call variables "surface" instead of "wl_surface"?
6 - Framebased input event delivery.
8 - Protocol for arbitrating access to scanout buffers (physically
9 contiguous memory). When a client goes fullscreen (or ideally as
10 the compositor starts the animation that will make it fullscreen)
11 we send a "give up your scanout buffer" to the current fullscreen
12 client (if any) and when the client acks that we send a "try to
13 allocate a scanout buffer now" event to the fullscreen-to-be
16 - Next steps based on EGL_WL_bind_display: create EGLImageKHR from
17 shm buffers? async auth in the implementation of the extension?
19 - wayland-egl: lazy-copy-back swapbuffer, sub-window, scanout flags
22 - configure should provide dx_left, dx_right, dy_top, dy_bottom, or
23 dx, dy, width and height.
25 - surface.set_grab_mode(GRAB_OWNER_EVENTS vs GRAB_SURFACE_EVENTS), to
26 make menus work right: click and drag in a menubar grabs the
27 pointer to the menubar (which we need for detecting motion into
28 another menu item), but we need events for the popup menu surface
31 - The message format has to include information about number of fds
32 in the message so we can skip a message correctly. Or we should
33 just give up on trying to recover from unknown messages. We need
34 to make sure you never get a message from an interface you don't
35 know about (using per-client id space and subscribe) or include
36 information on number of fds, so marshalling logic can skip.
38 - generate pointer_focus (and drag focus) on raise/lower, move
39 windows, all kinds of changes in surface stacking.
43 buffer = drm.create_buffer(); /* buffer with stuff in it */
45 cache.upload(buffer, x, y, width, height, int hash)
47 drm.buffer: id, name, stride etc /* event to announce cache buffer */
49 cache.image: hash, buffer, x, y, stride /* event to announce
50 * location in cache */
52 cache.reject: hash /* no upload for you! */
54 cache.retire: buffer /* cache has stopped using buffer, please
55 * reupload whatever you had in that buffer */
59 - Drag should not be tied to a source surface, just the client.
60 the grab will break if the surface goes away, but the wl_drag
61 struct doesn't need to hold on to the source surface.
63 - Root window must send NULL type (to decline drop) or
64 x-wayland/root-something type if the source offers that. But
65 the target deletes the drag_offer object when drag.pointer_focus
68 - How do we animate the drag icon back to the drag origin in case
69 of a failed drag? Client should set drag icon separately,
70 compositor can do it then.
72 - How to handle surfaces from clients that don't know about dnd or
73 don't care? Maybe the dnd object should have a
74 dnd.register_surface() method so clients can opt-in the surfaces
75 that will participate in dnd. Or just assume client is not
76 participating until we receive an accept request.
78 - Selection/copy+paste issues: is it sufficient to only introduce
79 the selection offer when a client receives kb focus? Or maybe
80 it is actually a security feature? Clipboard manager in server
81 for retained selections?
83 - Pointer image issue:
85 - A direct touch input device (eg touch screen) doesn't have a
86 pointer; indicate that somehow.
88 - Cursor themes, tie in with glyph/image cache.
90 - A "please suspend" event from the compositor, to indicate to an
91 application that it's no longer visible/active. Or maybe discard
92 buffer, as in "wayland discarded your buffer, it's no longer
93 visible, you can stop updating it now.", reattach, as in "oh hey,
94 I'm about to show your buffer that I threw away, what was it
95 again?". for wayland system compositor vt switcing, for example,
96 to be able to throw away the surfaces in the session we're
97 switching away from. for minimized windows that we don't want live
100 - Per client id space. Each client has an entire 32 bit id namespace
101 to itself. On the server side, each struct wl_client has an object
102 hash table. Object announcements use a server id space and clients
103 must respond with subscribe request with a client id for the
104 object. Part of wl_proxy_create_for_id():
106 wl_display_subscribe(display, id, new_id, my_version);
110 wl_display_bind(display, id, new_id, my_version);
114 - Maps the global object into the client id space, lets client
115 allocate the id. All ids are allocated by the client this way,
116 which fixes the range protocol problem.
118 - Tells the server that the client is interested in events from
119 the object. Lets the server know that a client participates in a
120 certain protocol (like drag and drop), so the server can account
121 for whether or not the client is expected to reply
123 - Server emits initial object state event(s) in reponse to
124 receiving the subscribe request. Introduces an extra round trip
125 at initialization time, but the server will still announces all
126 objects in one burst and the client can subscribe in a burst as
129 - Separates client resources, since each client will have it's own
130 hash table. It's not longer possible to guess the id of another
131 surface and access it.
133 - Server must track the client id for each client an object is
134 exposed to. In some cases we know this (a surface is always
135 only owned by one client), in other cases it provides a way to
136 track who's interested in the object events. For input device
137 events, we can look up the client name when it receives pointer
138 focus or keyboard focus and cache it in the device.
140 - Server must know which id to send when passing object references
141 in events. We could say that any object we're passing to a
142 client must have a server id, and each client has a server id ->
145 - Event when a surface moves from one output to another.
147 - input device discovery, hotplug
149 - Advertise axes as part of the discovery, use something like
150 "org.wayland.input.x" to identify the axes.
152 - keyboard state, layout events at connect time and when it
153 changes, keyboard leds
159 - synaptics, 3-button emulation, scim
161 - drm bo access control, authentication, flink_to
163 - Range protocol may not be sufficient... if a server cycles through
164 2^32 object IDs we don't have a way to handle wrapping. And since
165 we hand out a range of 256 IDs to each new clients, we're just
166 talking about 2^24 clients. That's 31 years with a new client
167 every minute... Maybe just use bigger ranges, then it's feasible
168 to track and garbage collect them when a client dies.
170 - Add protocol to let applications specify the effective/logical
171 surface rectangle, that is, the edge of the window, ignoring drop
172 shadows and other padding. The compositor needs this for snapping
173 and constraining window motion. Also, maybe communicate the opaque
174 region of the window (or just a conservative, simple estimate), to
175 let the compositor reduce overdraw.
177 - Protocol for specifying title bar rectangle (for moving
178 unresponsive apps) and a rectangle for the close button (for
179 detecting ignored close clicks).
181 - multi gpu, needs queue and seqno to wait on in requests
185 - pull in actions logic from xserver
187 - pull in keycode to keysym logic from libX11
189 - expose alloc functions in libxkbcommon, drop xserver funcs?
191 - pull the logic to write the xkb file from xkb_desc and names
192 into libxkbcommon and just build up the new xkb_desc instead of
193 dump+parse? (XkbWriteXKBKeymapForNames followed by
194 xkb_compile_keymap_from_string in XkbDDXLoadKeymapByNames)
196 - pull in keysym defs as XKB_KEY_BackSpace
198 - figure out what other X headers we can get rid of, make it not
199 need X at all (except when we gen the keysyms).
201 - Sort out namespace pollution (XkbFoo macros, atom funcs etc).
203 - Sort out 32 bit vmods and serialization
210 - draw window decorations in gtkwindow.c
212 - Details about pointer grabs. wayland doesn't have active grabs,
213 menus will behave subtly different. Under X, clicking a menu
214 open grabs the pointer and clicking outside the window pops down
215 the menu and swallows the click. without active grabs we can't
216 swallow the click. I'm sure there much more...
220 - Investigate DirectFB on Wayland (or is that Wayland on DirectFB?)
222 - SDL port, bnf has work in progress here:
223 http://cgit.freedesktop.org/~bnf/sdl-wayland/
225 - libva + eglimage + kms integration
229 - map multiple wayland input devices to MPX in Xorg.
231 - rootless; avoid allocating and setting the front buffer, draw
232 window decorations in the X server (!), how to map input?
237 - A wayland settings protocol to tell clients about themes (icons,
238 cursors, widget themes), fonts details (family, hinting
239 preferences) etc. Just send all settings at connect time, send
240 updates when a setting change. Getting a little close to gconf
241 here, but could be pretty simple:
243 interface "settings":
244 event int_value(string name, int value)
245 event string_value(string name, string value)
247 but maybe it's better to just require that clients get that from
248 somewhere else (gconf/dbus).
253 - AF_WAYLAND - A new socket type. Eliminate compositor context
254 switch by making kernel understand enough of wayland that it can
255 forward input events as wayland events and do page flipping in
256 response to surface_attach requests:
258 - ioctl(wayland_fd, "surface_attach to object 5 should do a kms page
261 - what about multiple crtcs? what about frame event for other
264 - forward these input devices to the client
266 - "scancode 124 pressed or released with scan codes 18,22 and 30
267 held down gives control back to userspace wayland.
269 - what about maintaining cursor position? what about pointer
270 acceleration? maybe this only works in "client cursor mode",
271 where wayland hides the cursor and only sends relative events?
272 Solves the composited cursor problem. How does X show its
275 - Probably not worth it.