video/out/wayland_common: support IME usage via text-input-v3

This is useful for text input in, for example, console.lua. Each
character in the commit string gets turned into an mpv key press.
Pre-edit strings are not handled, since there's currently no good way to
handle that or make it useful to text input scripts. Like win32, which I
tested in wine, another limitation is that the composition window is
always positioned at the top left of the window, since we cannot get
useful positioning hints from mpv scripts. It allows the composition
window to be within the window and avoids obstructing the console
prompt.

This can be enabled/disabled with --input-ime=<yes|no> (default: yes).

DOCS/interface-changes/input.txt

@@ -0,0 +1 @@
+add `--input-ime` to enable or disable the IME on supported VOs (Wayland)

DOCS/man/options.rst

@@ -4460,6 +4460,17 @@ Input
     ``N`` pixels while the mouse button is being held down (default: 3). This only
     affects VOs which support the ``begin-vo-dragging`` command.
 
+``--input-ime=<yes|no>``
+    Enable keyboard input via an active input method (IME) connected to the VO.
+    (default: yes). The input popup window, if there is any, is always
+    positioned at the top left of the window. Whether pre-edit text is drawn
+    depends on the platform. You may need to configure your IME to display the
+    pre-edit inside of the input popup window if you cannot read the pre-edit
+    text in the mpv window.
+
+    Wayland only. On Windows, the IME is currently always enabled. This option
+    is not applicable to terminal input.
+
 OSD
 ---
 

options/options.c

@@ -150,6 +150,7 @@ static const m_option_t mp_vo_opt_list[] = {
     {"fs", OPT_ALIAS("fullscreen")},
     {"input-cursor-passthrough", OPT_BOOL(cursor_passthrough)},
     {"native-keyrepeat", OPT_BOOL(native_keyrepeat)},
+    {"input-ime", OPT_BOOL(input_ime)},
     {"panscan", OPT_FLOAT(panscan), M_RANGE(0.0, 1.0)},
     {"video-zoom", OPT_FLOAT(zoom), M_RANGE(-20.0, 20.0)},
     {"video-pan-x", OPT_FLOAT(pan_x)},
@@ -259,6 +260,7 @@ const struct m_sub_options vo_sub_opts = {
         .keepaspect = true,
         .keepaspect_window = true,
         .native_fs = true,
+        .input_ime = true,
         .taskbar_progress = true,
         .show_in_taskbar = true,
         .border = true,

options/options.h

@@ -34,6 +34,7 @@ typedef struct mp_vo_opts {
     bool x11_wid_title;
     bool cursor_passthrough;
     bool native_keyrepeat;
+    bool input_ime;
 
     int wl_configure_bounds;
     int wl_content_type;

video/out/meson.build

@@ -3,6 +3,7 @@ protocols = [[wl_protocol_dir, 'stable/presentation-time/presentation-time.xml']
              [wl_protocol_dir, 'stable/viewporter/viewporter.xml'],
              [wl_protocol_dir, 'stable/xdg-shell/xdg-shell.xml'],
              [wl_protocol_dir, 'unstable/idle-inhibit/idle-inhibit-unstable-v1.xml'],
+             [wl_protocol_dir, 'unstable/text-input/text-input-unstable-v3.xml'],
              [wl_protocol_dir, 'unstable/linux-dmabuf/linux-dmabuf-unstable-v1.xml'],
              [wl_protocol_dir, 'unstable/xdg-decoration/xdg-decoration-unstable-v1.xml'],
              [wl_protocol_dir, 'staging/content-type/content-type-v1.xml'],

video/out/wayland_common.c

@@ -37,6 +37,7 @@
 
 // Generated from wayland-protocols
 #include "idle-inhibit-unstable-v1.h"
+#include "text-input-unstable-v3.h"
 #include "linux-dmabuf-unstable-v1.h"
 #include "presentation-time.h"
 #include "xdg-decoration-unstable-v1.h"
@@ -191,6 +192,7 @@ struct vo_wayland_seat {
     struct wl_pointer  *pointer;
     struct wl_touch    *touch;
     struct wl_data_device *dnd_ddev;
+    struct vo_wayland_text_input *text_input;
     /* TODO: unvoid this if required wayland protocols is bumped to 1.32+ */
     void *cursor_shape_device;
     uint32_t pointer_enter_serial;
@@ -230,6 +232,13 @@ struct vo_wayland_data_offer {
     bool offered_plain_text;
 };
 
+struct vo_wayland_text_input {
+    struct zwp_text_input_v3 *text_input;
+    uint32_t serial;
+    char *commit_string;
+    bool has_focus;
+};
+
 static bool single_output_spanned(struct vo_wayland_state *wl);
 
 static int check_for_resize(struct vo_wayland_state *wl, int edge_pixels,
@@ -877,6 +886,96 @@ static const struct wl_data_device_listener data_device_listener = {
     data_device_handle_selection,
 };
 
+static void enable_ime(struct vo_wayland_text_input *ti)
+{
+    if (!ti->has_focus)
+        return;
+
+    zwp_text_input_v3_enable(ti->text_input);
+    zwp_text_input_v3_set_content_type(
+        ti->text_input,
+        ZWP_TEXT_INPUT_V3_CONTENT_HINT_NONE,
+        ZWP_TEXT_INPUT_V3_CONTENT_PURPOSE_NORMAL);
+    zwp_text_input_v3_set_cursor_rectangle(ti->text_input, 0, 0, 0, 0);
+    zwp_text_input_v3_commit(ti->text_input);
+    ti->serial++;
+}
+
+static void disable_ime(struct vo_wayland_text_input *ti)
+{
+    zwp_text_input_v3_disable(ti->text_input);
+    zwp_text_input_v3_commit(ti->text_input);
+    ti->serial++;
+}
+
+static void text_input_enter(void *data, struct zwp_text_input_v3 *zwp_text_input_v3,
+                             struct wl_surface *surface)
+{
+    struct vo_wayland_seat *s = data;
+    struct vo_wayland_text_input *ti = s->text_input;
+    struct vo_wayland_state *wl = s->wl;
+
+    ti->has_focus = true;
+
+    if (!wl->opts->input_ime)
+        return;
+
+    enable_ime(ti);
+}
+
+static void text_input_leave(void *data, struct zwp_text_input_v3 *zwp_text_input_v3,
+                             struct wl_surface *surface)
+{
+    struct vo_wayland_seat *s = data;
+    struct vo_wayland_text_input *ti = s->text_input;
+
+    ti->has_focus = false;
+    disable_ime(ti);
+}
+
+static void text_input_preedit_string(void *data, struct zwp_text_input_v3 *zwp_text_input_v3,
+                                      const char *text, int32_t cursor_begin, int32_t cursor_end)
+{
+}
+
+static void text_input_commit_string(void *data, struct zwp_text_input_v3 *zwp_text_input_v3,
+                                     const char *text)
+{
+    struct vo_wayland_seat *s = data;
+    struct vo_wayland_text_input *ti = s->text_input;
+
+    talloc_replace(ti, ti->commit_string, text);
+}
+
+static void text_input_delete_surrounding_text(void *data,
+                                               struct zwp_text_input_v3 *zwp_text_input_v3,
+                                               uint32_t before_length, uint32_t after_length)
+{
+}
+
+static void text_input_done(void *data, struct zwp_text_input_v3 *zwp_text_input_v3,
+                            uint32_t serial)
+{
+    struct vo_wayland_seat *s = data;
+    struct vo_wayland_text_input *ti = s->text_input;
+    struct vo_wayland_state *wl = s->wl;
+
+    if (ti->serial == serial && ti->commit_string)
+        mp_input_put_key_utf8(wl->vo->input_ctx, 0, bstr0(ti->commit_string));
+
+    if (ti->commit_string)
+        TA_FREEP(&ti->commit_string);
+}
+
+static const struct zwp_text_input_v3_listener text_input_listener = {
+    text_input_enter,
+    text_input_leave,
+    text_input_preedit_string,
+    text_input_commit_string,
+    text_input_delete_surrounding_text,
+    text_input_done,
+};
+
 static void output_handle_geometry(void *data, struct wl_output *wl_output,
                                    int32_t x, int32_t y, int32_t phys_width,
                                    int32_t phys_height, int32_t subpixel,
@@ -1805,6 +1904,11 @@ static void registry_handle_add(void *data, struct wl_registry *reg, uint32_t id
         wl->idle_inhibit_manager = wl_registry_bind(reg, id, &zwp_idle_inhibit_manager_v1_interface, ver);
     }
 
+    if (!strcmp(interface, zwp_text_input_manager_v3_interface.name) && found++) {
+        ver = 1;
+        wl->text_input_manager = wl_registry_bind(reg, id, &zwp_text_input_manager_v3_interface, ver);
+    }
+
     if (found > 1)
         MP_VERBOSE(wl, "Registered interface %s at version %d\n", interface, ver);
 }
@@ -2433,6 +2537,8 @@ static void remove_seat(struct vo_wayland_seat *seat)
         wl_touch_destroy(seat->touch);
     if (seat->dnd_ddev)
         wl_data_device_destroy(seat->dnd_ddev);
+    if (seat->text_input)
+        zwp_text_input_v3_destroy(seat->text_input->text_input);
 #if HAVE_WAYLAND_PROTOCOLS_1_32
     if (seat->cursor_shape_device)
         wp_cursor_shape_device_v1_destroy(seat->cursor_shape_device);
@@ -2744,6 +2850,21 @@ static void toggle_fullscreen(struct vo_wayland_state *wl)
     }
 }
 
+static void toggle_ime(struct vo_wayland_state *wl)
+{
+    struct vo_wayland_seat *seat;
+    wl_list_for_each(seat, &wl->seat_list, link) {
+        struct vo_wayland_text_input *ti = seat->text_input;
+        if (!ti)
+            continue;
+        if (wl->opts->input_ime) {
+            enable_ime(ti);
+        } else {
+            disable_ime(ti);
+        }
+    }
+}
+
 static void toggle_maximized(struct vo_wayland_state *wl)
 {
     if (wl->opts->window_maximized) {
@@ -2935,6 +3056,8 @@ int vo_wayland_control(struct vo *vo, int *events, int request, void *arg)
                 set_content_type(wl);
             if (opt == &opts->cursor_passthrough)
                 set_input_region(wl, opts->cursor_passthrough);
+            if (opt == &opts->input_ime)
+                toggle_ime(wl);
             if (opt == &opts->fullscreen)
                 toggle_fullscreen(wl);
             if (opt == &opts->window_maximized)
@@ -3248,6 +3371,18 @@ bool vo_wayland_init(struct vo *vo)
                    zwp_idle_inhibit_manager_v1_interface.name);
     }
 
+    if (wl->text_input_manager) {
+        struct vo_wayland_seat *seat;
+        wl_list_for_each(seat, &wl->seat_list, link) {
+            seat->text_input = talloc_zero(seat, struct vo_wayland_text_input);
+            seat->text_input->text_input = zwp_text_input_manager_v3_get_text_input(wl->text_input_manager, seat->seat);
+            zwp_text_input_v3_add_listener(seat->text_input->text_input, &text_input_listener, seat);
+        }
+    } else {
+        MP_VERBOSE(wl, "Compositor doesn't support the %s protocol!\n",
+                    zwp_text_input_manager_v3_interface.name);
+    }
+
     wl->display_fd = wl_display_get_fd(wl->display);
 
     update_app_id(wl);
@@ -3406,6 +3541,9 @@ void vo_wayland_uninit(struct vo *vo)
     if (wl->idle_inhibit_manager)
         zwp_idle_inhibit_manager_v1_destroy(wl->idle_inhibit_manager);
 
+    if (wl->text_input_manager)
+        zwp_text_input_manager_v3_destroy(wl->text_input_manager);
+
     if (wl->presentation)
         wp_presentation_destroy(wl->presentation);
 

video/out/wayland_common.h

@@ -121,6 +121,9 @@ struct vo_wayland_state {
     struct zwp_idle_inhibit_manager_v1 *idle_inhibit_manager;
     struct zwp_idle_inhibitor_v1 *idle_inhibitor;
 
+    /* text-input */
+    struct zwp_text_input_manager_v3 *text_input_manager;
+
     /* linux-dmabuf */
     struct wl_list tranche_list;
     struct vo_wayland_tranche *current_tranche;