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 --wayland-ime=<yes|no> (default: yes).
mpv-player · Jan 22, 2025 · 7b9ad9f · 7b9ad9f
1 parent c438732
commit 7b9ad9f
Show file tree

Hide file tree

Showing 6 changed files with 117 additions and 0 deletions.
diff --git a/DOCS/man/options.rst b/DOCS/man/options.rst
@@ -6116,6 +6116,12 @@ them.
     Defines the size of an edge border (default: 32) to initiate client side
     resizes events in the wayland contexts with touch events.
 
+``--wayland-ime=<yes|no>``
+    Enable keyboard input via an active input method (IME) (default: yes). The
+    input popup window, if there is any, is always positioned at the top left of the
+    window. There is no support for displaying pre-edit text, so you may want to set
+    your IME to display pre-edit inside of the popup window.
+
 ``--wayland-present=<yes|no>``
     Enable the use of wayland's presentation time protocol for more accurate
     frame presentation if it is supported by the compositor (default: ``yes``).

diff --git a/options/options.c b/options/options.c
@@ -209,6 +209,7 @@ static const m_option_t mp_vo_opt_list[] = {
         M_RANGE(0, INT_MAX)},
     {"wayland-edge-pixels-touch", OPT_INT(wl_edge_pixels_touch),
         M_RANGE(0, INT_MAX)},
+    {"wayland-ime", OPT_BOOL(wl_ime)},
     {"wayland-present", OPT_BOOL(wl_present)},
 #endif
 #if HAVE_WIN32_DESKTOP
@@ -272,6 +273,7 @@ const struct m_sub_options vo_sub_opts = {
         .wl_content_type = -1,
         .wl_edge_pixels_pointer = 16,
         .wl_edge_pixels_touch = 32,
+        .wl_ime = true,
         .wl_present = true,
         .mmcss_profile = "Playback",
         .ontop_level = -1,

diff --git a/options/options.h b/options/options.h
@@ -40,6 +40,7 @@ typedef struct mp_vo_opts {
     bool wl_disable_vsync;
     int wl_edge_pixels_pointer;
     int wl_edge_pixels_touch;
+    bool wl_ime;
     bool wl_present;
 
     float panscan;

diff --git a/video/out/meson.build b/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'],

diff --git a/video/out/wayland_common.c b/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 +232,12 @@ 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;
+};
+
 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 +885,80 @@ static const struct wl_data_device_listener data_device_listener = {
     data_device_handle_selection,
 };
 
+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;
+
+    if (!wl->opts->wl_ime)
+        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 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;
+
+    zwp_text_input_v3_disable(ti->text_input);
+    zwp_text_input_v3_commit(ti->text_input);
+    ti->serial++;
+}
+
+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 +1887,11 @@ 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 +2520,8 @@ 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);
@@ -3248,6 +3337,18 @@ 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 +3507,9 @@ 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);
 

diff --git a/video/out/wayland_common.h b/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;