Virtual input #251
Virtual input #251
Conversation
seat.c
Outdated
| @@ -6,11 +6,14 @@ | |||
| * See the LICENSE file accompanying this file. | |||
| */ | |||
|
|
|||
| #define _XOPEN_SOURCE 500 | |||
There was a problem hiding this comment.
If this is for strdup, then #define _POSIX_C_SOURCE 200809L should be enough.
seat.c
Outdated
| */ | ||
| device->name = strdup(event->suggested_output->name); | ||
| /* event->suggested_seat should be checked if we handle multiple seats */ | ||
| handle_new_pointer(seat, wlr_pointer_from_input_device(device)); |
There was a problem hiding this comment.
No need to use wlr_pointer_from_input_device() here, we can just pass pointer.
seat.c
Outdated
| * its name so we do like other devices | ||
| */ | ||
| device->name = strdup(event->suggested_output->name); | ||
| /* event->suggested_seat should be checked if we handle multiple seats */ |
There was a problem hiding this comment.
Is that a TODO? If so, please mark it as such.
seat.c
Outdated
| * sub-optimal (we could just keep the suggested_output), but just copy | ||
| * its name so we do like other devices | ||
| */ | ||
| device->name = strdup(event->suggested_output->name); |
There was a problem hiding this comment.
Hm, I don't understand this line. device->name is the input device name, it's unrelated to the output name.
Did you mean pointer->output_name?
seat.c
Outdated
| { | ||
| if (virtual) |
seat.c
Outdated
| /* We apparently should not group virtual keyboards, | ||
| * so create a new group with it | ||
| */ | ||
| goto create_new; |
There was a problem hiding this comment.
Style: can we avoid this goto? e.g. by wrapping the code below in an if.
seat.c
Outdated
| /* If multiple seats are supported, check keyboard->seat | ||
| * to select the appropriate one */ | ||
|
|
||
| handle_new_keyboard(seat, wlr_keyboard_from_input_device(device), true); |
seat.c
Outdated
|
|
||
| seat->virtual_pointer = wlr_virtual_pointer_manager_v1_create(server->wl_display); | ||
| wl_signal_add(&seat->virtual_pointer->events.new_virtual_pointer, &seat->new_virtual_pointer); | ||
| seat->new_virtual_pointer.notify = handle_virtual_pointer; |
There was a problem hiding this comment.
Here we're creating one virtual-keyboard and virtual-pointer manager per seat. We shouldn't do this, instead we should only create one per server.
seat.c
Outdated
| * sub-optimal (we could just keep the suggested_output), but just copy | ||
| * its name so we do like other devices | ||
| */ | ||
| wlr_pointer->output_name = strdup(event->suggested_output->name); |
There was a problem hiding this comment.
event->suggested_output might be NULL, we should handle this case instead of crashing.
|
huh.. nice finding out about this when it is merged and the related issue gets closed :/ I'd have appreciated a heads up, would have been happy to rebase/fix things, it's not like I hadn't been responsive or anything... oh, well, thanks for taking this in anyway. Closing the other PR. |
|
Eeeh, sorry, I completely missed that other PR... It would've been nicer to at least record the proper attribution in the Git commit history @ChemicalXandco, next time please add yourself with Co-authored-by. |
|
I do not think it fixes #21 or at least an explanation how to enable such a virtual keyboard would be necessary cause I have no idea where I should start if I want to enable on-screen keyboards. It seems like #95 as well as some method to configure which virtual keyboard to invoke would be also required for #21 to be closed. |
|
The virtual keyboard protocol is used by on-screen keyboards, but is not enough: as you've figured out OSKs need layer-shell. This virtual keyboard patch can be used with CLI tools or remote desktop tools. |
Could you please reopen #21 then? |
|
The title may suggest so, though looking at the comments themselves it seems that the participants (including myself) are under the impression that "virtual keyboards" referred to OSKs. While technically the protocol of that name is now implemented, also Hjdskes seems to refer to OSKs as there would otherwise be no reason to link #95 for him. Could you maybe add "on-screen keyboard" to the title of #95 then such that people looking after this are not as confused as me or rename #21 and reopen it as the comments are about OSKs? |
|
Will there be a release with this feature? |
Successor to #189. Fixed conflicts and updated.
Fixes #21
Fixes #133
Tested with:
WLR_BACKENDS=headless WLR_NO_HARDWARE_CURSORS=1 WLR_LIBINPUT_NO_DEVICES=1 WLR_HEADLESS_OUTPUTS=1 ./cage