~leon_plickat

Germany

Physics student, hobby programmer, Free Software advocate.

Conducting user interface experiments, trying to fill holes in the Wayland desktop landscape, creating opinionated tools and moreā€¦

~leon_plickat/lavalauncher

Last active 28 days ago

~leon_plickat/public-inbox

Last active a month ago
View more

Recent activity

Re: I'm tired of this anti-Wayland horseshit 22 days ago

From Leon Plickat to ~sircmpwn/public-inbox

On Tue Feb 2, 2021 at 5:31 PM CET, Adrien Glauser wrote:
> [...] which makes me think that at least these two Wayland sessions
> have "objective" issues, visible from many use cases.
> [...] Does that make me and others users frustrated at Plasma 5
> Wayland issues conspiracy theorists?

No one is denying that there may be some issues. What matters is
what people do when they find that their use case is apparently not
supported in Wayland.

If Wayland or rather some implementation does not support your use
case and you feel like it should, you could just contact the developers,
/respectfully/ making them aware of your problem. If you are willing to
listen if the answer isn't a direct solution but rather suggesting

Re: [PATCH v2] Patch with feedback taken into consideration 28 days ago

From Leon Plickat to ~leon_plickat/lavalauncher

Thanks for the patch, review below.

On Tue Jan 26, 2021 at 5:49 PM CET, Bryan Ndjeutcha wrote:
> Second patch for center position
>
> - Removed default orientations in favor of an orientation config option

I am still not too sure about this, though I am not totally opposed to an
"orientation" option.

However, the way hiding is currently implemented, the size of the hidden bar
depends on the orientation, making the hidden bar awkward to use if the
orentation of the bar does not match the orientation of the screen edge it is
docked to. This needs to be changed before I can accept this patch.

Re: [PATCH casa] Open icon grid with finger axis events a month ago

From Leon Plickat to ~sircmpwn/public-inbox

On Sun Jan 24, 2021 at 12:56 AM CET, Drew DeVault wrote:
> I actually just gave you push access to the casa repository; I'm not
> particularly interested in its maintenance so you can help yourself to
> applying any patches you see fit.

Thank for the trust.

I'll do it in multiple patches then.


Friendly greetings,
Leon Plickat

Re: [PATCH casa] Open icon grid with finger axis events a month ago

From Leon Plickat to ~sircmpwn/public-inbox

On Sun Jan 24, 2021 at 12:41 AM CET, Drew DeVault wrote:
> Why not support any kind of axis event? Though if we're not sure it's
> from a finger it may be better to interpret any amount of "up" delta as
> committing to completely opening the drawer, and vice versa.

I initially had this idea as well, but have not yet come up with a way
to integrate it with the current approach of using an additional
touchpoint and the casa_surface_touch_{up,down}() functions for the
axis stuff.

Although I guess for discrete input devices the progress can be
manipulated directly, bypassing all of that.

I am just not sure about continous devices which are not guaranteed

[PATCH casa] Open icon grid with finger axis events a month ago

From Leon Henrik Plickat to ~sircmpwn/public-inbox

Implemented by having a special touchpoint for the pointer.
This is the first step towards making casa usable on laptops.
---
Note that regular devices axis events, such as scroll wheels on mice,
are not handled by this, because I have not yet come up yet with a good
way to map discrete scroll steps to how casa operates.

WL_POINTER_AXIS_SOURCE_CONTINOUS is also not handled, because AFAIK
those devices are not guaranteed to emit axis_stop.

 include/casa.h    |   7 ++
 include/surface.h |   2 +-
 src/input.c       | 170 +++++++++++++++++++++++++++++++++++++++++++++-
 src/surface.c     |  12 +++-
[message trimmed]

Re: [PATCH 2/2] Updated wlr-layer-shell to v4 + corresponding keyboard interactivity change (fixed 2 includes I missed) a month ago

From Leon Plickat to ~leon_plickat/lavalauncher

Thanks for the patch!

Just dropping a mail letting you know that it'll be a few days
more until I can properly review it. I currently don't have
sway master installed and am locked onto an to old wlroots
version for that, since I need a specific version for river
development. I'll set up a dedicated sway environment, but
It'll be a few days until I have time for that.


Friendly greetings,
lhp

Re: [PATCH v2] fix to doc indentation a month ago

From Leon Plickat to ~leon_plickat/lavalauncher

Thanks for the patch, but it appears like you worked on a /considerably/
outdated version of the code. LavaLauncher has not had a "center"
position for ages. In the current man page source this indentation
error is apparently already corrected.


Friendly greetings,
Leon Plickat

Re: [PATCH] Fixed input issues when interaction type not defined for buttons a month ago

From Leon Plickat to ~leon_plickat/lavalauncher

Thanks for the patch! I applied it to master.
There were some small issues, but I fixed them for you.

On Fri Jan 22, 2021 at 8:34 PM CET, Robert Miller wrote:
> /* Try to find a universal command and overwrite it. If none has been
> * found, create a new one.
> */
> - struct Lava_item_command *cmd = find_item_command(button,
> + /* Set need_pointer and need_touch b/c no interaction is specified */
> + data->need_pointer = true;
> + data->need_touch = true;
> +
> + struct Lava_item_command *cmd = find_item_command(button,
> INTERACTION_UNIVERSAL, 0, 0, false);

Re: [ISSUE] possible bug with sway (latest git master) a month ago

From Leon Plickat to ~leon_plickat/lavalauncher

Thanks for sending a patch!

On Fri Jan 22, 2021 at 4:55 PM CET, Robert Raffaele Miller wrote:
> So I think I've fixed it - or at least found where the input bug
> happens. In seat_handle_capabilities in seat.c, data-
> >need_keyboard/mouse/touch would never return true on sway,

That's odd. These values are entirely based on the configuration file
parsing. Meaning they are false until a command is defined which uses
touch / pointer / keyboard. They are used to avoid LavaLauncher from
binding unneeded input devices and I'd like to keep that, because
it prevents the process waking up for and processing uneeded events.

> If I remove the "&& data-

Re: [ISSUE] possible bug with sway (latest git master) a month ago

From Leon Plickat to ~leon_plickat/lavalauncher

On Fri Jan 22, 2021 at 3:11 AM CET, Robert Raffaele Miller wrote:
> No worries! I was actually going take a crack at it myself, I've been
> wanting to work on something for wayland for a little while now. I can
> also confirm it works here with sway/wlroots master on the commit
> before
> https://git.sr.ht/~leon_plickat/lavalauncher/commit/e0590619a522e75012d11c606303f4e158121d12
> , so I've got somewhere to start with looking for the bug (the new
> version of the protocol seems to only change keyboard input, so that
> might only be part of the issue?).

Indeed its major change seems to be the updated keyboard focus mechanism.
So if you'd like to work on this, you will have to remove the keyboard
focus hack currently in place (see bar_instance_pointer_{enter, leave}())
and set the keyboard interactivity to "in demand" in the function that