Add box shadows #64

WillPower3309 · 2022-11-29T05:37:49Z

Adds box shadows in (probably?) the most efficient way we can following the method outlined in https://madebyevan.com/shaders/fast-rounded-rectangle-shadows/.

TODO:

Avoid rendering shadows behind surfaces (likely involves glStencil)
Add configuration options / commands for controlling shadows (color, blur_sigma)
Render shadows for all container layouts
Render shadows for floating containers

closes #6

WillPower3309 · 2022-11-29T05:39:46Z

Currently has an issue with the pos_attrib, will be fixed soon (TM). Right now only the non rounded corner shader is implemented. I think I may have a shader for the rounded case and for the non rounded case for maximum speed B)

Very much a draft, no config options added yet either.

WillPower3309 · 2022-12-07T05:17:58Z

Currently held back by

00:00:01.159 [wlr] [GLES2] GL_INVALID_VALUE in glVertexAttribPointerARB(idx)
00:00:01.159 [wlr] [GLES2] GL_INVALID_VALUE in glEnableVertexAttribArray(index)
00:00:01.159 [wlr] [GLES2] GL_INVALID_VALUE in glDisableVertexAttribArray(index)

I'm looking into this but any additional eyes would be amazing. Once this is done we just need #58 done (which is just an hours worth of work) and we can release 0.2!

ErikReider · 2022-12-07T21:54:51Z

Currently held back by
00:00:01.159 [wlr] [GLES2] GL_INVALID_VALUE in glVertexAttribPointerARB(idx)
00:00:01.159 [wlr] [GLES2] GL_INVALID_VALUE in glEnableVertexAttribArray(index)
00:00:01.159 [wlr] [GLES2] GL_INVALID_VALUE in glDisableVertexAttribArray(index)
I'm looking into this but any additional eyes would be amazing. Once this is done we just need #58 done (which is just an hours worth of work) and we can release 0.2!

Found it. My eyes need a vacation after trying to spot this...

 include @sysconfdir@/sway/config.d/*
diff --git a/sway/desktop/fx_renderer.c b/sway/desktop/fx_renderer.c
index f12a705f..1dea93cc 100644
--- a/sway/desktop/fx_renderer.c
+++ b/sway/desktop/fx_renderer.c
@@ -238,7 +238,7 @@ struct fx_renderer *fx_renderer_create(struct wlr_egl *egl) {
 	}
 	renderer->shaders.box_shadow.proj = glGetUniformLocation(prog, "proj");
 	renderer->shaders.box_shadow.color = glGetUniformLocation(prog, "color");
-	renderer->shaders.box_shadow.pos_attrib = glGetUniformLocation(prog, "pos");
+	renderer->shaders.box_shadow.pos_attrib = glGetAttribLocation(prog, "pos");
 	renderer->shaders.box_shadow.position = glGetUniformLocation(prog, "position");
 	renderer->shaders.box_shadow.size = glGetUniformLocation(prog, "size");
 	renderer->shaders.box_shadow.blur_sigma = glGetUniformLocation(prog, "blur_sigma");

The shadow doesn't seem to work though :/

WillPower3309 · 2022-12-08T03:10:54Z

Gah I'm so mad that the fix was that 💀 thanks man! I'll try to fix up the shader

WillPower3309 · 2022-12-11T20:37:10Z

Alright, can confirm that the issue is in the fx_render_box_shadow function, likely in the shader. Heres the result of just calling fx_render_rect, which renders a rectangle over the window region as expected:

WillPower3309 · 2022-12-11T20:37:45Z

LOL, can confirm shader seems off: https://www.shadertoy.com/view/DdlSD8

ErikReider · 2022-12-11T20:51:00Z

Tried messing with the shader a few nights ago but wasn't able to fix it... I was only able to get two edges working properly :/

WillPower3309 · 2022-12-11T21:18:00Z

Alright I fixed it, no one is allowed to look at the commit change or they'll lose all faith in me as a maintainer 🥲

ErikReider · 2022-12-11T21:57:23Z

Looks pretty good!

This is after messing around with some margins and moving the render logic order:

WillPower3309 · 2022-12-11T22:09:19Z

Awesome! Still need to do that myself but good to see it's working!

WillPower3309 · 2022-12-11T22:43:00Z

Looks pretty good!

This is after messing around with some margins and moving the render logic order:

I'll be having the damage tracking avoid the window region so the shadow isnt rendered where it won't be visible, so that should avoid the issue of needing to render it before the window so it appears behind (while giving us a lot of performance savings). I don't think many users want to see the shadow behind transparent regions either, as it appears unexpectedly dark with the shadow rendered as well, but I'll keep an eye on requests there

ErikReider · 2022-12-11T22:47:22Z

I don't think many users want to see the shadow behind transparent regions either, as it appears unexpectedly dark with the shadow rendered as well, but I'll keep an eye on requests there.

I was thinking the same thing! A CSS box-shadow would do the same.

I can throw my diff here in case you find anything interesting :)

Note that this doesn't include any "spread" control :/

diff --git a/sway/desktop/render.c b/sway/desktop/render.c
index 0e1bb647..91c6e4a8 100644
--- a/sway/desktop/render.c
+++ b/sway/desktop/render.c
@@ -331,7 +331,8 @@ damage_finish:
 // _box.x and .y are expected to be layout-local
 // _box.width and .height are expected to be output-buffer-local
 void render_box_shadow(struct sway_output *output, pixman_region32_t *output_damage,
-		const struct wlr_box *_box, const float color[static 4]) {
+		const struct wlr_box *_box, const float color[static 4],
+		float blur_sigma) {
 	struct wlr_output *wlr_output = output->wlr_output;
 	struct fx_renderer *renderer = output->server->renderer;
 
@@ -354,7 +355,8 @@ void render_box_shadow(struct sway_output *output, pixman_region32_t *output_dam
 	pixman_box32_t *rects = pixman_region32_rectangles(&damage, &nrects);
 	for (int i = 0; i < nrects; ++i) {
 		scissor_output(wlr_output, &rects[i]);
-		fx_render_box_shadow(renderer, &box, color, wlr_output->transform_matrix, 0, 0.3);
+		fx_render_box_shadow(renderer, &box, color, wlr_output->transform_matrix,
+				0, blur_sigma);
 	}
 
 damage_finish:
@@ -977,6 +979,19 @@ static void render_containers_linear(struct sway_output *output,
 				marks_texture = child->marks_unfocused;
 			}
 
+			// render shadow
+			const float color[4] = {0, 0, 0, 0.5};
+			// Mimic how a CSS box-shadow looks
+			float blur_sigma = 20;
+			struct wlr_box box = {
+				state->x - blur_sigma,
+				state->y - blur_sigma,
+				state->width + blur_sigma * 2,
+				state->height + blur_sigma * 2
+			};
+			scale_box(&box, output->wlr_output->scale);
+			render_box_shadow(output, damage, &box, color, blur_sigma);
+
 			bool has_titlebar = state->border == B_NORMAL;
 			struct decoration_data deco_data = {
 				.alpha = child->alpha,
@@ -998,12 +1013,6 @@ static void render_containers_linear(struct sway_output *output,
 			} else if (state->border == B_PIXEL) {
 				render_top_border(output, damage, state, colors, deco_data.alpha, deco_data.corner_radius);
 			}
-
-			// render shadow
-			const float color[4] = {0,0,0,0.7};
-			struct wlr_box box = {state->x, state->y, state->width, state->height};
-			scale_box(&box, output->wlr_output->scale);
-			render_box_shadow(output, damage, &box, color);
 		} else {
 			render_container(output, damage, child,
 					parent->focused || child->current.focused);
diff --git a/sway/desktop/shaders/box_shadow.frag b/sway/desktop/shaders/box_shadow.frag
index 5dd50a25..7b5178d7 100644
--- a/sway/desktop/shaders/box_shadow.frag
+++ b/sway/desktop/shaders/box_shadow.frag
@@ -29,7 +29,10 @@ float random() {
 }
 
 void main() {
-    float frag_alpha = v_color.a * box_shadow(position, position + size, gl_FragCoord.xy, blur_sigma);
+    float frag_alpha = v_color.a * box_shadow(
+            position + blur_sigma,
+            position + size - blur_sigma,
+            gl_FragCoord.xy, blur_sigma * 0.5);
 
     // dither the alpha to break up color bands
     frag_alpha += (random() - 0.5) / 128.0;

WillPower3309 · 2022-12-12T01:30:30Z

Thanks for sharing! Are the adjustments to the box arbitrary? Curious about why you set the sigma to be 0.5 it's original value in the shader

ErikReider · 2022-12-13T08:56:35Z

Are the adjustments to the box arbitrary?

Otherwise the shadow would be clipped to the window size :/

Curious about why you set the sigma to be 0.5 it's original value in the shader

Yeah me too... I was just testing different things and it seemed to work the best then for some reason????

ErikReider · 2022-12-17T23:19:44Z

Got the rounded shadows working. Should I push my changes?

ErikReider · 2022-12-17T23:40:36Z

Setting glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA); fixes alpha with non-black shadows

ErikReider · 2022-12-18T00:18:39Z

I've also been experimenting with the damage that you mentioned earlier. I got a decent result but without the rounding of the corner... :/

WillPower3309 · 2022-12-18T08:17:14Z

Nice work! Feel free to push the shadow changes if they implement the rounded shadow outlined in the blog post attached to the PR description

WillPower3309 · 2022-12-18T08:18:03Z

Otherwise I'd love to see the damage tracking changes, and I can implement the rounded corners. Apologies for leaving this up without much progress this week, aiming to finish this soon so i can work on blur over the holiday downtime

ErikReider · 2022-12-18T10:29:53Z

Here's the small damage refactor:

diff --git a/sway/desktop/render.c b/sway/desktop/render.c
index 27567b58..55eb2076 100644
--- a/sway/desktop/render.c
+++ b/sway/desktop/render.c
@@ -345,6 +345,17 @@ void render_box_shadow(struct sway_output *output, pixman_region32_t *output_dam
 	pixman_region32_init(&damage);
 	pixman_region32_union_rect(&damage, &damage, box.x, box.y,
 		box.width, box.height);
+	
+	pixman_region32_t inner_damage;
+	pixman_region32_init(&inner_damage);
+	pixman_region32_union_rect(&inner_damage, &inner_damage,
+			box.x + blur_sigma,
+			box.y + blur_sigma,
+			box.width - blur_sigma * 2.0,
+			box.height - blur_sigma * 2.0);
+	pixman_region32_subtract(&damage, &damage, &inner_damage);
+	pixman_region32_fini(&inner_damage);
+
 	pixman_region32_intersect(&damage, &damage, output_damage);
 	bool damaged = pixman_region32_not_empty(&damage);
 	if (!damaged) {

There are still some damage issues when moving a floating window with shadows though. Not really sure why... :/

fuegoio · 2022-12-18T22:33:20Z

I've also been experimenting with the damage that you mentioned earlier. I got a decent result but without the rounding of the corner... :/

I pulled this code to test locally and try to improve the damage tracking and ran into this issue as well.
Especially, without your last comment, the shadow was rendering on top of windows. It can be fixed by inverting the order:

diff --git a/sway/desktop/render.c b/sway/desktop/render.c
index 27567b58..a9d15018 100644
--- a/sway/desktop/render.c
+++ b/sway/desktop/render.c
@@ -992,14 +992,6 @@ static void render_containers_linear(struct sway_output *output,
                                .saturation = child->saturation,
                                .has_titlebar = has_titlebar,
                        };
-                       render_view(output, damage, child, colors, deco_data);
-                       if (has_titlebar) {
-                               render_titlebar(output, damage, child, floor(state->x), floor(state->y),
-                                               state->width, colors, deco_data.alpha, deco_data.corner_radius,
-                                               title_texture, marks_texture);
-                       } else if (state->border == B_PIXEL) {
-                               render_top_border(output, damage, state, colors, deco_data.alpha, deco_data.corner_radius);
-                       }

                        // render shadow
                        const float color[4] = {0.0, 0.0, 0.0, 0.5};
@@ -1013,6 +1005,18 @@ static void render_containers_linear(struct sway_output *output,
                        };
                        scale_box(&box, output->wlr_output->scale);
                        render_box_shadow(output, damage, &box, color, blur_sigma, deco_data.corner_radius);
+
+                       render_view(output, damage, child, colors, deco_data);
+
+                       if (has_titlebar) {
+                               render_titlebar(output, damage, child, floor(state->x), floor(state->y),
+                                               state->width, colors, deco_data.alpha, deco_data.corner_radius,
+                                               title_texture, marks_texture);
+                       } else if (state->border == B_PIXEL) {
+                               render_top_border(output, damage, state, colors, deco_data.alpha, deco_data.corner_radius);
+                       }
                } else {
                        render_container(output, damage, child,
                                        parent->focused || child->current.focused);

I also fixed the squared corner that you have that appears because of the damaging that doesn't take into account the corner radius:

diff --git a/sway/desktop/render.c b/sway/desktop/render.c
index 27567b58..624eb2d2 100644
--- a/sway/desktop/render.c
+++ b/sway/desktop/render.c
@@ -345,6 +345,17 @@ void render_box_shadow(struct sway_output *output, pixman_region32_t *output_dam
        pixman_region32_init(&damage);
        pixman_region32_union_rect(&damage, &damage, box.x, box.y,
                box.width, box.height);
+
+       pixman_region32_t inner_damage;
+       pixman_region32_init(&inner_damage);
+       pixman_region32_union_rect(&inner_damage, &inner_damage,
+                       box.x + blur_sigma + corner_radius,
+                       box.y + blur_sigma + corner_radius,
+                       box.width - (blur_sigma + corner_radius) * 2.0,
+                       box.height - (blur_sigma + corner_radius) * 2.0);
+       pixman_region32_subtract(&damage, &damage, &inner_damage);
+       pixman_region32_fini(&inner_damage);
+
        pixman_region32_intersect(&damage, &damage, output_damage);
        bool damaged = pixman_region32_not_empty(&damage);
        if (!damaged) {

I don't know if it is the best solution (I think the damaging is less precise :/ ?), but it works!

fuegoio · 2022-12-18T23:16:24Z

I also have an issue with the glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA); flag, which is on my machine creating a glitch with rounded corners, that I don't have without this flag:

But I don't know why ahah, I tried playing with the rounded corner shader but with no luck.

ErikReider · 2022-12-19T11:55:53Z

the shadow was rendering on top of windows. It can be fixed by inverting the order:

With a "fixed" damage tracker this wouldn't be necessary. It would probably also make it easier to exclude the shadow from any future blurring done to the window :)

I also fixed the squared corner that you have that appears because of the damaging that doesn't take into account the corner radius:

This was also one of my solutions until I looked at a transparent window...

I also have an issue with the glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA); flag, which is on my machine creating a glitch with rounded corners, that I don't have without this flag:

Strange... This shouldn't be an issue when a proper damage tracker is added. Without the blending, any other shadow color than black would look like this:

WillPower3309 · 2022-12-19T20:04:46Z

I suspect that in order to account for the rounded corner case, we will need to damage track similar to how @fuegoio demonstrated, however avoid rendering in the area the window encompasses with a gl stencil

A stencil of the surface will need to be done for blur and work in improving rounded borders in the future too

WillPower3309 · 2022-12-24T23:28:08Z

Are the adjustments to the box arbitrary?

Otherwise the shadow would be clipped to the window size :/

Curious about why you set the sigma to be 0.5 it's original value in the shader

Yeah me too... I was just testing different things and it seemed to work the best then for some reason????

Interesting find that I had, it seems as though the radius equates to roughly 2 * the blur sigma, so this makes sense

WillPower3309 · 2022-12-25T00:03:20Z

Instead of cutting the sigma in half, and adjusting the box with the original sigma, let's keep the default sigma and adjust the box with 2 * the sigma. This would better match what should be expected of the sigma

WillPower3309 · 2022-12-25T00:08:49Z

After this, we just need to implement the gl stencil (which I'm looking into), add the config options / commands (which would be a great first task for anyone looking to be a new contributor), and then apply the box shadows on all container layouts (right now it's just for linear laid out containers)

WillPower3309 · 2022-12-25T18:49:27Z

I also have an issue with the glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA); flag, which is on my machine creating a glitch with rounded corners, that I don't have without this flag:

But I don't know why ahah, I tried playing with the rounded corner shader but with no luck.

in regards to this, I suspect we could set the blend function back to glBlendFunc(GL_ONE, GL_ONE_MINUS_SRC_ALPHA); after shadows are done rendering to fix, but I can't reproduce your issue yet

ErikReider · 2023-01-09T08:40:24Z

Did your ignore windows with CSD commit help at all? I can reproduce the issue, and if the CSD commit aimed to fix it we would revert that if it didn't work

It fixed another issue where shadows were being drawn twice (one by gtk and one by swayfx). An example would be the GTK file picker in Firefox:

ErikReider · 2023-01-09T12:15:03Z

Have you tried removing the view color buffer rendering? Makes it a lot more noticeable. It only seems to show when dragging a window over a location where the stencil mask resides (inside of the view).

WillPower3309 · 2023-01-09T12:28:57Z

Have you tried removing the view color buffer rendering? Makes it a lot more noticeable. It only seems to show when dragging a window over a location where the stencil mask resides (inside of the view).

Not yet, the other thing I've observed is that this seems to specifically impact popup windows with CSD (where the stencil is a bit off and too large since it thinks the decorations are part of the window). Not sure if there are cases outside of CSD decorated floating popup windows though

head-gardener · 2023-01-09T13:39:35Z

Not sure if this is directly related, but telegram-desktop has similar issue with shadows being bigger than the window itself whenever applications draws its own window frame with rounded corners.

ErikReider · 2023-01-09T21:17:33Z

@WillPower3309 Reverting back to drawing a rounded box as stencil mask fixes the floating bug and some CSD issues like the @head-gardener's comment above but not the GTK file dialog.

Seems like Hyprland doesn't have the file picker CSD issue, so I'll have to investigate further tomorrow.

I'll push the revert :)

WillPower3309 · 2023-01-09T21:25:15Z

@WillPower3309 Reverting back to drawing a rounded box as stencil mask fixes the floating bug and some CSD issues like the @head-gardener's comment above but not the GTK file dialog.

Seems like Hyprland doesn't have the file picker CSD issue, so I'll have to investigate further tomorrow.

I'll push the revert :)

Awesome, good stuff. I'll focus on fixing the titlebar sizing in the meantime (#90)

shaunsingh · 2023-01-14T02:45:04Z

For me only part of the screen gets shadows, sorry if its hard to notice on the background

The top renders just fine. You can see a diagonal where the shadow stops rendering

ErikReider · 2023-01-14T12:02:14Z

For me only part of the screen gets shadows, sorry if its hard to notice on the background

The top renders just fine. You can see a diagonal where the shadow stops rendering

Does it disappear if you open another terminal? Any way of reproducing this?

shaunsingh · 2023-01-14T19:55:58Z

Happens with all applications all the time. I am running asahi linux with the experimental gpu driver, but that has full opengl/es 2 support so it should be fine

let me know if there's any kind of logs you'd like me to share

WillPower3309 · 2023-01-15T19:38:10Z

Happens with all applications all the time. I am running asahi linux with the experimental gpu driver, but that has full opengl/es 2 support so it should be fine

let me know if there's any kind of logs you'd like me to share

Good to know! Thanks for letting us know, I am tempted to believe it's an asahi problem. I'm going to try to get this merged today pending some more testing on my end. If it still exhibits the problem when merged please raise an issue with some logs. This will allow us to get some more visibility on if other people experience the issue prior to making a release with this feature.

shaunsingh · 2023-01-16T01:28:56Z

Will do! It can very well just be an issue with the asahi driver. I'll find another machine to test my config on just in case

ErikReider · 2023-01-16T21:58:55Z

What we could do regarding the CSD concern is to hide it behind a config option. Something like shadows_on_csd?

ErikReider

LGTM! We can always add the csd option later :)

WillPower3309 · 2023-01-17T02:11:08Z

LGTM! We can always add the csd option later :)

My thoughts exactly. May do some minor refactoring then I'll merge

WillPower3309 · 2023-01-17T07:10:03Z

Moved the shadow rendering back to render_view, as it already checks for CSD borders

WillPower3309 · 2023-01-17T07:12:20Z

@ErikReider looks good on my end but I'll let you try it out again just in case anything changed in my small refactor. I may rename the config options prior to a merge tomorrow.

Interestingly the bad stencilling from #82 only appears on tabbed / stacked layouts for me now, not on regularly tiled windows with a titlebar. Strange, as logically that wouldn't make a difference.

ErikReider · 2023-01-17T08:18:18Z

Strange... I'll take a look at it later today :)

ErikReider · 2023-01-17T09:44:59Z

I'll let you merge this in case you want to change the config names

WillPower3309 · 2023-01-18T06:49:08Z

There are a few changes I may make prior to a release, namely different config names and an adjustment to damage tracking. However, I think that this can be merged as is so that it can be more widely tested. Good work everyone!

WillPower3309 marked this pull request as draft November 29, 2022 05:37

WillPower3309 force-pushed the shadows branch from 1b2ea6d to d8b10f6 Compare November 29, 2022 05:40

Back to rendering rounded box as stencil mask

2a54861

fix: tabs

545511f

WillPower3309 requested a review from ErikReider January 17, 2023 00:58

ErikReider approved these changes Jan 17, 2023

View reviewed changes

WillPower3309 added 3 commits January 17, 2023 01:54

moved shadow rendering to render_view

d71e896

cleanup old stencil impl

42f9949

removed extraneous shader changes

832a2f0

WillPower3309 force-pushed the shadows branch from 7075761 to 832a2f0 Compare January 17, 2023 07:08

WillPower3309 requested a review from ErikReider January 17, 2023 07:12

ErikReider approved these changes Jan 17, 2023

View reviewed changes

WillPower3309 merged commit 588ea8e into master Jan 18, 2023

WillPower3309 mentioned this pull request Aug 2, 2023

Add box shadows wlrfx/scenefx#16

Merged

Add box shadows #64

Add box shadows #64

Conversation

WillPower3309 commented Nov 29, 2022 • edited Loading

WillPower3309 commented Nov 29, 2022 • edited Loading

WillPower3309 commented Dec 7, 2022

ErikReider commented Dec 7, 2022

WillPower3309 commented Dec 8, 2022

WillPower3309 commented Dec 11, 2022

WillPower3309 commented Dec 11, 2022

ErikReider commented Dec 11, 2022

WillPower3309 commented Dec 11, 2022

ErikReider commented Dec 11, 2022

WillPower3309 commented Dec 11, 2022

WillPower3309 commented Dec 11, 2022 • edited Loading

ErikReider commented Dec 11, 2022 • edited Loading

WillPower3309 commented Dec 12, 2022

ErikReider commented Dec 13, 2022

ErikReider commented Dec 17, 2022

ErikReider commented Dec 17, 2022

ErikReider commented Dec 18, 2022

WillPower3309 commented Dec 18, 2022

WillPower3309 commented Dec 18, 2022

ErikReider commented Dec 18, 2022

fuegoio commented Dec 18, 2022 • edited Loading

fuegoio commented Dec 18, 2022

ErikReider commented Dec 19, 2022 • edited Loading

WillPower3309 commented Dec 19, 2022

WillPower3309 commented Dec 24, 2022

WillPower3309 commented Dec 25, 2022

WillPower3309 commented Dec 25, 2022

WillPower3309 commented Dec 25, 2022

ErikReider commented Jan 9, 2023 • edited Loading

ErikReider commented Jan 9, 2023

WillPower3309 commented Jan 9, 2023 • edited Loading

head-gardener commented Jan 9, 2023 • edited Loading

ErikReider commented Jan 9, 2023

WillPower3309 commented Jan 9, 2023

shaunsingh commented Jan 14, 2023

ErikReider commented Jan 14, 2023 • edited Loading

shaunsingh commented Jan 14, 2023

WillPower3309 commented Jan 15, 2023

shaunsingh commented Jan 16, 2023

ErikReider commented Jan 16, 2023

ErikReider left a comment

Choose a reason for hiding this comment

WillPower3309 commented Jan 17, 2023

WillPower3309 commented Jan 17, 2023

WillPower3309 commented Jan 17, 2023

ErikReider commented Jan 17, 2023

ErikReider commented Jan 17, 2023

WillPower3309 commented Jan 18, 2023

WillPower3309 commented Nov 29, 2022 •

edited

Loading

WillPower3309 commented Nov 29, 2022 •

edited

Loading

WillPower3309 commented Dec 11, 2022 •

edited

Loading

ErikReider commented Dec 11, 2022 •

edited

Loading

fuegoio commented Dec 18, 2022 •

edited

Loading

ErikReider commented Dec 19, 2022 •

edited

Loading

ErikReider commented Jan 9, 2023 •

edited

Loading

WillPower3309 commented Jan 9, 2023 •

edited

Loading

head-gardener commented Jan 9, 2023 •

edited

Loading

ErikReider commented Jan 14, 2023 •

edited

Loading