[Issue] New commits caused a problem. The widgets have lost their styling.

[NextStep-IM]

• I have read the wiki
• Linux distro: Arco Linux

---

The issue

I just updated to the newest commits and every widget seems to have lost its styling. Changing wallpaper (which changes colorscheme) seems to solve the problem but the widgets lose their styling again if I restart ags.

Click to expand

  ```plain rm: cannot remove '/home/samtheradiant/.cache/ags/media/latex/*': No such file or directory (com.github.Aylur.ags:67052): Gjs-Console-WARNING **: 22:48:30.023: Window.popup is DEPRECATED. the click away functionality depends on a bug which was patched in Hyprland and it never worked on Sway anyway. to close on the esc key use self.keybind("Escape", () => App.closeWindow("window-name")) (com.github.Aylur.ags:67052): Gjs-Console-WARNING **: 22:48:30.112: Window.popup is DEPRECATED. the click away functionality depends on a bug which was patched in Hyprland and it never worked on Sway anyway. to close on the esc key use self.keybind("Escape", () => App.closeWindow("window-name")) (com.github.Aylur.ags:67052): Gjs-Console-WARNING **: 22:48:30.137: Window.popup is DEPRECATED. the click away functionality depends on a bug which was patched in Hyprland and it never worked on Sway anyway. to close on the esc key use self.keybind("Escape", () => App.closeWindow("window-name")) (com.github.Aylur.ags:67052): Gjs-Console-WARNING **: 22:48:30.167: Window.popup is DEPRECATED. the click away functionality depends on a bug which was patched in Hyprland and it never worked on Sway anyway. to close on the esc key use self.keybind("Escape", () => App.closeWindow("window-name")) (com.github.Aylur.ags:67052): Gjs-Console-WARNING **: 22:48:30.182: Window.popup is DEPRECATED. the click away functionality depends on a bug which was patched in Hyprland and it never worked on Sway anyway. to close on the esc key use self.keybind("Escape", () => App.closeWindow("window-name")) (com.github.Aylur.ags:67052): Gjs-Console-WARNING **: 22:48:30.193: Window.popup is DEPRECATED. the click away functionality depends on a bug which was patched in Hyprland and it never worked on Sway anyway. to close on the esc key use self.keybind("Escape", () => App.closeWindow("window-name")) GLib.SpawnError: Failed to execute child process “sass” (No such file or directory) (com.github.Aylur.ags:67052): Gtk-CRITICAL **: 22:48:30.208: gtk_widget_set_accel_path: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed (com.github.Aylur.ags:67052): Gtk-CRITICAL **: 22:48:30.208: gtk_widget_add_accelerator: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed (com.github.Aylur.ags:67052): Gtk-CRITICAL **: 22:48:30.208: gtk_widget_set_accel_path: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed (com.github.Aylur.ags:67052): Gtk-CRITICAL **: 22:48:30.208: gtk_widget_add_accelerator: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed (com.github.Aylur.ags:67052): Gtk-CRITICAL **: 22:48:30.208: gtk_widget_set_accel_path: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed (com.github.Aylur.ags:67052): Gtk-CRITICAL **: 22:48:30.208: gtk_widget_add_accelerator: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed (com.github.Aylur.ags:67052): GdkPixbuf-CRITICAL **: 22:48:30.556: gdk_pixbuf_new_from_file_at_scale: assertion 'width > 0 || width == -1' failed Traceback (most recent call last): File "/home/samtheradiant/.config/ags/scripts/color_generation/generate_colors_material.py", line 20, in img = Image.open(sys.argv[2]) ^^^^^^^^^^^^^^^^^^^^^^^ File "/usr/lib/python3.11/site-packages/PIL/Image.py", line 3247, in open fp = builtins.open(filename, "rb") ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ FileNotFoundError: [Errno 2] No such file or directory: 'undefined' Traceback (most recent call last): File "/home/samtheradiant/.config/ags/scripts/color_generation/generate_colors_material.py", line 20, in img = Image.open(sys.argv[2]) ^^^^^^^^^^^^^^^^^^^^^^^ File "/usr/lib/python3.11/site-packages/PIL/Image.py", line 3247, in open fp = builtins.open(filename, "rb") ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ FileNotFoundError: [Errno 2] No such file or directory: 'undefined' Error: Expected type filename for Array element but got type 'undefined' ```

[daUnknownCoder]

• I have read the wiki
• Linux distro: Arco Linux

The issue

I just updated to the newest commits and every widget seems to have lost its styling. Changing wallpaper (which changes colorscheme) seems to solve the problem but the widgets lose their styling again if I restart ags.

Click to expand

i got that for a portrait wallpaper, choose a landscape one

[NextStep-IM]

i got that for a portrait wallpaper, choose a landscape one

No luck. Changing the wallpaper doesn't even change the colorscheme.

[end-4]

yay -S dart-sass

[NextStep-IM]

Fixed. Thanks, @end-4!

[daUnknownCoder]

yay -S dart-sass

Hey @end-4, I have dart-sass but when I used a portrait wallpaper

This Is the same wallpaper I tried, my bar went nuts?

[end-4]

idk, give logs run ~/.config/ags/scripts/color_generation/switchwall.sh in terminal

[daUnknownCoder]

run ~/.config/ags/scripts/color_generation/switchwall.sh in terminal

Error: expected "{".
╷
2 │ $darkmode: true;
│                ^
╵
scss/_material.scss 2:16  @import
scss/main.scss 8:9        root stylesheet
undefined
Gtk.CssProviderError: style.css:9:5Invalid name of pseudo-class
[Gradience] INFO: Preset Material3_Generated applied successfully for Gtk 3 and Gtk 4 applications.
[Gradience] INFO: In order for changes to take full effect, you need to log out.

Only happens with this image, i checked with other images, they work this one doesn't, havent tried many images but why is that possibly happening?

[daUnknownCoder]

gradience or the colorgen script doesnt respect black colors idk why coz i hav a lot of amoled black wallpapers, which always give bluish color:

my eyes say the bg is black, white and red but there's no trace of em' except white, dropping wallpaper to see for urself:

i also see that the colors produced are not according to the background like they work for some and not for others like brown, white, blue are the most colors i see:

see the bg is sorta pinky but there's no trace of it

there's another one see:

[end-4]

can you send ~/.config/ags/scss/_material.scss

[daUnknownCoder]

for the hacked wallpaper:

$darkmode: true;
$primary: #acc7ff;
$onPrimary: #002e6c;
$primaryContainer: #004397;
$onPrimaryContainer: #d6e2ff;
$secondary: #bfc6dc;
$onSecondary: #283041;
$secondaryContainer: #3f4759;
$onSecondaryContainer: #dae2f9;
$tertiary: #debbdf;
$onTertiary: #402843;
$tertiaryContainer: #583e5b;
$onTertiaryContainer: #fbd7fb;
$error: #ffb4a9;
$onError: #680003;
$errorContainer: #930006;
$onErrorContainer: #ffb4a9;
$colorbarbg: #101012;
$background: #101012;
$onBackground: #e4e2e6;
$surface: #1b1b1e;
$onSurface: #e4e2e6;
$surfaceVariant: #44474f;
$onSurfaceVariant: #c4c6d0;
$outline: #8e9099;
$shadow: #000000;
$inverseSurface: #e4e2e6;
$inverseOnSurface: #2f3033;
$inversePrimary: #005ac5;

for the spacex image:

Warning: Image not in RGB|RGBA format - Converting...
$darkmode: true;
$primary: #acc7ff;
$onPrimary: #002e6c;
$primaryContainer: #004397;
$onPrimaryContainer: #d6e2ff;
$secondary: #bfc6dc;
$onSecondary: #283041;
$secondaryContainer: #3f4759;
$onSecondaryContainer: #dae2f9;
$tertiary: #debbdf;
$onTertiary: #402843;
$tertiaryContainer: #583e5b;
$onTertiaryContainer: #fbd7fb;
$error: #ffb4a9;
$onError: #680003;
$errorContainer: #930006;
$onErrorContainer: #ffb4a9;
$colorbarbg: #101012;
$background: #101012;
$onBackground: #e4e2e6;
$surface: #1b1b1e;
$onSurface: #e4e2e6;
$surfaceVariant: #44474f;
$onSurfaceVariant: #c4c6d0;
$outline: #8e9099;
$shadow: #000000;
$inverseSurface: #e4e2e6;
$inverseOnSurface: #2f3033;
$inversePrimary: #005ac5;

[clsty]

Ok, so the problem seems obvious: when generating that scss, a "Warning" message blended in and messed up the file.

For solution, we may redirect the STDERR i.e. the warning message to /dev/null, e.g.

    color_generation/generate_colors_material.py --color "$1" "$lightdark" 1> "$HOME"/.cache/ags/user/generated/material_colors.scss 2>/dev/null

Here 1> means redirecting STDOUT and 2> means redirecting STDERR. We output them to different places.

But the solution only works when the warning message is in STDERR.

If it does not work, then we need to find another method.

---

Plus, let's reopen this issue since talking in a closed issue is kinda weird lol

[NextStep-IM]

Experienced the problem again with the latest commits. Logs:

(com.github.Aylur.ags:14156): Gjs-Console-CRITICAL **: 05:00:44.694: CSS ERROR: Invalid name of pseudo-class at line 8 in /home/samtheradiant/.cache/ags/user/generated/style.css
Gtk.CssProviderError: style.css:9:5Invalid name of pseudo-class

(com.github.Aylur.ags:14156): Gtk-CRITICAL **: 05:00:44.968: gtk_widget_set_accel_path: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed

(com.github.Aylur.ags:14156): Gtk-CRITICAL **: 05:00:44.968: gtk_widget_add_accelerator: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed

(com.github.Aylur.ags:14156): Gtk-CRITICAL **: 05:00:44.968: gtk_widget_set_accel_path: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed

(com.github.Aylur.ags:14156): Gtk-CRITICAL **: 05:00:44.968: gtk_widget_add_accelerator: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed

(com.github.Aylur.ags:14156): Gtk-CRITICAL **: 05:00:44.969: gtk_widget_set_accel_path: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed

(com.github.Aylur.ags:14156): Gtk-CRITICAL **: 05:00:44.969: gtk_widget_add_accelerator: assertion 'GTK_IS_ACCEL_GROUP (accel_group)' failed
Traceback (most recent call last):
  File "/home/samtheradiant/.config/ags/scripts/color_generation/generate_colors_material.py", line 20, in <module>
    img = Image.open(sys.argv[2])
          ^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/site-packages/PIL/Image.py", line 3247, in open
    fp = builtins.open(filename, "rb")
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: 'undefined'

[NextStep-IM]

Aaaaand now its fixed.

[UcraNerd]

Aaaaand now its fixed.

If it's fixed don't you think that the issue should be closed?

[NextStep-IM]

If it's fixed don't you think that the issue should be closed?

It fixed itself and I have no idea how. So I left it to end_4 judgement.

[daUnknownCoder]

yo this still happens: #428, new tracker over there