gtk activity
Luna opened merge request !10597: css: fix data-image resource leak on textures at GNOME / gtk
Fixes #8516 (closed) Using the repro program from the issue before: luna@luna-pc:~/Documents/tmp/gtk$ ./_build/css_leak textures created=200 released=0 luna@luna-pc:~/Documents/tmp/gtk$ ./_build/css_leak direct textures created=200 released=200 after fix: [gtk] luna@luna-pc:~/Documents/tmp/gtk$ ./_build/css_leak textures created=200 released=200 [gtk] luna@luna-pc:~/Documents/tmp/gtk$ ./_build/css_leak direct textures created=200 released=200
Luna commented on issue #8516 at GNOME / gtk
Yup! I was just waiting to be able to fork since I just created my account and was getting the fork limit thing since I just added my SSH key, let me try now!
Matthias Clasen commented on issue #8516 at GNOME / gtk
I think you are right. Want to send a merge request?
Matthias Clasen commented on merge request !10595 at GNOME / gtk
Hmm, something goes wrong with the rendering now - making the stroke bounds larger or smaller should not really affect the position of the rendered path, it seems to do so.
Luna opened issue #8516: GTK Leaks image textures when they are loaded as data-urls at GNOME / gtk
Steps to reproduce Here's a small self contained repro: #include static const char png_base64[] = "iVBORw0KGgoAAAANSUhEUgAAABAAAAAQCAYAAAAf8/" "9hAAAAGUlEQVR4nGP4z8DwnxLMMGrAqAGjBgwXAwAwxP4QHCfkAAAAAABJRU5ErkJggg=="; static guint created, released; static void (*texture_constructed)(GObject *); static void texture_released(gpointer data, GObject *object) { ++released; } static void texture_constructed_override(GObject *object) { texture_constructed(object); ++created; g_object_weak_ref(object,…
Matthias Clasen commented on issue #8514 at GNOME / gtk
Yes, the default is butt, and that is what I see in your node
Matthias Clasen commented on issue #8514 at GNOME / gtk
Well, I was a little confused here, but nevertheless, adding overflow="hidden" to the might help.
João Sacramento commented on issue #8416 at GNOME / gtk
I ran into what looks like the same bug in Gajim (the "Edit Shortcut" dialog in Preferences flickered, gajim#12835) and narrowed it down a bit: The trigger seems to be the window having no border radius, not fullscreen/maximized as such. Gajim's macOS CSS had .csd { border-radius: 0; }, and its windows flickered whenever an AdwAlertDialog opened, also when not maximized. libadwaita sets --window-radius: 0px for .maximized and .fullscreen, which would explain why it only shows up there by…
Carlos Garnacho commented on issue #2636 at GNOME / gtk
This was fixed from the GTK perspective with !9958 (merged), for free to users of GtkSearchBar even. Anything else left, must be handled in the respective projects by using gtk_editable_set_input_interceptor() on the search entry, in order to intercept input automatically from a toplevel or parent container.
Benjamin Otte pushed to project branch wip/otte/rnt at GNOME / gtk
Benjamin Otte (d8a357ab) at 25 Oct 06:50 rendernode-tool: Improve error messages ... and 4 more commits
Matthias Clasen opened merge request !10566: rendernode tool: Allow simplify --all-nodes at GNOME / gtk
Add an option to make simplify operate on all the nodes in the stack, in addition to the directly passed node. The simplified results are put back on the stack, in the same order.
Benjamin Otte opened merge request !10565: More rendernode-tool refactoring at GNOME / gtk
This code mainly just moves code from gtk-rendernode-tool.c into rntcontext.c and gives it a proper API. It also vfunc-ifies the print functions so that we can capture the output. The goal is of course to be able to use the tool in other places (like the testsuite or the inspector) where we need some tighter control over things.
Benjamin Otte pushed new project branch wip/otte/rnt at GNOME / gtk
Benjamin Otte (c49ffc89) at 25 Oct 05:24 rendernode-tool: Improve error messages ... and 8 more commits
Marge Bot commented on merge request !10564 at GNOME / gtk
This branch couldn't be merged: CI failed! See pipeline https://gitlab.gnome.org/GNOME/gtk/-/pipelines/1133659.
Marge Bot accepted merge request !10563: gdk: Add gdk_color_apply_color_matrix at GNOME / gtk
Best to put this code in a central place. Also add a test.
Marge Bot pushed to project branch main at GNOME / gtk
Marge Bot (c706c00e) at 25 Oct 03:51 Merge branch 'for-main' into 'main' ... and 2 more commits
Matthias Clasen opened merge request !10564: Revert "ci: Allow the gnome-os job to fail" at GNOME / gtk
This reverts commit 87383719. The required pango change has landed in the gnome os image.
Matthias Clasen opened merge request !10563: gdk: Add gdk_color_apply_color_matrix at GNOME / gtk
Best to put this code in a central place. Also add a test.
yodel50 commented on issue #8499 at GNOME / gtk
I suspect it's maybe_rerender_style_sheet() in gtk/gtkcssprovider.c: it calls gtk_css_provider_reset(), which unregisters the theme's gtk.gresource, then re-parses the CSS, so the resource:// import fails. Keeping priv->resource and priv->path across the reset might fix it (untested): GResource *resource = g_steal_pointer (&priv->resource); char *path = g_steal_pointer (&priv->path); gtk_css_provider_reset (css_provider); /* ... */ priv->resource = resource; priv->path = path;
yodel50 opened issue #8499: Theme reload on color-scheme change fails to import CSS, windows become transparent at GNOME / gtk
Steps to reproduce Ubuntu 26.04, gtk-theme Yaru-sage (its gtk-4.0/gtk.css points to the theme's gtk.gresource) Run any plain GTK4 app gsettings set org.gnome.desktop.interface color-scheme prefer-dark (or default) Current behavior Color-scheme change logs: Theme parser error: gtk.css:1:1-68: Failed to import: The resource at “/com/ubuntu/themes/Yaru-sage/4.0/gtk.css” does not exist (or gtk-dark.css for apps preferring dark). The CSS doesn't get applied and windows render with no…
Marge Bot pushed to project branch main at GNOME / gtk
Marge Bot (1d273776) at 24 Oct 17:44 Merge branch 'arithmetic-clip' into 'main' ... and 4 more commits
Marge Bot accepted merge request !10562: gsk: Make arithmetic nodes keep clip separate at GNOME / gtk
Clarify that the rect passed to gsk_arithmetic_node_new is, in fact, a clip rect. Store it separately from the bounds, and add a getter.
Christian Hergert commented on merge request !10394 at GNOME / gtk
I think the objection has to do with expectations widgets are written with for state propagation. For example: hover the checkbox indicator move to the label indicator is no longer highlighted but the button remains hovered Probably from widgets relying on state-flags-changed a bit more than they should. Perhaps on one side, we should fix the widgets too, but that may end up having to be handled at the same time.
Matthias Clasen opened merge request !10562: gsk: Make arithmetic nodes keep clip separate at GNOME / gtk
Clarify that the rect passed to gsk_arithmetic_node_new is, in fact, a clip rect. Store it separately from the bounds, and add a getter.
Marge Bot accepted merge request !10561: gsk: Make component-transfer nodes keep clip separate at GNOME / gtk
Clarify that the rect passed to gsk_component_transfer_node_new2 is, in fact, a clip rect. Store it separately from the bounds, and add a getter.
Marge Bot pushed to project branch main at GNOME / gtk
Marge Bot (d09ebf0c) at 24 Oct 16:44 Merge branch 'for-main' into 'main' ... and 7 more commits
Matthias Clasen commented on merge request !10418 at GNOME / gtk
Please drop the excessive comment, add more details in the commit message if you feel they are necessary.
Marge Bot commented on merge request !10561 at GNOME / gtk
This branch couldn't be merged: CI failed! See pipeline https://gitlab.gnome.org/GNOME/gtk/-/pipelines/1133408.