Age | Commit message (Collapse) | Author |
|
|
|
|
|
This is important for password managers like keepass.
Unlike the previous version of this code, we no longer traverse widget
hierarchies. Instead, a pointer to the real tab label widget is stored
on whatever widget happens to be returned by gtk_notebook_get_tab_label().
Luckily, there is an easy way to do this via g_object.
|
|
|
|
Not ideal, because it doesn't work when the mouse hovers over empty
areas. If we connect that signal to the notebook itself, it will also
trigger when scrolling the web page -- if the web_view doesn't handle
the scroll. The event then propagates upwards in the hierarchy and ends
up in the notebook.
Means we would have to wrap the web_view in a dummy evbox that catches
all scroll events. Nah.
|
|
Unmaintained code, better alternatives exist.
|
|
This is just one file, there's no point in doing this.
|
|
|
|
|
|
|
|
|
|
|
|
GTK_ICON_SIZE_SMALL_TOOLBAR is 16px, according to the docs.
|
|
If the size already matches exactly, Gdk will do nothing.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This is a more meaningful path.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
By default, WebKit only responds to release events. The old code
listened for press events. That's probably not the right way to do
things. We should properly "override" the release events.
Closes #56.
|
|
I was probably thinking about multibyte encodings when I left that note,
but we explicitly check whether the first two bytes are ':' and '/', so
it's fine to skip them.
|
|
|
|
Yes, this needs the two void casts to make the compiler not complain
about unused parameters, but I think "explicit is better than implicit"
here. And none of the other callbacks use this trick.
|
|
Thanks to @jun7 for helping me out!
Closes #54.
|
|
|
|
Closes #47.
|
|
|
|
|
|
CC #36.
|
|
|
|
CC #35.
|
|
CC #32.
|
|
The main process and the process that sends the message might have
different working directories. As a result, the main process might not
be able to detect "foo.html" as a file when called as "lariza foo.html".
|
|
|
|
I was still seeing this warning whenever the GtkLevelBar was set to a
value of 0:
Negative content width -2 (allocation 0, extents 1x1) while
allocating gadget (node block, owner GtkLevelBar)
This happens even in a minimal test program:
#include <gtk/gtk.h>
int
main(int argc, char **argv)
{
GtkWidget *win, *progress;
gtk_init(&argc, &argv);
win = gtk_window_new(GTK_WINDOW_TOPLEVEL);
progress = gtk_level_bar_new();
gtk_level_bar_set_value(GTK_LEVEL_BAR(progress), 0);
gtk_container_add(GTK_CONTAINER(win), progress);
gtk_widget_show_all(win);
gtk_main();
}
It would appear that it's illegal to call gtk_level_bar_set_value() with
a value of 0. Or, and that's just as likely, I don't understand how a
GtkLevelBar is supposed to work. You don't even have to call
gtk_level_bar_set_value() at all, since 0 is the default value of such a
bar.
All of this doesn't really matter, though, since GtkEntry has a built-in
progress bar that we can use.
CC #20.
|
|
I think this is a good thing to do in any case. One crashed tab/window
should no longer be able to crash all other tabs as well.
CC #28: This change also appears to be a workaround for scenario number
two (`Alt+e` can crash WebKit/lariza).
|