Regression from bug 763358.
Steps from bug 762641:
1) Launch a webapp from the shell (ie, the dash or launcher in unity or the activities overlay in gnome shell)
2) Close the webapp and then launch it from a terminal window (ie, ~/.http\;www.lordofultima.com/webapprt-stub)
App launches with correct app icon.
App launches without an app icon.
Chris, was there a reason to choose gdk_set_program_class over g_set_prgname?
g_set_prgname would also set the name class hint, which is the "formal name of the application that should be used when retrieving the application's resources from the resource database", but it's hard to guess the modern effects of these ancient properties.
Often the name field of the class hint gets overridden by a windowtype attribute of the XUL window anyway, so perhaps it's more a question of whether the binary name or the app name is more suitable for GLib, if it matters.
There wasn't really any particular reason for choosing gdk_set_program_class over g_set_prgname, other than the former being a more direct way of achieving the same result. g_set_prgname should work too, but I want to have a quick look in glib at where this names is used, just to make sure it doesn't break anything else.
Created attachment 637826 [details] [diff] [review]
Looks like this fixes the problem.
However sometimes when you launch an application from the terminal, for a second you get two icons on the Launcher.
Yes, g_set_prgname should be fine here. The places it gets used in glib and gtk are either not used by us, or they don't care what the name is (or both)
Comment on attachment 637826 [details] [diff] [review]
(In reply to Marco Castelluccio from comment #4)
> Created attachment 637826 [details] [diff] [review]
> use g_set_prgname
> Looks like this fixes the problem.
> However sometimes when you launch an application from the terminal, for a
> second you get two icons on the Launcher.
Can you still reproduce this with this patch applied? If so, can you file a bug for this?
(In reply to Jason Smith [:jsmith] from comment #7)
> Can you still reproduce this with this patch applied? If so, can you file a
> bug for this?
Can't reproduce anymore. I'll file a new bug in case this will present again.