The execvp function usually doesn't return and exits by itself. But,
when having an error, it returns. A faulty dmenu setting could trigger
this and stop dunst receiving signals.
The execvp function usually doesn't return and exits by itself. But,
when having an error, it returns. As a faulty browser setting could
trigger this, dunst has to handle this, as the forked child wants to
access the X11 server too, resulting in XIO errors.
Fixes#476
When receiving a PropertyChange XEvent, it's mostly because the focus of
another window changed to another window. This isn't actually neccessary
for dunst. We need the event only, to know, when the screens have
switched.
As redrawing the screen became more expensive with the drop of GTK3, we
have to ignore this event and only really redraw the window, when the
focus has moved to another screen.
As per discussion in #334 drop dependency on gtk3 and instead depend
only on gdk-pixbuf2 for icon loading.
The only gtk3 function used was gdk_cairo_set_source_pixbuf in order to
convert the loaded pixbufs into cairo surfaces for us to render. In
order to drop the dependency this step was bypassed by using pixbufs
export to png capability and importing it into cairo via a pngstream.
When the DBus name lost function is called, either the whole DBus
connection is lost, or another rivaling FDN daemon acquired the name.
Without enough experience, this is impossible to trace back and examine
further.
To make it easier for new users, dunst prints out the PID of the
process, which currently acquired the FDN DBus name.
The hints given via DBus are not constants. Therefore the color fields
have to get freed during notification cleanup. As it's not possible to
disinguish, which field is constant, we have to duplicate the settings
on assignment.
While the notification spec allows tags like <a href="...">...</a> and
<img src="..." alt="...">, pango cannot parse these tags and therefore
these tags should be removed before passed to pango.
Also the method notification_extract_markup_urls is not needed anymore,
as markup_strip_a can return URLs optionally.
This implies, that URL replacement is now indicated via show_indicators
for URLs and the dmenu string is in the format of
'[text between a tags] URL\n'. This is similarly handled for images,
too.
The signal_notification_closed function is called even when a
notification has been pulled from history and re-emits the
NotificationClosed signal, this can confuse clients since multiple such
signals are not accounted for in the spec.
Distinguish between a failure to connect to DBus and a failure to
connect to the actual FDN DBus path.
This is possible, because `on_name_acquired` gets its DBus connection
passed as a parameter. If this parameter is NULL, there is no connection
established and the DBus name is not lost, because another notification
daemon is running.
When having a long running notification and an additional two shorter
notifications arrive. The short notification, which times out later,
does not timeout correctly and gets flushed only, when the long running
notification is flushed or a Focus* XEvent is sent.
The commit 7f335b7 introduced, that specific notifications did not time
out properly. Although these had been internally closed already, X11
still displayed them.
As GApplications expose their icon via image-path hint and not via the
actual app_icon field, GApplication icons haven't been displayed before.
According to the notification spec, image-path hint should also override
the app_icon as follows:
1. image-data hint
2. image-path hint
3. app_icon parameter
