Re: Auditing QEMU to replace NULL with &error_abort

[Daniel P . Berrangé]

On Wed, Jun 23, 2021 at 07:31:13PM +0400, Marc-André Lureau wrote:
> Hi
> 
> On Wed, Jun 23, 2021 at 7:10 PM Daniel P. Berrangé <berrange@redhat.com>
> wrote:
> 
> > On Wed, Jun 23, 2021 at 02:16:55PM +0200, Markus Armbruster wrote:
> > > &error_abort has been a clear win for us.  &error_fatal too, when used
> > > judiciously.  Marc-André tried to get both into GLib, unsuccessfully[2].
> >
> > ...snip...
> >
> > > [2] https://gitlab.gnome.org/GNOME/glib/-/issues/2288
> >
> > This doesn't actually suggest adding error_abort/fatal to GLib. Rather
> > it adds a general callback hook to GLib. Biggest complaints there
> > are around the callback concept and difficulty of safely using it,
> > which I can't disagree with.
> >
> > I wonder if we would have more luck if we explicitly proposed the
> > error_abort/fatal concept to GLib instead. At least that would not
> > hit any of the complaints raised about the callback.
> >
> >
> Without callbacks, it will be difficult to report errors back to the
> monitor, or prettify it the way we want (since we would be using extended
> GErrors for hints etc)
> 
> But we could have a more specific callback for that perhaps?
> 
> You are welcome to propose something else :)

I was thinking g_set_error would use g_warning/g_error to report
it, and thus involve the g_log_default_handler callback, which
we could provide to feed back into the monitor.

Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|