>
> On Wed, Mar 23, 2016 at 01:06:26PM +0000, Frediano Ziglio wrote:
> > All other messages (even empty ones) have a structure defined.
>
> This is purely for consistency? Or was this causing issues?
>
I don't think this patch is going to change much bits of code.
Pure consistency.
> >
> > Signed-off-by: Frediano Ziglio <fziglio@xxxxxxxxxx>
> > ---
> > server/red-qxl.c | 3 ++-
> > server/red-qxl.h | 3 +++
> > server/red-worker.c | 2 +-
> > 3 files changed, 6 insertions(+), 2 deletions(-)
> >
> > diff --git a/server/red-qxl.c b/server/red-qxl.c
> > index 7ef9fe1..4c1c3e6 100644
> > --- a/server/red-qxl.c
> > +++ b/server/red-qxl.c
> > @@ -853,6 +853,7 @@ void spice_qxl_gl_scanout(QXLInstance *qxl,
> > uint32_t stride, uint32_t format,
> > int y_0_top)
> > {
> > + RedWorkerMessageGlScanout payload;
> > spice_return_if_fail(qxl != NULL);
> >
> > QXLState *qxl_state = qxl->st;
> > @@ -877,7 +878,7 @@ void spice_qxl_gl_scanout(QXLInstance *qxl,
> >
> > /* FIXME: find a way to coallesce all pending SCANOUTs */
> > dispatcher_send_message(&qxl_state->dispatcher,
> > - RED_WORKER_MESSAGE_GL_SCANOUT, NULL);
> > + RED_WORKER_MESSAGE_GL_SCANOUT, &payload);
>
> I'd set payload to 0 with
> RedWorkerMessageGlScanout payload = { 0, };
> or something like that if that's possible. Or add a comment that payload
I could use payload = {};
However there are lot of similar structures without any comment or empty
initialization.
I'd go for a
payload = { /* empty */ };
> is a dummy empty struct. Otherwise it looks like we are sending
> uninitialized data.
>
> Acked-by: Christophe Fergeau <cfergeau@xxxxxxxxxx>
>
> Christophe
>
Frediano
_______________________________________________
Spice-devel mailing list
Spice-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/spice-devel