On 11/07/2011 01:39 PM, Peter Hutterer wrote:
> Does what it says on the box, compliments MakeInputMask.
>
> Signed-off-by: Peter Hutterer <[email protected]>
> ---
> Xi/exevents.c | 8 +++++++-
> 1 files changed, 7 insertions(+), 1 deletions(-)
>
> diff --git a/Xi/exevents.c b/Xi/exevents.c
> index 7f8e6db..6c19ab3 100644
> --- a/Xi/exevents.c
> +++ b/Xi/exevents.c
> @@ -1678,6 +1678,12 @@ MakeInputMasks(WindowPtr pWin)
> return TRUE;
> }
>
> +static void
> +FreeInputMask(OtherInputMasks *imask)
> +{
> + free(imask);
> +}
> +
FreeInputClient (from patch 05/12) takes a double pointer the object to
be freed, and sets the first pointer to NULL. If that's the paradigm we
want to use internally, shouldn't we do the same here?
> void
> RecalculateDeviceDeliverableEvents(WindowPtr pWin)
> {
> @@ -1738,7 +1744,7 @@ InputClientGone(WindowPtr pWin, XID id)
> } else if (!(other->next)) {
> if (ShouldFreeInputMasks(pWin, TRUE)) {
> wOtherInputMasks(pWin)->inputClients = other->next;
> - free(wOtherInputMasks(pWin));
> + FreeInputMask(wOtherInputMasks(pWin));
> pWin->optional->inputMasks = (OtherInputMasks *) NULL;
> CheckWindowOptionalNeed(pWin);
> FreeInputClient(&other);
_______________________________________________
[email protected]: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: http://lists.x.org/mailman/listinfo/xorg-devel
