[PATCH v2] xf86Xinput: Added the xf86Post(Proximity|Button|Key)EventP helper functions.

[Oliver McFadden]

On Tue, 2009-07-28 at 03:49 +0200, ext Peter Hutterer wrote:
> Thanks for the update.
> 
> On Mon, Jul 27, 2009 at 12:06:14PM +0300, oliver.mcfadden at nokia.com wrote:
> > From: Oliver McFadden <oliver.mcfadden at nokia.com>
> > 
> > I only intend to use xf86PostButtonEventP in xf86-input-evdev, however,
> > for the sake of symmetry (and possible future use) I have added pointer
> > versions of all the VA args functions.
> 
> I'd rather you skip this paragraph in the commit message since it doesn't
> contribute majorly. When you're sending a git patch you can add comments
> like this after the --- deliminator (before the diffstat) and it will get
> skipped when the patch is applied.
> 
> This space is also handy when you're resending a patch as it allows you to
> comment what has changed since the previous version.

Oh, thanks. I was sure that git had that feature but had forgotten about
it.

> Also one minor thing: I prefer "Add foo" rather than "Added foo" since
> you're describing a patch that actively does something. Feel free to
> disagree with that, I don't mind leaving it like that either.

I'm just used to saying "Added feature X" rather than "Add feature X"; I
don't mind if it's changed.

> anyway, the real comment is below.
>  
> > xf86PostKeyboardEvent also makes use of xf86PostKeyEventP to avoid code
> > duplication, and the valuator verification has been split into a
> > separate static inline function, xf86VerifyValuators.
> > ---
> >  hw/xfree86/common/xf86Xinput.c |  157 ++++++++++++++++++++++++++--------------
> >  hw/xfree86/common/xf86Xinput.h |    8 ++
> >  2 files changed, 109 insertions(+), 56 deletions(-)
> > 
> > diff --git a/hw/xfree86/common/xf86Xinput.c b/hw/xfree86/common/xf86Xinput.c
> > index b4169cf..3b0bf58 100644
> > --- a/hw/xfree86/common/xf86Xinput.c
> > +++ b/hw/xfree86/common/xf86Xinput.c
> > @@ -701,6 +701,18 @@ DeleteInputDeviceRequest(DeviceIntPtr pDev)
> >   * convenient functions to post events
> >   */
> >  
> > +static inline Bool
> > +xf86VerifyValuators(int num_valuators)
> > +{
> > +    if (num_valuators > MAX_VALUATORS) {
> > +	xf86Msg(X_ERROR, "%s: num_valuator %d is greater than"
> > +		" MAX_VALUATORS\n", __FUNCTION__, num_valuators);
> 
> wouldn't __FUNCTION__ resolve to xf86VerifyValuators? If so, that seems a
> tad useless. Better to make it a macro so it resolves to the actual
> xf86PostXYZ function when triggered.

I believe that gcc will inline the function with an optimization level
greater than -O0, but yes, I'm happy to convert that into a macro. I can
resend it later today.

> Cheers,
>   Peter