[Xcb] [PATCH] xfixes: Update to version 5

[Bart Massey]

I'm fine with whatever folks decide; Lord knows I'm out of the loop at
this point. Just pointing out that every time we make XCB's names
different from the spec names, we get to write a bit of documentation
noting and explaining this. The main reason we wanted the names to
match was so that we could be lazy and not write our own separate
documentation. :-)  --Bart

On Sun, Feb 17, 2013 at 7:02 AM, Daniel Martin <consume.noise at gmail.com> wrote:
> On Thu, Feb 14, 2013 at 04:45:39PM -0800, Bart Massey wrote:
>> On Tue, Feb 5, 2013 at 4:30 PM, Neil Roberts <neil at linux.intel.com> wrote:
>> > Thanks for the feedback. Here is a second attempt with the enum as
>> > suggested.
>> >
>> > In the spec the name of the enum is BarrierDirections and it has
>> > values like ‘BarrierPositiveX’. This would end up with painfully long
>> > names like this:
>> >
>> > XCB_XFIXES_BARRIER_DIRECTION_MASK_BARRIER_POSITIVE_X
>> >
>> > Therefore I went with the same name that you suggested, i.e.
>> > ‘DirectionMask’ with values like ‘PositiveX’.
>> >
>> > Regards,
>> > - Neil
>> >
>> > -- >8 --
>> >
>> > This adds the two new functions in XFixes version 5 for handling
>> > pointer barriers.
>> >
>> > Signed-off-by: Neil Roberts <neil at linux.intel.com>
>> > ---
>> >  src/xfixes.xml | 30 +++++++++++++++++++++++++++++-
>> >  1 file changed, 29 insertions(+), 1 deletion(-)
>> >
>> > diff --git a/src/xfixes.xml b/src/xfixes.xml
>> > index 9bbeaab..014a37c 100644
>> > --- a/src/xfixes.xml
>> > +++ b/src/xfixes.xml
>> > @@ -26,7 +26,7 @@ authorization from the authors.
>> >  -->
>> >  <!-- This file describes version 4 of XFixes. -->
>> >  <xcb header="xfixes" extension-xname="XFIXES" extension-name="XFixes"
>> > -    major-version="4" minor-version="0">
>> > +    major-version="5" minor-version="0">
>> >    <import>xproto</import>
>> >    <import>render</import>
>> >    <import>shape</import>
>> > @@ -331,4 +331,32 @@ authorization from the authors.
>> >    <request name="ShowCursor" opcode="30">
>> >      <field type="WINDOW" name="window" />
>> >    </request>
>> > +
>> > +  <!-- Version 5 -->
>> > +
>> > +  <xidtype name="BARRIER" />
>> > +
>> > +  <enum name="DirectionMask">
>> > +    <item name="PositiveX"><bit>0</bit></item>
>> > +    <item name="PositiveY"><bit>1</bit></item>
>> > +    <item name="NegativeX"><bit>2</bit></item>
>> > +    <item name="NegativeY"><bit>3</bit></item>
>> > +  </enum>
>> > +
>> > +  <request name="CreatePointerBarrier" opcode="31">
>> > +    <field type="BARRIER" name="barrier" />
>> > +    <field type="DRAWABLE" name="drawable" />
>> > +    <field type="CARD16" name="x1" />
>> > +    <field type="CARD16" name="y1" />
>> > +    <field type="CARD16" name="x2" />
>> > +    <field type="CARD16" name="y2" />
>> > +    <field type="CARD32" name="directions" mask="DirectionMask" />
>> > +    <pad bytes="2" />
>> > +    <field type="CARD16" name="num_devices" />
>> > +    <list type="CARD16" name="devices"><fieldref>num_devices</fieldref></list>
>> > +  </request>
>> > +
>> > +  <request name="DeletePointerBarrier" opcode="32">
>> > +    <field type="BARRIER" name="barrier" />
>> > +  </request>
>> >  </xcb>
>> > --
>> > 1.7.11.3.g3c3efa5
>
>> I would love to have the names in XCB match the names in the
>> spec--this was a design goal of XCB. Those are painful looking names,
>> however. I wonder if there's some compromise? --Bart
>
> The enum could've been called 'BarrierDirections' as in the spec. But,
> removing the 'Barrier' prefix from the item names is imho a good
> decision.
>
> As this commit hasn't landed in a release yet, we could change the enum
> name. Would you be fine with that compromise?