latest patch

[Roman Kennke]

Hi Glen,

Sorry for the late reply. The change seems reasonable, I will commit it
in a bit.

Regarding the platform differences: it appears that the behavior of some
of the things that we do here is dependend on the platform. For example,
when I run your popup-trigger test on Linux, I get a popup trigger on
MOUSE_PRESSED (and yes, when I rightclick somewhere, popups appears on
pressing, not on releasing the mouse). You observed a similar difference
earlier regarding the mouse-clicking emulation. I see that as a good
thing: it gives us some freedoms to implement these things. I think in
the long run we have the following options:

- Implement whatever we think is correct. This will result in a
hodgepodge of little behavioural differences, i.e. some things will
behave like on Windows, some like on Linux, some like on the Mac
(because the patches contributed come from people on those platforms).
This is not necessarily a bad thing, applications should not depend on
those differences. This is what we have now.
- We decide for one platform and try to emulate behaviour of this. For
example, we could decide that Windows is our reference (ugh).
- We introduce a little abstraction layer and make the exact behaviour
configurable, i.e. decide by a system property which emulation to use.

The last option would clearly be the coolest, but require most work, the
first one we already have, the 2nd one will probably never happen (which
platform to choose?).

After the patch is in, can you give it a little more testing and send me
a notice when you think it's fine for a release?

Cheers,
Roman

> Hi there, I've found a regression which is causing me some grief with
> the popup boolean again.  I believe it should only be true for right
> click on mouse release (in Windows anyway).  I've attached a patch
> with test.
> 
> Thanks,
> Glen
> 
> On Wed, Apr 18, 2012 at 10:19 AM, Glen Schrader <gschrader at gmail.com> wrote:
> > Thanks!  Just out of curiosity I ran ClickEventTestFESTTest without
> > the cacio runner, the tests pass on Windows but the last 3 tests in
> > fail on Mac.  The Mac still considers it a mouse click when you move
> > the mouse a pixel, I'm kind of baffled on what's going on in the last
> > 2, no events are added to the list in those.  This won't impact me
> > though so I'm not going to worry about it.
> >
> > I have nothing else at the moment for the release.
> >
> > Cheers,
> > Glen
> >
> >
> > On Wed, Apr 18, 2012 at 9:16 AM, Roman Kennke <roman at kennke.org> wrote:
> >>> Thanks for finishing what I had!  I think you missed adding
> >>> MouseClickSupport in your commit though.
> >>
> >> Damn! Fixed now...
> >>
> >> Thanks,
> >> Roman
> >>
> >>
> >>>
> >>> On Wed, Apr 18, 2012 at 6:41 AM, Roman Kennke <roman at kennke.org> wrote:
> >>> > Hi Glen,
> >>> >
> >>> > Am Dienstag, den 17.04.2012, 10:17 -0600 schrieb Glen Schrader:
> >>> >> Here is a patch for the click event support along with test (hopefully
> >>> >> this patch works better!)
> >>> >
> >>> > Thanks for all your efforts!
> >>> >
> >>> > I took your testcase, added some tests and improved the implementation.
> >>> > It now supports the following:
> >>> >
> >>> > - Multi-click (double-click).
> >>> > - Click event in correct order (after release event).
> >>> > - Press and release also receive correct click count.
> >>> > - When press and release are on different location -> no click.
> >>> >
> >>> > I'll commit the attached patch shortly.
> >>> >
> >>> > Please let me know if you would like to see anything else before the
> >>> > release!
> >>> >
> >>> > Again: thanks for your efforts!
> >>> >
> >>> > Cheers,
> >>> > Roman
> >>> >
> >>> >
> >>>
> >>
> >>