harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Oleg Khaschansky" <oleg.v.khaschan...@gmail.com>
Subject Re: [classlib][awt] clipping area
Date Wed, 13 Dec 2006 17:37:21 GMT
> Because I'm not sure that MultuRectArea is fully implemented shape. We
> have special handling for it everywhere.

It is implemented. But you are absolutely right, MultiRectArea is a
special case! I created a Shape with the same implementation as
MultiRectArea has but with the different class name (to avoid entering
the MRA-specific code on instanceof checks) and used it instead of
MultiRectArea in this test. And Harmony demonstrated the same behavior
as RI.

So, the reason is that MultiRectArea class is handled in a special way
in Harmony. It is not a bug then.

On 12/13/06, Alexey Petrenko <alexey.a.petrenko@gmail.com> wrote:
> 2006/12/13, Oleg Khaschansky <oleg.v.khaschansky@gmail.com>:
> > Hmm, I don't think that things are so simple. Everyone can implement
> > their own Shape and MultiRectArea is simply a Shape implementation. So
> > the testcase with it is pretty legal, why not?
> Because I'm not sure that MultuRectArea is fully implemented shape. We
> have special handling for it everywhere.
>
> SY, Alexey
>
> > But we know that RI demonstrates a different behavior if we pass the
> > same clip rect as a Rectangle and as a MultiRectArea. Taking this into
> > account I'd say that RI behavior is inconsistent.
> >
> > On the other hand, we don't have pixel-by-pixel compatibility with RI
> > in java2d, e.g. I think, we use different shape rasterization
> > algorythms.
> >
> > I'd suggest to close this bug as a non-bug difference.
> >
> > On 12/13/06, Shipilov, Alexander D <alexander.d.shipilov@intel.com> wrote:
> > > >Why do you need Area here? You can use Rectangle alone.
> > > Yes I can. I can also use setClip without even Rectangle:
> > > setClip(10, 10, 2, 10);
> > > But I did it for 1:1 correspondence with original test :)
> > >
> > > Thanks,
> > > Alexander Shipilov
> > >
> > > >-----Original Message-----
> > > >From: Alexey Petrenko [mailto:alexey.a.petrenko@gmail.com]
> > > >Sent: Wednesday, December 13, 2006 5:05 PM
> > > >To: dev@harmony.apache.org
> > > >Subject: Re: [classlib][awt] clipping area
> > > >
> > > >2006/12/13, Shipilov, Alexander D <alexander.d.shipilov@intel.com>:
> > > >> >I think that it is better to use java.awt.Rectangle here.
> > > >> >Please add your modified test case with the results to JIRA.
> > > >> Sure, it is :). I meant that I used it both:
> > > >>        Area area = new Area(new Rectangle(10, 10, 2, 10));
> > > >Why do you need Area here? You can use Rectangle alone.
> > > >
> > > >SY, Alexey
> > > >
> > > >> >-----Original Message-----
> > > >> >From: Alexey Petrenko [mailto:alexey.a.petrenko@gmail.com]
> > > >> >Sent: Wednesday, December 13, 2006 4:57 PM
> > > >> >To: dev@harmony.apache.org
> > > >> >Subject: Re: [classlib][awt] clipping area
> > > >> >
> > > >> >2006/12/13, Shipilov, Alexander D <alexander.d.shipilov@intel.com>:
> > > >> >> Thank you, it is very appropriate action!
> > > >> >> So, if we change MultiRectArea for a class Area from public
API,
> > > >> Harmony
> > > >> >> works the same as RI.
> > > >> >I think that it is better to use java.awt.Rectangle here.
> > > >> >Please add your modified testcase with the results to JIRA.
> > > >> >
> > > >> >> Is that means that bug
> > > >> http://issues.apache.org/jira/browse/HARMONY-2032
> > > >> >> is incorrect and we should close it?
> > > >> >If it is not reproducible with public API this bug is incorrect,
> > > IMHO.
> > > >> >
> > > >> >SY, Alexey
> > > >> >
> > > >> >> >-----Original Message-----
> > > >> >> >From: Alexey Petrenko [mailto:alexey.a.petrenko@gmail.com]
> > > >> >> >Sent: Wednesday, December 13, 2006 4:13 PM
> > > >> >> >To: dev@harmony.apache.org
> > > >> >> >Subject: Re: [classlib][awt] clipping area
> > > >> >> >
> > > >> >> >Test case from this issue uses Harmony specific class
> > > >> >> >org.apache.harmony.awt.gl.MultiRectArea. I'm not sure
how it was
> > > >> >> >integrated to RI and does it work ok on RI.
> > > >> >> >
> > > >> >> >I think that first of all this test should be modified
to use
> > > only
> > > >> >> public
> > > >> >> >API.
> > > >> >> >
> > > >> >> >SY, Alexey
> > > >> >>
> > > >>
> > >
> >
>

Mime
View raw message