Re: [Shapefusion-devel] ShapeFusion progress

[tiennou <tie...@gm...>]

Le 28 f=C3=A9vr. 07 =C3=A0 15:11, ti...@da... a =C3=A9crit :

> Sorry for the scarce partecipation, I'm next to an importan
> exam and I'm doing  brain time-sharing between multiple
> disciplines...

No problem about that... I wish I could let you
>
>> Are you satisfied with my naming conventions of the code I
>> revamped ? Using mKeypointObscured as a member name and
>> KeypointObscured() and   SetKeypointObscured() as methods
>> ?
> I agree.

Okay, so that's what I'll use...
>
>> I'm trying to get ShapesEditor to work again, and as they
>> are   numerous references to the old Shapes.h structs, I'm
>> thinking about   making all members public, because it
>> would make it work with no   change. On the other hand
>> this would mean that my accessor stuff is   useless, so
>> I'm wondering ;-)
> I say let's make all them public. It's not ideal, but we can
> fix it later.

I think I will change everything right now, because I've already =20
changed members name, so I'm stuck.

>> I'm thinking about refactoring code from View/Browser,
>> because there   is numerous methods related to turning
>> ShapesFrame/ShapesBitmap into   images/thumbnails for
>> viewing. I added these capabilities to   ShapesBitmap (you
>> can now call ShapeBitmapToImage and ImageThumbnail
>> directly on them), but I need to add a pointer to a
>> ShapesChunk in   ShapesFrame in order to be able to access
>> the associated ShapeBitmap-
>>> ShapeBitmapToImage call... What do you think about it ?
> I don't like this. IMO a ShapesFrame should not be able to
> generate its own
> wxImage, because it's linked to a ShapesBitmap in a "weak"
> way (that is, you need to access the full chunk to realize
> the link). I'm thinking about a new class that knows how to
> convert ShapesBitmaps and ShapesFrames to wxImage. It could
> be called for example ShapesRenderDevice or ShapesRenderer.
> The nice thing about that would be just ONE SETTING for
> "show transparent pixels", "view using color table", "smooth
> thumbnails" and so on. We could just ask it "I need to
> convert this bitmap/frame for viewing, could you do it?" and
> it would create a wxImage with the correct display settings.
> It could even cache thumbnails I guess. We could store a
> pointer to this converter in ShapesView. Could it work?

I think there is some need for a refactor in those parts of the code, =20=

but yes, it can be delayed until we have a working wxDocument approach.
I was just going overshot ;-)

tiennou