[Public WebGL] WebGL back buffer contents

Patrick Baggett [email protected]
Thu Jan 28 13:04:00 PST 2010


I've been watching this thread, and I'd say that most developers who are
used to XWindows and Win32 environments know that when there is a resizing
of the window, you need to call glViewport(), and I think the semantics Ken
described mirror that familiarity. When in doubt, I think the simplest
solution should be the best, and Ken is right, describing the semantics of
an automatic-resize-but-only-if-you-dont-do-glViewport (and making sure the
implementation does so properly) is just more difficult than it needs to be.
It also doesn't take into account that the author could call a 3rd party
piece of code that does glViewport() and then disables this automatic
behavior, and the author is left wondering why automatic viewport handling
doesn't work, etc. etc. Law of least surprises, anyone?

That being said, I think that automatic calls of glViewport() are just
wrong. The viewport is just how NDC are transformed, and have little to do
with the actual back buffer contents or dimensions -- they are GL state, not
part of the underlying windowing system state, therefore it should not be
the windowing system's [i.e. canvas in this case] job to modify them. I
think that just crosses some system separation line.

I use glViewport() to draw a subrect in 3D and use the remaining space for
2D elements. I would not want an implementation to waste it's time calling
glViewport() for me, just so I can call it with what I really want, or for
that matter, to clobber what I've done.

Patrick Baggett



On Thu, Jan 28, 2010 at 1:30 PM, Kenneth Russell <[email protected]> wrote:

>  On Thu, Jan 28, 2010 at 10:39 AM, Tim Johansson <[email protected]> wrote:
> > On 2010-01-28 18:27, Chris Marrin wrote:
> >>
> >> On Jan 28, 2010, at 1:43 AM, Tim Johansson wrote:
> >>
> >>>>>
> >>>> In recent conversations with people at NVIDIA they were pretty adamant
> >>>> that we not mess with the OpenGL state, in particular the viewport,
> >>>> upon changing the canvas size. Gregg's render target example is a good
> >>>> reason to not automatically call glViewport() behind the scenes.
> >>>> Another point is that it isn't obvious exactly when the implicit call
> >>>> to  glViewport() would be made -- whether it would be synchronous with
> >>>> the setting of the width / height properties or whether it would
> >>>> somehow be done lazily, before the first draw call after the canvas is
> >>>> resized.
> >>>>
> >>>> I used to think that doing this automatically would be a good idea but
> >>>> because of the semantic details I don't any more.
> >>>>
> >>>>
> >>>
> >>> We recently said that resizing should clear the canvas, which means we
> >>> have to mess with the gl sate anyway. We will at least need to change
> bound
> >>> FBO to 0 and disable scissor testing. Does it not qualify as messing
> with
> >>> the state if we mess with it and then try to restore it?
> >>
> >> Why would we have to change the FBO or scissor state? If you have an FBO
> >> set, then changing the size of the drawing buffer doesn't affect it,
> does
> >> it?
> >>
> > When resizing the canvas what you have to do is first resize the
> backbuffer
> > which is probably an FBO or pbuffer (unless you created a bigger
> backbuffer
> > than you needed to avoid having to resize it).
> >
> > After that you need to clear the backbuffer, otherwise resizing and
> drawing
> > a small triangle could give undefined results. Assuming you want to call
> > glClear to clear the backbuffer you have to switch FBO or you will clear
> the
> > applications FBO rather than the backbuffer.
> >
> > If the application has a scissor set and you clear without disabling
> scissor
> > test you will only clear the area defined by that scissor rect, which
> might
> > not be the entire new backbuffer.
> >
> > Changing the size of the drawing buffer will not affect the FBO, but it
> will
> > affect the drawing buffer in a way that is difficult to do without
> changing
> > the GL state.
> >
> >>>
> >>> I agree that in the case of rendering to an FBO you would not want to
> >>> change the viewport automatically, but I would rather have apps handle
> that
> >>> case themselves and set the viewport back than have apps handle the
> most
> >>> common case themselves. If the spec says that changing width or height
> >>> behaves as if it calls viewport as soon as the width/height attribute
> is
> >>> changed it would be clear what will happen. I would not even call it
> messing
> >>> with the state as it just does what the app tells it when the app tells
> it
> >>> to, that is change the size of the buffer and the viewport.
> >>>
> >>> We could also do something more magical and say that if you have not
> >>> changed the viewport yourself it will automatically change. If you ever
> did
> >>> call viewport automatic viewport adjustment is disabled. That way it
> would
> >>> never change the viewport for apps actually modifying the viewport and
> it
> >>> would just work for apps who does not care about the viewport. So
> initial
> >>> value for viewport would be "automatic", but there is no way to set it
> back
> >>> to "automatic" if you change it yourself.
> >>
> >> I'm not sure of your concern. What is the problem with requiring that
> the
> >> author calls ctx.viewport() after changing the width and height? The
> author
> >> has to know about viewports and keeping track of them. So why is that an
> >> issue?
> >>
> > If you have to call viewport every time you resize the canvas then yes,
> you
> > have to know about it. If it changes automatically you would not have to
> > know about viewports if you just want to render to the full canvas and
> don't
> > want to change the viewport.
> >
> > It is not a big issue, just seems like an inconvenience that can be
> fixed.
>
> It's going to be very complicated to explain the semantics to
> developers if we sometimes automatically resize the viewport and
> sometimes don't. Given the lack of per-element resize events, I think
> we should be consistent and require the developer to always set the
> viewport. Most developers will probably just set the viewport to
> (canvas.width, canvas.height) in their draw callback, which will allow
> them to handle resizing easily, and in one place.
>
> -Ken
>
> -----------------------------------------------------------
> You are currently subscribe to [email protected]
> To unsubscribe, send an email to [email protected] with
> the following command in the body of your email:
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://khronos.org/pipermail/public_webgl_khronos.org/attachments/20100128/dab8fb4b/attachment.html>


More information about the public_webgl mailing list