Hi Guys,
My changes are now checked in, the createGraphicsContext() call now
automatically catches any undefined displayNum and screenNum settings,
resetting them automatically to 0 respectively.
This is catch for when user code handles undefined displayNum and
screenNum settings that can be set by readDisplay(). It should be that
the user code actually detects the -1 values that can be set and
explicitly handle these, so if you wish to assume a screenNum of 0
when the screenNum is undefined you can do this, or set it to another
value if appropriate. Personally I rather than user code does decide
explicitly what the value should be, rather than the OSG implicitly
assume some default setting when values are undefined.
FYI, These changes didn't make it into the OSG-2.3.1 release I made
this morning, and where checked in about two hours after I tagged.
Robert.
On Jan 4, 2008 12:55 PM, Robert Osfield <[EMAIL PROTECTED]> wrote:
> Hi Guys,
>
> I'm currently compiling a couple of change that will hopefully catch
> undefined screen settings slightly higher up - in the
> GraphicsContext::createGraphicsContext(Traits*) call.
>
> I've also added a new GraphicsContext::ScreenIdentifier method, and
> extra details about the undefined screen num policy that readDisplay()
> uses. The new method looks like:
>
> /** Set any undefined displayNum or screenNum values (i.e.
> -1) to the default display & screen of 0 respectively.*/
> void setUndefinedScreenDetailsToDefaultScreen()
> {
> if (displayNum<0) displayNum = 0;
> if (screenNum<0) screenNum = 0;
> }
>
>
> I'll check this in once things are working.
> Robert.
>
>
> On Jan 4, 2008 11:28 AM, Robert Osfield <[EMAIL PROTECTED]> wrote:
> > Hi Tim,
> >
> > On Jan 4, 2008 11:20 AM, Tim Moore <[EMAIL PROTECTED]> wrote:
> > > I couldn't find one. There is some code in osgViewer/View.cpp to deal
> > > with the case in
> > > setUpViewAcrossAllScreens(), which is probably the common code path for
> > > osgviewer. I
> > > suspect that you hit the problem if your application initializes a Traits
> > > structure and
> > > creates a graphics context itself.
> >
> > If the catch blocks in osgViewer::View aren't been used then the only
> > place to catch this case is within GrpahicsWindowX11, Win32 and
> > Carbon, this should be a straight forward addition though. I believe
> > added entries to these is better than hacking the catch for when no
> > screen number is defined, -1 is deliberately used to tell the
> > developer than the DISPLAY string didn't explicitly specify a screen
> > number, remove this and you also remove the ability for the
> > application developer to do their own handling of the fallback.
> >
> > Robert.
> >
>
_______________________________________________
osg-submissions mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
