On 2010-12-20, at 3:18 PM, Kenneth Russell wrote:
On Mon, Dec 20, 2010 at 12:22 PM, Steve Baker <firstname.lastname@example.org
On 12/20/2010 12:42 PM, Kenneth Russell wrote:
On Sat, Dec 18, 2010 at 8:32 PM, Steve Baker <email@example.com> wrote:
On my son's EeePC netbook (which has some god-awful Intel processor) -
using Chrome/ANGLE - I'm getting a GL "Invalid Operation" error right after:
var gl = canvas.getContext ( "experimental-webgl",
alpha : false ,
antialias : true ,
depth : true ,
stencil : false ,
premultipliedAlpha: false } ) ;
...but it's returning a non-null value for "gl" and if I clear the error
- everything seems to run OK. I don't see this problem on other systems.
I thought these parameters were just advisory - there shouldn't be any
illegal values -right? Sure the machine doesn't support antialiassing
- but it's OK to ask...right?
Yes, this sounds like a bug. Does it also happen if you pass the
command line argument --use-gl=desktop ? If not, please file a bug
with the ANGLE project -- sounds like a GL error is being synthesized
when it shouldn't be. Otherwise, please file a bug at
http://crbug.com/ and CC: kbr at chromium.org.
This netbook doesn't have OpenGL support (crappy Intel chipset only
supports OpenGL 1.2!)...so we can't do that test.
I see. Well, feel free to log bugs against both projects and point
each to the other bug. We'll probably have to make them low priority
but if we can get our hands on similar hardware can try to track down
where the error is being synthesized.
You could always build your own ANGLE debug libraries (all you need is Visual Studio 2008 Express) and run with those.
The resulting "debug.txt" file should allow you to figure out if the if ANGLE generated the error or not.
Hope this helps,