Open
Bug 860163
Opened 12 years ago
Updated 2 years ago
Tegra 3 fails 1.0.1/conformance/more/functions/uniformi and uniformiBadArgs
Categories
(Core :: Graphics: CanvasWebGL, defect, P5)
Tracking
()
NEW
People
(Reporter: jgilbert, Unassigned)
Details
(Whiteboard: webgl-driver)
`int`s coming out of a `ivec4` seem to max out at a value of 2.
GLSL precision for `int` should default to mediump, and indeed adding `precision mediump int` doesn't change anything. Add `precision highp int` and it works, though. It would be spec-compliant to supply highp when mediump (or lowp!) is required, so we *could* just do that.
FWIW, even lowp ints should have a range of (2^-8, 2^8). Only lowp floats can have a range as small as (-2, 2). Why the driver would supply lowp float functionality for mediump ints is anyone's guess.
We should see about injecting `precision highp int` on affected devices. (where available)
Reporter | ||
Updated•11 years ago
|
Whiteboard: webgl-driver
Reporter | ||
Updated•6 years ago
|
Priority: -- → P5
Updated•2 years ago
|
Severity: normal → S3
You need to log in
before you can comment on or make changes to this bug.
Description
•