On 10/01/2013 02:48 PM, Matt Turner wrote:
On Fri, Sep 20, 2013 at 7:02 PM, Chad Versace
<[email protected]> wrote:
Test eglCreateContext with with EGL_CONTEXT_OPENGL_BIT_KHR set for GL,
GLES1, GLES2, and GLES3. If context creation succeeds, then verify the
context is really a debug context by verifying GL_CONTEXT_FLAGS contains
GL_CONTEXT_FLAG_DEBUG_BIT. If context creation fails, then verify that
EGL_BAD_MATCH is emitted.
On Mesa 9.2 with Intel Haswell, the GL test passes, but the GLES tests fail
because eglCreateContext emits EGL_BAD_ATTRIBUTE.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=69624
Signed-off-by: Chad Versace <[email protected]>
---
tests/all.tests | 2 +
.../spec/egl_khr_create_context/CMakeLists.gl.txt | 1 +
.../spec/egl_khr_create_context/valid-flag-debug.c | 184 +++++++++++++++++++++
3 files changed, 187 insertions(+)
create mode 100644 tests/egl/spec/egl_khr_create_context/valid-flag-debug.c
+#include "piglit-util-egl.h"
+#include "piglit-util-gl-common.h"
+#include "common.h"
+
+const char *usage = "%s gl|gles1|gles2|gles3\n";
I don't like seeing printf format specifiers in strings like this.
Since usage is only used once, just inline the string into the fprintf
call.
Will fix in v2.
+ switch (context_bit) {
+ case EGL_OPENGL_BIT:
+ dispatch_api = PIGLIT_DISPATCH_GL;
+ break;
+ case EGL_OPENGL_ES_BIT:
+ dispatch_api = PIGLIT_DISPATCH_ES1;
piglit doesn't actually have ES1 dispatch. Is this a problem?
With the one fix and assuming this isn't a problem,
Reviewed-by: Matt Turner <[email protected]>
Argh. The lack of ES1 dispatch makes the test skip on ES1. Oops.
Seeing that the only GL functions used here are glGetString and glGetIntegerv,
I think it's safe to lie to piglit_dispatch_init by requesting ES2 dispatch.
I tried this, and it worked for me. What do you think?
v2 coming now, with the lie.
_______________________________________________
Piglit mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/piglit
