From: José Fonseca
Adapt the primitive-id-no-gs-first test case to first provoking vertex
convention, exposing a bug in llvmpipe.
---
tests/shaders/shader_runner.c | 15 ++
.../primitive-id-no-gs-first-vertex.shader_test| 63 ++
2 files changed, 7
From: José Fonseca
Deprecated by JUnit backend.
Tested with json backend + summary html to ensure nothing's broken.
---
framework/junit.py| 377 --
framework/programs/summary.py | 111 -
piglit| 4 -
pigli
From: José Fonseca
Use os.path instead of posixpath, which handles both / and \ on Windows.
---
framework/backends/junit.py | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/framework/backends/junit.py b/framework/backends/junit.py
index 59d06c6..3647601 100644
--- a/frame
From: José Fonseca
Otherwise subprocess will raise "preexec_fn is not supported on Windows
platforms" ValueError exception.
---
framework/test/base.py | 8 +++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/framework/test/base.py b/framework/test/base.py
index c3234d6..7aab45f
From: José Fonseca
Showing the command line of the test is quite useful, specially when
diagnosing failures.
Unfortunately JUnit format allows no better place to have it other than
stdout/stderr.
---
framework/backends/junit.py | 3 +++
1 file changed, 3 insertions(+)
diff --git a/framework/ba
From: José Fonseca
Like done with other drivers. To make it easier for people to test
llvmpipe.
---
tests/llvmpipe.py | 16
1 file changed, 16 insertions(+)
create mode 100644 tests/llvmpipe.py
diff --git a/tests/llvmpipe.py b/tests/llvmpipe.py
new file mode 100644
index
From: José Fonseca
Tested with NVIDIA drivers.
---
.../glx_arb_create_context/invalid-es-version.c| 23 +++---
1 file changed, 7 insertions(+), 16 deletions(-)
diff --git a/tests/spec/glx_arb_create_context/invalid-es-version.c
b/tests/spec/glx_arb_create_context/invalid-es
From: José Fonseca
https://www.opengl.org/sdk/docs/man2/xhtml/glXChooseFBConfig.xml states
that
"GLX_DOUBLEBUFFER Must be followed by True or False."
No idea how this ever passed given the buffer overflow that ensued.
---
tests/glx/glx-close-display.c | 2 +-
1 file changed, 1 insertion(+),
From: José Fonseca
As these are not supported on MSVC, not even MSVC 2013 since it only has
library support for C99, it doesn't actually support C99 syntax.
Suggested by Jan Vesely.
---
CMakeLists.txt | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/CMakeLists.txt b/CMak
From: José Fonseca
Verified that cmake aborts with MSVS 2012.
https://bugs.freedesktop.org/show_bug.cgi?id=85528
---
CMakeLists.txt | 13 +-
README | 6 +-
include/msvc/c99/inttypes.h | 305 -
From: José Fonseca
---
.../spec/arb_texture_cube_map_array/cubemap-lod.c | 21 ---
tests/spec/arb_texture_cube_map_array/cubemap.c| 21 ---
tests/texturing/cubemap-shader.c | 21 ---
tests/texturing/cubemap.c
From: José Fonseca
---
tests/shaders/shader_runner.c | 8 ++--
tests/util/piglit-util.h | 7 ---
2 files changed, 2 insertions(+), 13 deletions(-)
diff --git a/tests/shaders/shader_runner.c b/tests/shaders/shader_runner.c
index 1f93216..628ef44 100644
--- a/tests/shaders/shader_run
From: José Fonseca
undefining max/min seems unavoidable, but at least it can be done in a
single place.
---
tests/spec/glsl-1.30/execution/fs-texelFetch-2D.c | 3 ---
tests/spec/glsl-1.30/execution/fs-texelFetchOffset-2D.c | 3 ---
tests/spec/glsl-1.30/execution/fs-textureOffset-2D.c|
From: José Fonseca
As the semantics of sprintf_s's arguments are quite different, so code
that relies on it might misbehave.
---
tests/util/piglit-util.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tests/util/piglit-util.h b/tests/util/piglit-util.h
index 0238b7f..84e2066
From: José Fonseca
---
tests/util/piglit-util.h | 8
1 file changed, 8 deletions(-)
diff --git a/tests/util/piglit-util.h b/tests/util/piglit-util.h
index 24ee405..0238b7f 100755
--- a/tests/util/piglit-util.h
+++ b/tests/util/piglit-util.h
@@ -60,14 +60,6 @@ extern "C" {
#define sn
From: José Fonseca
It's already defined in tests/util/piglit-util.h if necessary, furthermore
it's a macro on MSVC, so redefining it causes errors.
Trivial.
---
tests/spec/gl-1.0/readpix.c | 7 ---
1 file changed, 7 deletions(-)
diff --git a/tests/spec/gl-1.0/readpix.c b/tests/spec/gl-1.0/
From: José Fonseca
Non-standard GCC extension.
---
tests/spec/gl-1.0/readpix.c | 64 ++---
1 file changed, 37 insertions(+), 27 deletions(-)
diff --git a/tests/spec/gl-1.0/readpix.c b/tests/spec/gl-1.0/readpix.c
index 7ee0be9..26f44a6 100644
--- a/tests/s
From: José Fonseca
It seems that Mako templates don't handle CRLF properly, so force them
to retain Unix line endings on Windows.
Fixes https://bugs.freedesktop.org/show_bug.cgi?id=80489
---
.gitattributes | 1 +
1 file changed, 1 insertion(+)
create mode 100644 .gitattributes
diff --git a/.g
From: José Fonseca
IIRC they are not a standard C feature. At any rate, they aren't
accepted by MSVC.
---
tests/texturing/bptc-float-modes.c | 8
tests/texturing/bptc-modes.c | 18 +-
2 files changed, 13 insertions(+), 13 deletions(-)
diff --git a/tests/texturin
From: José Fonseca
Otherwise the build fails when libwaffle-1.so is not on a standard
library directory (ie., outside /usr/lib or /usr/local/lib).
---
tests/util/CMakeLists.txt | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tests/util/CMakeLists.txt b/tests/util/CMakeLists.t
From: José Fonseca
It's fine for URLs to have forward slash directory separators when viewing
the HTML on Windows. But if one moves the HTML report to a Linux machine
or puts it on a web-server and accesses it from a Linux machine then URLs
with forward slashes become dead.
This change prevents
From: José Fonseca
Just like piglit_report_result(PIGLIT_SKIP) does.
Otherwise these tests are considered failures or -- with my commit
058e0f8a1e536b68ef43d27ada80645845a39e19 -- crashes.
---
framework/exectest.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/framework/ex
From: José Fonseca
Also return the exception code as the program return code, which matches
the normal behaviour of uncaught exceptions.
Currently I see several tests failing on Windows, but with no clue
(in stdout, stderr, or exit code) as to why. This should help.
Trivial. Only affects Wind
From: José Fonseca
Should fix https://bugs.freedesktop.org/show_bug.cgi?id=79745
---
cmake/target_api/gl/CMakeLists.txt | 11 +++
tests/util/piglit-framework-gl/piglit_glut_framework.c | 8 +---
2 files changed, 16 insertions(+), 3 deletions(-)
diff --git a/cmak
From: José Fonseca
Jenkins doesn't make a huge distinction between failures and errors though.
---
framework/programs/summary.py | 2 ++
1 file changed, 2 insertions(+)
diff --git a/framework/programs/summary.py b/framework/programs/summary.py
index 28fcffb..da4e040 100644
--- a/framework/progr
From: José Fonseca
From what I gathered, both on Linux and Windows, uncaught
exceptions/signals lead to a negative return code.
---
framework/exectest.py | 16 +---
1 file changed, 1 insertion(+), 15 deletions(-)
diff --git a/framework/exectest.py b/framework/exectest.py
index 5cfa0
From: José Fonseca
- Prevent the use of reserved characters in filenames (e.g., '>' and '<'
which are used in several cases.)
- Handle mixture of backward and forward slashes gracefuly.
---
framework/summary.py | 19 +++
1 file changed, 15 insertions(+), 4 deletions(-)
diff -
From: José Fonseca
Because test paths end up having a mixture of backward and forward
slashes.
Trivial.
---
framework/programs/summary.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/framework/programs/summary.py b/framework/programs/summary.py
index db3a83e..28fcffb 1006
From: José Fonseca
Otherwise Jenkins won't show any progress while piglit is running.
PS: Another alternative would be to detect the presence of
JENKINS_HOME/HUDSON_HOME environment variables. I'm happy either way.
---
framework/log.py | 25 ++---
1 file changed, 18 inserti
From: José Fonseca
Therefore preventing tests from hanging. More details in the code comment.
---
tests/util/piglit-util.c | 61
1 file changed, 61 insertions(+)
diff --git a/tests/util/piglit-util.c b/tests/util/piglit-util.c
index 9f09ba2..6fe
From: José Fonseca
Windows doesn't support #!. Instead it knows how to invoke scripts
based off the file extension.
---
CMakeLists.txt | 7 ++-
piglit | 4
2 files changed, 10 insertions(+), 1 deletion(-)
diff --git a/CMakeLists.txt b/CMakeLists.txt
index a77a2a8..fed58bb 1006
From: José Fonseca
This (plus a corresponding Mesa change) enables to run the piglit
testsuite through piglit-run without human intervention.
---
framework/programs/run.py| 18 ++
tests/util/piglit-framework-gl.h | 2 ++
tests/util/piglit-util.c | 37
From: José Fonseca
Mesa didn't get this quite right, causing problems to ApiTrace --
https://github.com/apitrace/apitrace/issues/261 .
---
tests/all.py | 1 +
tests/spec/arb_vertex_array_bgra/CMakeLists.gl.txt | 1 +
tests/spec/arb_vertex_array_bgra/get.c
From: José Fonseca
Copy'n'paste typo. Trivial.
---
CMakeLists.txt | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/CMakeLists.txt b/CMakeLists.txt
index cbdccf1..3b051f7 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -220,7 +220,7 @@ if (NOT MSVC)
# Unfortunately
From: José Fonseca
max/min were being used in several places, without including
algorithm header, and on MSVC 2013 the header is not being included
internally.
---
tests/glean/tbinding.cpp | 2 +-
tests/glean/tblend.cpp | 1 +
tests/glean/tlogicop.cpp | 1 +
tests/glean/tpgos.cpp| 1 +
tes
From: José Fonseca
MSVC does not consider `assert(TRUE)` as never returning.
---
tests/spec/arb_shader_atomic_counters/common.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/tests/spec/arb_shader_atomic_counters/common.c
b/tests/spec/arb_shader_atomic_counters/common.c
index ac7627f..4590
From: José Fonseca
---
tests/util/piglit-dispatch.h | 14 ++
1 file changed, 6 insertions(+), 8 deletions(-)
diff --git a/tests/util/piglit-dispatch.h b/tests/util/piglit-dispatch.h
index 669aecc..c24c186 100644
--- a/tests/util/piglit-dispatch.h
+++ b/tests/util/piglit-dispatch.h
@
From: José Fonseca
---
CMakeLists.txt | 8 +++-
include/msvc/c99/stdbool.h | 4 ++--
2 files changed, 9 insertions(+), 3 deletions(-)
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 3b051f7..6f45d2e 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -230,7 +230,13 @@ if (NOT
From: José Fonseca
It should have been SENTINEL and not -SENTINEL in piglit_test_float,
as that what's checked below.
Also take in consideration that in general `(GLfoo)SENTINEL != SENTINEL`,
specially for GLfloat.
---
tests/util/minmax-test.c | 6 +++---
1 file changed, 3 insertions(+), 3 dele
From: José Fonseca
For MSVC `foo()` and `foo(void)` are not the same thing.
---
tests/spec/gl-1.0/beginend-coverage.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/tests/spec/gl-1.0/beginend-coverage.c
b/tests/spec/gl-1.0/beginend-coverage.c
index 8befd30..71c69f5 10
From: José Fonseca
Very few people are building piglit with MSVC, and the current avalanche
of insignificant warnings ends up being a disservice for those who do, as
it only ends up hiding potential problems.
---
CMakeLists.txt | 11 ++-
1 file changed, 6 insertions(+), 5 deletions(-)
d
From: José Fonseca
MSVC 2013 provides all math.h functions that were missing in previous
versions.
---
tests/texturing/shaders/common.c | 7 ---
tests/util/piglit-util.h | 18 ++
2 files changed, 14 insertions(+), 11 deletions(-)
diff --git a/tests/texturing/shaders
From: José Fonseca
Not supported by MSVC.
---
tests/texturing/shaders/textureGather.c | 7 +--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/tests/texturing/shaders/textureGather.c
b/tests/texturing/shaders/textureGather.c
index cb7fef0..bd48790 100644
--- a/tests/texturing/
From: José Fonseca
Handy until we have waffle everywhere.
---
.../piglit-framework-gl/piglit_glut_framework.c| 30 +++---
1 file changed, 27 insertions(+), 3 deletions(-)
diff --git a/tests/util/piglit-framework-gl/piglit_glut_framework.c
b/tests/util/piglit-framework-gl/pi
From: José Fonseca
---
tests/fbo/CMakeLists.gl.txt | 6 +-
tests/fbo/fbo-depth-sample-compare.c | 17 -
2 files changed, 17 insertions(+), 6 deletions(-)
diff --git a/tests/fbo/CMakeLists.gl.txt b/tests/fbo/CMakeLists.gl.txt
index 2f33481..7a24147 100644
--- a/test
From: José Fonseca
When testing with llvmpipe, somehow the draw() function is returning
constant elapsed time of 0.000222 approximately, regardless how big
iters is.
This is probably a bug in llvmpipe, but it seems a good idea to fix the
test to prevent integer overflow and the consequent infini
From: José Fonseca
Which was lost as result of commit 5a12f3b9f0d6d93da1ac7a72c9806b1a831a1744.
---
tests/quick.py | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/tests/quick.py b/tests/quick.py
index 939aded..d603265 100644
--- a/tests/quick.py
+++ b/tests/quick.py
@@
From: José Fonseca
These tests were relying on the generated tests discovery to create
empty groups from empty directories. However empty directories
were not being packaging, causing Python KeyError exceptions.
This change adds the Group initialization, like done elsewhere.
---
tests/all.tests
From: José Fonseca
It doesn't make much sense to package piglit or install on machines, due
to the frequency piglit tests are added/changed.
However, binary packages are still useful for distributed testing where
the binaries are built on one machine, and then passed to (sometimes
other teams) f
From: José Fonseca
Not actually necessary. rgbTriStrip test used it but it was just a proof
of concept, never part of any test list.
This greatly simplifies building piglit for Windows.
---
CMakeLists.txt|4 +-
README| 35 +---
tests/CMakeLists.
From: José Fonseca
Signed types have less precision due to the sign bit.
Without this it is impossible to devise an efficient way of lerping
signed normalized types that makes this test happy -- it would require
32bit float/int multiply and divides to get exact results.
---
tests/fbo/fbo-blendi
From: José Fonseca
---
src/glean/tstencil2.cpp | 154 ---
src/glean/tstencil2.h | 19 +-
2 files changed, 151 insertions(+), 22 deletions(-)
diff --git a/src/glean/tstencil2.cpp b/src/glean/tstencil2.cpp
index 7db3a06..9440c6f 100644
--- a/sr
From: José Fonseca
---
.../glsl2/comment-continuation-preprocessor.frag | 34
1 file changed, 34 insertions(+)
create mode 100644
tests/glslparsertest/glsl2/comment-continuation-preprocessor.frag
diff --git a/tests/glslparsertest/glsl2/comment-continuation-preprocesso
From: José Fonseca
A variation of
generated_tests/spec/glsl-1.20/execution/uniform-initializer/vs-bool-array.shader_test
which exercises a bug in Mesa gallium state tracker.
---
.../execution/vs-all-equal-bool-array.shader_test | 32
1 file changed, 32 insertions(+)
cre
From: José Fonseca
ARB_instanced_arrays specification says:
Dependencies on ARB_draw_instanced
If neither ARB_draw_instanced nor EXT_draw_instanced is
supported, all references to instanceID should be removed from
section 2.8.
If ARB_draw_instanced is not su
From: José Fonseca
---
tests/util/gen_dispatch.py |7 +++
1 file changed, 7 insertions(+)
diff --git a/tests/util/gen_dispatch.py b/tests/util/gen_dispatch.py
index 96a7f77..1e19c61 100644
--- a/tests/util/gen_dispatch.py
+++ b/tests/util/gen_dispatch.py
@@ -474,6 +474,13 @@ def generat
From: José Fonseca
Based of fbo-blit.
It's not as strict as it could be, because OpenGL's rules are a bit
ambigous in respect to clamping to edges, and because Mesa clamps blit
with integer precision.
---
tests/fbo/CMakeLists.gl.txt|1 +
tests/fbo/fbo-blit-stretch.cpp | 503 +++
From: José Fonseca
Verified on NVIDIA.
---
.../execution/fs-textureGrad-miplevels.shader_test | 193
1 file changed, 193 insertions(+)
create mode 100644
tests/spec/glsl-1.30/execution/fs-textureGrad-miplevels.shader_test
diff --git
a/tests/spec/glsl-1.30/execution/fs-t
From: José Fonseca
---
CMakeLists.txt | 14 +++---
1 file changed, 11 insertions(+), 3 deletions(-)
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 7202417..0aad196 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -76,16 +76,24 @@ if(NOT DEFINED python)
message(FATAL_ERR
From: José Fonseca
Stencil buffer is not normally visible, so it's much easier to debug this
if the color inversion is done through blending.
---
tests/general/triangle-rasterization-overdraw.cpp | 24 +++--
1 file changed, 8 insertions(+), 16 deletions(-)
diff --git a/tests/g
60 matches
Mail list logo