| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Rename waffle_dlsym_gl to waffle_dl_sym, and place it into a new module
waffle_dl. A future commit will add a new function to this module:
waffle_dl_can_open.
Rename all analogous internal functions PLATFORM_dlsym_gl to
PLATFORM_dl_sym, and place it into a new module PLATFORM_dl.
I'm trying to maintain a consistent naming scheme where functions are
prefixed according to module and where the public api closely matches the
internal api, at least until the public api freezes at 1.0.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
The 'dl' param selects which GL library on which to call dlsym.
This is a prerequisite for allowing the creation multiple contexts with
different GL API's simultaneously. To achieve that goal, we must eliminate
the WAFFLE_GL_API attribute from waffle_init.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Add the following enums:
WAFFLE_DL_OPENGL
WAFFLE_DL_OPENGL_ES1
WAFFLE_DL_OPENGL_ES2
A future commit will unbind waffle_dlsym_gl from the API given to
waffle_init. Then, these enums will be passed to waffle_dlsym_gl to speciy
the GL library.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
This makes the name consistent with other functions in the waffle_error
module.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Change the signature of waffle_get_error_m to simply return a `const
char*` that points to the error message. Before this patch, the function
required a string buffer argument and had tricky calling conventions.
While we're touching the function, rename it to waffle_error_get_info.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Of course, we don't reset the error state at api entry points that
*retrieve* the error state.
Update the doxygen in waffle_error.h to reflect this change.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
This whitespace inconsitency was causing spurious merge conflicts.
This patch was created with
for f in $(find $WAFFLE_TOP -type f | grep -E "\.(c|h|txt)$")
do
if ! diff <(tail -c1 "$f") <(echo) >&/dev/null
then
echo >> "$f"
fi
done
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
To ensure this, each file is enclosed in one of two blocks: @defgroup or
@addtogroup.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Now, all functions in the wcore_error use thread-local storage.
In public doxygen, document that waffle_get_error() now returns the error
code of the most recent failure of *the current thread*.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
WAFFLE_UNSUPPORTED_ON_PLATFORM
This error is for requested actions that are unsupported on the
current platform.
WAFFLE_NOT_IMPLEMENTED
The error is for features whoe implementation is not yet complete.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
* Describe attributes in a table.
* Fix incorrect Doxygen syntax.
Doxygen interprets indented blocks as code, and waffle_init() had
all paragraphs wrongly indented.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
* Describe attributes in a table.
* Give a code example.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
This name is more consistent with GLX and EGL, where the enum name is None
and EGL_NONE, respectively.
Also, switch positions of WAFFLE_NONE(0) and WAFFLE_DONT_CARE(-1), because
-1 < 0.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
The @defgroup directives never worked here. Future commits will document
these enums in the doxygen for the function that uses them.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
This matches the value of EGL_DONT_CARE and GLX_DONT_CARE.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Change the default value of WAFFE_{RED,GREEN,BLUE,ALPHA}_SIZE to
WAFFLE_DONT_CARE. The egl module translates WAFFLE_DONT_CARE to
EGL_DONT_CARE.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Nothing uses the enum yet.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
WAFFLE_GL_API -> WAFFLE_OPENGL_API
WAFFLE_GL -> WAFFLE_OPENGL
WAFFLE_GLES1 -> WAFFLE_OPENGL_ES1
WAFFLE_GLES2 -> WAFFLE_OPENGL_ES2
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
The default values were chosen during CMake configuration time. This was
a recipe for disaster. libwaffle built by different people would have
significant behavioral differences.
- Remove the CMake options waffle_default_{platform,gl_api}.
- Remove the macros WAFFLE_{PLATFORM,GL_API}_DEFAULT.
- Require the user provide attributes WAFFLE_{PLATFORM,GL_API} to
waffle_init().
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Install libwaffle.so to ${waffle_install_libdir}.
Install headers to ${waffle_install_includedir}/waffle.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Add a `display` parameter to waffle_make_current in order to allow users
to release the current context. To release the current context, the user
calls it like this: waffle_make_current(dpy, window=0, ctx=0).
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
All tests pass.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
The API is mostly complete. As far as my future-looking eye can tell, it
only lacks:
- a way to choose the GL version and profile
- a way to handle key and mouse input
The API likely lacks much more. My future-looking eye is quite blind.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
