Global variable `grd_curcanv` is set to a variety of canvases, some of
which are local stack variables. Use of global variables in this way is
fragile, but works as long as the global is not used beyond the life of
the backing local.
Unfortunately, some existing uses do access the canvas beyond the
lifetime of the backing local. Playing movies sets the font of the
current canvas. If the current canvas is an expired stack variable,
setting the font overwrites other stack data. This data corruption
causes various symptoms, such as inability to play the escape tunnel
movie.
Prior to 03cca2b3dc, the corruption on
playing the endlevel movie had no user-visible effect. That commit
created a large local variable, which changed stack layout. Starting
with that commit, the corruption causes the movie to play as all black.
Fix this, and protect against some other data corruption possiblities,
by clearing the global when the local goes out of scope.
Reported-by: Havner <https://github.com/dxx-rebirth/dxx-rebirth/issues/345> (only as cutscene failure to play, not as the underlying corruption issue)
Rather than use an inline wrapper and rely on the compiler optimizer to
redirect gr_set_current_canvas(nullptr) to gr_set_default_canvas,
rewrite all relevant calls directly in the source.
git grep -l 'gr_set_current_canvas' | xargs sed -i -e 's:gr_set_current_canvas(\s*NULL\|nullptr\s*);:gr_set_default_canvas();:'
When reporting use of SDL_mixer, report its loaded version. When
running in verbose mode, report both the compile-time and load-time
versions of PhysFS, libSDL, and SDL_mixer.
Some Linux libraries print their own messages to stdout/stderr,
particularly in case of severe errors. Decorate messages generated by
Rebirth to distinguish them from library generated messages.
For specific blacklisted renderers, add a message informing the user
that the blacklist matched and changed settings.
x86_64-pc-linux-gnu-g++-6.3 issues a -Wodr diagnostic for the mismatch
between
common/include/3d.h: void _g3_draw_poly(grs_canvas &, uint_fast32_t nv, cg3s_point *const *pointlist, uint8_t color);
similar/arch/ogl/ogl.cpp: void _g3_draw_poly(grs_canvas &canvas, const uint_fast32_t nv, const g3s_point *const *const pointlist, const uint8_t palette_color_index)
The declaration and definition are compatible because `cg3s_point` is a
type alias for `const g3s_point`, so the code worked as it was. Switch
the definition to use `cg3s_point` like the declaration to prevent this
warning and to protect against any future (unlikely) changes to the
definition of `cg3s_point`.
Macro LINE_SPACING previously used global grd_curcanv implicitly.
Change it to take a canvas argument. Change all callers to pass
grd_curcanv, so that usage is explicit.
Split ogl_get_verinfo() into ogl_tune_for_current() and
ogl_extensions_init(), and consolidate all the OpenGL extension handling
into ogl_extensions.cpp. Unify the code paths for texture anisotropy and
GPU synchronization for OpenGL and OpenGL ES.
Currently, our renderer only uses GLES 1.0, so no real world implementation
will support sync objects for such an old context, but the logic is valid,
and this way, the GLES specific code paths are reduced.
This patch also fixes an issue where the old ogl_get_verinfo() did modify
the texture filtering mode if no anisotropic filter was available. This
was some leftover from the time when the anisotropic filter was a just
a specific CGameConfig.TexFilt mode, and not a separate, orthogonal setting
CGameCfg.TexAnisotropy that it is now.
derhass reports that OpenGL ES builds warn for unused parameter
`texanis`. This parameter controls a feature that is not available in
current Rebirth OpenGL ES builds. Cast the parameter to void to silence
the warning. Long term, the code ought to be changed to advise the user
when their platform lacks this feature. In the short term, merge this
fix so that the build will succeed.
https://github.com/dxx-rebirth/dxx-rebirth/pull/256