The GL queries are needed only if the user-provided pixel storage
doesn't contain enough information, in particular dimensions and byte
size of compression block.
So in case they are present, no query for compressed image size is done
for full image queries (just saving one API call) and no queries for
block dimensions and byte size for subimage queries (saving four API
calls and not requiring ARB_internalformat_query2).
The documentation was fixed and tests were updated to take the
ARB_internalformat_query2 requirement into account.
There were various issues with the tests at the moment:
* In order to test non-default pixel storage parameters, out-of-bound
array slices were passed to GL functions. That of course cause ASan
to complain (and maybe also some drivers to crash).
* Only the non-default pixel storage parameters were tested.
Unfortunately this is a heavily buggy area in all the drivers, so
when such a test failed, it was not immediately clear whether the
failure is related just to driver not understanding pixel storage
properly or something deeper and made debugging much much harder than
it should be.
* For compressed textures, default and non-default pixel storage has
grossly different code paths and these should be tested both by
default.
* Some test cases had a patched-in code testing for both default and
non-default pixel storage, but the code was so hidden that I didn't
even know it was there -- it should be separate test cases.
All the tests are now instanced, testing both the simple case and
some pixel storage case, they are also reordered and cleaned up a bit.
There will be numerous additions to this one so it made sense to make it
a static library instead of a header-only library. That also allows
CMake users to just link to Magnum::OpenGLTester instead of going
through the pain of a huge branching in order to find a correct
windowless application just to run their tests. It could have been done
even without the static library using a INTERFACE target, but that
wouldn't work on CMake < 3.0 (which, unfortunately, quite a few people
are still stuck with).
Unfortunately it's already heavily used elsewhere so I had to go through
the pain of deprecating the old implementation. The old implementation
was header-only so it can't be just typedef'd to the new one as there
would be linker failures. So the old header is just kept as it was, with
only the macros reduced.
To be more consistent with GLSL naming. Also, the original naming was
quite misleading, as normalize() is used in GLSL for something
completely different.
If building with deprecated APIs, the Functions.h header includes the
new Packing.h header and the {de,}normalize() functions are defined as
deprecated aliases to the new functions. This will be removed at some
point in the future.
Each Windowless*Application has now a companion Windowless*Context that
manages just the GL context creation and nothing else, with the ability
to just create the context and not make it current, so it doesn't affect
current thread state and can be moved to another thread and make current
there.
Other things that were done:
* Using `NoCreateT` instead of `nullptr` for creating the application
without creating GL context.
* Properly handling failed creation of Magnum context instance -- if it
errors out, also the GL context is destroyed to make it possible to
create the context with a different configuration.
* Reworked AMD and NVidia binary driver workaround, where core context
created with specific version doesn't automatically choose the newest
available (creating compatibility context on the other hand causes
the version to get stuck on 2.1 on Mesa and OSX).
* Added the above workaround also for WindowlessWglApplication to avoid
driver issues in the future.
* Reworked WindowlessWglApplication to not be so crazily entangled. It
was a misunderstanding on my side about how WINAPI works. Much
simpler now (and I hope still working :D).
Vladimír Vondruš