Fixes failing opengl32:sw_extensions test.
The test expects version `1.1.0` which is what Windows reports,
and ReactOS Mesa3D was reporting just `1.1`.
Instead of messing with global variables and the like, we introduce two target properties:
- WITH_CXX_EXCEPTIONS: if you want to use C++ exceptions
- WITH_CXX_RTTI: if you need RTTI in your module
You can use the newly introduced set_target_cpp_properties function, with WITH_EXCEPTIONS and WITH_RTTI arguments
We also introduce two libraries :
- cpprt: for C++ runtime routines
- cppstl: for the C++ standard template library
NB: On GCC, this requires to create imported libraries with the related built-in libraries:libsupc++, limingwex, libstdc++
Finally, we manage the relevant flags with the ad-hoc generator expressions
So, if you don't need exceptions, nor RTTI, nor use any runtime at all: you simply have nothing else to do than add your C++ file to your module
clang-cl 10:
'.../attrib.h(32,9): warning: 'ATTRIB_H' is used as a header guard here, followed by #define of a different macro [-Wheader-guard]'
Addendum to 5f2bebf.
CORE-7499 CORE-14306
Since 5f2bebf7, the Mesa library in dll/opengl/ has been downgraded from 8.0.4 to 2.6,
while the related documents were not updated to reflect the change.
Fixes GCC 8 warnings:
dll/opengl/glu32/src/libnurbs/internals/subdivider.cc:852:2: error: this 'if' clause does not guard... [-Werror=misleading-indentation]
if( jarc->pwlArc ) jarc->pwlArc->deleteMe( pwlarcpool ); jarc->pwlArc = 0;
^~
dll/opengl/glu32/src/libnurbs/internals/subdivider.cc:852:59: note: ...this statement, but the latter is misleadingly indented as if it were guarded by the 'if'
if( jarc->pwlArc ) jarc->pwlArc->deleteMe( pwlarcpool ); jarc->pwlArc = 0;
^~~~
There is no need to compile our DLLs as shared libraries since we are
managing symbols exports and imports through spec files.
On my system, this reduces the configure-time by a factor of two.
With this commit, we now use a forked version of MESA which only supports OpenGL 1.1, like the windows implementation does.
It exposes :
- The same pixel formats
- The same set of extensions
- Nothing more
All of this without taking 10% of your build time.
If you need a more modern option, look at the MESA package from Rapps, which is (and must be) maintained outside of this code tree.
CORE-7499
To check that these changes are correct, checkout in a directory (let's call it "ros_svn") the /trunk/reactos/ of our read-only SVN repo r76032 and in /trunk/reactos/modules/, the rosapps, rostests and wallpapers.
In a second directory (let's call it "ros_git"), clone the corresponding Git-converted ReactOS directory.
Before applying this patch (and the previous one that added back the empty directories), you should see additional files in ros_git that are not in ros_svn, corresponding to these files I'm deleting here (plus some .gitignore files),
and you should also see additional files in ros_svn that do not appear in ros_git: these are the empty directories I've restored in my previous patch.
Now, after the application of both the previous patch that restores the empty directories (and deletes the .gitignore files), and this patch that removes the ghost files, you should only see that the only differences
between ros_git and ros_svn are the extra .keep files in the empty directories, and that's all!
Command-line for the tests:
diff --strip-trailing-cr -r ros_svn ros_git > diff_svn2git.txt
"-r" means recursive, and "--strip-trailing-cr" ignores the CR-LF vs. LF (or CR) EOLs.
(*): by "ghost" old(*) files I understand files that existed previously in the far past, that then were deleted long ago in SVN, and that popped out back during the Git migration.
