Instead of using qmake, use config-tool. This is no different than when
we deprecated the other per-dependency config-tool types (sdl2-config,
llvm-config, etc) for just config-tool
Currently the Qt Dependencies still use the old "combined" method for
dependencies with multiple ways to be found. This is problematic as it
means that `get_variable()` and friends don't work, as the dependency
can't implement any of those methods. The correct solution is to make
use of multiple Dependency instances, and a factory to tie them
together. This does that.
To handle QMake, I've leveraged the existing config-tool mechanism,
which allows us to save a good deal of code, and use well tested code
instead of rolling more of our own code.
The one thing this doesn't do, but we probably should, is expose the
macOS ExtraFrameworks directly, instead of forcing them to be found
through QMake. That is a problem for another series, and someone who
cares more about macOS than I do.
We don't always have qmake installed (and it's good to test failure
paths too!) so we can't expect this to succeed in all cases. With the
following commit we'll use a test.json to test both pkg-config and
qmake, so we need to be able to skip.
It's a method on the QtDependeny that exists purely for the consumption
of the qt module (in the form, return some stuff the module makes into
an instance variable). So put it where it actually belongs, and pass the
qt dependency into it.
Dependencies is already a large and complicated package without adding
programs to the list. This also allows us to untangle a bit of spaghetti
that we have.
If a custom_target output is a directory, we install it as a directory,
not as a file. And, we try to track subdirectories which are created so
uninstalling works. But one directory creation did not go through
DirMaker, in the case where the output directory does not have any
further subdirectories.
Consolidate on makedirs, since I don't see much point in using os.mkdir
right here.
This requires quite a complex and messy logic.
As @dcbaker suggested in #8491, this could be replaced by
an abstraction over linker flags instead of having GNU flags
translated.
get_non_matching_default_options is checking a string from
project_default_options against a validated value from
coredata.options.
Passing the string to validate_value ensures that the comparison
is sound; otherwise, "false" might be compared against False
and a bogus difference is printed.
When reverting from 0.57 to 0.56, one can see an error like this:
File /Users/pm215/src/qemu-for-merges/meson/mesonbuild/coredata.py,
line 1016, in load
obj = pickle.load(f)
ModuleNotFoundError: No module named 'mesonbuild.mesonlib.universal';
'mesonbuild.mesonlib' is not a package
FAILED: build.ninja
The reason is that the old version fails to resolve mesonbuild.mesonlib,
which is a similar situation to the existing AttributeError check. Raise
a MesonException for ModuleNotFoundError as well, so that reconfiguration
proceeds using cmd_line.txt.
Previously builds would *potentially* get sammed with messaging at
configure time that duplicate entries in an array would be an error in
the future, and the cause was because the same entries were getting
added over and over to pkg_config_path.p
Dylan Baker
HRXN
Eli Schwartz
Xavier Claessens
Remi Thebault
Jussi Pakkanen
Paolo Bonzini
Tristan Partin
Laurin-Luis Lehning