Fixed manually promoting wrap files with a full path, e.g.
`meson wrap promote subprojects/s1/subprojects/projname.wrap`,
which resulted in an error before (new test added:
`./run_unittests.py AllPlatformTests.test_subproject_promotion_wrap`).
Additionally, running promote with an invalid subproject path now fails
properly. Before, it just silently did nothing (added to test:
`./run_unittests.py AllPlatformTests.test_subproject_promotion`).
g-ir-scanner is very picky about the flags that it can accept, so the
build fails on macOS if you have Framework external dependencies,
which add -F and -framework arguments.
Also fix incorrect de-duping of -framework arguments for gtkdoc.
Earlier, we would replace the subproject option with the parent
project's option, which is incorrect if the types are not the same.
Now we retain the subproject's option and print a warning. It's not
advisable to issue an error in this case because subproject option
yielding is involuntary for the parent project (option names can match
because of coincidences).
I believe the intent (from 30d0c2292f) is
that `[binaries]` isn't needed just for "target-only cross" (build ==
host != target). This fixes the code to match that, hopefully clarifying
the control flow in the process, and also improves the message to make
that clear.
Shared modules may be resource-only DLLs, or might automatically
self-initialize using C constructors or WinMain at DLL load time.
When an import library is not found for a shared module, just print
a message about it instead of erroring out.
Fixes#3965
* Lookup is now performed according to the following order:
1. use `pkg-config` with `PKG_CONFIG_LIBDIR=LIBPC` from python
2. use plain `PKG_CONFIG_{LIBDIR,PATH}` from the environment
3. try to extract the information from SYSCONFIG
The second step is necessary for relocated python installations
and cross compilation scenarios, where the value of `LIBPC` might
be wrong.
Otherwise we get an error while checking the subproject version:
Uncomparable version string 'none'.
If the dependency was found as a not-found dependency in the
subproject and is not required, just take it.
Needs a `mock` kwarg to Interpreter to not do any parsing of build
files, but only setup the builtins and functions.
Also consolidate the documentation and data tests into one class.
The code was adding the library paths to LD_LIBRARY_PATH, but that
doesn't work on Windows where they need to be added to PATH instead.
Move the environ handling into gtkdoc_run_check() and add paths to PATH
instead of LD_LIBRARY_PATH when on Windows.
This fixes the gtk-doc build for glib on Windows
(in case glib isn't installed already)
Use mesonlib.for_windows or mesonlib.for_cygwin instead of
reimplementing them.
Add CrossBuildInfo.get_host_system to shorten the repeated the code in
the mesonlib.for_<platform> methods.
If we aren't cross-building, this distinction isn't very interesting
(Most dependencies didn't bother reporting this. pkgconfig did it
unconditionally, and Qt reported as 'native' or 'cross' depending on
is_cross_build(), i.e. ignoring the native: keyword)
If successful, we should identify the method which was successful
If successful, we should report the version found (if known)
If failing, we should identify the methods we tried
Some dependency detectors which had no reporting now gain it
There's all kinds of complexities, inconsistencies and special cases hidden
in the existing behaviour, e.g.:
- boost reports modules requested, and BOOST_ROOT (if set)
- gtest/gmock report if they are a prebuilt library or header only
- mpi reports the language
- qt reports modules requested, and the config tool used or tried
- configtool reports the config tool used
- llvm reports if missing modules are optional (one per line)
We add some simple hooks to allow the dependency object to expose the
currently reported information into the consolidated reporting
Note that PkgConfigDependency() takes a silent: keyword which is used
internallly to suppress reporting. This behaviour isn't needed in
find_external_dependency().
Give ConfigToolDependency() a finish_init callback, so that tool-specific
initialization can be called from the constructor, rather than after
construction in the factory class.
v2:
finalize -> finish_init for clarity
find_external_dependency() now makes and iterates over a list of callables
which are constructors with bound arguments for the dependency objects we
are going to attempt to make, so we can consolidate reporting on these
attempts and handling failures in that function.
dub dependencies allegedly support method:pkgconfig, but this is meaningless
as it's only ever used when the dub method is explicitly requested.
Possibly cut-and-pasted from elsewhere?
Ninja buffers all commands and prints them only after they are
complete. Because of this, long-running commands such as `cargo
build` show no output at all and it's impossible to know if the
command is merely taking too long or is stuck somewhere.
To cater to such use-cases, Ninja has a 'pool' with depth 1 called
'console', and all processes in this pool have the following
properties:
1. stdout is connected to the program, so output can be seen in
real-time
2. The output of all other commands is buffered and displayed after
a command in this pool finishes running
3. Commands in this pool are executed serially (normal commands
continue to run in the background)
This feature is available since Ninja v1.5
https://ninja-build.org/manual.html#_the_literal_console_literal_pool
Meson tries to find the interpreter path through the "py" launcher on Windows in all
cases which breaks if meson is run under MSYS2 and an official CPython is installed as well.
MSYS2 Python doesn't install a py launcher which results in meson finding the system one instead
even though python2/python3 is in PATH.
Always check if the interpreter name is in PATH before falling back to checking the py launcher.