This seems to be related to deleting the current working directory.
Simply deleting all of the trees inside the build directory instead
seems to fix it. This only appears with some combination of generated
targets, running the test case against say "1 trivial" doesn't show the
bug.
See this mesa bug: https://bugs.freedesktop.org/show_bug.cgi?id=109071
It is possible that the subproject has been downloaded already, in that
case there is no reason to not use it. If the subproject has not been
downlaoded already it will fail do_subproject().
This comes up now and again when people try do do something like:
meson.build:
```meson
my_sources = ['foo.c']
subdir('subdir')
executable('foo', my_sources)
```
subdir/meson.build:
```meson
my_sources += ['bar.c']
```
The code our projects care about verifying coverage for mostly lives in
the source_root with the exception of the generated source files in
build_root. This change cleans up the output so we don't have prefixed
paths on our source files anymore.
This doesn't touch everything as it's just based on the python3 module
tests, ported to the python module. It's still better than the one very
basic test in the unit test module.
This would normally be exposed by the ExternalProgramHolder, but wasn't
due to the implementation of the PythonInstallation module. Because of
that I've duplicated the method so that we can add the FeatureNew
decorator.
Fixes#4070
Into a generic interpreter object. This isn't a module, it's an object
returned by a module, it also happens to be a special case of an
ExternalProgram, which is a normal interpreter object. Let's treat it
like one.
Since `_process_libs` appends the lib's dependencies this list already,
the final return value of `_process_libs` will end up after its
dependencies, which is the wrong way around. (The lib must come first,
then its dependencies)
The easiest solution is to simply pre-pend the return value of
`_process_libs` rather than appending it, so that its dependencies come
after the library itself.
Closes#4091.
Windows requires things to be linked, on macOS distutils doesn't link by default.
On Linux etc. things are not so clear, some distros like Debian patch distutils to not link,
some don't. In addition the manylinux wheels spec prohibits linking against libpython
and upstream is thinking about changing the default:
https://bugs.python.org/issue34814
Call into distutils to figure out what distutils does and in case it doesn't link
against libpython replace the passed in Python dependency with a partial one.
Daniel Mensinger
Jussi Pakkanen
Xavier Claessens
Nirbheek Chauhan
lzutao
John Ericson
Dylan Baker
Alistair Thomas
John Ericson
Konstantin Kharlamov
Thibault Saunier
William A. Kennington III
Luca Weiss
Niklas Haas
Christoph Reiter