The 'Visual Studio 2017' image does not include VS2010, so we have to use
the 'Visual Studio 2015' image by default and only use the 2017 image for
msvc2017 compiler tests.
VS2017 requires the 'WindowsTargetPlatformVersion' property to be set.
We gather the version to use from the environment variable
'WindowsSDKVersion' that will be set by the VS developer command prompt.
VS2015 automatically picks up outputs from CustomBuild commands, which we
now use instead of CustomBuildStep commands, which do not get picked up
automatically.
This changes how generated files are added to the VS project.
Previously, they were all added as a single CustomBuildStep with all
generator commands, inputs and outputs merged together.
Now, each input file is added separately to the project and is given a
CustomBuild command. This adds all generator input files to the files list
in the VS gui and allows to run only some of the generator commands if
only some of the input files have changed.
Because we are using check_output, if the command fails no output will
be printed at all. So, we use subprocess.run instead.
Also, on configure failures, print the meson-log.txt instead of stdout.
Also forcibly undefine __has_include and test that the fallback include
check in cc.has_header() works.
This is important because all the latest compilers support it now
and we might have no test coverage at all by accident. GCC 5, ICC 17,
Clang 3.8, and VS2015 Update 2 already support it.
We differ from cmake in the following manner:
* We only set the major version (SOVERSION) in the dylib
* If SOVERSION is not specified, we deduce it from version (VERSION)
See installed_files.txt for a list of output dylib names.
In this test, we try to manually link against the generated library to
create an executable and then run it to verify that it works.
Also test for all possible library versioning in the versioning tests on
Windows. Even though they yield the same dll naming, we should still
test it.
We check for the existence of PDB files in the install script, so we
don't need to do all this mucking about here. That's more robust too
because we don't need to parse build arguments in buildtype=plain
and decide if the PDB file would be generated.
We automatically convert that to use sys.executable now which is
always available on all platforms (because we're running with it).
On some platforms like NetBSD, `python` doesn't exist, and you must
use a specific python version. On most other distros, `python` is
Python 2, and we don't want to depend on that.
Closes https://github.com/mesonbuild/meson/issues/695
All these scripts were being used as `find_program()`, so we do not
lose any test coverage by doing this.
There is no need to do obj.get_command() and in fact it's wrong
because the VS backends need to resolve each object to absolute paths
and get_command() does not do that.
This should fix invocation of GNOME module helpers with the VS backends
For the record, absolute paths for programs are needed because the
same PATH environment won't necessarily be available to Visual Studio
when it builds the generated solution.
Related to https://github.com/mesonbuild/meson/issues/1419
This used to produce a warning, but then would crash anyway. It's
simpler if we just error out and have the user disable gir generation or
install gobject-introspection.
Nicolas Schneider
Patrick Griffis
Matthias Klumpp
Tim-Philipp Müller
Jussi Pakkanen
Thibault Saunier
Nirbheek Chauhan
Haakon Sporsheim
Michal Sojka
Marc Becker
Elliott Sales de Andrade