This fixes issues where a new option is added, an option is removed, the
constraints of an option are changed, an option file is added where one
didn't previously exist, an option file is deleted, or it is renamed
between meson_options.txt and meson.options
There is one case that is known to not work, but it's probably a less
common case, which is setting options for an unconfigured subproject.
We could probably make that work in some cases, but I don't think it
makes sense to download a wrap during meson configure.
When we load the option file in the interpreter record which file it
was, and what the hash of that file was. This will let `meson configure`
know that the options have changed since the last re-configure.
This demonstrates a limitation of Meson that can happen in the following
situation:
`meson setup builddir`
<update to meson.options> (such as git pull)
`meson configure -Dnew_option=value builddir`
<Error>
Since Meson is unaware of the updates to the meson.options file. The
purposes of these tests is to fix this behavior.
As far as we can ascertain, the idea here was supposed to be that this version
at the bottom gets updated as a last-resort for development LLVMs on Debian.
Signed-off-by: Sam James <sam@gentoo.org>
Signed-off-by: Eli Schwartz <eschwartz93@gmail.com>
The test.json format currently has three keys related to skipping tests:
* `skip_on_jobname`
* `skip_on_os`
* `skip_on_env`
While `skip_on_env` marks the test itself as skipped, i.e. they don't get run when the
conditions are met, the other two skip options are just marking the test as "expected to be skipped"
if the conditions apply, i.e. they want to see `MESON_SKIP_TEST` in the output and things will
fail if that doesn't happen. They don't actually skip the tests as the names imply.
To make this clearer rename the keys:
* `skip_on_jobname` -> `expect_skip_on_jobname`
* `skip_on_os` -> `expect_skip_on_os`
`skip_on_env` stays the same, since that actually skips.
The docs were also confused about this, so adjust that too.
On Arch's shiny new 48-core/96-thread build server, the
`test_install_log_content` test fails because of an unexpected
`invalid-symlink.txt` file. Apparently the test runs in parallel with
`test_install_subdir_symlinks`, which modifies the `59 install subdir`
source directory.
To fix this, make `install_subdir_invalid_symlinks` copy the entire test
into a tmpdir before modifying it.
By adding the option name to UserOption object, it is now possible to
display the name of the affected option when the given option value is
not valid.
Fixes#12635
Otherwise internal dependencies have auto-generated names that are not
human readable. Instead, use the name that the dependency overrides. For
example:
```meson
meson.override_dependency('zlib', declare_dependency())
dep_zlib = dependency('zlib')
assert(dep_zlib.name() == 'zlib')
```
Fixes: #12967
When rustc cannot run host binaries it does an early return which
skipped getting native-static-libs. Move that code earlier to always run
it.
While at it, failing to determine those libs is a fatal error. We would
crash later when trying to access rustc.native_static_libs attribute
otherwise.
Places where compiler needs it already have access to Environment object
and can use it directly.
This fixes mypy complaining that not all compilers have self.exe_wrapper
in run() method that got moved to base class.
When doing a compiler check with a File object, there is no need to dump
the file into logs, we can just print the path. The file could
potentially be big.
The previous approach wasn't great because you couldn't control what the
detected C++ stdlib impl was. We just had a preference list we tweaked the
searched order for per OS. That doesn't work great for e.g. Gentoo with libc++
or Gentoo Prefix on macOS where we might be using libstdc++ even though the host
is libc++.
Jonathan Wakely, the libstdc++ maintainer, gave a helpful answer on how
to best detect libc++ vs libstdc++ via macros on SO [0]. Implement it.
TL;DR: Use <version> from C++20 if we can, use <ciso646> otherwise. Check for
_LIBCPP_VERSION as libstdc++ doesn't always define a macro.
[0] https://stackoverflow.com/a/31658120
Signed-off-by: Sam James <sam@gentoo.org>
Signed-off-by: Eli Schwartz <eschwartz93@gmail.com>
Version 1.9 removes support for python 3.7, so we either need to pin the
version to 1.8 as long as we're support Python 3.7, or we need to drop
linting for 3.7
The cuda compiler also executes the host compiler and generally needs to know
the host compiler flags. We did this for regular args but not for error args.
We use `-Xcompiler=` since that's how nvcc tells the difference between args
it uses itself for device code and args it passes to the host compiler.
Signed-off-by: Eli Schwartz <eschwartz93@gmail.com>
[eli: fix quoting style]
When linking with a Rust rlib, we should also link with its external
system dependencies. This was currently done only for C ABI crates, do
it for both rlib and staticlib now.
musl 1.2.5 exposes this symbol in the default profile, and in future
will expose it unconditionally, as it is on track to becoming part of
POSIX.
So rather than maintaining a list of systems to run this on, let's
just skip testing the feature test macro if we find ppoll in the
default profile.
When only a single input file shows up in an arglist, it makes no sense
to inject `-W,--start-group -lone -Wl,--end-group`, since there is
nothing being grouped together. It's just longer command lines for
nothing.
This is just a bad regex match, because it matches *.so as a prospective
filename input even though it is actually not an input but a linker
flag. --start-group is only relevant to input files...
For instance, on Windows, if the terminal is in cp65001, the subprocess
output is not encoded correctly and it results in error when running
unit test.
When the proc-macro rlib is in a different subdir, it would miss the
needed -L argument and rustc would not find it. Meson was assuming that
proc-macros are only needed when building libraries that uses it, but it
turns out that was wrong, as show by the unit test.
* `-Wpedantic` creates useless churn due to its use of gcc-line directives:
../foo.cu:1:3: warning: style of line directive is a GCC extension
1 | namespace Foo {
| ^~
https://stackoverflow.com/a/31001220
Cargo implements this configuration option to override the entry point
of the library.
To get test coverage, I've modified one of the two Cargo subprojects
in the test to use a non-default library entrypoint. The other still
uses the default.
Dylan Baker
Stephan Lachnit
Sam James
Christoph Reiter
Jan Alexander Steffens (heftig)
Charles Brunet
Xavier Claessens
Yegor Yefremov
Jussi Pakkanen
Petr Machacek
Elliot
David Seifert
Alyssa Ross
LIU Hao
Eli Schwartz