- ModuleState is now a real class that will have methods in the future
for actions modules needs, instead of using interpreter internal API.
- New ModuleObject base class, similar to InterpreterObject, that should
be used by all objects returned by modules. Its methods gets the
ModuleState passed as first argument. It has a `methods` dictionary to
define what is public API that can be called from build definition.
- Method return value is not required to be a ModuleReturnValue any
more, it can be any type that interpreter can holderify, including
ModuleObject.
- Legacy module API is maintained until we port all modules.
In the future modules should be updated:
- Use methods dict.
- Remove snippets.
- Custom objects returned by modules should all be subclass of
ModuleObject to get the state iface in their methods.
- Modules should never call into interpreter directly and instead state
object should have wrapper API.
- Stop using ModuleReturnValue in methods that just return simple
objects like strings. Possibly remove ModuleReturnValue completely
since all objects that needs to be processed by interpreter (e.g.
CustomTarget) should be created through ModuleState API.
Reusing the runners for multiple repeats of the test run gets in the
way of the progress report, which stores runners in an OrderedSet.
Instead, create a separate SingleTestRunner object for each repeat.
While at it, fix the "duplicate suite" assertion as it can fire
with TAP tests and --repeat=N.
Fixes: #8405
Rather than having to manually build the locale aware man paths with
`install_data('foo.fr.1', install_dir: join_paths(get_option('mandir'), 'fr', 'man1'), rename: 'foo.1')`
Support doing
`install_man('foo.fr.1', locale: 'fr')`
hdf5's config-tools will not show compile arguments (including the
include directory) if called without `-c`. Make sure to get both the
compile and link arguments.
When building on Linux, I see:
rustc -C linker=cc --color=always --crate-type rlib --crate-name rs_math -g --emit dep-info=rs_math.d --emit link -L /usr/lib/x86_64-linux-gnu -o librs_math.rlib -l static=m ../rs_math.rs
error: failed to add native library /usr/lib/x86_64-linux-gnu/libm.a: file too small to be an archive
I think the "file too small to be an archive" message is coming from
libLLVM, and is a case of LLVM not handling this type of "script
archive". So, possibly this is just LLVM not handling a linking case.
The rust_args usage in meson.build is invalid, but required to
reproduce the issue in the test case. Perhaps meson should
automatically add the library include path via the dep_m meson object,
or maybe the meson.build can add the link path in a better way.
Changing '-l static=m' to '-l dylib=m' appears to fix this case. (See
comments in meson.build.)
The way the tracking is currently done it works if no new subprojects
are added to a configured build directory. For cases where we want to
add a new subproject, it fails because we don't initialize builtins for
that subproject. This corrects that by checking to see if the subproject
already exists, and if it doesn't initializes the bultins for it.
Fixes: #8421
Useful in case of boolean values to distinguish between a boolean
value having been set in the native/cross file and not having been
provided, which can't be achieved by passing a fallback parameter
to .get_external_property().
It's a pure subset of `get_external_property`, and has odd behavior in
host == build configurations. `get_external_property` is clear, and uses
the standard `native : bool` syntax to control host vs build properties
meson compile itself doesn't permit GNU-style argument permutation, i.e.
TARGET to precede options, so why should we expect ninja to?
And indeed, ninja doesn't document support for this -- but it does
accept it anyway, which is confusing and results in people thinking it's
"supposed to" work.
However, if NINJA=samu then this is in fact enforced. samu does not
permit GNU-style argument permutation. As a result, the arguments passed
to mcompile are actively re-ordered before being passed to the
subprocess, and samu dies with a fatal error.
Fix ordering in mcompile.py and add a comment to warn future readers
that the order does, in fact, matter.
Reworks:
* Add MESON_SKIP_TEST error for darwin from polyglot test
* Remove dependencies from executable
Dylan's changes:
* Add support for both static and shared testing
* Test with pkg-config, cmake, and system
* reformat to meson's style
Co-authored-by: Dylan Baker <dylan@pnwbakers.com>
As a Meson developer it's often frustrating to have a single functional
test with a regression. These tests can be awkward to reproduce,
especially when they make use of a test.json file. This script provides
a simmple interface to call functional tests 1 at a time, regardless of
whether they use a test.json or not. If they do use a test.json, and
have a matrix, then the `--subtest` option can be used to select spcific
combinations, for example:
```sh
./run_single_test.py "test cases/frameworks/15 llvm" --subtest 1
```
will run only the second (zero indexed of course) subtest from the llvm
test cases.
This is not a super elegent script, but this is super useful.
Enables -Db_sanitize=undefined and company.
Also serves as a testcase for NVCC comma-shielding: Because the test-
case declares `b_sanitize=address,undefined`, the host GCC compiler
needs `-fsanitize=address,undefined`, but this stands a danger of being
split by NVCC when wrapped with `-Xcompiler=args,args`. Special,
already-existing comma-shielding codepaths activate to prevent this
splitting.
Closes#8394.
By default all subprojects are installed. If --skip-subprojects is given
with no value only the main project is installed. If --skip-subprojects
is given with a value, it should be a coma separated list of subprojects
to skip and all others will be installed.
Fixes: #2550.
Paolo Bonzini
Xavier Claessens
Remi Thebault
Leif Middelschulte
Dylan Baker
Jussi Pakkanen
Jason Woodward
Jeff Moguillansky
Michael Kuhn
Jordan Justen
Stephen Gregoratto
Tim-Philipp Müller
Riccardo Schirone
Matthias Klumpp
Eli Schwartz
Olexa Bilaniuk
Patryk Obara