This is another toolchain also called `armclang`, but it is not a cross
compiler like Keil's `armclang`. It is essentially the same as `clang`
based on its interface and CMake's support of the toolchain.
Use an `armltd` prefix for the compiler ID.
Fixes: #7255
Removed errant "type: ignore".
Fixed issue with "fetch" call. This issue was the following:
Dict::get() and Dict::pop() have the following signature:
T.Callable[[_T, _U], _U | None] OR T.Callable[[_T], _U | None]
Note how the return type is _U here. When the fetch() function was
actually being called, it had the following signature:
T.Callable[[_T, T.List[_U]], T.Union[T.List[_U], _U]]
This is incompatible with the previous definitions. The solution is
simply to move where the default value is introduced if fetch() produces
None.
If the LTO threads == 0 clang will default to the same argument we
manually pass, which meant we dropped support for admittedly ancient
versions of clang that didn't yet add that option.
Drop the extraneous argument, and add a specific error condition when
too old versions of clang are detected.
Fixes#9569
Clippy is a compiler wrapper for rust that provides an extra layer of
linting. It's quite popular, but unfortunately doesn't provide the
output of the compiler that it's wrapping in it's output, so we don't
detect that clippy is rustc. This small patch adds a new compiler class
(that is the Rustc class with a different id) and the necessary logic to
detect that clippy is in fact rustc)
Fixes: #8767
Currently this implements 3 warning levels, 1 and 2 are just the
"default" set by rustc, 3, is "everything is a warning", and 0 is
"nothign is a warning".
This should be done in all cases of language_stdlib_only_link_flags, but
I don't have access to all of the compilers to test it.
This is required in cases where object files created by gfortran are
linked using another compiler with a differen default search path, such
as gfortran and clang together.
This patch adds a new meson built-in option for cython, allowing it to
target C++ instead of C as the intermediate language. This can, of
course, be done on a per-target basis using the `override_options`
keyword argument, or for the entire project in the project function.
There are some things in this patch that are less than ideal. One of
them is that we have to add compilers in the build layer, but there
isn't a better place to do it because of per target override_options.
There's also some design differences between Meson and setuptools, in
that Meson only allows options on a per-target rather than a per-file
granularity.
Fixes#9015
I ran into one of these from LGTM, and it would be nice if pylint could
warn me as part of my local development process instead of waiting for
the CI to tell me.
-Oz is the appropriate flag to use when you want to produce the smallest
possible binary, and is one would expect when setting optimization to s
or using the minsize build type.
It is imported from a subpackage in __init__ alongside a big list of
other things which are all exported. And elsewhere, this import is
re-imported by other code. It's pretty clearly an oversight that it
didn't get added to __all__
The clang compiler now reimplements and re-checks the c_std and cpp_std
options in order to use them for objc as well, but it didn't
consistently support the same options. First it completely excluded all
the gnu ones, and then it added a handful of them but not for C++.
Be fully consistent -- or at least as consistent as we can be, given a
minimally working fix. (The C/C++ compiler mixin actually gates
different stds depending on detected clang version, we do not do that
here.)
Fixes regression in c54dd63547
Fixes incomplete fix from #8766 (which didn't fix objcpp at all)
Fixes#9237
Allow using the links method to test that the C++ driver (e.g. g++) can be used to
link C objects. One usecase is that the C compiler's libsanitizer might not be
compatible with the one included by the C++ driver.
This is theoretically backwards-incompatible, but it should be treated as a
bugfix in my opinion. There is no way in Meson to compile a .c file with the
C++ driver as part of a build target, therefore there would be no reason to
do something like meson.get_compiler(meson.get_compiler('cpp').links(files('main.c')).
Fixes: #7703
In some cases, link tests would like to use objects provided by a compiler
for a different language, for example linking a C object file with a C++
compiler. This kind of scenario is what link_language is for, but it is
impossible to test that it works with a linker test.
This patch implements the required support in the Compiler class. The
source code compiler is passed to the Compiler.links method as an
argument.
* compilers: improve docstring to `get_compiler_check_args()`
There was an incomplete list, which wasn't useful as it now takes an
enum anyway. Also add a new entry to the list of reasons to use this
function.
* clang: Add -Werror=implicit-function-declarations to check_args
Unlike GCC, clang warns but doesn't error when an implicit function
declaration happens. This means in checks like
`compiler.has_header_symbol('string.h', 'strlcat')` (on Linux, at least)
that GCC will fail, as there is no such function; clang will emit a
warning, but since it exists with a 0 status Meson interprets that as
success. To fix this, add `-Werror=implicit-function-declarations` to
clang's check arguments.
There seems to be something specific about functions that _may_ exist in
a header on a given system, as `cc.has_header_symbol('string.h',
'foobar')` will return false with clang, but `strlcat` will return true,
even though it's not defined. It is however, defined in some OSes, like
Solaris and the BSDs.
Fixes#9140
We have a lot of these. Some of them are harmless, if unidiomatic, such
as `if (condition)`, others are potentially dangerous `assert(...)`, as
`assert(condtion)` works as expected, but `assert(condition, message)`
will result in an assertion that never triggers, as what you're actually
asserting is `bool(tuple[2])`, which will always be true.
There are two changes here, one is to remove an `elif` that is
effectively an `else`, that helps the type checker and provides a small
speedup potentially. The second is a potentially unbound variable, that
currently isn't hit, but very much could be.
When mutable items are stored in an lru cache, changing the returned
items changes the cached items as well. Therefore we want to ensure that
we're not mutating them. Using the ImmutableListProtocol allows mypy to
find mutations and reject them. This doesn't solve the problem of
mutable values inside the values, so you could have to do things like:
```python
ImmutableListProtocol[ImmutableListProtocol[str]]
```
or equally hacky. It can also be used for input types and acts a bit
like C's const:
```python
def foo(arg: ImmutableListProtocol[str]) -> T.List[str]:
arg[1] = 'foo' # works while running, but mypy errors
```
* Allow address sanitizer for Visual Studio 2019 version 16.9
Address Sanitizer was first supported with the current syntax in Visual
Studio 16.9.0 (cl version 19.28.29910).
* VS: Convert /fsanitize=address to project file setting
This gets rid of compile warnings, and simplifies the code.
Note that `work_dir` in sanity_check_impl was incorrect,
it was used both to prepend to file names and as cwd=work_dir
argument to Popen. This is fixed here.
Closes gh-7344
They are supposed to have different behavior. The environment variables
apply to both the compiler and linker when the compiler acts as a
linker, but the command line ones do not.
Fixes#8345
Dependencies is already a large and complicated package without adding
programs to the list. This also allows us to untangle a bit of spaghetti
that we have.
This requires quite a complex and messy logic.
As @dcbaker suggested in #8491, this could be replaced by
an abstraction over linker flags instead of having GNU flags
translated.
All changes were created by running
"pyupgrade --py3-only --keep-percent-format"
and committing the results. I have not touched string formatting for
now.
- use set literals
- simplify .format() parameter naming
- remove __future__
- remove default "r" mode for open()
- use OSError rather than compatibility aliases
- remove stray parentheses in function(generator) scopes
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.
Currently we don't handle things correctly if we get a string we should
split, and the linker and needs compiler arguments. It would result in
two unsplit strings in a list, instead of the split arguments in a list
Fixes: #8348
Clang has a hand `-Wunused-command-line-argument` switch, which when
turned to an error, gets very grump about `-flto-jobs=0` being set in
the compiler arguments (although `-flto=` belongs there). We'll refactor
a bit to put that only in the link arguments.
GCC doesn't have this probably because, a) it doesn't have an equivalent
warning, and b) it uses `-flto=<$numthreads.
Fixes: #8347
Both Clang and GCC support using multiple threads for preforming link
time optimizaions, and they can now be configured using the
`-Db_lto_threads` option.
Fixes#7820
Currently mesonlib does some import tricks to figure out whether it
needs to use windows or posix specific functions. This is a little
hacky, but works fine. However, the way the typing stubs are implemented
for the msvcrt and fnctl modules will cause mypy to fail on the other
platform, since the functions are not implemented.
To aleviate this (and for slightly cleaner design), I've split mesonlib
into a pacakge with three modules. A universal module contains all of
the platform agnositc code, a win32 module contains window specific
code, a posix module contains the posix specific code, and a platform
module contains no-op implementations. Then the package's __init__ file
imports all of the universal functions and all of the functions from the
approriate platform module, or the no-op versions as fallbacks. This
makes mypy happy, and avoids `if`ing all over the code to switch between
the platform specific code.
This commit performs some cleanup for the msvc and clang-cl arguments.
* "Disable Debug" (`/Od`) is no longer manually specified for optimization levels {`0`,`g`} (it is already the default for MSVC).
* "Run Time Checking" (`/RTC1`) removed from `debug` buildtype by default
* Clang-CL `debug` buildtype arguments now match MSVC arguments
* There is now no difference between `buildtype` flags and `debug` + `optimization` flags
Compiler tests, such as checking for atomics support, could fail
when compiling to WebAssembly multithreaded targets because the
compiler tests got compiled to 'output.wasm'.
Using the '.wasm' suffix in recent versions of emscripten engages
STANDALONE_WASM mode, which disables features that require a JS
runtime like shared memory.
This created false negatives on support of those features when
building a library to be linked into an executable that is not
in STANDALONE_WASM mode.
Changing these to 'output.o' will continue to produce WebAssembly
object files, but they will no longer be configured for standalone
runtime mode.
The /ZI flag adds in "Edit and Continue" debug information, which will
cause massive slowdown. It is not a flag that we should be adding by
default to debug builds.
/Zi will still be added.
This has a bunch of nice features. It obviously centralizes everything,
which is nice. It also means that env is only re-read at `meson --wipe`,
not `meson --reconfigure`. And it's going to allow more cleanups.
This is PEP8 convention for a const variable. Also, make the type
Mapping, which doesn't have mutation methods. This means mypy will warn
us if someone tries to change this.
This patches takes the options work to it's logical conclusion: A single
flat dictionary of OptionKey: UserOptions. This allows us to simplify a
large number of cases, as we don't need to check if an option is in this
dict or that one (or any of 5 or 6, actually).
I would have prefered to do these seperatately, but they are combined in
some cases, so it was much easier to convert them together.
this eliminates the builtins_per_machine dict, as it's duplicated with
the OptionKey's machine parameter.
Adds TemporaryDirectoryWinProof which calls windows_proof_rmtree() on
error.
Use instead of hacky error handling (which might shadow other OSError)
in Compiler.compile().
Since the current approach of demoting to the nearest C standard *might*
work, but might not. For projects like Glib that detect which standard
is used and fall back this is fine. For projects like libdrm that only
work with gnu standards, this wont. We're nog tusing a warning because
this shouldn't be fatal if --meson-fatal-warnings is used. Also demote a
similar message in IntelCl from warning to log.
And then update the choices in each leaf class. This way we don't end up
with another case where we implicitly allow an invalid standard to be
set on a compiler that doesn't have a 'std' setting currently.
As far as I can Tell, rust just handles this for us (it's always worked
with no special arguments from us). However, since we're going to add
support for base options for rust, we need to add the method.
When TemporaryDirectory() cleans up on __exit__ it sometimes throws
OSError noting that the dir isn't empty. This happens after the
first yield in this generator and leads to the exception being handled
which leads to a second yield.
contextlib.contextmanager() fails then since the function it wraps is only
allowed to yield once.
Fix this by not yielding again in the error case.
Fixes#7947
Using the std option, so now `rust_std=..` will work. I've chosen to use
"std" even though rust calls these "editions", as meson refers to
language versions as "standards", which makes meson feel more uniform,
and be less surprising.
Fixes: #5100
It is much easier to not try to parse options into complicated
structures until we actually collected all options: machine files,
command line, project()'s default_options, environment.
Arsen Arsenović