Forcing serialization on when writing out the build rule makes very
little sense. It was always "forced" on because we mandated a couple of
environment variables due to legacy reasons.
Add an attribute to RunTarget to say that a given target doesn't *need*
those environment variables, and let ninja optimize them away and run
the command directly if set.
That method had nothing specific to the backend, it's purely a Target
method. This allows to cache the OptionOverrideProxy object on the
Target instance instead of creating a new one for each option lookup.
There is no need to go through all sources again, we already did that to
populate self.compilers. When cs or java compilers are in the list, then
there must be only one compiler.
The code was also not considering generate sources any way.
Previously subprojects inherited languages already added by main
project, or any previous subproject. This change to have a list of
compilers per interpreters, which means that if a subproject does not
add 'c' language it won't be able to compile .c files any more, even if
main project added the 'c' language.
This delays processing list of compilers until the interpreter adds the
BuildTarget into its list of targets. That way the interpreter can add
missing languages instead of duplicating that logic into BuildTarget for
the cython case.
We currently don't handle subdirectories correctly in
structured_sources, which is problematic. To make this easier to handle
correctly, I've simply changed `structured_sources` to only use Files
and not strings as an implementation detail.
Using future annotations, type annotations become strings at runtime and
don't impact performance. This is not possible to do with T.cast though,
because it is a function argument instead of an annotation.
Quote the type argument everywhere in order to have the same effect as
future annotations. This also allows linters to better detect in some
cases that a given import is typing-only.
A backslash-escape of the last newline before a run of whitespace
leading to the indented string ending and function termination `)` does
not actually escape the entire line and make it do nothing. In fact what
it does is cause all that whitespace to be part of the preceding line,
and get printed.
Meanwhile the textwrap.dedent documentation states that lines with only
whitespace get normalized. When you *don't* mess with that final line,
dedent actually does the right thing and makes the output message end
with a single newline after the important text.
Android requires shared modules that use symbols from other shared
modules to be linked before they can be dlopen()ed in the correct
order. Not doing so leads to a missing symbol error:
https://github.com/android/ndk/issues/201
We need to always allow linking for this. Also add a soname, although
it's not confirmed that it's needed, and it doesn't really hurt if it
isn't needed.
Use a derived type when passing `subproject` around, so that mypy knows
it's actually a SubProject, not a str. This means that passing anything
other than a handle to the interpreter state's subproject attribute
becomes a type violation, specifically when the order of the *four*
different str arguments is typoed.
Because we don't want to pass the Interpreter kwargs into the build
layer. This turned out to be a mega commit, as there's really on elegant
way to make this change in an incremental way. On the nice side, mypy
made this change super easy, as nearly all of the calls to
`CustomTarget` are fully type checked!
It also turns out that we're not handling install_tags in custom_target
correctly, since we're not converting the boolean values into Optional
values!
In some cases, init variables that accept None as a sentinel and
immediately overwrite with [], are migrated to dataclass field
factories. \o/
Note: dataclasses by default cannot provide eq methods, as they then
become unhashable. In the future we may wish to opt into declaring them
frozen, instead/additionally.
When generating pkgconfig file for a library that links to an
uninstalled static library built by custom_target() Meson was crashing
when trying to access some attributes that does not exist on that class.
Also fix is_internal() implementation, it only really make sense on a
CustomTargetIndex or if CustomTarget has only a single output.
FeatureCheck always immediately sets extra_message to '' if it isn't
explicitly passed, so there is really no point in using None as a
sentinel that is never used.
Names used in init functions are sometimes pointlessly different from
the class instance attributes they are immediately assigned to. They
would make more sense if defined properly.
Emit a detailed deprecation warning that explains what to do instead.
Also add a unittest.
```
DEPRECATION: target prog links against shared module mymod, which is incorrect.
This will be an error in the future, so please use shared_library() for mymod instead.
If shared_module() was used for mymod because it has references to undefined symbols,
use shared_libary() with `override_options: ['b_lundef=false']` instead.
```
Fixes https://github.com/mesonbuild/meson/issues/9492
It works fine as-is with an empty list, and since that's easier to get
using our typed_kwargs, and thus is what we're passing, go ahead and
simplify the class to only take a list of strings.
It's supposed to emit an error message, but instead it did a traceback.
It used to be, if no install_dir was specified then it was simply not in
kwargs, but due to typed_kwargs it will now be there, but not have
viable contents, so the dict membership check got skipped.
Fixes#9522
Eli Schwartz
Xavier Claessens
Tristan Partin
Dylan Baker
Remi Thebault
Nirbheek Chauhan
William Toohey
Daniel Mensinger
Ben Boeckel
Paolo Bonzini
Pablo Correa Gómez
Mike Gilbert