Previously, when mixing Zig and C/C++ code for windows-gnu targets, zig
would get codeview format but the C/C++ code would not get any debug
info. Now, C/C++ code properly emits debug info in codeview format and
everything just works.
This then allows for proper resolution of names via runpath search
path list, i.e., `-rpath @loader_path` will correctly resolve
to `@rpath/libxxx.dylib (...)` in the linked binary.
This commit version-gates appending `-syslibroot` flag to lld. This
is predicated upon the fact that for versions of macOS lower than
11, lld would fail to find and link against frameworks with this
flag specified.
Co-authored-by: Andrew Kelley <andrew@ziglang.org>
* Fix floating point parsing on BE systems
* Load the appropriate endian.h files for macOS and BSD
* Add endian definition for Windows and extra check for ldshape selection
* Fix endian macro definition for macOS
Apparently their macros are defined without a leading __.
* Define new macro for endian checking purposes
This is gross and I really do not like the lack of standardization
around this part, but what can I do?
Applications supplying their own custom stack to pthread_create are not
allowed to free the allocated memory after pthread_join returns as,
according to the specification, the thread is not guaranteed to be dead
after the join call returns.
Avoid this class of problems by avoiding the use of a custom stack
altogether, let pthread handle its own resources.
Allocations made on the child stack are now done on the C heap.
Thanks @semarie for noticing the problem on OpenBSD and suggesting a
fix.
Closes#7275
If we enter the `if` because `comp.disable_c_depfile` is false the `man`
object has no manifest and calling .final on it will trip an assertion.
Closes#7096
commit 9d1816111d1d30e18b8cb43a4aa31c194fb204c4 used the "output path"
as the path for passing shared library artifact paths to the Zig CLI.
For Windows, this was incorrect because it would pass the .dll instead
of the .lib file. This commit passes the "output lib path" instead,
which makes it pass the .lib path in case of a .dll on Windows.
This way the linker does not complain and say, "bad file type. Did you
specify a DLL instead of an import library?"
See #5854
Some tiny tweaks too:
* Use `wasm-freestanding-musl` instead of `wasm32-freestanding-musl`,
making it pointer-size-agnostic.
* Fix trying to build non-existent wasm musl start files.
Positional shared library arguments were not being detected as causing
dynamic linking, resulting in invalid linker lines. LLD did not have an
error message for this when targeting x86_64-linux but it did emit an
error message when targeting aarch64-linux, which is how I noticed the
problem.
This surfaced an error having to do with fifo.pipe() in the cat example
which I did not diagnose but solved the issue by doing the revamp that
was already overdue for that example.
It appears that the zig-window project was exploiting the previous
behavior for it to function properly, so this prompts the question, is
there some kind of static/dynamic executable hybrid that the compiler
should recognize? Unclear - but we can discuss that in #7240.
ZigTypeIdOptional types have a different way of specifying their payload
value depending on whether the child type is a pointer or not (plus some
other special cases).
Fixes#7251
stderr_mutex may still be held when we reach the signal handler, grab
our own stderr handle to print the error messages and avoid deadlocking.
Closes#7247
Make gen_const_ptr_struct_recursive aware of the possible presence of
some trailing padding by always bitcasting the pointer to its expected
type.
Not an elegant solution but makes LLVM happy and is consistent with how
the other callsites are handling this case.
Fixes#5398
Recent versions of cmake complain about it:
<<
CMake Deprecation Warning at CMakeLists.txt:1 (cmake_minimum_required):
Compatibility with CMake < 2.8.12 will be removed from a future version of
CMake.
>>
We don't require anything from version 2.8.5, and version 2.8.12 was
released in 2011, so it is very unlikely that anyone still uses 2.8.5.
