It is the job of codegen backends to mark Decls that are referenced as
alive so that the frontend does not sweep them with the garbage. This
commit unifies the code between the backends with an added method on
Decl.
The implementation is more complete than before, switching on the Decl
val tag and recursing into sub-values.
As a result, two more array tests are passing.
also use the common naming convention for glibc versions ("2.33" rather
than "2-33").
I also verified that these files are exactly identical to the previous
files from before zig updated to glibc 2.34.
__libc_start_main() from glibc.2.33.so or older needs to have a __libc_csu_init function callback parameter.
glibc-2.34 on the other hand has a different __libc_start_main() that does not use it,
and the start.S file from glibc-2.34 no longer construct the init function and pass null when calling __libc_start_main.
So, When targetting an older glibc, use the start.s files as they were in glibc-2.33 and construct the __libc_csu_init function.
fixes#10386#10512
Introduce `validate_array_init_comptime`, similar to
`validate_struct_init_comptime` introduced in
713d2a9b3883942491b40738245232680877cc66.
`zirValidateArrayInit` is improved to detect comptime array literals and
emit AIR accordingly. This code is very similar to the changes
introduced in that same commit for `zirValidateStructInit`.
The C backend needed some improvements to continue passing the same set
of tests:
* `resolveInst` for arrays now will add a local `static const` with the
array value and so then `elem_val` instructions reference that local.
It memoizes accesses using `value_map`, which is changed to use
`Air.Inst.Ref` as the key rather than `Air.Inst.Index`.
* This required a mechanism for writing to a "header" which is lines
that appear at the beginning of a function body, before everything
else.
* dbg_stmt output comments rather than `#line` directives.
TODO comment reproduced here:
We need to re-evaluate whether to emit these or not. If we naively emit
these directives, the output file will report bogus line numbers because
every newline after the #line directive adds one to the line.
We also don't print the filename yet, so the output is strictly unhelpful.
If we wanted to go this route, we would need to go all the way and not output
newlines until the next dbg_stmt occurs.
Perhaps an additional compilation option is in order?
`Value.elemValue` is improved to support `elem_ptr` values.
Handle `__DATA,.rustc` section containing `rustc` metadata - this
is required to get crates like `serde_derive` link properly.
Note to self: this special section has to be copied __verbatim__
from the relocatable object file - this includes preserving its size
even though unpadded according the section's required alignment.
AIR:
* `array_elem_val` is now allowed to be used with a vector as the array
type.
* New instructions: splat, vector_init
AstGen:
* The splat ZIR instruction uses coerced_ty for the ResultLoc, avoiding
an unnecessary `as` instruction, since the coercion will be performed
in Sema.
* Builtins that accept vectors now ignore the type parameter. Comment
from this commit reproduced here:
The accepted proposal #6835 tells us to remove the type parameter from
these builtins. To stay source-compatible with stage1, we still observe
the parameter here, but we do not encode it into the ZIR. To implement
this proposal in stage2, only AstGen code will need to be changed.
Sema:
* `clz` and `ctz` ZIR instructions are now handled by the same function
which accept AIR tag and comptime eval function pointer to
differentiate.
* `@typeInfo` for vectors is implemented.
* `@splat` is implemented. It takes advantage of `Value.Tag.repeated` 😎
* `elemValue` is implemented for vectors, when the index is a scalar.
Handling a vector index is still TODO.
* Element-wise coercion is implemented for vectors. It could probably
be optimized a bit, but it is at least complete & correct.
* `Type.intInfo` supports vectors, returning int info for the element.
* `Value.ctz` initial implementation. Needs work.
* `Value.eql` is implemented for arrays and vectors.
LLVM backend:
* Implement vector support when lowering `array_elem_val`.
* Implement vector support when lowering `ctz` and `clz`.
* Implement `splat` and `vector_init`.
By placing the stack at the start of the memory section, we prevent the runtime
from silently overwriting the global declarations and instead trap.
We do however, allow users to overwrite this behavior by setting the global-base,
which puts the stack at the end of the memory section and the static data at the base that was specified.
The reason a user would want to do this, is when they are sure the stack will not overflow and they want
to decrease the binary size as the offsets to the static memory are generally smaller.
(Having the stack in front, means that accessing the memory after the stack has a bigger offset when loading/storing from memory).
This allows users to add file paths to device paths, which is often used
in methods like `boot_services.loadImage` and `boot_services.startImage`,
which take a device path with an additional file path appended to locate
the image.
UEFI uses `\` for paths exclusively. This changes std.fs.path to use `\`
for UEFI path joining. Also adds a few tests regarding it, specifically
in making sure double-separators do not result from path joining, as the
UEFI spec says to convert any that result from joining into single
separators (UEFI Spec Version 2.7, pg. 448).
Prior to this change, even if the use specified the sysroot on the
compiler line like so
```
zig cc --sysroot=/path/to/sdk
```
it would only be used as a prefix to include paths and not as a prefix
for `zig ld` linker.
- Add an `Metadata.isFree` helper method.
- Implement `Metadata.isTombstone` and `Metadata.isFree` with `@bitCast` then comparing to a constant. I assume `@bitCast`-then-compare is faster than the old method because it only involves one comparison, and doesn't require bitmasking.
- Summary of benchmarked changes (`gotta-go-fast`, run locally, compared to master):
- 3/4 of the hash map benchmarks used ~10% fewer cycles
- The last one (project Euler) shows 4% fewer cycles.
