We don't pass no-omit-frame-pointer in release safe by default, so it
also makes sense to not try to collect stack trace frames by default in
release safe mode.
This makes `@returnAddress()` return 0 for WebAssembly (when not using
the Emscripten OS) and avoids trying to capture stack traces for the
general purpose allocator on that target.
The high level Allocator interface API functions will now do a
`@returnAddress()` so that stack traces captured by allocator
implementations have a return address that does not include the
Allocator overhead functions. This makes `4` a more reasonable default
for how many stack frames to capture.
* std.Mutex API is improved to not have init() deinit(). This API is
designed to support static initialization and does not require any
resource cleanup. This also happens to work around some kind of
stage1 behavior that wasn't letting the new allocator mutex code
get compiled.
* the general purpose allocator now returns a bool from deinit()
which tells if there were any leaks. This value is used by the test
runner to fail the tests if there are any.
* self-hosted compiler is updated to use the general purpose allocator
when not linking against libc.
`std.GeneralPurposeAllocator` is now available. It is a function that
takes a configuration struct (with default field values) and returns an
allocator. There is a detailed description of this allocator in the
doc comments at the top of the new file.
The main feature of this allocator is that it is *safe*. It
prevents double-free, use-after-free, and detects leaks.
Some deprecation compile errors are removed.
The Allocator interface gains `old_align` as a new parameter to
`resizeFn`. This is useful to quickly look up allocations.
`std.heap.page_allocator` is improved to use mmap address hints to avoid
obtaining the same virtual address pages when unmapping and mapping
pages. The new general purpose allocator uses the page allocator as its
backing allocator by default.
`std.testing.allocator` is replaced with usage of this new allocator,
which does leak checking, and so the LeakCheckAllocator is retired.
stage1 is improved so that the `@typeInfo` of a pointer has a lazy value
for the alignment of the child type, to avoid false dependency loops
when dealing with pointers to async function frames.
The `std.mem.Allocator` interface is refactored to be in its own file.
`std.Mutex` now exposes the dummy mutex with `std.Mutex.Dummy`.
This allocator is great for debug mode, however it needs some work to
have better performance in release modes. The next step will be setting
up a series of tests in ziglang/gotta-go-fast and then making
improvements to the implementation.
This commit reimagines `std.os.windows.GetFinalPathNameByHandle`
using `DeviceIoControl` to query the OS mount manager for the DOS
(symlink) paths for the given NT volume name. In particular,
it uses `IOCTL_MOUNTMGR_QUERY_POINTS` ioctl opcode to query the
manager for the available moount points.
Favour newer API which uses `NtQueryInformationFile` with class flags
`FileNormalizedNameInformation` and `FileVolumeNameInformation`
instead of lower-level `NtQueryObject`. `NtQueryObject` is still
used as a fallback in case the former are unavailable.
This commit proposes an initial draft of `GetPathNameByHandle` function
which wraps NT syscalls and strives to emulate (currently only
partially) the `kernel32.GetFinalPathNameByHandleW` function.
Supports writing the first function. Still TODO is:
* handling the .debug_line header growing too large
* adding a new file to an existing compilation
* adding an additional function to an existing file
* handling incremental updates
* adding the main IR debug ops for IR instructions
There are also issues to work out:
* readelf --debug-dump=rawline is saying there is no .debug_str section
even though there is
* readelf --debug-dump=decodedline is saying the file index 0 is bad
and reporting some other kind of corruption.
* the .debug_line header is written properly
* link.File.Elf gains:
- SrcFn, which is now a field in Module.Fn
- SrcFile, which is now a field in Module.Scope.File
* link.File.Elf gets a whole *Package field rather than only
root_src_dir_path.
* the fields first_dbg_line_file and last_dbg_line_file tell where the
Line Number Program begins and ends, which alows moving files when
the header gets too big, and allows appending files to the end.
* codegen is passed a buffer for emitting .debug_line
Line Number Program opcodes for functions.
See #5963
There is some work-in-progress code here, but I need to go make some
experimental changes to changing how to represent source locations and I
want to do that in a separate commit.
This shaves off one syscall (we use one instead of two if we were to
use `windows.OpenFile` wrapper). Clean up flag generation in `OpenFile`.
Hopefully, we're in a much better place to *almost* support `openW`
and `openatW`.
This way, we can remove more `kernel32` calls such as `RemoveDirectoryW`
or `DeleteFileW`, and use `std.os.windows.DeleteFile` instead which
is purely NT-based.
Replace them with `std.os.windows.OpenFile` instead. To allow
creation/opening of directories, `std.os.windows.OpenFileOptions`
now features a `.expect_dir: bool` member which is meant to emualate
POSIX's `O_DIRECTORY` flag.
As discussed in the previous commit, it would be better to avoid
function pointers to syscalls and explicitly split the control
path into two function calls instead. This commit addresses that
for `std.os.windows.DeviceIoControl`.
When there are no format parameters, it simply calls `writeAll`. This
has the effect of no longer emitting a compile error for using `{}` and
not having any parameters, however, at this point in the development
process of Zig I think that tradeoff is worthwhile.
On the other hand, it might be OK to simply define formatting to work
this way. It's a common pattern to use the formatting function's format
string `"like this", .{}` instead of `"{}", .{"like this"}`, which can
lead to accidentally putting control characters in the formatting
string, however, with this change that works just fine.
This commit adds a Zig wrapper for `kernel32.DeviceIoControl` which
applies ReactOS logic for deciding whether to use
`ntdll.NtDeviceIoControlFile` or `ntdll.NtFsControlFile` based on the
value of passed `IO_CONTROL_CODE`. The decision logic is based on the
logic found in ReactOS found in the following [link].
Thanks to Daurnimator for finding this bit in ReactOS!
[link]: https://doxygen.reactos.org/d3/d74/deviceio_8c.html
* Added `replace` for simple pre-allocated search-and-replace
* Added `replacementSize` for calculating ahead-of-time buffer sizes for
performing a safe search-and-replace
* Added `replaceOwned` for automatically allocating and performing
replacement
* AST: flatten ControlFlowExpression into Continue, Break, and Return.
* AST: unify identifiers and literals into the same AST type: OneToken
* AST: ControlFlowExpression uses TrailerFlags to optimize storage
space.
* astgen: support `var` as well as `const` locals, and support
explicitly typed locals. Corresponding Module and codegen code is not
implemented yet.
* astgen: support result locations.
* ZIR: add the following instructions (see the corresponding doc
comments for explanations of semantics):
- alloc
- alloc_inferred
- bitcast_result_ptr
- coerce_result_block_ptr
- coerce_result_ptr
- coerce_to_ptr_elem
- ensure_result_used
- ensure_result_non_error
- ret_ptr
- ret_type
- store
- param_type
* the skeleton structure for result locations is set up. It's looking
pretty clean so far.
* add compile error for unused result and compile error for discarding
errors.
* astgen: split builtin calls up to implemented manually, and implement
`@as`, `@bitCast` (and others) with respect to result locations.
* add CLI support for hex and raw object formats. They are not
supported by the self-hosted compiler yet, and emit errors.
* rename `--c` CLI to `-ofmt=[objectformat]` which can be any of the
object formats. Only ELF and C are supported so far. Also added missing
help to the help text.
* Remove hard tabs from C backend test cases. Shame on you Noam, you
are grounded, you should know better, etc. Bad boy.
* Delete C backend code and test case that relied on comptime_int
incorrectly making it all the way to codegen.
Otherwise, the behaviour can lead to unexpected results, resulting
in removing an entire tree that's not necessarily under the root.
Furthermore, this change is needed if are to properly handle dir
symlinks on Windows. Without explicitly requiring that a directory
or file is opened with `FILE_OPEN_REPARSE_POINT`, Windows automatically
dereferences all symlinks along the way. This commit adds another
option to `OpenDirOptions`, namely `.no_follow`, which defaults to
`false` and can be used to specifically open a directory symlink on
Windows or call `openat` with `O_NOFOLLOW` flag in POSIX.
This way `std.fs.symlinkAbsolute` becomes cross-platform and we can
legally include `SymlinkFlags` as an argument that's only used on
Windows. Also, now `std.os.symlink` generates a compile error on
Windows with a message to instead use `std.os.windows.CreateSymbolicLink`.
Finally, this PR also reshuffles the tests between `std.os.test` and
`std.fs.test`.
