Previously, we'd overwrite the errors in a circular buffer. Now that
error return traces are intended to follow a stack discipline, we no
longer have to support the index rolling over. By treating the trace
like a saturating stack, any pop/restore code still behaves correctly
past-the-end of the trace.
As a bonus, this adds a small blurb to let the user know when the trace
saturated and x number of frames were dropped.
This change extends the "lifetime" of the error return trace associated
with an error to continue throughout the block of a `const` variable
that it is assigned to.
This is necessary to support patterns like this one in test_runner.zig:
```zig
const result = foo();
if (result) |_| {
// ... success logic
} else |err| {
// `foo()` should be included in the error trace here
return error.TestFailed;
}
```
To make this happen, the majority of the error return trace popping logic
needed to move into Sema, since `const x = foo();` cannot be examined
syntactically to determine whether it modifies the error return trace. We
also have to make sure not to delete pertinent block information before it
makes it to Sema, so that Sema can pop/restore around blocks correctly.
* Why do this only for `const` and not `var`? *
There is room to relax things for `var`, but only a little bit. We could
do the same thing we do for const and keep the error trace alive for the
remainder of the block where the *assignment* happens. Any wider scope
would violate the stack discipline for traces, so it's not viable.
In the end, I decided the most consistent behavior for the user is just
to kill all error return traces assigned to a mutable `var`.
In order to enforce a strict stack discipline for error return traces,
we cannot track error return traces that are stored in variables:
```zig
const x = errorable(); // errorable()'s error return trace is killed here
// v-- error trace starts here instead
return x catch error.UnknownError;
```
In order to propagate error return traces, function calls need to be passed
directly to an error-handling expression (`if`, `catch`, `try` or `return`):
```zig
// When passed directly to `catch`, the return trace is propagated
return errorable() catch error.UnknownError;
// Using a break also works
return blk: {
// code here
break :blk errorable();
} catch error.UnknownError;
```
Why do we need this restriction? Without it, multiple errors can co-exist
with their own error traces. Handling that situation correctly means either:
a. Dynamically allocating trace memory and tracking lifetimes, OR
b. Allowing the production of one error to interfere with the trace of another
(which is the current status quo)
This is piece (3/3) of https://github.com/ziglang/zig/issues/1923#issuecomment-1218495574
This makes possible to query the memory map size from EFI firmware
without making any allocation beforehand. This makes possible to be
precise about the size of the allocation which will own a copy of
the memory map from the UEFI application.
I have noticed this causing my terminal to stop accepting input
sometimes. The previous implementation with all of its flaws was better
in the sense that it never caused this to happen.
This commit has multiple reverts in it:
Revert "Merge pull request #13148 from r00ster91/progressfollowup"
This reverts commit cb257d59f9, reversing
changes made to f5f28e0d2c.
Revert "`std.Progress`: fix inaccurate line truncation and use optimal
max terminal width (#12079)"
This reverts commit cd3d8f3a4e.
Currenty copy_file_range always uses at least two syscalls:
1. As many as it needs to do the initial copy (always 1 during my
testing)
2. The last one is always when offset is the size of the file.
The second syscall is used to detect the terminating condition. However,
because we do a stat for other reasons, we know the size of the file,
and we can skip the syscall.
Sparse files: since copy_file_range expands holes of sparse files, I
conclude that this layer was not intended to work with sparse files. In
other words, this commit does not make it worse for sparse file society.
Test program
------------
const std = @import("std");
pub fn main() !void {
const arg1 = std.mem.span(std.os.argv[1]);
const arg2 = std.mem.span(std.os.argv[2]);
try std.fs.cwd().copyFile(arg1, std.fs.cwd(), arg2, .{});
}
Test output (current master)
----------------------------
Observe two `copy_file_range` syscalls: one with 209 bytes, one with
zero:
$ zig build-exe cp.zig
$ strace ./cp ./cp.zig ./cp2.zig |& grep copy_file_range
copy_file_range(3, [0], 5, [0], 4294967295, 0) = 209
copy_file_range(3, [209], 5, [209], 4294967295, 0) = 0
$
Test output (this diff)
-----------------------
Observe a single `copy_file_range` syscall with 209 bytes:
$ /code/zig/build/zig build-exe cp.zig
$ strace ./cp ./cp.zig ./cp2.zig |& grep copy_file_range
copy_file_range(3, [0], 5, [0], 4294967295, 0) = 209
$
When creating an `EmulatableRunStep`, it now correctly depends
on its own step rather than only the executable that was created.
This means we do not need to add extra `dependOn` statements on
both the emulatable step as well as the check object step.
same change as [68e26a2cee] "std: check for overflow in writeCurrentStackTrace"
On arm64 macOS, the address of the last frame is 0x0 rather than
a positive value like 0x1 on x86_64 macOS, therefore, we overflow
an integer trying to subtract 1 when printing the stack trace. This
patch fixes it by first checking for this condition before trying
to subtract 1.
Same behaviour on i386-windows-msvc.
Note that we do not need to signal the `SignalIterator` about this
as it will correctly detect this condition on the subsequent iteration
and return `null`, thus terminating the loop.
* Sema: implement linksection on functions
* Implement function linksection in Sema.
* Don't clobber function linksection/align/addrspace in Sema.
* Fix copy-paste typo in tests.
* Add a bunch of missing test_step.dependOn.
* Fix checkInSymtab match.
Closes#12546
* Fix for: DefaultRwLock accumulates write-waiters, eventually fails to write lock #13163
* Comment out debug.print at the end of the last test.
* Code formatting
* - use equality test after lock/unlock rather than peeking into internals.
however, this is still implementation specific and only done for
DefaultRwLock.
- add num_reads maximum to ensure that reader threads stop if writer threads are
starved
- use relaxed orderings for the read atomic counter
- don't check at the end for non-zero read ops, since the reader threads may
only run once if they are starved
* More review changes
- Monotonic is sufficient for incrementing the reads counter
Commit f14cc75 accidentally added a const when grepping for assignments
to `std.builtin.Type.StructField.default_value`, however when looking
into it further, I noticed that even though this default_value field is
emitted into the .data section, the value it points to is actually
emitted into the .rodata section, so it seems correct to use const here.
- For ALU operations, src should be allowed to be an explicit Reg.
- Expose AluOp and JmpOp as public types.
This makes code generation using BPF as a backend easier,
as AluOp and JmpOp can be used directly as part of an IR
Due to the unavailability of fchdir in Windows, a call for setting the
CWD needs to either call chdir with the path string or call
SetCurrentDirectory.
Either way, since we are dealing with a Handle in Windows, a call for
GetFinalPathNameByHandle is necessary for getting the file path first.
This moves functions that LLVM generates calls to,
to the compiler_rt implementation itself, rather than c.zig.
This is a prerequisite for native backends to link with compiler-rt.
This also allows native backends to generate calls to `memcpy` and the like.
The definition of HKEY__ as a struct with an unused int field is only the case in the Windows headers when `STRICT` is defined. From https://learn.microsoft.com/en-us/windows/win32/winprog/enabling-strict:
> When STRICT is defined, data type definitions change as follows:
>
> - Specific handle types are defined to be mutually exclusive; for example, you will not be able to pass an HWND where an HDC type argument is required. Without STRICT, all handles are defined as HANDLE, so the compiler does not prevent you from using one type of handle where another type is expected.
Zig's `opaque {}` already gives this benefit to us, so the usage of a struct with an unused field is unnecessary, and it was causing HKEY to have an alignment of 4, which is a problem because there are HKEY constants like HKEY_LOCAL_MACHINE (0x80000002) that are not 4-byte aligned. Without this change, the compiler would not allow something like HKEY_LOCAL_MACHINE to be defined since it enforces pointer alignment.
This makes the following changes for i386:
long long and unsigned long long have 4 byte alignment on non-Windows
f64 (double) has 4-byte alignment on non-Windows
long double is 80 bits and has 4 byte alignment on mingw
long double on android is 64 bits, not 80: https://www.uclibc.org/docs/psABI-i386.pdfFixes#12453Fixes#12987
