Fixed an issue where external local variable references in C code were translated to incorrect zig code

[flyfish30]

Fixed issue #23275

If can not find extern variable in current block scope, then recusively call function getLocalExternAlias with parent block scope until reach the root scope.
When trans a implicit cast expression with kind ArrayToPointerDecay, do not add address_of tag.

[Vexu]

If can not find extern variable in current block scope, then recusively call function getLocalExternAlias with parent block scope until reach the root scope.

This needs a test, here is a reduced example:

int foo(int bar) {
    extern int a;
    if (bar) {
        return a;
    }
    return 0;
}

When trans a implicit cast expression with kind ArrayToPointerDecay, do not add address_of tag.

This especially needs a test since I don't know what the issue is.

[flyfish30]

@Vexu,

How to create a test for:

int foo(int bar) {
    extern int a;
    if (bar) {
        return a;
    }
    return 0;
}

The "extern int a" need a external variable "a" that declarated in other C file, but the test case of translate-c can only add one C source file.
The add case function in test/src/RunTranslatedC.zig is show in bellow:

pub fn add(
    self: *RunTranslatedCContext,
    name: []const u8,
    source: []const u8,
    expected_stdout: []const u8,
) void {
    const tc = self.create(false, "source.c", name, source, expected_stdout);
    self.addCase(tc);
}

[flyfish30]

If can not find extern variable in current block scope, then recusively call function getLocalExternAlias with parent block scope until reach the root scope.

This needs a test, here is a reduced example:

int foo(int bar) {
    extern int a;
    if (bar) {
        return a;
    }
    return 0;
}

When trans a implicit cast expression with kind ArrayToPointerDecay, do not add address_of tag.

This especially needs a test since I don't know what the issue is.

I had update the PR. Add a test for your first problem as bellow:

    cases.add("extern local variable referenced in sub block scope, Isssue #23275",
        \\#include <stdlib.h>
        \\int a = 42;
        \\int foo(int bar) {
        \\    extern int a;
        \\    if (bar) {
        \\        return a;
        \\    }
        \\    return 0;
        \\}
        \\int main() {
        \\    int result1 = foo(0);
        \\    if (result1 != 0) abort();
        \\    int result2 = foo(1);
        \\    if (result2 != 42) abort();
        \\    a = 100;
        \\    int result3 = foo(1);
        \\    if (result3 != 100) abort();
        \\    return 0;
        \\}
    , "");

For your second question, I modified the method to convert array to pointer, this new method fixes the "array to pointer decay" test case failure issue. Please recheck.

[Vexu]

A simpler translate test like test/cases/translate_c/_Static_assert.c would suffice, also note the big notice at the top of test/run_translated_c.zig.

For your second question, I modified the method to convert array to pointer, this new method fixes the "array to pointer decay" test case failure issue. Please recheck.

&arr and &arr[0] should do the same thing, what were you trying to fix with the change?

[flyfish30]

For your second question, I modified the method to convert array to pointer, this new method fixes the "array to pointer decay" test case failure issue. Please recheck.

&arr and &arr[0] should do the same thing, what were you trying to fix with the change?

I have a C function in bellow:

int foo(int *bar) {
    extern int array_ref[];
    return 1 + (short)(bar - array_ref);
}

The translate-c will generate a zig function in bellow:

pub export fn foo(arg_bar: [*c]c_int) c_int {
    var bar = arg_bar;
    _ = &bar;
    const ExternLocal_array_ref = struct {
        extern var array_ref: [*c]c_int;
    };
    _ = &ExternLocal_array_ref;
    return @as(c_int, 1) + @as(c_int, @bitCast(@as(c_int, @as(c_short, @bitCast(@as(c_short, @truncate(@divExact(@as(c_long, @bitCast(@intFromPtr(bar) -% @intFromPtr(@as([*c]c_int, @ptrCast(@alignCast(&ExternLocal_array_ref.array_ref)))))), @sizeOf(c_int)))))))));
}

In return statement, "&ExternLocal_array_ref.array_ref" will get a pointer that point to address of ExternLocal_array_ref.array_ref, it is incorrect. I expected it generate a zig code "&ExternLocal_array_ref.array_ref[0]", this is a pointer that point to ExternLocal_array_ref.array_ref self.

[Vexu]

In that case your original fix of omitting the & would have worked if you only applied it to arrays with unknown size. Always adding [0] will also work but I suspect it will require updating more existing tests.

[flyfish30]

I also test the bellow test case by use &arr[0], it is pass the testing.

    cases.add("pointer difference: scalar array w/ size truncation or negative result. Issue #7216",
        \\#include <stdlib.h>
        \\#include <stddef.h>
        \\#define SIZE 10
        \\int main() {
        \\    int foo[SIZE];
        \\    int *start = &foo[0];
        \\    int *one_past_end = start + SIZE;
        \\    ptrdiff_t diff = one_past_end - start;
        \\    char diff_char = one_past_end - start;
        \\    if (diff != SIZE || diff_char != SIZE) abort();
        \\    diff = start - one_past_end;
        \\    if (diff != -SIZE) abort();
        \\    if (one_past_end - foo != SIZE) abort();
        \\    if ((one_past_end - 1) - foo != SIZE - 1) abort();
        \\    if ((start + 1) - foo != 1) abort();
        \\    return 0;
        \\}
    , "");

[Vexu]

You also need // link_libc=true to use libc.

[flyfish30]

You also need // link_libc=true to use libc.

I had add // link_libc=true in to sub_scope_extern_local_var_ref.c file, why it is not needed // link_libc=true in macOS?

[mlugg]

macOS tests always link libc, regardless of whether you ask them to, since libc is the only way to interact with the OS on macOS (it doesn't have stable syscalls).