tally/engine/test/struct_oracle.zig

//! Holds `struct_layout`'s profiles to a C compiler (NFR-3).
//!
//! For each profile, writes one C file asserting, with `_Static_assert`, what the
//! engine computed: the size and in-struct alignment of every scalar kind, then the
//! size, alignment, and every member's offset and size for each test struct below.
//! `build.zig` compiles each file for every clang target the profile claims to
//! reproduce, so a wrong table entry or a wrong layout rule fails `zig build test`
//! with the struct and the claim in the message. Nothing beyond Zig is needed: clang
//! is the one Zig ships.
//!
//! The test structs are C, read by the engine's own parser, so the compiler and the
//! engine see the same text.
//!
//! Usage: struct-oracle <abi-id> <output.c> [<abi-id> <output.c> ...]
//! Every profile in the engine must be named exactly once, so a profile cannot be
//! added without being checked.

const std = @import("std");
const engine = @import("engine");
const struct_layout = engine.struct_layout;
const CStructParser = engine.CStructParser;
const Primitive = struct_layout.Primitive;

const Case = struct {
    /// C that the engine does not read: declarations a test struct refers to.
    prelude: []const u8 = "",
    /// One struct definition, named, as the engine parses it.
    source: []const u8,
};

/// Each one exists to pin a rule. Any `push` a case makes it also pops: clang warns
/// about an unterminated one, and the generator resets the pack after every case
/// with `#pragma pack()` so a plain `pack(N)` cannot leak into the next.
const cases = [_]Case{
    // The design document's example: one byte of interior padding.
    .{ .source = "struct spec { u32 x; u8 flags; u16 id; u64 ts; };" },
    // Tail padding.
    .{ .source = "struct tail { uint64_t a; uint8_t b; };" },
    // The data model: what `long`, pointers and `wchar_t` are.
    .{ .source = "struct model { char c; long l; void *p; wchar_t w; size_t n; char e; };" },
    // Where 8-byte types align.
    .{ .source = "struct doubles { char c; double d; char e; long long ll; char f; };" },
    // Alignment rising and falling through a struct.
    .{ .source = "struct mixed { char a; short b; char c; int d; char e; long long f; char g; };" },
    // Every spelling the parser accepts maps to the kind the compiler gives it.
    .{ .source =
    \\struct spellings {
    \\    signed char sc; unsigned char uc; char c; _Bool b; bool b2;
    \\    short s; unsigned short us; short int si; int i; unsigned u; signed sg;
    \\    long l; unsigned long ul; long int li; long long ll; unsigned long long int ull;
    \\    float f; double d;
    \\    int8_t i8_; uint8_t u8_; int16_t i16_; uint16_t u16_; int32_t i32_; uint32_t u32_;
    \\    int64_t i64_; uint64_t u64_;
    \\    intptr_t ip; uintptr_t up; size_t sz; ssize_t ssz; ptrdiff_t pd;
    \\    wchar_t wc; char16_t c16; char32_t c32;
    \\    u8 s1; i8 s2; u16 s3; i16 s4; u32 s5; i32 s6; u64 s7; i64 s8;
    \\    f32 s9; f64 s10; usize s11; isize s12; ptr s13;
    \\    const volatile int cv;
    \\};
    },
    // Sixteen-byte integers, on the profiles that have them.
    .{ .source =
    \\struct wide {
    \\    char c; __int128 big; char d; unsigned __int128 ubig;
    \\    __int128_t t; __uint128_t ut; u128 x; i128 y;
    \\};
    },
    // Arrays, including C's octal and hex lengths.
    .{ .source = "struct arrays { char tag[3]; int grid[2][3]; short s[0x4]; char oct[010]; double d[1]; };" },
    // A flexible array member: no size, but it aligns, and the struct rounds up.
    .{ .source = "struct flex { short len; char c; double data[]; };" },
    .{ .source = "struct flex_tail { int n; char c; char data[]; };" },
    // Pointers of every shape.
    .{ .prelude = "enum color { red, green };", .source =
    \\struct ptrs {
    \\    char c; const char *name, **more; void (*cb)(int);
    \\    void (*handlers[3])(void); char (*row)[16]; struct ptrs *next;
    \\    struct elsewhere *opaque; enum color k; char *const *const pp;
    \\};
    },
    // A typedef'd anonymous struct.
    .{ .source = "typedef struct { char c; double d; } typedef_s;" },
    // `packed` on the struct.
    .{ .source = "struct __attribute__((packed)) packed_s { char c; int i; double d; short s; };" },
    // ... with a member that asks for alignment anyway.
    .{ .source = "struct __attribute__((packed)) packed_aligned { char c; int i __attribute__((aligned(4))); short s; };" },
    // `packed` on one member.
    .{ .source = "struct packed_member { char c; int i __attribute__((packed)); short s; };" },
    // `alignas` and `aligned` on members.
    .{ .source = "struct aligned_member { char c; alignas(16) int i; _Alignas(8) char d; };" },
    .{ .source = "struct attr_aligned { char c; int i __attribute__((aligned(8))); char d; };" },
    // An `aligned` smaller than the natural alignment.
    .{ .source = "struct aligned_lower { char c; double d __attribute__((aligned(2))); };" },
    // `aligned` on the struct.
    .{ .source = "struct __attribute__((aligned(32))) aligned_struct { int a; char b; };" },
    .{ .source = "struct trailing_aligned { int a; char b; } __attribute__((aligned(16)));" },
    // `#pragma pack`.
    .{ .source = "#pragma pack(push, 1)\nstruct pack1 { char c; int i; double d; };\n#pragma pack(pop)" },
    .{ .source = "#pragma pack(2)\nstruct pack2 { char c; double d; short s; char e; };" },
    .{ .source = "#pragma pack(4)\nstruct pack4 { char c; double d; long long ll; };" },
    .{ .source = "#pragma pack(16)\nstruct pack16 { char c; double d; };" },
    // How pack meets the attributes: the interactions most worth checking.
    .{ .source = "#pragma pack(push, 2)\nstruct pack_attr_aligned { char c; int i __attribute__((aligned(8))); };\n#pragma pack(pop)" },
    .{ .source = "#pragma pack(push, 2)\nstruct pack_alignas { char c; alignas(8) int i; };\n#pragma pack(pop)" },
    .{ .source = "#pragma pack(push, 4)\nstruct __attribute__((packed)) pack_packed { char c; double d; };\n#pragma pack(pop)" },
    .{ .source = "#pragma pack(push, 1)\nstruct __attribute__((aligned(8))) pack_struct_aligned { char c; int i; };\n#pragma pack(pop)" },
    .{ .source = "#pragma pack(push, 2)\nstruct pack_flex { char c; double data[]; };\n#pragma pack(pop)" },
};

pub fn main(init: std.process.Init) !void {
    const allocator = init.arena.allocator();
    const io = init.io;
    const args = try init.minimal.args.toSlice(allocator);

    if (args.len < 3 or (args.len - 1) % 2 != 0) {
        std.log.err("usage: struct-oracle <abi-id> <output.c> [...]", .{});
        return error.BadArguments;
    }

    var seen = [_]bool{false} ** struct_layout.abis.len;
    var i: usize = 1;
    while (i < args.len) : (i += 2) {
        const index = for (struct_layout.abis, 0..) |abi, n| {
            if (std.mem.eql(u8, abi.id, args[i])) break n;
        } else {
            std.log.err("no profile named '{s}'", .{args[i]});
            return error.UnknownAbi;
        };
        if (seen[index]) {
            std.log.err("'{s}' named twice", .{args[i]});
            return error.DuplicateAbi;
        }
        seen[index] = true;

        const text = try generate(allocator, &struct_layout.abis[index]);
        try std.Io.Dir.cwd().writeFile(io, .{ .sub_path = args[i + 1], .data = text });
    }
    for (struct_layout.abis, seen) |abi, checked| {
        if (!checked) {
            std.log.err("profile '{s}' has no compiler target in build.zig", .{abi.id});
            return error.UncheckedAbi;
        }
    }
}

/// The C spelling of each kind, for the scalar probes.
const spelling: std.EnumArray(Primitive, []const u8) = .init(.{
    .char = "char",
    .bool = "_Bool",
    .short = "short",
    .int = "int",
    .long = "long",
    .long_long = "long long",
    .int128 = "__int128",
    .float = "float",
    .double = "double",
    .pointer = "void *",
    .wchar = "wchar_t",
    .enumeration = "enum probe_enum",
});

const prelude =
    \\#include <stddef.h>
    \\#include <stdint.h>
    \\#include <stdbool.h>
    \\#include <stdalign.h>
    \\
    \\/* What the engine's parser accepts that C does not define for itself. */
    \\typedef __CHAR16_TYPE__ char16_t;
    \\typedef __CHAR32_TYPE__ char32_t;
    \\typedef __PTRDIFF_TYPE__ ssize_t;
    \\typedef uint8_t u8;
    \\typedef int8_t i8;
    \\typedef uint16_t u16;
    \\typedef int16_t i16;
    \\typedef uint32_t u32;
    \\typedef int32_t i32;
    \\typedef uint64_t u64;
    \\typedef int64_t i64;
    \\typedef float f32;
    \\typedef double f64;
    \\typedef size_t usize;
    \\typedef ptrdiff_t isize;
    \\typedef void *ptr;
    \\#ifdef __SIZEOF_INT128__
    \\typedef unsigned __int128 u128;
    \\typedef __int128 i128;
    \\#endif
    \\
    \\enum probe_enum { probe_enum_value };
    \\
;

fn generate(allocator: std.mem.Allocator, abi: *const struct_layout.Abi) ![]u8 {
    var out: std.Io.Writer.Allocating = .init(allocator);
    const w = &out.writer;

    try w.print(
        \\/* Generated by engine/test/struct_oracle.zig. Every assertion is a claim the
        \\ * engine makes about the {s} profile ({s}); a failure here means
        \\ * the engine disagrees with the compiler. */
        \\
    , .{ abi.id, abi.label });
    try w.writeAll(prelude);

    // Every scalar kind: its size, and its alignment inside a struct (which is not
    // `_Alignof`'s answer on i386, where a double aligns to 8 alone and 4 in a struct).
    try w.writeAll("/* Scalars */\n");
    for (std.enums.values(Primitive)) |p| {
        const shape = abi.shapes.get(p) orelse {
            if (p != .int128) return error.ProbeMissing;
            try w.print(
                \\#ifdef __SIZEOF_INT128__
                \\#error "{s}: the engine says there is no __int128, and this target has one"
                \\#endif
                \\
            , .{abi.id});
            continue;
        };
        if (p == .int128) try w.print(
            \\#ifndef __SIZEOF_INT128__
            \\#error "{s}: the engine says there is an __int128, and this target has none"
            \\#endif
            \\
        , .{abi.id});
        const c_type = spelling.get(p);
        try w.print(
            \\struct probe_{t} {{ char c; {s} v; }};
            \\_Static_assert(sizeof({s}) == {d}, "{s}: the engine says {s} is {d} bytes");
            \\_Static_assert(__builtin_offsetof(struct probe_{t}, v) == {d}, "{s}: the engine says {s} aligns to {d} in a struct");
            \\
        , .{
            p, c_type,          c_type, shape.size, abi.id,          c_type, shape.size,
            p, shape.alignment, abi.id, c_type,     shape.alignment,
        });
    }

    for (cases) |case| try writeCase(allocator, w, abi, case);
    return out.toOwnedSlice();
}

fn writeCase(allocator: std.mem.Allocator, w: *std.Io.Writer, abi: *const struct_layout.Abi, case: Case) !void {
    var diag: CStructParser.Diagnostic = .{};
    var parsed = CStructParser.parse(allocator, case.source, &diag) catch |err| {
        const at = CStructParser.locate(case.source, diag.span.start);
        std.log.err("a test struct does not parse ({t}, line {d}, column {d}):\n{s}", .{
            err, at.line, at.column, case.source,
        });
        return err;
    };
    defer parsed.deinit();
    const def = parsed.definition;
    const name = def.name orelse return error.UnnamedCase;
    var ref_buf: [64]u8 = undefined;
    const ref = switch (def.name_kind) {
        .tag => try std.fmt.bufPrint(&ref_buf, "struct {s}", .{name}),
        .typedef => name,
    };

    var layout = struct_layout.compute(allocator, def, abi, null) catch |err| switch (err) {
        // Left out entirely, since the compiler would refuse the struct. Whether the
        // type really is absent is the scalar probes' assertion.
        error.TypeUnavailable => {
            try w.print("\n/* {s}: uses a type {s} does not have */\n", .{ name, abi.id });
            return;
        },
        else => {
            std.log.err("{s} does not lay out on {s}: {t}", .{ name, abi.id, err });
            return err;
        },
    };
    defer layout.deinit(allocator);

    try w.print("\n/* {s} */\n{s}\n{s}\n#pragma pack()\n", .{ name, case.prelude, case.source });
    try w.print(
        \\_Static_assert(sizeof({s}) == {d}, "{s}: the engine says {s} is {d} bytes");
        \\_Static_assert(_Alignof({s}) == {d}, "{s}: the engine says {s} aligns to {d}");
        \\
    , .{ ref, layout.size, abi.id, ref, layout.size, ref, layout.alignment, abi.id, ref, layout.alignment });
    for (def.members, layout.placements) |member, placement| {
        try w.print(
            \\_Static_assert(__builtin_offsetof({s}, {s}) == {d}, "{s}: the engine says {s}.{s} is at {d}");
            \\
        , .{ ref, member.name, placement.offset, abi.id, ref, member.name, placement.offset });
        if (!member.flexible) try w.print(
            \\_Static_assert(sizeof((({s} *)0)->{s}) == {d}, "{s}: the engine says {s}.{s} is {d} bytes");
            \\
        , .{ ref, member.name, placement.size, abi.id, ref, member.name, placement.size });
    }
}