detectAbiAndDynamicLinker

fn detectAbiAndDynamicLinker(
    cpu: Target.Cpu,
    os: Target.Os,
    query: Target.Query,
) DetectError!Target {
    const native_target_has_ld = comptime Target.DynamicLinker.kind(builtin.os.tag) != .none;
    const is_linux = builtin.target.os.tag == .linux;
    const is_solarish = builtin.target.os.tag.isSolarish();
    const is_darwin = builtin.target.os.tag.isDarwin();
    const have_all_info = query.dynamic_linker.get() != null and
        query.abi != null and (!is_linux or query.abi.?.isGnu());
    const os_is_non_native = query.os_tag != null;
    // The Solaris/illumos environment is always the same.
    if (!native_target_has_ld or have_all_info or os_is_non_native or is_solarish or is_darwin) {
        return defaultAbiAndDynamicLinker(cpu, os, query);
    }
    if (query.abi) |abi| {
        if (abi.isMusl()) {
            // musl implies static linking.
            return defaultAbiAndDynamicLinker(cpu, os, query);
        }
    }
    // The current target's ABI cannot be relied on for this. For example, we may build the zig
    // compiler for target riscv64-linux-musl and provide a tarball for users to download.
    // A user could then run that zig compiler on riscv64-linux-gnu. This use case is well-defined
    // and supported by Zig. But that means that we must detect the system ABI here rather than
    // relying on `builtin.target`.
    const all_abis = comptime blk: {
        assert(@intFromEnum(Target.Abi.none) == 0);
        const fields = std.meta.fields(Target.Abi)[1..];
        var array: [fields.len]Target.Abi = undefined;
        for (fields, 0..) |field, i| {
            array[i] = @field(Target.Abi, field.name);
        }
        break :blk array;
    };
    var ld_info_list_buffer: [all_abis.len]LdInfo = undefined;
    var ld_info_list_len: usize = 0;

    switch (Target.DynamicLinker.kind(os.tag)) {
        // The OS has no dynamic linker. Leave the list empty and rely on `Abi.default()` to pick
        // something sensible in `abiAndDynamicLinkerFromFile()`.
        .none => {},
        // The OS has a system-wide dynamic linker. Unfortunately, this implies that there's no
        // useful ABI information that we can glean from it merely being present. That means the
        // best we can do for this case (for now) is also `Abi.default()`.
        .arch_os => {},
        // The OS can have different dynamic linker paths depending on libc/ABI. In this case, we
        // need to gather all the valid arch/OS/ABI combinations. `abiAndDynamicLinkerFromFile()`
        // will then look for a dynamic linker with a matching path on the system and pick the ABI
        // we associated it with here.
        .arch_os_abi => for (all_abis) |abi| {
            const ld = Target.DynamicLinker.standard(cpu, os, abi);

            // Does the generated target triple actually have a standard dynamic linker path?
            if (ld.get() == null) continue;

            ld_info_list_buffer[ld_info_list_len] = .{
                .ld = ld,
                .abi = abi,
            };
            ld_info_list_len += 1;
        },
    }

    const ld_info_list = ld_info_list_buffer[0..ld_info_list_len];

    // Best case scenario: the executable is dynamically linked, and we can iterate
    // over our own shared objects and find a dynamic linker.
    const elf_file = elf_file: {
        // This block looks for a shebang line in /usr/bin/env,
        // if it finds one, then instead of using /usr/bin/env as the ELF file to examine, it uses the file it references instead,
        // doing the same logic recursively in case it finds another shebang line.

        var file_name: []const u8 = switch (os.tag) {
            // Since /usr/bin/env is hard-coded into the shebang line of many portable scripts, it's a
            // reasonably reliable path to start with.
            else => "/usr/bin/env",
            // Haiku does not have a /usr root directory.
            .haiku => "/bin/env",
        };

        // According to `man 2 execve`:
        //
        // The kernel imposes a maximum length on the text
        // that follows the "#!" characters at the start of a script;
        // characters beyond the limit are ignored.
        // Before Linux 5.1, the limit is 127 characters.
        // Since Linux 5.1, the limit is 255 characters.
        //
        // Tests show that bash and zsh consider 255 as total limit,
        // *including* "#!" characters and ignoring newline.
        // For safety, we set max length as 255 + \n (1).
        var buffer: [255 + 1]u8 = undefined;
        while (true) {
            // Interpreter path can be relative on Linux, but
            // for simplicity we are asserting it is an absolute path.
            const file = fs.openFileAbsolute(file_name, .{}) catch |err| switch (err) {
                error.NoSpaceLeft => unreachable,
                error.NameTooLong => unreachable,
                error.PathAlreadyExists => unreachable,
                error.SharingViolation => unreachable,
                error.InvalidUtf8 => unreachable, // WASI only
                error.InvalidWtf8 => unreachable, // Windows only
                error.BadPathName => unreachable,
                error.PipeBusy => unreachable,
                error.FileLocksNotSupported => unreachable,
                error.WouldBlock => unreachable,
                error.FileBusy => unreachable, // opened without write permissions
                error.AntivirusInterference => unreachable, // Windows-only error

                error.IsDir,
                error.NotDir,
                error.AccessDenied,
                error.PermissionDenied,
                error.NoDevice,
                error.FileNotFound,
                error.NetworkNotFound,
                error.FileTooBig,
                error.Unexpected,
                => |e| {
                    std.log.warn("Encountered error: {s}, falling back to default ABI and dynamic linker.", .{@errorName(e)});
                    return defaultAbiAndDynamicLinker(cpu, os, query);
                },

                else => |e| return e,
            };
            var is_elf_file = false;
            defer if (is_elf_file == false) file.close();

            // Shortest working interpreter path is "#!/i" (4)
            // (interpreter is "/i", assuming all paths are absolute, like in above comment).
            // ELF magic number length is also 4.
            //
            // If file is shorter than that, it is definitely not ELF file
            // nor file with "shebang" line.
            const min_len: usize = 4;

            const len = preadAtLeast(file, &buffer, 0, min_len) catch |err| switch (err) {
                error.UnexpectedEndOfFile,
                error.UnableToReadElfFile,
                error.ProcessNotFound,
                => return defaultAbiAndDynamicLinker(cpu, os, query),

                else => |e| return e,
            };
            const content = buffer[0..len];

            if (mem.eql(u8, content[0..4], std.elf.MAGIC)) {
                // It is very likely ELF file!
                is_elf_file = true;
                break :elf_file file;
            } else if (mem.eql(u8, content[0..2], "#!")) {
                // We detected shebang, now parse entire line.

                // Trim leading "#!", spaces and tabs.
                const trimmed_line = mem.trimStart(u8, content[2..], &.{ ' ', '\t' });

                // This line can have:
                // * Interpreter path only,
                // * Interpreter path and arguments, all separated by space, tab or NUL character.
                // And optionally newline at the end.
                const path_maybe_args = mem.trimEnd(u8, trimmed_line, "\n");

                // Separate path and args.
                const path_end = mem.indexOfAny(u8, path_maybe_args, &.{ ' ', '\t', 0 }) orelse path_maybe_args.len;

                file_name = path_maybe_args[0..path_end];
                continue;
            } else {
                // Not a ELF file, not a shell script with "shebang line", invalid duck.
                return defaultAbiAndDynamicLinker(cpu, os, query);
            }
        }
    };
    defer elf_file.close();

    // TODO: inline this function and combine the buffer we already read above to find
    // the possible shebang line with the buffer we use for the ELF header.
    return abiAndDynamicLinkerFromFile(elf_file, cpu, os, ld_info_list, query) catch |err| switch (err) {
        error.FileSystem,
        error.SystemResources,
        error.SymLinkLoop,
        error.ProcessFdQuotaExceeded,
        error.SystemFdQuotaExceeded,
        error.ProcessNotFound,
        => |e| return e,

        error.UnableToReadElfFile,
        error.InvalidElfClass,
        error.InvalidElfVersion,
        error.InvalidElfEndian,
        error.InvalidElfFile,
        error.InvalidElfMagic,
        error.Unexpected,
        error.UnexpectedEndOfFile,
        error.NameTooLong,
        error.StaticElfFile,
        // Finally, we fall back on the standard path.
        => |e| {
            std.log.warn("Encountered error: {s}, falling back to default ABI and dynamic linker.", .{@errorName(e)});
            return defaultAbiAndDynamicLinker(cpu, os, query);
        },
    };
}