//! Originally, we tried using LIEF to inject the module graph into a MachO segment //! But this incurred a fixed 350ms overhead on every build, which is unacceptable //! so we give up on codesigning support on macOS for now until we can find a better solution const bun = @import("root").bun; const std = @import("std"); const Schema = bun.Schema.Api; const strings = bun.strings; const Output = bun.Output; const Global = bun.Global; const Environment = bun.Environment; const Syscall = bun.sys; const SourceMap = bun.sourcemap; const StringPointer = bun.StringPointer; const w = std.os.windows; pub const StandaloneModuleGraph = struct { bytes: []const u8 = "", files: bun.StringArrayHashMap(File), entry_point_id: u32 = 0, // We never want to hit the filesystem for these files // We use the `/$bunfs/` prefix to indicate that it's a virtual path // It is `/$bunfs/` because: // // - `$` makes it unlikely to collide with a real path // - `/$bunfs/` is 8 characters which is fast to compare for 64-bit CPUs pub const base_path = switch (Environment.os) { else => "/$bunfs/", // Special case for windows because of file URLs being invalid // if they do not have a drive letter. B drive because 'bun' but // also because it's more unlikely to collide with a real path. .windows => "B:\\~BUN\\", }; pub const base_public_path = targetBasePublicPath(Environment.os, ""); pub const base_public_path_with_default_suffix = targetBasePublicPath(Environment.os, "root/"); pub fn targetBasePublicPath(target: Environment.OperatingSystem, comptime suffix: [:0]const u8) [:0]const u8 { return switch (target) { .windows => "B:/~BUN/" ++ suffix, else => "/$bunfs/" ++ suffix, }; } pub fn isBunStandaloneFilePath(str: []const u8) bool { return bun.strings.hasPrefixComptime(str, base_path) or (Environment.isWindows and bun.strings.hasPrefixComptime(str, base_public_path)); } pub fn entryPoint(this: *const StandaloneModuleGraph) *File { return &this.files.values()[this.entry_point_id]; } // by normalized file path pub fn find(this: *const StandaloneModuleGraph, name: []const u8) ?*File { if (!isBunStandaloneFilePath(base_path)) { return null; } return this.findAssumeStandalonePath(name); } pub fn findAssumeStandalonePath(this: *const StandaloneModuleGraph, name: []const u8) ?*File { if (Environment.isWindows) { var normalized_buf: bun.PathBuffer = undefined; const normalized = bun.path.platformToPosixBuf(u8, name, &normalized_buf); return this.files.getPtr(normalized); } return this.files.getPtr(name); } pub const CompiledModuleGraphFile = struct { name: Schema.StringPointer = .{}, contents: Schema.StringPointer = .{}, sourcemap: Schema.StringPointer = .{}, bytecode: Schema.StringPointer = .{}, encoding: Encoding = .latin1, loader: bun.options.Loader = .file, module_format: ModuleFormat = .none, }; pub const Encoding = enum(u8) { binary = 0, latin1 = 1, // Not used yet. utf8 = 2, }; pub const ModuleFormat = enum(u8) { none = 0, esm = 1, cjs = 2, }; pub const File = struct { name: []const u8 = "", loader: bun.options.Loader, contents: [:0]const u8 = "", sourcemap: LazySourceMap, cached_blob: ?*bun.JSC.WebCore.Blob = null, encoding: Encoding = .binary, wtf_string: bun.String = bun.String.empty, bytecode: []u8 = "", module_format: ModuleFormat = .none, pub fn lessThanByIndex(ctx: []const File, lhs_i: u32, rhs_i: u32) bool { const lhs = ctx[lhs_i]; const rhs = ctx[rhs_i]; return bun.strings.cmpStringsAsc({}, lhs.name, rhs.name); } pub fn toWTFString(this: *File) bun.String { if (this.wtf_string.isEmpty()) { switch (this.encoding) { .binary, .utf8 => { this.wtf_string = bun.String.createUTF8(this.contents); }, .latin1 => { this.wtf_string = bun.String.createStaticExternal(this.contents, true); }, } } // We don't want this to free. return this.wtf_string.dupeRef(); } pub fn blob(this: *File, globalObject: *bun.JSC.JSGlobalObject) *bun.JSC.WebCore.Blob { if (this.cached_blob == null) { const store = bun.JSC.WebCore.Blob.Store.init(@constCast(this.contents), bun.default_allocator); // make it never free store.ref(); const b = bun.JSC.WebCore.Blob.initWithStore(store, globalObject).new(); b.allocator = bun.default_allocator; if (bun.http.MimeType.byExtensionNoDefault(bun.strings.trimLeadingChar(std.fs.path.extension(this.name), '.'))) |mime| { store.mime_type = mime; b.content_type = mime.value; b.content_type_was_set = true; b.content_type_allocated = false; } // The real name goes here: store.data.bytes.stored_name = bun.PathString.init(this.name); // The pretty name goes here: if (strings.hasPrefixComptime(this.name, base_public_path_with_default_suffix)) { b.name = bun.String.createUTF8(this.name[base_public_path_with_default_suffix.len..]); } else if (this.name.len > 0) { b.name = bun.String.createUTF8(this.name); } this.cached_blob = b; } return this.cached_blob.?; } }; pub const LazySourceMap = union(enum) { serialized: SerializedSourceMap, parsed: *SourceMap.ParsedSourceMap, none, /// It probably is not possible to run two decoding jobs on the same file var init_lock: bun.Lock = .{}; pub fn load(this: *LazySourceMap) ?*SourceMap.ParsedSourceMap { init_lock.lock(); defer init_lock.unlock(); return switch (this.*) { .none => null, .parsed => |map| map, .serialized => |serialized| { var stored = switch (SourceMap.Mapping.parse( bun.default_allocator, serialized.mappingVLQ(), null, std.math.maxInt(i32), std.math.maxInt(i32), )) { .success => |x| x, .fail => { this.* = .none; return null; }, }; const source_files = serialized.sourceFileNames(); const slices = bun.default_allocator.alloc(?[]u8, source_files.len * 2) catch bun.outOfMemory(); const file_names: [][]const u8 = @ptrCast(slices[0..source_files.len]); const decompressed_contents_slice = slices[source_files.len..][0..source_files.len]; for (file_names, source_files) |*dest, src| { dest.* = src.slice(serialized.bytes); } @memset(decompressed_contents_slice, null); const data = bun.new(SerializedSourceMap.Loaded, .{ .map = serialized, .decompressed_files = decompressed_contents_slice, }); stored.external_source_names = file_names; stored.underlying_provider = .{ .data = @truncate(@intFromPtr(data)) }; stored.is_standalone_module_graph = true; const parsed = stored.new(); // allocate this on the heap parsed.ref(); // never free this.* = .{ .parsed = parsed }; return parsed; }, }; } }; pub const Offsets = extern struct { byte_count: usize = 0, modules_ptr: bun.StringPointer = .{}, entry_point_id: u32 = 0, }; const trailer = "\n---- Bun! ----\n"; pub fn fromBytes(allocator: std.mem.Allocator, raw_bytes: []u8, offsets: Offsets) !StandaloneModuleGraph { if (raw_bytes.len == 0) return StandaloneModuleGraph{ .files = bun.StringArrayHashMap(File).init(allocator), }; const modules_list_bytes = sliceTo(raw_bytes, offsets.modules_ptr); const modules_list: []align(1) const CompiledModuleGraphFile = std.mem.bytesAsSlice(CompiledModuleGraphFile, modules_list_bytes); if (offsets.entry_point_id > modules_list.len) { return error.@"Corrupted module graph: entry point ID is greater than module list count"; } var modules = bun.StringArrayHashMap(File).init(allocator); try modules.ensureTotalCapacity(modules_list.len); for (modules_list) |module| { modules.putAssumeCapacity( sliceToZ(raw_bytes, module.name), File{ .name = sliceToZ(raw_bytes, module.name), .loader = module.loader, .contents = sliceToZ(raw_bytes, module.contents), .sourcemap = if (module.sourcemap.length > 0) .{ .serialized = .{ .bytes = @alignCast(sliceTo(raw_bytes, module.sourcemap)), } } else .none, .bytecode = if (module.bytecode.length > 0) @constCast(sliceTo(raw_bytes, module.bytecode)) else &.{}, .module_format = module.module_format, }, ); } modules.lockPointers(); // make the pointers stable forever return StandaloneModuleGraph{ .bytes = raw_bytes[0..offsets.byte_count], .files = modules, .entry_point_id = offsets.entry_point_id, }; } fn sliceTo(bytes: []const u8, ptr: bun.StringPointer) []const u8 { if (ptr.length == 0) return ""; return bytes[ptr.offset..][0..ptr.length]; } fn sliceToZ(bytes: []const u8, ptr: bun.StringPointer) [:0]const u8 { if (ptr.length == 0) return ""; return bytes[ptr.offset..][0..ptr.length :0]; } pub fn toBytes(allocator: std.mem.Allocator, prefix: []const u8, output_files: []const bun.options.OutputFile, output_format: bun.options.Format) ![]u8 { var serialize_trace = bun.tracy.traceNamed(@src(), "StandaloneModuleGraph.serialize"); defer serialize_trace.end(); var entry_point_id: ?usize = null; var string_builder = bun.StringBuilder{}; var module_count: usize = 0; for (output_files) |output_file| { string_builder.countZ(output_file.dest_path); string_builder.countZ(prefix); if (output_file.value == .buffer) { if (output_file.output_kind == .sourcemap) { // This is an over-estimation to ensure that we allocate // enough memory for the source-map contents. Calculating // the exact amount is not possible without allocating as it // involves a JSON parser. string_builder.cap += output_file.value.buffer.bytes.len * 2; } else if (output_file.output_kind == .bytecode) { // Allocate up to 256 byte alignment for bytecode string_builder.cap += (output_file.value.buffer.bytes.len + 255) / 256 * 256 + 256; } else { if (entry_point_id == null) { if (output_file.output_kind == .@"entry-point") { entry_point_id = module_count; } } string_builder.countZ(output_file.value.buffer.bytes); module_count += 1; } } } if (module_count == 0 or entry_point_id == null) return &[_]u8{}; string_builder.cap += @sizeOf(CompiledModuleGraphFile) * output_files.len; string_builder.cap += trailer.len; string_builder.cap += 16; string_builder.cap += @sizeOf(Offsets); try string_builder.allocate(allocator); var modules = try std.ArrayList(CompiledModuleGraphFile).initCapacity(allocator, module_count); var source_map_header_list = std.ArrayList(u8).init(allocator); defer source_map_header_list.deinit(); var source_map_string_list = std.ArrayList(u8).init(allocator); defer source_map_string_list.deinit(); var source_map_arena = bun.ArenaAllocator.init(allocator); defer source_map_arena.deinit(); for (output_files) |output_file| { if (!output_file.output_kind.isFileInStandaloneMode()) { continue; } if (output_file.value != .buffer) { continue; } const dest_path = bun.strings.removeLeadingDotSlash(output_file.dest_path); const bytecode: StringPointer = brk: { if (output_file.bytecode_index != std.math.maxInt(u32)) { // Use up to 256 byte alignment for bytecode // Not aligning it correctly will cause a runtime assertion error, or a segfault. const bytecode = output_files[output_file.bytecode_index].value.buffer.bytes; const aligned = std.mem.alignInSlice(string_builder.writable(), 128).?; @memcpy(aligned[0..bytecode.len], bytecode[0..bytecode.len]); const unaligned_space = aligned[bytecode.len..]; const offset = @intFromPtr(aligned.ptr) - @intFromPtr(string_builder.ptr.?); const len = bytecode.len + @min(unaligned_space.len, 128); string_builder.len += len; break :brk StringPointer{ .offset = @truncate(offset), .length = @truncate(len) }; } else { break :brk .{}; } }; var module = CompiledModuleGraphFile{ .name = string_builder.fmtAppendCountZ("{s}{s}", .{ prefix, dest_path, }), .loader = output_file.loader, .contents = string_builder.appendCountZ(output_file.value.buffer.bytes), .encoding = switch (output_file.loader) { .js, .jsx, .ts, .tsx => .latin1, else => .binary, }, .module_format = if (output_file.loader.isJavaScriptLike()) switch (output_format) { .cjs => .cjs, .esm => .esm, else => .none, } else .none, .bytecode = bytecode, }; if (output_file.source_map_index != std.math.maxInt(u32)) { defer source_map_header_list.clearRetainingCapacity(); defer source_map_string_list.clearRetainingCapacity(); _ = source_map_arena.reset(.retain_capacity); try serializeJsonSourceMapForStandalone( &source_map_header_list, &source_map_string_list, source_map_arena.allocator(), output_files[output_file.source_map_index].value.buffer.bytes, ); module.sourcemap = string_builder.addConcat(&.{ source_map_header_list.items, source_map_string_list.items, }); } modules.appendAssumeCapacity(module); } const offsets = Offsets{ .entry_point_id = @as(u32, @truncate(entry_point_id.?)), .modules_ptr = string_builder.appendCount(std.mem.sliceAsBytes(modules.items)), .byte_count = string_builder.len, }; _ = string_builder.append(std.mem.asBytes(&offsets)); _ = string_builder.append(trailer); const output_bytes = string_builder.ptr.?[0..string_builder.len]; if (comptime Environment.isDebug) { // An expensive sanity check: var graph = try fromBytes(allocator, @alignCast(output_bytes), offsets); defer { graph.files.unlockPointers(); graph.files.deinit(); } bun.assert_eql(graph.files.count(), modules.items.len); } return output_bytes; } const page_size = if (Environment.isLinux and Environment.isAarch64) // some linux distros do 64 KB pages on aarch64 64 * 1024 else std.mem.page_size; pub fn inject(bytes: []const u8, self_exe: [:0]const u8) bun.FileDescriptor { var buf: bun.PathBuffer = undefined; var zname: [:0]const u8 = bun.span(bun.fs.FileSystem.instance.tmpname("bun-build", &buf, @as(u64, @bitCast(std.time.milliTimestamp()))) catch |err| { Output.prettyErrorln("error: failed to get temporary file name: {s}", .{@errorName(err)}); Global.exit(1); }); const cleanup = struct { pub fn toClean(name: [:0]const u8, fd: bun.FileDescriptor) void { _ = Syscall.close(fd); _ = Syscall.unlink(name); } }.toClean; const cloned_executable_fd: bun.FileDescriptor = brk: { if (comptime Environment.isWindows) { // copy self and then open it for writing var in_buf: bun.WPathBuffer = undefined; strings.copyU8IntoU16(&in_buf, self_exe); in_buf[self_exe.len] = 0; const in = in_buf[0..self_exe.len :0]; var out_buf: bun.WPathBuffer = undefined; strings.copyU8IntoU16(&out_buf, zname); out_buf[zname.len] = 0; const out = out_buf[0..zname.len :0]; bun.copyFile(in, out).unwrap() catch |err| { Output.prettyErrorln("error: failed to copy bun executable into temporary file: {s}", .{@errorName(err)}); Global.exit(1); }; const file = bun.sys.openFileAtWindows( bun.invalid_fd, out, // access_mask w.SYNCHRONIZE | w.GENERIC_WRITE | w.DELETE, // create disposition w.FILE_OPEN, // create options w.FILE_SYNCHRONOUS_IO_NONALERT | w.FILE_OPEN_REPARSE_POINT, ).unwrap() catch |e| { Output.prettyErrorln("error: failed to open temporary file to copy bun into\n{}", .{e}); Global.exit(1); }; break :brk file; } if (comptime Environment.isMac) { // if we're on a mac, use clonefile() if we can // failure is okay, clonefile is just a fast path. if (Syscall.clonefile(self_exe, zname) == .result) { switch (Syscall.open(zname, bun.O.RDWR | bun.O.CLOEXEC, 0)) { .result => |res| break :brk res, .err => {}, } } } // otherwise, just copy the file const fd = brk2: { var tried_changing_abs_dir = false; for (0..3) |retry| { switch (Syscall.open(zname, bun.O.CLOEXEC | bun.O.RDWR | bun.O.CREAT, 0)) { .result => |res| break :brk2 res, .err => |err| { if (retry < 2) { // they may not have write access to the present working directory // // but we want to default to it since it's the // least likely to need to be copied due to // renameat() across filesystems // // so in the event of a failure, we try to // we retry using the tmp dir // // but we only do that once because otherwise it's just silly if (!tried_changing_abs_dir) { tried_changing_abs_dir = true; const zname_z = bun.strings.concat(bun.default_allocator, &.{ bun.fs.FileSystem.instance.fs.tmpdirPath(), std.fs.path.sep_str, zname, &.{0}, }) catch bun.outOfMemory(); zname = zname_z[0..zname_z.len -| 1 :0]; continue; } switch (err.getErrno()) { // try again .PERM, .AGAIN, .BUSY => continue, else => break, } Output.prettyErrorln("error: failed to open temporary file to copy bun into\n{}", .{err}); Global.exit(1); } }, } } unreachable; }; const self_fd = brk2: { for (0..3) |retry| { switch (Syscall.open(self_exe, bun.O.CLOEXEC | bun.O.RDONLY, 0)) { .result => |res| break :brk2 res, .err => |err| { if (retry < 2) { switch (err.getErrno()) { // try again .PERM, .AGAIN, .BUSY => continue, else => {}, } } Output.prettyErrorln("error: failed to open bun executable to copy from as read-only\n{}", .{err}); cleanup(zname, fd); Global.exit(1); }, } } unreachable; }; defer _ = Syscall.close(self_fd); bun.copyFile(self_fd.cast(), fd.cast()).unwrap() catch |err| { Output.prettyErrorln("error: failed to copy bun executable into temporary file: {s}", .{@errorName(err)}); cleanup(zname, fd); Global.exit(1); }; break :brk fd; }; var total_byte_count: usize = undefined; if (Environment.isWindows) { total_byte_count = bytes.len + 8 + (Syscall.setFileOffsetToEndWindows(cloned_executable_fd).unwrap() catch |err| { Output.prettyErrorln("error: failed to seek to end of temporary file\n{}", .{err}); cleanup(zname, cloned_executable_fd); Global.exit(1); }); } else { const seek_position = @as(u64, @intCast(brk: { const fstat = switch (Syscall.fstat(cloned_executable_fd)) { .result => |res| res, .err => |err| { Output.prettyErrorln("{}", .{err}); cleanup(zname, cloned_executable_fd); Global.exit(1); }, }; break :brk @max(fstat.size, 0); })); total_byte_count = seek_position + bytes.len + 8; // From https://man7.org/linux/man-pages/man2/lseek.2.html // // lseek() allows the file offset to be set beyond the end of the // file (but this does not change the size of the file). If data is // later written at this point, subsequent reads of the data in the // gap (a "hole") return null bytes ('\0') until data is actually // written into the gap. // switch (Syscall.setFileOffset(cloned_executable_fd, seek_position)) { .err => |err| { Output.prettyErrorln( "{}\nwhile seeking to end of temporary file (pos: {d})", .{ err, seek_position, }, ); cleanup(zname, cloned_executable_fd); Global.exit(1); }, else => {}, } } var remain = bytes; while (remain.len > 0) { switch (Syscall.write(cloned_executable_fd, bytes)) { .result => |written| remain = remain[written..], .err => |err| { Output.prettyErrorln("error: failed to write to temporary file\n{}", .{err}); cleanup(zname, cloned_executable_fd); Global.exit(1); }, } } // the final 8 bytes in the file are the length of the module graph with padding, excluding the trailer and offsets _ = Syscall.write(cloned_executable_fd, std.mem.asBytes(&total_byte_count)); if (comptime !Environment.isWindows) { _ = bun.C.fchmod(cloned_executable_fd.int(), 0o777); } return cloned_executable_fd; } pub const CompileTarget = @import("./compile_target.zig"); pub fn download(allocator: std.mem.Allocator, target: *const CompileTarget, env: *bun.DotEnv.Loader) ![:0]const u8 { var exe_path_buf: bun.PathBuffer = undefined; var version_str_buf: [1024]u8 = undefined; const version_str = try std.fmt.bufPrintZ(&version_str_buf, "{}", .{target}); var needs_download: bool = true; const dest_z = target.exePath(&exe_path_buf, version_str, env, &needs_download); if (needs_download) { try target.downloadToPath(env, allocator, dest_z); } return try allocator.dupeZ(u8, dest_z); } pub fn toExecutable( target: *const CompileTarget, allocator: std.mem.Allocator, output_files: []const bun.options.OutputFile, root_dir: std.fs.Dir, module_prefix: []const u8, outfile: []const u8, env: *bun.DotEnv.Loader, output_format: bun.options.Format, ) !void { const bytes = try toBytes(allocator, module_prefix, output_files, output_format); if (bytes.len == 0) return; const fd = inject( bytes, if (target.isDefault()) bun.selfExePath() catch |err| { Output.err(err, "failed to get self executable path", .{}); Global.exit(1); } else download(allocator, target, env) catch |err| { Output.err(err, "failed to download cross-compiled bun executable", .{}); Global.exit(1); }, ); fd.assertKind(.system); if (Environment.isWindows) { var outfile_buf: bun.OSPathBuffer = undefined; const outfile_slice = brk: { const outfile_w = bun.strings.toWPathNormalized(&outfile_buf, std.fs.path.basenameWindows(outfile)); bun.assert(outfile_w.ptr == &outfile_buf); const outfile_buf_u16 = bun.reinterpretSlice(u16, &outfile_buf); outfile_buf_u16[outfile_w.len] = 0; break :brk outfile_buf_u16[0..outfile_w.len :0]; }; bun.C.moveOpenedFileAtLoose(fd, bun.toFD(root_dir.fd), outfile_slice, true).unwrap() catch |err| { if (err == error.EISDIR) { Output.errGeneric("{} is a directory. Please choose a different --outfile or delete the directory", .{bun.fmt.utf16(outfile_slice)}); } else { Output.err(err, "failed to move executable to result path", .{}); } _ = bun.C.deleteOpenedFile(fd); Global.exit(1); }; return; } var buf: bun.PathBuffer = undefined; const temp_location = bun.getFdPath(fd, &buf) catch |err| { Output.prettyErrorln("error: failed to get path for fd: {s}", .{@errorName(err)}); Global.exit(1); }; if (comptime Environment.isMac) { if (target.os == .mac) { var signer = std.process.Child.init( &.{ "codesign", "--remove-signature", temp_location, }, bun.default_allocator, ); if (bun.logger.Log.default_log_level.atLeast(.verbose)) { signer.stdout_behavior = .Inherit; signer.stderr_behavior = .Inherit; signer.stdin_behavior = .Inherit; } else { signer.stdout_behavior = .Ignore; signer.stderr_behavior = .Ignore; signer.stdin_behavior = .Ignore; } _ = signer.spawnAndWait() catch {}; } } bun.C.moveFileZWithHandle( fd, bun.FD.cwd(), bun.sliceTo(&(try std.posix.toPosixPath(temp_location)), 0), bun.toFD(root_dir.fd), bun.sliceTo(&(try std.posix.toPosixPath(std.fs.path.basename(outfile))), 0), ) catch |err| { if (err == error.IsDir) { Output.prettyErrorln("error: {} is a directory. Please choose a different --outfile or delete the directory", .{bun.fmt.quote(outfile)}); } else { Output.prettyErrorln("error: failed to rename {s} to {s}: {s}", .{ temp_location, outfile, @errorName(err) }); } _ = Syscall.unlink( &(try std.posix.toPosixPath(temp_location)), ); Global.exit(1); }; } pub fn fromExecutable(allocator: std.mem.Allocator) !?StandaloneModuleGraph { // Do not invoke libuv here. const self_exe = openSelf() catch return null; defer _ = Syscall.close(self_exe); var trailer_bytes: [4096]u8 = undefined; std.posix.lseek_END(self_exe.cast(), -4096) catch return null; var read_amount: usize = 0; while (read_amount < trailer_bytes.len) { switch (Syscall.read(self_exe, trailer_bytes[read_amount..])) { .result => |read| { if (read == 0) return null; read_amount += read; }, .err => { return null; }, } } if (read_amount < trailer.len + @sizeOf(usize) + 32) // definitely missing data return null; var end = @as([]u8, &trailer_bytes).ptr + read_amount - @sizeOf(usize); const total_byte_count: usize = @as(usize, @bitCast(end[0..8].*)); if (total_byte_count > std.math.maxInt(u32) or total_byte_count < 4096) { // sanity check: the total byte count should never be more than 4 GB // bun is at least like 30 MB so if it reports a size less than 4096 bytes then something is wrong return null; } end -= trailer.len; if (!bun.strings.hasPrefixComptime(end[0..trailer.len], trailer)) { // invalid trailer return null; } end -= @sizeOf(Offsets); const offsets: Offsets = std.mem.bytesAsValue(Offsets, end[0..@sizeOf(Offsets)]).*; if (offsets.byte_count >= total_byte_count) { // if we hit this branch then the file is corrupted and we should just give up return null; } var to_read = try bun.default_allocator.alloc(u8, offsets.byte_count); var to_read_from = to_read; // Reading the data and making sure it's page-aligned + won't crash due // to out of bounds using mmap() is very complicated. // we just read the whole thing into memory for now. // at the very least // if you have not a ton of code, we only do a single read() call if (Environment.allow_assert or offsets.byte_count > 1024 * 3) { const offset_from_end = trailer_bytes.len - (@intFromPtr(end) - @intFromPtr(@as([]u8, &trailer_bytes).ptr)); std.posix.lseek_END(self_exe.cast(), -@as(i64, @intCast(offset_from_end + offsets.byte_count))) catch return null; if (comptime Environment.allow_assert) { // actually we just want to verify this logic is correct in development if (offsets.byte_count <= 1024 * 3) { to_read_from = try bun.default_allocator.alloc(u8, offsets.byte_count); } } var remain = to_read_from; while (remain.len > 0) { switch (Syscall.read(self_exe, remain)) { .result => |read| { if (read == 0) return null; remain = remain[read..]; }, .err => { bun.default_allocator.free(to_read); return null; }, } } } if (offsets.byte_count <= 1024 * 3) { // we already have the bytes end -= offsets.byte_count; @memcpy(to_read[0..offsets.byte_count], end[0..offsets.byte_count]); if (comptime Environment.allow_assert) { bun.assert(bun.strings.eqlLong(to_read, end[0..offsets.byte_count], true)); } } return try StandaloneModuleGraph.fromBytes(allocator, to_read, offsets); } /// heuristic: `bun build --compile` won't be supported if the name is "bun", "bunx", or "node". /// this is a cheap way to avoid the extra overhead of opening the executable, and also just makes sense. fn isBuiltInExe(comptime T: type, argv0: []const T) bool { if (argv0.len == 0) return false; if (argv0.len == 3) { if (bun.strings.eqlComptimeCheckLenWithType(T, argv0, bun.strings.literal(T, "bun"), false)) { return true; } } if (argv0.len == 4) { if (bun.strings.eqlComptimeCheckLenWithType(T, argv0, bun.strings.literal(T, "bunx"), false)) { return true; } if (bun.strings.eqlComptimeCheckLenWithType(T, argv0, bun.strings.literal(T, "node"), false)) { return true; } } if (comptime Environment.isDebug) { if (bun.strings.eqlComptimeCheckLenWithType(T, argv0, bun.strings.literal(T, "bun-debug"), true)) { return true; } if (bun.strings.eqlComptimeCheckLenWithType(T, argv0, bun.strings.literal(T, "bun-debugx"), true)) { return true; } } return false; } fn openSelf() std.fs.OpenSelfExeError!bun.FileDescriptor { if (!Environment.isWindows) { const argv = bun.argv; if (argv.len > 0) { if (isBuiltInExe(u8, argv[0])) { return error.FileNotFound; } } } switch (Environment.os) { .linux => { if (std.fs.openFileAbsoluteZ("/proc/self/exe", .{})) |easymode| { return bun.toFD(easymode.handle); } else |_| { if (bun.argv.len > 0) { // The user doesn't have /proc/ mounted, so now we just guess and hope for the best. var whichbuf: bun.PathBuffer = undefined; if (bun.which( &whichbuf, bun.getenvZ("PATH") orelse return error.FileNotFound, "", bun.argv[0], )) |path| { return bun.toFD((try std.fs.cwd().openFileZ(path, .{})).handle); } } return error.FileNotFound; } }, .mac => { // Use of MAX_PATH_BYTES here is valid as the resulting path is immediately // opened with no modification. const self_exe_path = try bun.selfExePath(); const file = try std.fs.openFileAbsoluteZ(self_exe_path.ptr, .{}); return bun.toFD(file.handle); }, .windows => { const image_path_unicode_string = std.os.windows.peb().ProcessParameters.ImagePathName; const image_path = image_path_unicode_string.Buffer.?[0 .. image_path_unicode_string.Length / 2]; var nt_path_buf: bun.WPathBuffer = undefined; const nt_path = bun.strings.addNTPathPrefix(&nt_path_buf, image_path); const basename_start = std.mem.lastIndexOfScalar(u16, nt_path, '\\') orelse return error.FileNotFound; const basename = nt_path[basename_start + 1 .. nt_path.len - ".exe".len]; if (isBuiltInExe(u16, basename)) { return error.FileNotFound; } return bun.sys.openFileAtWindows( bun.FileDescriptor.cwd(), nt_path, // access_mask w.SYNCHRONIZE | w.GENERIC_READ, // create disposition w.FILE_OPEN, // create options w.FILE_SYNCHRONOUS_IO_NONALERT | w.FILE_OPEN_REPARSE_POINT, ).unwrap() catch { return error.FileNotFound; }; }, else => @compileError("TODO"), } } /// Source map serialization in the bundler is specially designed to be /// loaded in memory as is. Source contents are compressed with ZSTD to /// reduce the file size, and mappings are stored as uncompressed VLQ. pub const SerializedSourceMap = struct { bytes: []const u8, /// Following the header bytes: /// - source_files_count number of StringPointer, file names /// - source_files_count number of StringPointer, zstd compressed contents /// - the mapping data, `map_vlq_length` bytes /// - all the StringPointer contents pub const Header = extern struct { source_files_count: u32, map_bytes_length: u32, }; pub fn header(map: SerializedSourceMap) *align(1) const Header { return @ptrCast(map.bytes.ptr); } pub fn mappingVLQ(map: SerializedSourceMap) []const u8 { const head = map.header(); const start = @sizeOf(Header) + head.source_files_count * @sizeOf(StringPointer) * 2; return map.bytes[start..][0..head.map_bytes_length]; } pub fn sourceFileNames(map: SerializedSourceMap) []align(1) const StringPointer { const head = map.header(); return @as([*]align(1) const StringPointer, @ptrCast(map.bytes[@sizeOf(Header)..]))[0..head.source_files_count]; } fn compressedSourceFiles(map: SerializedSourceMap) []align(1) const StringPointer { const head = map.header(); return @as([*]align(1) const StringPointer, @ptrCast(map.bytes[@sizeOf(Header)..]))[head.source_files_count..][0..head.source_files_count]; } /// Once loaded, this map stores additional data for keeping track of source code. pub const Loaded = struct { map: SerializedSourceMap, /// Only decompress source code once! Once a file is decompressed, /// it is stored here. Decompression failures are stored as an empty /// string, which will be treated as "no contents". decompressed_files: []?[]u8, pub fn sourceFileContents(this: Loaded, index: usize) ?[]const u8 { if (this.decompressed_files[index]) |decompressed| { return if (decompressed.len == 0) null else decompressed; } const compressed_codes = this.map.compressedSourceFiles(); const compressed_file = compressed_codes[@intCast(index)].slice(this.map.bytes); const size = bun.zstd.getDecompressedSize(compressed_file); const bytes = bun.default_allocator.alloc(u8, size) catch bun.outOfMemory(); const result = bun.zstd.decompress(bytes, compressed_file); if (result == .err) { bun.Output.warn("Source map decompression error: {s}", .{result.err}); bun.default_allocator.free(bytes); this.decompressed_files[index] = ""; return null; } const data = bytes[0..result.success]; this.decompressed_files[index] = data; return data; } }; }; pub fn serializeJsonSourceMapForStandalone( header_list: *std.ArrayList(u8), string_payload: *std.ArrayList(u8), arena: std.mem.Allocator, json_source: []const u8, ) !void { const out = header_list.writer(); const json_src = bun.logger.Source.initPathString("sourcemap.json", json_source); var log = bun.logger.Log.init(arena); defer log.deinit(); // the allocator given to the JS parser is not respected for all parts // of the parse, so we need to remember to reset the ast store bun.JSAst.Expr.Data.Store.reset(); bun.JSAst.Stmt.Data.Store.reset(); defer { bun.JSAst.Expr.Data.Store.reset(); bun.JSAst.Stmt.Data.Store.reset(); } var json = bun.JSON.ParseJSON(&json_src, &log, arena, false) catch return error.InvalidSourceMap; const mappings_str = json.get("mappings") orelse return error.InvalidSourceMap; if (mappings_str.data != .e_string) return error.InvalidSourceMap; const sources_content = switch ((json.get("sourcesContent") orelse return error.InvalidSourceMap).data) { .e_array => |arr| arr, else => return error.InvalidSourceMap, }; const sources_paths = switch ((json.get("sources") orelse return error.InvalidSourceMap).data) { .e_array => |arr| arr, else => return error.InvalidSourceMap, }; if (sources_content.items.len != sources_paths.items.len) { return error.InvalidSourceMap; } const map_vlq: []const u8 = mappings_str.data.e_string.slice(arena); try out.writeInt(u32, sources_paths.items.len, .little); try out.writeInt(u32, @intCast(map_vlq.len), .little); const string_payload_start_location = @sizeOf(u32) + @sizeOf(u32) + @sizeOf(bun.StringPointer) * sources_content.items.len * 2 + // path + source map_vlq.len; for (sources_paths.items.slice()) |item| { if (item.data != .e_string) return error.InvalidSourceMap; const decoded = try item.data.e_string.stringDecodedUTF8(arena); const offset = string_payload.items.len; try string_payload.appendSlice(decoded); const slice = bun.StringPointer{ .offset = @intCast(offset + string_payload_start_location), .length = @intCast(string_payload.items.len - offset), }; try out.writeInt(u32, slice.offset, .little); try out.writeInt(u32, slice.length, .little); } for (sources_content.items.slice()) |item| { if (item.data != .e_string) return error.InvalidSourceMap; const utf8 = try item.data.e_string.stringDecodedUTF8(arena); defer arena.free(utf8); const offset = string_payload.items.len; const bound = bun.zstd.compressBound(utf8.len); try string_payload.ensureUnusedCapacity(bound); const unused = string_payload.unusedCapacitySlice(); const compressed_result = bun.zstd.compress(unused, utf8, 1); if (compressed_result == .err) { bun.Output.panic("Unexpected error compressing sourcemap: {s}", .{bun.span(compressed_result.err)}); } string_payload.items.len += compressed_result.success; const slice = bun.StringPointer{ .offset = @intCast(offset + string_payload_start_location), .length = @intCast(string_payload.items.len - offset), }; try out.writeInt(u32, slice.offset, .little); try out.writeInt(u32, slice.length, .little); } try out.writeAll(map_vlq); bun.assert(header_list.items.len == string_payload_start_location); } };