From 11cc1c16fa36a7eb13cba1c43fb153ee6aca7b58 Mon Sep 17 00:00:00 2001 From: Andrew Kelley Date: Sat, 4 Feb 2023 16:14:43 -0700 Subject: make `@embedFile` support module-mapped names the same way as `@import` closes #14553 --- src/Module.zig | 54 ++++++++++++++++++++++++++++++++++++++++-------------- 1 file changed, 40 insertions(+), 14 deletions(-) (limited to 'src/Module.zig') diff --git a/src/Module.zig b/src/Module.zig index 3243c36ee5..e4cf0189cc 100644 --- a/src/Module.zig +++ b/src/Module.zig @@ -4865,14 +4865,31 @@ pub fn importFile( }; } -pub fn embedFile(mod: *Module, cur_file: *File, rel_file_path: []const u8) !*EmbedFile { +pub fn embedFile(mod: *Module, cur_file: *File, import_string: []const u8) !*EmbedFile { const gpa = mod.gpa; - // The resolved path is used as the key in the table, to detect if - // a file refers to the same as another, despite different relative paths. + if (cur_file.pkg.table.get(import_string)) |pkg| { + const resolved_path = try std.fs.path.resolve(gpa, &[_][]const u8{ + pkg.root_src_directory.path orelse ".", pkg.root_src_path, + }); + var keep_resolved_path = false; + defer if (!keep_resolved_path) gpa.free(resolved_path); + + const gop = try mod.embed_table.getOrPut(gpa, resolved_path); + errdefer assert(mod.embed_table.remove(resolved_path)); + if (gop.found_existing) return gop.value_ptr.*; + + const sub_file_path = try gpa.dupe(u8, pkg.root_src_path); + errdefer gpa.free(sub_file_path); + + return newEmbedFile(mod, pkg, sub_file_path, resolved_path, &keep_resolved_path, gop); + } + + // The resolved path is used as the key in the table, to detect if a file + // refers to the same as another, despite different relative paths. const cur_pkg_dir_path = cur_file.pkg.root_src_directory.path orelse "."; const resolved_path = try std.fs.path.resolve(gpa, &[_][]const u8{ - cur_pkg_dir_path, cur_file.sub_file_path, "..", rel_file_path, + cur_pkg_dir_path, cur_file.sub_file_path, "..", import_string, }); var keep_resolved_path = false; defer if (!keep_resolved_path) gpa.free(resolved_path); @@ -4881,9 +4898,6 @@ pub fn embedFile(mod: *Module, cur_file: *File, rel_file_path: []const u8) !*Emb errdefer assert(mod.embed_table.remove(resolved_path)); if (gop.found_existing) return gop.value_ptr.*; - const new_file = try gpa.create(EmbedFile); - errdefer gpa.destroy(new_file); - const resolved_root_path = try std.fs.path.resolve(gpa, &[_][]const u8{cur_pkg_dir_path}); defer gpa.free(resolved_root_path); @@ -4902,7 +4916,23 @@ pub fn embedFile(mod: *Module, cur_file: *File, rel_file_path: []const u8) !*Emb }; errdefer gpa.free(sub_file_path); - var file = try cur_file.pkg.root_src_directory.handle.openFile(sub_file_path, .{}); + return newEmbedFile(mod, cur_file.pkg, sub_file_path, resolved_path, &keep_resolved_path, gop); +} + +fn newEmbedFile( + mod: *Module, + pkg: *Package, + sub_file_path: []const u8, + resolved_path: []const u8, + keep_resolved_path: *bool, + gop: std.StringHashMapUnmanaged(*EmbedFile).GetOrPutResult, +) !*EmbedFile { + const gpa = mod.gpa; + + const new_file = try gpa.create(EmbedFile); + errdefer gpa.destroy(new_file); + + var file = try pkg.root_src_directory.handle.openFile(sub_file_path, .{}); defer file.close(); const actual_stat = try file.stat(); @@ -4915,10 +4945,6 @@ pub fn embedFile(mod: *Module, cur_file: *File, rel_file_path: []const u8) !*Emb const bytes = try file.readToEndAllocOptions(gpa, std.math.maxInt(u32), size_usize, 1, 0); errdefer gpa.free(bytes); - log.debug("new embedFile. resolved_root_path={s}, resolved_path={s}, sub_file_path={s}, rel_file_path={s}", .{ - resolved_root_path, resolved_path, sub_file_path, rel_file_path, - }); - if (mod.comp.whole_cache_manifest) |whole_cache_manifest| { const copied_resolved_path = try gpa.dupe(u8, resolved_path); errdefer gpa.free(copied_resolved_path); @@ -4927,13 +4953,13 @@ pub fn embedFile(mod: *Module, cur_file: *File, rel_file_path: []const u8) !*Emb try whole_cache_manifest.addFilePostContents(copied_resolved_path, bytes, stat); } - keep_resolved_path = true; // It's now owned by embed_table. + keep_resolved_path.* = true; // It's now owned by embed_table. gop.value_ptr.* = new_file; new_file.* = .{ .sub_file_path = sub_file_path, .bytes = bytes, .stat = stat, - .pkg = cur_file.pkg, + .pkg = pkg, .owner_decl = undefined, // Set by Sema immediately after this function returns. }; return new_file; -- cgit v1.2.3 From 9cb52ca6ce7043ba0ce08d5650ac542075f10685 Mon Sep 17 00:00:00 2001 From: Andrew Kelley Date: Sun, 5 Feb 2023 19:39:04 -0700 Subject: move the cache system from compiler to std lib --- CMakeLists.txt | 5 +- lib/std/Build.zig | 2 + lib/std/Build/Cache.zig | 1276 ++++++++++++++++++++++++++++++++++ lib/std/Build/Cache/DepTokenizer.zig | 1069 ++++++++++++++++++++++++++++ src/Cache.zig | 1265 --------------------------------- src/Compilation.zig | 92 ++- src/DepTokenizer.zig | 1069 ---------------------------- src/Module.zig | 2 +- src/Package.zig | 2 +- src/glibc.zig | 2 +- src/link.zig | 2 +- src/link/Coff/lld.zig | 2 +- src/link/Elf.zig | 2 +- src/link/MachO.zig | 2 +- src/link/MachO/zld.zig | 2 +- src/link/Wasm.zig | 2 +- src/main.zig | 4 +- src/mingw.zig | 2 +- 18 files changed, 2404 insertions(+), 2398 deletions(-) create mode 100644 lib/std/Build/Cache.zig create mode 100644 lib/std/Build/Cache/DepTokenizer.zig delete mode 100644 src/Cache.zig delete mode 100644 src/DepTokenizer.zig (limited to 'src/Module.zig') diff --git a/CMakeLists.txt b/CMakeLists.txt index f8029fdcde..448fb400b6 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -216,6 +216,9 @@ set(ZIG_STAGE2_SOURCES "${CMAKE_SOURCE_DIR}/lib/std/atomic/stack.zig" "${CMAKE_SOURCE_DIR}/lib/std/base64.zig" "${CMAKE_SOURCE_DIR}/lib/std/buf_map.zig" + "${CMAKE_SOURCE_DIR}/lib/std/Build.zig" + "${CMAKE_SOURCE_DIR}/lib/std/Build/Cache.zig" + "${CMAKE_SOURCE_DIR}/lib/std/Build/Cache/DepTokenizer.zig" "${CMAKE_SOURCE_DIR}/lib/std/builtin.zig" "${CMAKE_SOURCE_DIR}/lib/std/c.zig" "${CMAKE_SOURCE_DIR}/lib/std/c/linux.zig" @@ -523,9 +526,7 @@ set(ZIG_STAGE2_SOURCES "${CMAKE_SOURCE_DIR}/lib/std/zig/tokenizer.zig" "${CMAKE_SOURCE_DIR}/src/Air.zig" "${CMAKE_SOURCE_DIR}/src/AstGen.zig" - "${CMAKE_SOURCE_DIR}/src/Cache.zig" "${CMAKE_SOURCE_DIR}/src/Compilation.zig" - "${CMAKE_SOURCE_DIR}/src/DepTokenizer.zig" "${CMAKE_SOURCE_DIR}/src/Liveness.zig" "${CMAKE_SOURCE_DIR}/src/Module.zig" "${CMAKE_SOURCE_DIR}/src/Package.zig" diff --git a/lib/std/Build.zig b/lib/std/Build.zig index a67ded03e8..428a11948d 100644 --- a/lib/std/Build.zig +++ b/lib/std/Build.zig @@ -19,6 +19,8 @@ const NativeTargetInfo = std.zig.system.NativeTargetInfo; const Sha256 = std.crypto.hash.sha2.Sha256; const Build = @This(); +pub const Cache = @import("Build/Cache.zig"); + /// deprecated: use `CompileStep`. pub const LibExeObjStep = CompileStep; /// deprecated: use `Build`. diff --git a/lib/std/Build/Cache.zig b/lib/std/Build/Cache.zig new file mode 100644 index 0000000000..e316322c68 --- /dev/null +++ b/lib/std/Build/Cache.zig @@ -0,0 +1,1276 @@ +//! Manages `zig-cache` directories. +//! This is not a general-purpose cache. It is designed to be fast and simple, +//! not to withstand attacks using specially-crafted input. + +pub const Directory = struct { + /// This field is redundant for operations that can act on the open directory handle + /// directly, but it is needed when passing the directory to a child process. + /// `null` means cwd. + path: ?[]const u8, + handle: std.fs.Dir, + + pub fn join(self: Directory, allocator: Allocator, paths: []const []const u8) ![]u8 { + if (self.path) |p| { + // TODO clean way to do this with only 1 allocation + const part2 = try std.fs.path.join(allocator, paths); + defer allocator.free(part2); + return std.fs.path.join(allocator, &[_][]const u8{ p, part2 }); + } else { + return std.fs.path.join(allocator, paths); + } + } + + pub fn joinZ(self: Directory, allocator: Allocator, paths: []const []const u8) ![:0]u8 { + if (self.path) |p| { + // TODO clean way to do this with only 1 allocation + const part2 = try std.fs.path.join(allocator, paths); + defer allocator.free(part2); + return std.fs.path.joinZ(allocator, &[_][]const u8{ p, part2 }); + } else { + return std.fs.path.joinZ(allocator, paths); + } + } + + /// Whether or not the handle should be closed, or the path should be freed + /// is determined by usage, however this function is provided for convenience + /// if it happens to be what the caller needs. + pub fn closeAndFree(self: *Directory, gpa: Allocator) void { + self.handle.close(); + if (self.path) |p| gpa.free(p); + self.* = undefined; + } +}; + +gpa: Allocator, +manifest_dir: fs.Dir, +hash: HashHelper = .{}, +/// This value is accessed from multiple threads, protected by mutex. +recent_problematic_timestamp: i128 = 0, +mutex: std.Thread.Mutex = .{}, + +/// A set of strings such as the zig library directory or project source root, which +/// are stripped from the file paths before putting into the cache. They +/// are replaced with single-character indicators. This is not to save +/// space but to eliminate absolute file paths. This improves portability +/// and usefulness of the cache for advanced use cases. +prefixes_buffer: [3]Directory = undefined, +prefixes_len: usize = 0, + +pub const DepTokenizer = @import("Cache/DepTokenizer.zig"); + +const Cache = @This(); +const std = @import("std"); +const builtin = @import("builtin"); +const crypto = std.crypto; +const fs = std.fs; +const assert = std.debug.assert; +const testing = std.testing; +const mem = std.mem; +const fmt = std.fmt; +const Allocator = std.mem.Allocator; +const log = std.log.scoped(.cache); + +pub fn addPrefix(cache: *Cache, directory: Directory) void { + if (directory.path) |p| { + log.debug("Cache.addPrefix {d} {s}", .{ cache.prefixes_len, p }); + } + cache.prefixes_buffer[cache.prefixes_len] = directory; + cache.prefixes_len += 1; +} + +/// Be sure to call `Manifest.deinit` after successful initialization. +pub fn obtain(cache: *Cache) Manifest { + return Manifest{ + .cache = cache, + .hash = cache.hash, + .manifest_file = null, + .manifest_dirty = false, + .hex_digest = undefined, + }; +} + +pub fn prefixes(cache: *const Cache) []const Directory { + return cache.prefixes_buffer[0..cache.prefixes_len]; +} + +const PrefixedPath = struct { + prefix: u8, + sub_path: []u8, +}; + +fn findPrefix(cache: *const Cache, file_path: []const u8) !PrefixedPath { + const gpa = cache.gpa; + const resolved_path = try fs.path.resolve(gpa, &[_][]const u8{file_path}); + errdefer gpa.free(resolved_path); + return findPrefixResolved(cache, resolved_path); +} + +/// Takes ownership of `resolved_path` on success. +fn findPrefixResolved(cache: *const Cache, resolved_path: []u8) !PrefixedPath { + const gpa = cache.gpa; + const prefixes_slice = cache.prefixes(); + var i: u8 = 1; // Start at 1 to skip over checking the null prefix. + while (i < prefixes_slice.len) : (i += 1) { + const p = prefixes_slice[i].path.?; + if (mem.startsWith(u8, resolved_path, p)) { + // +1 to skip over the path separator here + const sub_path = try gpa.dupe(u8, resolved_path[p.len + 1 ..]); + gpa.free(resolved_path); + return PrefixedPath{ + .prefix = @intCast(u8, i), + .sub_path = sub_path, + }; + } else { + log.debug("'{s}' does not start with '{s}'", .{ resolved_path, p }); + } + } + + return PrefixedPath{ + .prefix = 0, + .sub_path = resolved_path, + }; +} + +/// This is 128 bits - Even with 2^54 cache entries, the probably of a collision would be under 10^-6 +pub const bin_digest_len = 16; +pub const hex_digest_len = bin_digest_len * 2; +pub const BinDigest = [bin_digest_len]u8; + +const manifest_file_size_max = 50 * 1024 * 1024; + +/// The type used for hashing file contents. Currently, this is SipHash128(1, 3), because it +/// provides enough collision resistance for the Manifest use cases, while being one of our +/// fastest options right now. +pub const Hasher = crypto.auth.siphash.SipHash128(1, 3); + +/// Initial state, that can be copied. +pub const hasher_init: Hasher = Hasher.init(&[_]u8{0} ** Hasher.key_length); + +pub const File = struct { + prefixed_path: ?PrefixedPath, + max_file_size: ?usize, + stat: Stat, + bin_digest: BinDigest, + contents: ?[]const u8, + + pub const Stat = struct { + inode: fs.File.INode, + size: u64, + mtime: i128, + }; + + pub fn deinit(self: *File, gpa: Allocator) void { + if (self.prefixed_path) |pp| { + gpa.free(pp.sub_path); + self.prefixed_path = null; + } + if (self.contents) |contents| { + gpa.free(contents); + self.contents = null; + } + self.* = undefined; + } +}; + +pub const HashHelper = struct { + hasher: Hasher = hasher_init, + + /// Record a slice of bytes as an dependency of the process being cached + pub fn addBytes(hh: *HashHelper, bytes: []const u8) void { + hh.hasher.update(mem.asBytes(&bytes.len)); + hh.hasher.update(bytes); + } + + pub fn addOptionalBytes(hh: *HashHelper, optional_bytes: ?[]const u8) void { + hh.add(optional_bytes != null); + hh.addBytes(optional_bytes orelse return); + } + + pub fn addListOfBytes(hh: *HashHelper, list_of_bytes: []const []const u8) void { + hh.add(list_of_bytes.len); + for (list_of_bytes) |bytes| hh.addBytes(bytes); + } + + /// Convert the input value into bytes and record it as a dependency of the process being cached. + pub fn add(hh: *HashHelper, x: anytype) void { + switch (@TypeOf(x)) { + std.builtin.Version => { + hh.add(x.major); + hh.add(x.minor); + hh.add(x.patch); + }, + std.Target.Os.TaggedVersionRange => { + switch (x) { + .linux => |linux| { + hh.add(linux.range.min); + hh.add(linux.range.max); + hh.add(linux.glibc); + }, + .windows => |windows| { + hh.add(windows.min); + hh.add(windows.max); + }, + .semver => |semver| { + hh.add(semver.min); + hh.add(semver.max); + }, + .none => {}, + } + }, + else => switch (@typeInfo(@TypeOf(x))) { + .Bool, .Int, .Enum, .Array => hh.addBytes(mem.asBytes(&x)), + else => @compileError("unable to hash type " ++ @typeName(@TypeOf(x))), + }, + } + } + + pub fn addOptional(hh: *HashHelper, optional: anytype) void { + hh.add(optional != null); + hh.add(optional orelse return); + } + + /// Returns a hex encoded hash of the inputs, without modifying state. + pub fn peek(hh: HashHelper) [hex_digest_len]u8 { + var copy = hh; + return copy.final(); + } + + pub fn peekBin(hh: HashHelper) BinDigest { + var copy = hh; + var bin_digest: BinDigest = undefined; + copy.hasher.final(&bin_digest); + return bin_digest; + } + + /// Returns a hex encoded hash of the inputs, mutating the state of the hasher. + pub fn final(hh: *HashHelper) [hex_digest_len]u8 { + var bin_digest: BinDigest = undefined; + hh.hasher.final(&bin_digest); + + var out_digest: [hex_digest_len]u8 = undefined; + _ = std.fmt.bufPrint( + &out_digest, + "{s}", + .{std.fmt.fmtSliceHexLower(&bin_digest)}, + ) catch unreachable; + return out_digest; + } +}; + +pub const Lock = struct { + manifest_file: fs.File, + + pub fn release(lock: *Lock) void { + if (builtin.os.tag == .windows) { + // Windows does not guarantee that locks are immediately unlocked when + // the file handle is closed. See LockFileEx documentation. + lock.manifest_file.unlock(); + } + + lock.manifest_file.close(); + lock.* = undefined; + } +}; + +pub const Manifest = struct { + cache: *Cache, + /// Current state for incremental hashing. + hash: HashHelper, + manifest_file: ?fs.File, + manifest_dirty: bool, + /// Set this flag to true before calling hit() in order to indicate that + /// upon a cache hit, the code using the cache will not modify the files + /// within the cache directory. This allows multiple processes to utilize + /// the same cache directory at the same time. + want_shared_lock: bool = true, + have_exclusive_lock: bool = false, + // Indicate that we want isProblematicTimestamp to perform a filesystem write in + // order to obtain a problematic timestamp for the next call. Calls after that + // will then use the same timestamp, to avoid unnecessary filesystem writes. + want_refresh_timestamp: bool = true, + files: std.ArrayListUnmanaged(File) = .{}, + hex_digest: [hex_digest_len]u8, + /// Populated when hit() returns an error because of one + /// of the files listed in the manifest. + failed_file_index: ?usize = null, + /// Keeps track of the last time we performed a file system write to observe + /// what time the file system thinks it is, according to its own granularity. + recent_problematic_timestamp: i128 = 0, + + /// Add a file as a dependency of process being cached. When `hit` is + /// called, the file's contents will be checked to ensure that it matches + /// the contents from previous times. + /// + /// Max file size will be used to determine the amount of space the file contents + /// are allowed to take up in memory. If max_file_size is null, then the contents + /// will not be loaded into memory. + /// + /// Returns the index of the entry in the `files` array list. You can use it + /// to access the contents of the file after calling `hit()` like so: + /// + /// ``` + /// var file_contents = cache_hash.files.items[file_index].contents.?; + /// ``` + pub fn addFile(self: *Manifest, file_path: []const u8, max_file_size: ?usize) !usize { + assert(self.manifest_file == null); + + const gpa = self.cache.gpa; + try self.files.ensureUnusedCapacity(gpa, 1); + const prefixed_path = try self.cache.findPrefix(file_path); + errdefer gpa.free(prefixed_path.sub_path); + + log.debug("Manifest.addFile {s} -> {d} {s}", .{ + file_path, prefixed_path.prefix, prefixed_path.sub_path, + }); + + self.files.addOneAssumeCapacity().* = .{ + .prefixed_path = prefixed_path, + .contents = null, + .max_file_size = max_file_size, + .stat = undefined, + .bin_digest = undefined, + }; + + self.hash.add(prefixed_path.prefix); + self.hash.addBytes(prefixed_path.sub_path); + + return self.files.items.len - 1; + } + + pub fn addOptionalFile(self: *Manifest, optional_file_path: ?[]const u8) !void { + self.hash.add(optional_file_path != null); + const file_path = optional_file_path orelse return; + _ = try self.addFile(file_path, null); + } + + pub fn addListOfFiles(self: *Manifest, list_of_files: []const []const u8) !void { + self.hash.add(list_of_files.len); + for (list_of_files) |file_path| { + _ = try self.addFile(file_path, null); + } + } + + /// Check the cache to see if the input exists in it. If it exists, returns `true`. + /// A hex encoding of its hash is available by calling `final`. + /// + /// This function will also acquire an exclusive lock to the manifest file. This means + /// that a process holding a Manifest will block any other process attempting to + /// acquire the lock. If `want_shared_lock` is `true`, a cache hit guarantees the + /// manifest file to be locked in shared mode, and a cache miss guarantees the manifest + /// file to be locked in exclusive mode. + /// + /// The lock on the manifest file is released when `deinit` is called. As another + /// option, one may call `toOwnedLock` to obtain a smaller object which can represent + /// the lock. `deinit` is safe to call whether or not `toOwnedLock` has been called. + pub fn hit(self: *Manifest) !bool { + const gpa = self.cache.gpa; + assert(self.manifest_file == null); + + self.failed_file_index = null; + + const ext = ".txt"; + var manifest_file_path: [self.hex_digest.len + ext.len]u8 = undefined; + + var bin_digest: BinDigest = undefined; + self.hash.hasher.final(&bin_digest); + + _ = std.fmt.bufPrint( + &self.hex_digest, + "{s}", + .{std.fmt.fmtSliceHexLower(&bin_digest)}, + ) catch unreachable; + + self.hash.hasher = hasher_init; + self.hash.hasher.update(&bin_digest); + + mem.copy(u8, &manifest_file_path, &self.hex_digest); + manifest_file_path[self.hex_digest.len..][0..ext.len].* = ext.*; + + if (self.files.items.len == 0) { + // If there are no file inputs, we check if the manifest file exists instead of + // comparing the hashes on the files used for the cached item + while (true) { + if (self.cache.manifest_dir.openFile(&manifest_file_path, .{ + .mode = .read_write, + .lock = .Exclusive, + .lock_nonblocking = self.want_shared_lock, + })) |manifest_file| { + self.manifest_file = manifest_file; + self.have_exclusive_lock = true; + break; + } else |open_err| switch (open_err) { + error.WouldBlock => { + self.manifest_file = try self.cache.manifest_dir.openFile(&manifest_file_path, .{ + .lock = .Shared, + }); + break; + }, + error.FileNotFound => { + if (self.cache.manifest_dir.createFile(&manifest_file_path, .{ + .read = true, + .truncate = false, + .lock = .Exclusive, + .lock_nonblocking = self.want_shared_lock, + })) |manifest_file| { + self.manifest_file = manifest_file; + self.manifest_dirty = true; + self.have_exclusive_lock = true; + return false; // cache miss; exclusive lock already held + } else |err| switch (err) { + error.WouldBlock => continue, + else => |e| return e, + } + }, + else => |e| return e, + } + } + } else { + if (self.cache.manifest_dir.createFile(&manifest_file_path, .{ + .read = true, + .truncate = false, + .lock = .Exclusive, + .lock_nonblocking = self.want_shared_lock, + })) |manifest_file| { + self.manifest_file = manifest_file; + self.have_exclusive_lock = true; + } else |err| switch (err) { + error.WouldBlock => { + self.manifest_file = try self.cache.manifest_dir.openFile(&manifest_file_path, .{ + .lock = .Shared, + }); + }, + else => |e| return e, + } + } + + self.want_refresh_timestamp = true; + + const file_contents = try self.manifest_file.?.reader().readAllAlloc(gpa, manifest_file_size_max); + defer gpa.free(file_contents); + + const input_file_count = self.files.items.len; + var any_file_changed = false; + var line_iter = mem.tokenize(u8, file_contents, "\n"); + var idx: usize = 0; + while (line_iter.next()) |line| { + defer idx += 1; + + const cache_hash_file = if (idx < input_file_count) &self.files.items[idx] else blk: { + const new = try self.files.addOne(gpa); + new.* = .{ + .prefixed_path = null, + .contents = null, + .max_file_size = null, + .stat = undefined, + .bin_digest = undefined, + }; + break :blk new; + }; + + var iter = mem.tokenize(u8, line, " "); + const size = iter.next() orelse return error.InvalidFormat; + const inode = iter.next() orelse return error.InvalidFormat; + const mtime_nsec_str = iter.next() orelse return error.InvalidFormat; + const digest_str = iter.next() orelse return error.InvalidFormat; + const prefix_str = iter.next() orelse return error.InvalidFormat; + const file_path = iter.rest(); + + cache_hash_file.stat.size = fmt.parseInt(u64, size, 10) catch return error.InvalidFormat; + cache_hash_file.stat.inode = fmt.parseInt(fs.File.INode, inode, 10) catch return error.InvalidFormat; + cache_hash_file.stat.mtime = fmt.parseInt(i64, mtime_nsec_str, 10) catch return error.InvalidFormat; + _ = std.fmt.hexToBytes(&cache_hash_file.bin_digest, digest_str) catch return error.InvalidFormat; + const prefix = fmt.parseInt(u8, prefix_str, 10) catch return error.InvalidFormat; + if (prefix >= self.cache.prefixes_len) return error.InvalidFormat; + + if (file_path.len == 0) { + return error.InvalidFormat; + } + if (cache_hash_file.prefixed_path) |pp| { + if (pp.prefix != prefix or !mem.eql(u8, file_path, pp.sub_path)) { + return error.InvalidFormat; + } + } + + if (cache_hash_file.prefixed_path == null) { + cache_hash_file.prefixed_path = .{ + .prefix = prefix, + .sub_path = try gpa.dupe(u8, file_path), + }; + } + + const pp = cache_hash_file.prefixed_path.?; + const dir = self.cache.prefixes()[pp.prefix].handle; + const this_file = dir.openFile(pp.sub_path, .{ .mode = .read_only }) catch |err| switch (err) { + error.FileNotFound => { + try self.upgradeToExclusiveLock(); + return false; + }, + else => return error.CacheUnavailable, + }; + defer this_file.close(); + + const actual_stat = this_file.stat() catch |err| { + self.failed_file_index = idx; + return err; + }; + const size_match = actual_stat.size == cache_hash_file.stat.size; + const mtime_match = actual_stat.mtime == cache_hash_file.stat.mtime; + const inode_match = actual_stat.inode == cache_hash_file.stat.inode; + + if (!size_match or !mtime_match or !inode_match) { + self.manifest_dirty = true; + + cache_hash_file.stat = .{ + .size = actual_stat.size, + .mtime = actual_stat.mtime, + .inode = actual_stat.inode, + }; + + if (self.isProblematicTimestamp(cache_hash_file.stat.mtime)) { + // The actual file has an unreliable timestamp, force it to be hashed + cache_hash_file.stat.mtime = 0; + cache_hash_file.stat.inode = 0; + } + + var actual_digest: BinDigest = undefined; + hashFile(this_file, &actual_digest) catch |err| { + self.failed_file_index = idx; + return err; + }; + + if (!mem.eql(u8, &cache_hash_file.bin_digest, &actual_digest)) { + cache_hash_file.bin_digest = actual_digest; + // keep going until we have the input file digests + any_file_changed = true; + } + } + + if (!any_file_changed) { + self.hash.hasher.update(&cache_hash_file.bin_digest); + } + } + + if (any_file_changed) { + // cache miss + // keep the manifest file open + self.unhit(bin_digest, input_file_count); + try self.upgradeToExclusiveLock(); + return false; + } + + if (idx < input_file_count) { + self.manifest_dirty = true; + while (idx < input_file_count) : (idx += 1) { + const ch_file = &self.files.items[idx]; + self.populateFileHash(ch_file) catch |err| { + self.failed_file_index = idx; + return err; + }; + } + try self.upgradeToExclusiveLock(); + return false; + } + + if (self.want_shared_lock) { + try self.downgradeToSharedLock(); + } + + return true; + } + + pub fn unhit(self: *Manifest, bin_digest: BinDigest, input_file_count: usize) void { + // Reset the hash. + self.hash.hasher = hasher_init; + self.hash.hasher.update(&bin_digest); + + // Remove files not in the initial hash. + for (self.files.items[input_file_count..]) |*file| { + file.deinit(self.cache.gpa); + } + self.files.shrinkRetainingCapacity(input_file_count); + + for (self.files.items) |file| { + self.hash.hasher.update(&file.bin_digest); + } + } + + fn isProblematicTimestamp(man: *Manifest, file_time: i128) bool { + // If the file_time is prior to the most recent problematic timestamp + // then we don't need to access the filesystem. + if (file_time < man.recent_problematic_timestamp) + return false; + + // Next we will check the globally shared Cache timestamp, which is accessed + // from multiple threads. + man.cache.mutex.lock(); + defer man.cache.mutex.unlock(); + + // Save the global one to our local one to avoid locking next time. + man.recent_problematic_timestamp = man.cache.recent_problematic_timestamp; + if (file_time < man.recent_problematic_timestamp) + return false; + + // This flag prevents multiple filesystem writes for the same hit() call. + if (man.want_refresh_timestamp) { + man.want_refresh_timestamp = false; + + var file = man.cache.manifest_dir.createFile("timestamp", .{ + .read = true, + .truncate = true, + }) catch return true; + defer file.close(); + + // Save locally and also save globally (we still hold the global lock). + man.recent_problematic_timestamp = (file.stat() catch return true).mtime; + man.cache.recent_problematic_timestamp = man.recent_problematic_timestamp; + } + + return file_time >= man.recent_problematic_timestamp; + } + + fn populateFileHash(self: *Manifest, ch_file: *File) !void { + const pp = ch_file.prefixed_path.?; + const dir = self.cache.prefixes()[pp.prefix].handle; + const file = try dir.openFile(pp.sub_path, .{}); + defer file.close(); + + const actual_stat = try file.stat(); + ch_file.stat = .{ + .size = actual_stat.size, + .mtime = actual_stat.mtime, + .inode = actual_stat.inode, + }; + + if (self.isProblematicTimestamp(ch_file.stat.mtime)) { + // The actual file has an unreliable timestamp, force it to be hashed + ch_file.stat.mtime = 0; + ch_file.stat.inode = 0; + } + + if (ch_file.max_file_size) |max_file_size| { + if (ch_file.stat.size > max_file_size) { + return error.FileTooBig; + } + + const contents = try self.cache.gpa.alloc(u8, @intCast(usize, ch_file.stat.size)); + errdefer self.cache.gpa.free(contents); + + // Hash while reading from disk, to keep the contents in the cpu cache while + // doing hashing. + var hasher = hasher_init; + var off: usize = 0; + while (true) { + // give me everything you've got, captain + const bytes_read = try file.read(contents[off..]); + if (bytes_read == 0) break; + hasher.update(contents[off..][0..bytes_read]); + off += bytes_read; + } + hasher.final(&ch_file.bin_digest); + + ch_file.contents = contents; + } else { + try hashFile(file, &ch_file.bin_digest); + } + + self.hash.hasher.update(&ch_file.bin_digest); + } + + /// Add a file as a dependency of process being cached, after the initial hash has been + /// calculated. This is useful for processes that don't know all the files that + /// are depended on ahead of time. For example, a source file that can import other files + /// will need to be recompiled if the imported file is changed. + pub fn addFilePostFetch(self: *Manifest, file_path: []const u8, max_file_size: usize) ![]const u8 { + assert(self.manifest_file != null); + + const gpa = self.cache.gpa; + const prefixed_path = try self.cache.findPrefix(file_path); + errdefer gpa.free(prefixed_path.sub_path); + + log.debug("Manifest.addFilePostFetch {s} -> {d} {s}", .{ + file_path, prefixed_path.prefix, prefixed_path.sub_path, + }); + + const new_ch_file = try self.files.addOne(gpa); + new_ch_file.* = .{ + .prefixed_path = prefixed_path, + .max_file_size = max_file_size, + .stat = undefined, + .bin_digest = undefined, + .contents = null, + }; + errdefer self.files.shrinkRetainingCapacity(self.files.items.len - 1); + + try self.populateFileHash(new_ch_file); + + return new_ch_file.contents.?; + } + + /// Add a file as a dependency of process being cached, after the initial hash has been + /// calculated. This is useful for processes that don't know the all the files that + /// are depended on ahead of time. For example, a source file that can import other files + /// will need to be recompiled if the imported file is changed. + pub fn addFilePost(self: *Manifest, file_path: []const u8) !void { + assert(self.manifest_file != null); + + const gpa = self.cache.gpa; + const prefixed_path = try self.cache.findPrefix(file_path); + errdefer gpa.free(prefixed_path.sub_path); + + log.debug("Manifest.addFilePost {s} -> {d} {s}", .{ + file_path, prefixed_path.prefix, prefixed_path.sub_path, + }); + + const new_ch_file = try self.files.addOne(gpa); + new_ch_file.* = .{ + .prefixed_path = prefixed_path, + .max_file_size = null, + .stat = undefined, + .bin_digest = undefined, + .contents = null, + }; + errdefer self.files.shrinkRetainingCapacity(self.files.items.len - 1); + + try self.populateFileHash(new_ch_file); + } + + /// Like `addFilePost` but when the file contents have already been loaded from disk. + /// On success, cache takes ownership of `resolved_path`. + pub fn addFilePostContents( + self: *Manifest, + resolved_path: []u8, + bytes: []const u8, + stat: File.Stat, + ) error{OutOfMemory}!void { + assert(self.manifest_file != null); + const gpa = self.cache.gpa; + + const ch_file = try self.files.addOne(gpa); + errdefer self.files.shrinkRetainingCapacity(self.files.items.len - 1); + + log.debug("Manifest.addFilePostContents resolved_path={s}", .{resolved_path}); + + const prefixed_path = try self.cache.findPrefixResolved(resolved_path); + errdefer gpa.free(prefixed_path.sub_path); + + log.debug("Manifest.addFilePostContents -> {d} {s}", .{ + prefixed_path.prefix, prefixed_path.sub_path, + }); + + ch_file.* = .{ + .prefixed_path = prefixed_path, + .max_file_size = null, + .stat = stat, + .bin_digest = undefined, + .contents = null, + }; + + if (self.isProblematicTimestamp(ch_file.stat.mtime)) { + // The actual file has an unreliable timestamp, force it to be hashed + ch_file.stat.mtime = 0; + ch_file.stat.inode = 0; + } + + { + var hasher = hasher_init; + hasher.update(bytes); + hasher.final(&ch_file.bin_digest); + } + + self.hash.hasher.update(&ch_file.bin_digest); + } + + pub fn addDepFilePost(self: *Manifest, dir: fs.Dir, dep_file_basename: []const u8) !void { + assert(self.manifest_file != null); + + const dep_file_contents = try dir.readFileAlloc(self.cache.gpa, dep_file_basename, manifest_file_size_max); + defer self.cache.gpa.free(dep_file_contents); + + var error_buf = std.ArrayList(u8).init(self.cache.gpa); + defer error_buf.deinit(); + + var it: DepTokenizer = .{ .bytes = dep_file_contents }; + + // Skip first token: target. + switch (it.next() orelse return) { // Empty dep file OK. + .target, .target_must_resolve, .prereq => {}, + else => |err| { + try err.printError(error_buf.writer()); + log.err("failed parsing {s}: {s}", .{ dep_file_basename, error_buf.items }); + return error.InvalidDepFile; + }, + } + // Process 0+ preqreqs. + // Clang is invoked in single-source mode so we never get more targets. + while (true) { + switch (it.next() orelse return) { + .target, .target_must_resolve => return, + .prereq => |file_path| try self.addFilePost(file_path), + else => |err| { + try err.printError(error_buf.writer()); + log.err("failed parsing {s}: {s}", .{ dep_file_basename, error_buf.items }); + return error.InvalidDepFile; + }, + } + } + } + + /// Returns a hex encoded hash of the inputs. + pub fn final(self: *Manifest) [hex_digest_len]u8 { + assert(self.manifest_file != null); + + // We don't close the manifest file yet, because we want to + // keep it locked until the API user is done using it. + // We also don't write out the manifest yet, because until + // cache_release is called we still might be working on creating + // the artifacts to cache. + + var bin_digest: BinDigest = undefined; + self.hash.hasher.final(&bin_digest); + + var out_digest: [hex_digest_len]u8 = undefined; + _ = std.fmt.bufPrint( + &out_digest, + "{s}", + .{std.fmt.fmtSliceHexLower(&bin_digest)}, + ) catch unreachable; + + return out_digest; + } + + /// If `want_shared_lock` is true, this function automatically downgrades the + /// lock from exclusive to shared. + pub fn writeManifest(self: *Manifest) !void { + assert(self.have_exclusive_lock); + + const manifest_file = self.manifest_file.?; + if (self.manifest_dirty) { + self.manifest_dirty = false; + + var contents = std.ArrayList(u8).init(self.cache.gpa); + defer contents.deinit(); + + const writer = contents.writer(); + var encoded_digest: [hex_digest_len]u8 = undefined; + + for (self.files.items) |file| { + _ = std.fmt.bufPrint( + &encoded_digest, + "{s}", + .{std.fmt.fmtSliceHexLower(&file.bin_digest)}, + ) catch unreachable; + try writer.print("{d} {d} {d} {s} {d} {s}\n", .{ + file.stat.size, + file.stat.inode, + file.stat.mtime, + &encoded_digest, + file.prefixed_path.?.prefix, + file.prefixed_path.?.sub_path, + }); + } + + try manifest_file.setEndPos(contents.items.len); + try manifest_file.pwriteAll(contents.items, 0); + } + + if (self.want_shared_lock) { + try self.downgradeToSharedLock(); + } + } + + fn downgradeToSharedLock(self: *Manifest) !void { + if (!self.have_exclusive_lock) return; + + // WASI does not currently support flock, so we bypass it here. + // TODO: If/when flock is supported on WASI, this check should be removed. + // See https://github.com/WebAssembly/wasi-filesystem/issues/2 + if (builtin.os.tag != .wasi or std.process.can_spawn or !builtin.single_threaded) { + const manifest_file = self.manifest_file.?; + try manifest_file.downgradeLock(); + } + + self.have_exclusive_lock = false; + } + + fn upgradeToExclusiveLock(self: *Manifest) !void { + if (self.have_exclusive_lock) return; + assert(self.manifest_file != null); + + // WASI does not currently support flock, so we bypass it here. + // TODO: If/when flock is supported on WASI, this check should be removed. + // See https://github.com/WebAssembly/wasi-filesystem/issues/2 + if (builtin.os.tag != .wasi or std.process.can_spawn or !builtin.single_threaded) { + const manifest_file = self.manifest_file.?; + // Here we intentionally have a period where the lock is released, in case there are + // other processes holding a shared lock. + manifest_file.unlock(); + try manifest_file.lock(.Exclusive); + } + self.have_exclusive_lock = true; + } + + /// Obtain only the data needed to maintain a lock on the manifest file. + /// The `Manifest` remains safe to deinit. + /// Don't forget to call `writeManifest` before this! + pub fn toOwnedLock(self: *Manifest) Lock { + const lock: Lock = .{ + .manifest_file = self.manifest_file.?, + }; + + self.manifest_file = null; + return lock; + } + + /// Releases the manifest file and frees any memory the Manifest was using. + /// `Manifest.hit` must be called first. + /// Don't forget to call `writeManifest` before this! + pub fn deinit(self: *Manifest) void { + if (self.manifest_file) |file| { + if (builtin.os.tag == .windows) { + // See Lock.release for why this is required on Windows + file.unlock(); + } + + file.close(); + } + for (self.files.items) |*file| { + file.deinit(self.cache.gpa); + } + self.files.deinit(self.cache.gpa); + } +}; + +/// On operating systems that support symlinks, does a readlink. On other operating systems, +/// uses the file contents. Windows supports symlinks but only with elevated privileges, so +/// it is treated as not supporting symlinks. +pub fn readSmallFile(dir: fs.Dir, sub_path: []const u8, buffer: []u8) ![]u8 { + if (builtin.os.tag == .windows) { + return dir.readFile(sub_path, buffer); + } else { + return dir.readLink(sub_path, buffer); + } +} + +/// On operating systems that support symlinks, does a symlink. On other operating systems, +/// uses the file contents. Windows supports symlinks but only with elevated privileges, so +/// it is treated as not supporting symlinks. +/// `data` must be a valid UTF-8 encoded file path and 255 bytes or fewer. +pub fn writeSmallFile(dir: fs.Dir, sub_path: []const u8, data: []const u8) !void { + assert(data.len <= 255); + if (builtin.os.tag == .windows) { + return dir.writeFile(sub_path, data); + } else { + return dir.symLink(data, sub_path, .{}); + } +} + +fn hashFile(file: fs.File, bin_digest: *[Hasher.mac_length]u8) !void { + var buf: [1024]u8 = undefined; + + var hasher = hasher_init; + while (true) { + const bytes_read = try file.read(&buf); + if (bytes_read == 0) break; + hasher.update(buf[0..bytes_read]); + } + + hasher.final(bin_digest); +} + +// Create/Write a file, close it, then grab its stat.mtime timestamp. +fn testGetCurrentFileTimestamp() !i128 { + var file = try fs.cwd().createFile("test-filetimestamp.tmp", .{ + .read = true, + .truncate = true, + }); + defer file.close(); + + return (try file.stat()).mtime; +} + +test "cache file and then recall it" { + if (builtin.os.tag == .wasi) { + // https://github.com/ziglang/zig/issues/5437 + return error.SkipZigTest; + } + + const cwd = fs.cwd(); + + const temp_file = "test.txt"; + const temp_manifest_dir = "temp_manifest_dir"; + + try cwd.writeFile(temp_file, "Hello, world!\n"); + + // Wait for file timestamps to tick + const initial_time = try testGetCurrentFileTimestamp(); + while ((try testGetCurrentFileTimestamp()) == initial_time) { + std.time.sleep(1); + } + + var digest1: [hex_digest_len]u8 = undefined; + var digest2: [hex_digest_len]u8 = undefined; + + { + var cache = Cache{ + .gpa = testing.allocator, + .manifest_dir = try cwd.makeOpenPath(temp_manifest_dir, .{}), + }; + cache.addPrefix(.{ .path = null, .handle = fs.cwd() }); + defer cache.manifest_dir.close(); + + { + var ch = cache.obtain(); + defer ch.deinit(); + + ch.hash.add(true); + ch.hash.add(@as(u16, 1234)); + ch.hash.addBytes("1234"); + _ = try ch.addFile(temp_file, null); + + // There should be nothing in the cache + try testing.expectEqual(false, try ch.hit()); + + digest1 = ch.final(); + try ch.writeManifest(); + } + { + var ch = cache.obtain(); + defer ch.deinit(); + + ch.hash.add(true); + ch.hash.add(@as(u16, 1234)); + ch.hash.addBytes("1234"); + _ = try ch.addFile(temp_file, null); + + // Cache hit! We just "built" the same file + try testing.expect(try ch.hit()); + digest2 = ch.final(); + + try testing.expectEqual(false, ch.have_exclusive_lock); + } + + try testing.expectEqual(digest1, digest2); + } + + try cwd.deleteTree(temp_manifest_dir); + try cwd.deleteFile(temp_file); +} + +test "check that changing a file makes cache fail" { + if (builtin.os.tag == .wasi) { + // https://github.com/ziglang/zig/issues/5437 + return error.SkipZigTest; + } + const cwd = fs.cwd(); + + const temp_file = "cache_hash_change_file_test.txt"; + const temp_manifest_dir = "cache_hash_change_file_manifest_dir"; + const original_temp_file_contents = "Hello, world!\n"; + const updated_temp_file_contents = "Hello, world; but updated!\n"; + + try cwd.deleteTree(temp_manifest_dir); + try cwd.deleteTree(temp_file); + + try cwd.writeFile(temp_file, original_temp_file_contents); + + // Wait for file timestamps to tick + const initial_time = try testGetCurrentFileTimestamp(); + while ((try testGetCurrentFileTimestamp()) == initial_time) { + std.time.sleep(1); + } + + var digest1: [hex_digest_len]u8 = undefined; + var digest2: [hex_digest_len]u8 = undefined; + + { + var cache = Cache{ + .gpa = testing.allocator, + .manifest_dir = try cwd.makeOpenPath(temp_manifest_dir, .{}), + }; + cache.addPrefix(.{ .path = null, .handle = fs.cwd() }); + defer cache.manifest_dir.close(); + + { + var ch = cache.obtain(); + defer ch.deinit(); + + ch.hash.addBytes("1234"); + const temp_file_idx = try ch.addFile(temp_file, 100); + + // There should be nothing in the cache + try testing.expectEqual(false, try ch.hit()); + + try testing.expect(mem.eql(u8, original_temp_file_contents, ch.files.items[temp_file_idx].contents.?)); + + digest1 = ch.final(); + + try ch.writeManifest(); + } + + try cwd.writeFile(temp_file, updated_temp_file_contents); + + { + var ch = cache.obtain(); + defer ch.deinit(); + + ch.hash.addBytes("1234"); + const temp_file_idx = try ch.addFile(temp_file, 100); + + // A file that we depend on has been updated, so the cache should not contain an entry for it + try testing.expectEqual(false, try ch.hit()); + + // The cache system does not keep the contents of re-hashed input files. + try testing.expect(ch.files.items[temp_file_idx].contents == null); + + digest2 = ch.final(); + + try ch.writeManifest(); + } + + try testing.expect(!mem.eql(u8, digest1[0..], digest2[0..])); + } + + try cwd.deleteTree(temp_manifest_dir); + try cwd.deleteTree(temp_file); +} + +test "no file inputs" { + if (builtin.os.tag == .wasi) { + // https://github.com/ziglang/zig/issues/5437 + return error.SkipZigTest; + } + const cwd = fs.cwd(); + const temp_manifest_dir = "no_file_inputs_manifest_dir"; + defer cwd.deleteTree(temp_manifest_dir) catch {}; + + var digest1: [hex_digest_len]u8 = undefined; + var digest2: [hex_digest_len]u8 = undefined; + + var cache = Cache{ + .gpa = testing.allocator, + .manifest_dir = try cwd.makeOpenPath(temp_manifest_dir, .{}), + }; + cache.addPrefix(.{ .path = null, .handle = fs.cwd() }); + defer cache.manifest_dir.close(); + + { + var man = cache.obtain(); + defer man.deinit(); + + man.hash.addBytes("1234"); + + // There should be nothing in the cache + try testing.expectEqual(false, try man.hit()); + + digest1 = man.final(); + + try man.writeManifest(); + } + { + var man = cache.obtain(); + defer man.deinit(); + + man.hash.addBytes("1234"); + + try testing.expect(try man.hit()); + digest2 = man.final(); + try testing.expectEqual(false, man.have_exclusive_lock); + } + + try testing.expectEqual(digest1, digest2); +} + +test "Manifest with files added after initial hash work" { + if (builtin.os.tag == .wasi) { + // https://github.com/ziglang/zig/issues/5437 + return error.SkipZigTest; + } + const cwd = fs.cwd(); + + const temp_file1 = "cache_hash_post_file_test1.txt"; + const temp_file2 = "cache_hash_post_file_test2.txt"; + const temp_manifest_dir = "cache_hash_post_file_manifest_dir"; + + try cwd.writeFile(temp_file1, "Hello, world!\n"); + try cwd.writeFile(temp_file2, "Hello world the second!\n"); + + // Wait for file timestamps to tick + const initial_time = try testGetCurrentFileTimestamp(); + while ((try testGetCurrentFileTimestamp()) == initial_time) { + std.time.sleep(1); + } + + var digest1: [hex_digest_len]u8 = undefined; + var digest2: [hex_digest_len]u8 = undefined; + var digest3: [hex_digest_len]u8 = undefined; + + { + var cache = Cache{ + .gpa = testing.allocator, + .manifest_dir = try cwd.makeOpenPath(temp_manifest_dir, .{}), + }; + cache.addPrefix(.{ .path = null, .handle = fs.cwd() }); + defer cache.manifest_dir.close(); + + { + var ch = cache.obtain(); + defer ch.deinit(); + + ch.hash.addBytes("1234"); + _ = try ch.addFile(temp_file1, null); + + // There should be nothing in the cache + try testing.expectEqual(false, try ch.hit()); + + _ = try ch.addFilePost(temp_file2); + + digest1 = ch.final(); + try ch.writeManifest(); + } + { + var ch = cache.obtain(); + defer ch.deinit(); + + ch.hash.addBytes("1234"); + _ = try ch.addFile(temp_file1, null); + + try testing.expect(try ch.hit()); + digest2 = ch.final(); + + try testing.expectEqual(false, ch.have_exclusive_lock); + } + try testing.expect(mem.eql(u8, &digest1, &digest2)); + + // Modify the file added after initial hash + try cwd.writeFile(temp_file2, "Hello world the second, updated\n"); + + // Wait for file timestamps to tick + const initial_time2 = try testGetCurrentFileTimestamp(); + while ((try testGetCurrentFileTimestamp()) == initial_time2) { + std.time.sleep(1); + } + + { + var ch = cache.obtain(); + defer ch.deinit(); + + ch.hash.addBytes("1234"); + _ = try ch.addFile(temp_file1, null); + + // A file that we depend on has been updated, so the cache should not contain an entry for it + try testing.expectEqual(false, try ch.hit()); + + _ = try ch.addFilePost(temp_file2); + + digest3 = ch.final(); + + try ch.writeManifest(); + } + + try testing.expect(!mem.eql(u8, &digest1, &digest3)); + } + + try cwd.deleteTree(temp_manifest_dir); + try cwd.deleteFile(temp_file1); + try cwd.deleteFile(temp_file2); +} diff --git a/lib/std/Build/Cache/DepTokenizer.zig b/lib/std/Build/Cache/DepTokenizer.zig new file mode 100644 index 0000000000..8f9f2f81cd --- /dev/null +++ b/lib/std/Build/Cache/DepTokenizer.zig @@ -0,0 +1,1069 @@ +const Tokenizer = @This(); + +index: usize = 0, +bytes: []const u8, +state: State = .lhs, + +const std = @import("std"); +const testing = std.testing; +const assert = std.debug.assert; + +pub fn next(self: *Tokenizer) ?Token { + var start = self.index; + var must_resolve = false; + while (self.index < self.bytes.len) { + const char = self.bytes[self.index]; + switch (self.state) { + .lhs => switch (char) { + '\t', '\n', '\r', ' ' => { + // silently ignore whitespace + self.index += 1; + }, + else => { + start = self.index; + self.state = .target; + }, + }, + .target => switch (char) { + '\t', '\n', '\r', ' ' => { + return errorIllegalChar(.invalid_target, self.index, char); + }, + '$' => { + self.state = .target_dollar_sign; + self.index += 1; + }, + '\\' => { + self.state = .target_reverse_solidus; + self.index += 1; + }, + ':' => { + self.state = .target_colon; + self.index += 1; + }, + else => { + self.index += 1; + }, + }, + .target_reverse_solidus => switch (char) { + '\t', '\n', '\r' => { + return errorIllegalChar(.bad_target_escape, self.index, char); + }, + ' ', '#', '\\' => { + must_resolve = true; + self.state = .target; + self.index += 1; + }, + '$' => { + self.state = .target_dollar_sign; + self.index += 1; + }, + else => { + self.state = .target; + self.index += 1; + }, + }, + .target_dollar_sign => switch (char) { + '$' => { + must_resolve = true; + self.state = .target; + self.index += 1; + }, + else => { + return errorIllegalChar(.expected_dollar_sign, self.index, char); + }, + }, + .target_colon => switch (char) { + '\n', '\r' => { + const bytes = self.bytes[start .. self.index - 1]; + if (bytes.len != 0) { + self.state = .lhs; + return finishTarget(must_resolve, bytes); + } + // silently ignore null target + self.state = .lhs; + }, + '/', '\\' => { + self.state = .target_colon_reverse_solidus; + self.index += 1; + }, + else => { + const bytes = self.bytes[start .. self.index - 1]; + if (bytes.len != 0) { + self.state = .rhs; + return finishTarget(must_resolve, bytes); + } + // silently ignore null target + self.state = .lhs; + }, + }, + .target_colon_reverse_solidus => switch (char) { + '\n', '\r' => { + const bytes = self.bytes[start .. self.index - 2]; + if (bytes.len != 0) { + self.state = .lhs; + return finishTarget(must_resolve, bytes); + } + // silently ignore null target + self.state = .lhs; + }, + else => { + self.state = .target; + }, + }, + .rhs => switch (char) { + '\t', ' ' => { + // silently ignore horizontal whitespace + self.index += 1; + }, + '\n', '\r' => { + self.state = .lhs; + }, + '\\' => { + self.state = .rhs_continuation; + self.index += 1; + }, + '"' => { + self.state = .prereq_quote; + self.index += 1; + start = self.index; + }, + else => { + start = self.index; + self.state = .prereq; + }, + }, + .rhs_continuation => switch (char) { + '\n' => { + self.state = .rhs; + self.index += 1; + }, + '\r' => { + self.state = .rhs_continuation_linefeed; + self.index += 1; + }, + else => { + return errorIllegalChar(.continuation_eol, self.index, char); + }, + }, + .rhs_continuation_linefeed => switch (char) { + '\n' => { + self.state = .rhs; + self.index += 1; + }, + else => { + return errorIllegalChar(.continuation_eol, self.index, char); + }, + }, + .prereq_quote => switch (char) { + '"' => { + self.index += 1; + self.state = .rhs; + return Token{ .prereq = self.bytes[start .. self.index - 1] }; + }, + else => { + self.index += 1; + }, + }, + .prereq => switch (char) { + '\t', ' ' => { + self.state = .rhs; + return Token{ .prereq = self.bytes[start..self.index] }; + }, + '\n', '\r' => { + self.state = .lhs; + return Token{ .prereq = self.bytes[start..self.index] }; + }, + '\\' => { + self.state = .prereq_continuation; + self.index += 1; + }, + else => { + self.index += 1; + }, + }, + .prereq_continuation => switch (char) { + '\n' => { + self.index += 1; + self.state = .rhs; + return Token{ .prereq = self.bytes[start .. self.index - 2] }; + }, + '\r' => { + self.state = .prereq_continuation_linefeed; + self.index += 1; + }, + else => { + // not continuation + self.state = .prereq; + self.index += 1; + }, + }, + .prereq_continuation_linefeed => switch (char) { + '\n' => { + self.index += 1; + self.state = .rhs; + return Token{ .prereq = self.bytes[start .. self.index - 1] }; + }, + else => { + return errorIllegalChar(.continuation_eol, self.index, char); + }, + }, + } + } else { + switch (self.state) { + .lhs, + .rhs, + .rhs_continuation, + .rhs_continuation_linefeed, + => return null, + .target => { + return errorPosition(.incomplete_target, start, self.bytes[start..]); + }, + .target_reverse_solidus, + .target_dollar_sign, + => { + const idx = self.index - 1; + return errorIllegalChar(.incomplete_escape, idx, self.bytes[idx]); + }, + .target_colon => { + const bytes = self.bytes[start .. self.index - 1]; + if (bytes.len != 0) { + self.index += 1; + self.state = .rhs; + return finishTarget(must_resolve, bytes); + } + // silently ignore null target + self.state = .lhs; + return null; + }, + .target_colon_reverse_solidus => { + const bytes = self.bytes[start .. self.index - 2]; + if (bytes.len != 0) { + self.index += 1; + self.state = .rhs; + return finishTarget(must_resolve, bytes); + } + // silently ignore null target + self.state = .lhs; + return null; + }, + .prereq_quote => { + return errorPosition(.incomplete_quoted_prerequisite, start, self.bytes[start..]); + }, + .prereq => { + self.state = .lhs; + return Token{ .prereq = self.bytes[start..] }; + }, + .prereq_continuation => { + self.state = .lhs; + return Token{ .prereq = self.bytes[start .. self.index - 1] }; + }, + .prereq_continuation_linefeed => { + self.state = .lhs; + return Token{ .prereq = self.bytes[start .. self.index - 2] }; + }, + } + } + unreachable; +} + +fn errorPosition(comptime id: std.meta.Tag(Token), index: usize, bytes: []const u8) Token { + return @unionInit(Token, @tagName(id), .{ .index = index, .bytes = bytes }); +} + +fn errorIllegalChar(comptime id: std.meta.Tag(Token), index: usize, char: u8) Token { + return @unionInit(Token, @tagName(id), .{ .index = index, .char = char }); +} + +fn finishTarget(must_resolve: bool, bytes: []const u8) Token { + return if (must_resolve) .{ .target_must_resolve = bytes } else .{ .target = bytes }; +} + +const State = enum { + lhs, + target, + target_reverse_solidus, + target_dollar_sign, + target_colon, + target_colon_reverse_solidus, + rhs, + rhs_continuation, + rhs_continuation_linefeed, + prereq_quote, + prereq, + prereq_continuation, + prereq_continuation_linefeed, +}; + +pub const Token = union(enum) { + target: []const u8, + target_must_resolve: []const u8, + prereq: []const u8, + + incomplete_quoted_prerequisite: IndexAndBytes, + incomplete_target: IndexAndBytes, + + invalid_target: IndexAndChar, + bad_target_escape: IndexAndChar, + expected_dollar_sign: IndexAndChar, + continuation_eol: IndexAndChar, + incomplete_escape: IndexAndChar, + + pub const IndexAndChar = struct { + index: usize, + char: u8, + }; + + pub const IndexAndBytes = struct { + index: usize, + bytes: []const u8, + }; + + /// Resolve escapes in target. Only valid with .target_must_resolve. + pub fn resolve(self: Token, writer: anytype) @TypeOf(writer).Error!void { + const bytes = self.target_must_resolve; // resolve called on incorrect token + + var state: enum { start, escape, dollar } = .start; + for (bytes) |c| { + switch (state) { + .start => { + switch (c) { + '\\' => state = .escape, + '$' => state = .dollar, + else => try writer.writeByte(c), + } + }, + .escape => { + switch (c) { + ' ', '#', '\\' => {}, + '$' => { + try writer.writeByte('\\'); + state = .dollar; + continue; + }, + else => try writer.writeByte('\\'), + } + try writer.writeByte(c); + state = .start; + }, + .dollar => { + try writer.writeByte('$'); + switch (c) { + '$' => {}, + else => try writer.writeByte(c), + } + state = .start; + }, + } + } + } + + pub fn printError(self: Token, writer: anytype) @TypeOf(writer).Error!void { + switch (self) { + .target, .target_must_resolve, .prereq => unreachable, // not an error + .incomplete_quoted_prerequisite, + .incomplete_target, + => |index_and_bytes| { + try writer.print("{s} '", .{self.errStr()}); + if (self == .incomplete_target) { + const tmp = Token{ .target_must_resolve = index_and_bytes.bytes }; + try tmp.resolve(writer); + } else { + try printCharValues(writer, index_and_bytes.bytes); + } + try writer.print("' at position {d}", .{index_and_bytes.index}); + }, + .invalid_target, + .bad_target_escape, + .expected_dollar_sign, + .continuation_eol, + .incomplete_escape, + => |index_and_char| { + try writer.writeAll("illegal char "); + try printUnderstandableChar(writer, index_and_char.char); + try writer.print(" at position {d}: {s}", .{ index_and_char.index, self.errStr() }); + }, + } + } + + fn errStr(self: Token) []const u8 { + return switch (self) { + .target, .target_must_resolve, .prereq => unreachable, // not an error + .incomplete_quoted_prerequisite => "incomplete quoted prerequisite", + .incomplete_target => "incomplete target", + .invalid_target => "invalid target", + .bad_target_escape => "bad target escape", + .expected_dollar_sign => "expecting '$'", + .continuation_eol => "continuation expecting end-of-line", + .incomplete_escape => "incomplete escape", + }; + } +}; + +test "empty file" { + try depTokenizer("", ""); +} + +test "empty whitespace" { + try depTokenizer("\n", ""); + try depTokenizer("\r", ""); + try depTokenizer("\r\n", ""); + try depTokenizer(" ", ""); +} + +test "empty colon" { + try depTokenizer(":", ""); + try depTokenizer("\n:", ""); + try depTokenizer("\r:", ""); + try depTokenizer("\r\n:", ""); + try depTokenizer(" :", ""); +} + +test "empty target" { + try depTokenizer("foo.o:", "target = {foo.o}"); + try depTokenizer( + \\foo.o: + \\bar.o: + \\abcd.o: + , + \\target = {foo.o} + \\target = {bar.o} + \\target = {abcd.o} + ); +} + +test "whitespace empty target" { + try depTokenizer("\nfoo.o:", "target = {foo.o}"); + try depTokenizer("\rfoo.o:", "target = {foo.o}"); + try depTokenizer("\r\nfoo.o:", "target = {foo.o}"); + try depTokenizer(" foo.o:", "target = {foo.o}"); +} + +test "escape empty target" { + try depTokenizer("\\ foo.o:", "target = { foo.o}"); + try depTokenizer("\\#foo.o:", "target = {#foo.o}"); + try depTokenizer("\\\\foo.o:", "target = {\\foo.o}"); + try depTokenizer("$$foo.o:", "target = {$foo.o}"); +} + +test "empty target linefeeds" { + try depTokenizer("\n", ""); + try depTokenizer("\r\n", ""); + + const expect = "target = {foo.o}"; + try depTokenizer( + \\foo.o: + , expect); + try depTokenizer( + \\foo.o: + \\ + , expect); + try depTokenizer( + \\foo.o: + , expect); + try depTokenizer( + \\foo.o: + \\ + , expect); +} + +test "empty target linefeeds + continuations" { + const expect = "target = {foo.o}"; + try depTokenizer( + \\foo.o:\ + , expect); + try depTokenizer( + \\foo.o:\ + \\ + , expect); + try depTokenizer( + \\foo.o:\ + , expect); + try depTokenizer( + \\foo.o:\ + \\ + , expect); +} + +test "empty target linefeeds + hspace + continuations" { + const expect = "target = {foo.o}"; + try depTokenizer( + \\foo.o: \ + , expect); + try depTokenizer( + \\foo.o: \ + \\ + , expect); + try depTokenizer( + \\foo.o: \ + , expect); + try depTokenizer( + \\foo.o: \ + \\ + , expect); +} + +test "prereq" { + const expect = + \\target = {foo.o} + \\prereq = {foo.c} + ; + try depTokenizer("foo.o: foo.c", expect); + try depTokenizer( + \\foo.o: \ + \\foo.c + , expect); + try depTokenizer( + \\foo.o: \ + \\ foo.c + , expect); + try depTokenizer( + \\foo.o: \ + \\ foo.c + , expect); +} + +test "prereq continuation" { + const expect = + \\target = {foo.o} + \\prereq = {foo.h} + \\prereq = {bar.h} + ; + try depTokenizer( + \\foo.o: foo.h\ + \\bar.h + , expect); + try depTokenizer( + \\foo.o: foo.h\ + \\bar.h + , expect); +} + +test "multiple prereqs" { + const expect = + \\target = {foo.o} + \\prereq = {foo.c} + \\prereq = {foo.h} + \\prereq = {bar.h} + ; + try depTokenizer("foo.o: foo.c foo.h bar.h", expect); + try depTokenizer( + \\foo.o: \ + \\foo.c foo.h bar.h + , expect); + try depTokenizer( + \\foo.o: foo.c foo.h bar.h\ + , expect); + try depTokenizer( + \\foo.o: foo.c foo.h bar.h\ + \\ + , expect); + try depTokenizer( + \\foo.o: \ + \\foo.c \ + \\ foo.h\ + \\bar.h + \\ + , expect); + try depTokenizer( + \\foo.o: \ + \\foo.c \ + \\ foo.h\ + \\bar.h\ + \\ + , expect); + try depTokenizer( + \\foo.o: \ + \\foo.c \ + \\ foo.h\ + \\bar.h\ + , expect); +} + +test "multiple targets and prereqs" { + try depTokenizer( + \\foo.o: foo.c + \\bar.o: bar.c a.h b.h c.h + \\abc.o: abc.c \ + \\ one.h two.h \ + \\ three.h four.h + , + \\target = {foo.o} + \\prereq = {foo.c} + \\target = {bar.o} + \\prereq = {bar.c} + \\prereq = {a.h} + \\prereq = {b.h} + \\prereq = {c.h} + \\target = {abc.o} + \\prereq = {abc.c} + \\prereq = {one.h} + \\prereq = {two.h} + \\prereq = {three.h} + \\prereq = {four.h} + ); + try depTokenizer( + \\ascii.o: ascii.c + \\base64.o: base64.c stdio.h + \\elf.o: elf.c a.h b.h c.h + \\macho.o: \ + \\ macho.c\ + \\ a.h b.h c.h + , + \\target = {ascii.o} + \\prereq = {ascii.c} + \\target = {base64.o} + \\prereq = {base64.c} + \\prereq = {stdio.h} + \\target = {elf.o} + \\prereq = {elf.c} + \\prereq = {a.h} + \\prereq = {b.h} + \\prereq = {c.h} + \\target = {macho.o} + \\prereq = {macho.c} + \\prereq = {a.h} + \\prereq = {b.h} + \\prereq = {c.h} + ); + try depTokenizer( + \\a$$scii.o: ascii.c + \\\\base64.o: "\base64.c" "s t#dio.h" + \\e\\lf.o: "e\lf.c" "a.h$$" "$$b.h c.h$$" + \\macho.o: \ + \\ "macho!.c" \ + \\ a.h b.h c.h + , + \\target = {a$scii.o} + \\prereq = {ascii.c} + \\target = {\base64.o} + \\prereq = {\base64.c} + \\prereq = {s t#dio.h} + \\target = {e\lf.o} + \\prereq = {e\lf.c} + \\prereq = {a.h$$} + \\prereq = {$$b.h c.h$$} + \\target = {macho.o} + \\prereq = {macho!.c} + \\prereq = {a.h} + \\prereq = {b.h} + \\prereq = {c.h} + ); +} + +test "windows quoted prereqs" { + try depTokenizer( + \\c:\foo.o: "C:\Program Files (x86)\Microsoft Visual Studio\foo.c" + \\c:\foo2.o: "C:\Program Files (x86)\Microsoft Visual Studio\foo2.c" \ + \\ "C:\Program Files (x86)\Microsoft Visual Studio\foo1.h" \ + \\ "C:\Program Files (x86)\Microsoft Visual Studio\foo2.h" + , + \\target = {c:\foo.o} + \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\foo.c} + \\target = {c:\foo2.o} + \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\foo2.c} + \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\foo1.h} + \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\foo2.h} + ); +} + +test "windows mixed prereqs" { + try depTokenizer( + \\cimport.o: \ + \\ C:\msys64\home\anon\project\zig\master\zig-cache\o\qhvhbUo7GU5iKyQ5mpA8TcQpncCYaQu0wwvr3ybiSTj_Dtqi1Nmcb70kfODJ2Qlg\cimport.h \ + \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\stdio.h" \ + \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt.h" \ + \\ "C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vcruntime.h" \ + \\ "C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\sal.h" \ + \\ "C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\concurrencysal.h" \ + \\ C:\msys64\opt\zig\lib\zig\include\vadefs.h \ + \\ "C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vadefs.h" \ + \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_wstdio.h" \ + \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_stdio_config.h" \ + \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\string.h" \ + \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_memory.h" \ + \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_memcpy_s.h" \ + \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\errno.h" \ + \\ "C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vcruntime_string.h" \ + \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_wstring.h" + , + \\target = {cimport.o} + \\prereq = {C:\msys64\home\anon\project\zig\master\zig-cache\o\qhvhbUo7GU5iKyQ5mpA8TcQpncCYaQu0wwvr3ybiSTj_Dtqi1Nmcb70kfODJ2Qlg\cimport.h} + \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\stdio.h} + \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt.h} + \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vcruntime.h} + \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\sal.h} + \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\concurrencysal.h} + \\prereq = {C:\msys64\opt\zig\lib\zig\include\vadefs.h} + \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vadefs.h} + \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_wstdio.h} + \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_stdio_config.h} + \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\string.h} + \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_memory.h} + \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_memcpy_s.h} + \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\errno.h} + \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vcruntime_string.h} + \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_wstring.h} + ); +} + +test "windows funky targets" { + try depTokenizer( + \\C:\Users\anon\foo.o: + \\C:\Users\anon\foo\ .o: + \\C:\Users\anon\foo\#.o: + \\C:\Users\anon\foo$$.o: + \\C:\Users\anon\\\ foo.o: + \\C:\Users\anon\\#foo.o: + \\C:\Users\anon\$$foo.o: + \\C:\Users\anon\\\ \ \ \ \ foo.o: + , + \\target = {C:\Users\anon\foo.o} + \\target = {C:\Users\anon\foo .o} + \\target = {C:\Users\anon\foo#.o} + \\target = {C:\Users\anon\foo$.o} + \\target = {C:\Users\anon\ foo.o} + \\target = {C:\Users\anon\#foo.o} + \\target = {C:\Users\anon\$foo.o} + \\target = {C:\Users\anon\ foo.o} + ); +} + +test "windows drive and forward slashes" { + try depTokenizer( + \\C:/msys64/what/zig-cache\tmp\48ac4d78dd531abd-cxa_thread_atexit.obj: \ + \\ C:/msys64/opt/zig3/lib/zig/libc/mingw/crt/cxa_thread_atexit.c + , + \\target = {C:/msys64/what/zig-cache\tmp\48ac4d78dd531abd-cxa_thread_atexit.obj} + \\prereq = {C:/msys64/opt/zig3/lib/zig/libc/mingw/crt/cxa_thread_atexit.c} + ); +} + +test "error incomplete escape - reverse_solidus" { + try depTokenizer("\\", + \\ERROR: illegal char '\' at position 0: incomplete escape + ); + try depTokenizer("\t\\", + \\ERROR: illegal char '\' at position 1: incomplete escape + ); + try depTokenizer("\n\\", + \\ERROR: illegal char '\' at position 1: incomplete escape + ); + try depTokenizer("\r\\", + \\ERROR: illegal char '\' at position 1: incomplete escape + ); + try depTokenizer("\r\n\\", + \\ERROR: illegal char '\' at position 2: incomplete escape + ); + try depTokenizer(" \\", + \\ERROR: illegal char '\' at position 1: incomplete escape + ); +} + +test "error incomplete escape - dollar_sign" { + try depTokenizer("$", + \\ERROR: illegal char '$' at position 0: incomplete escape + ); + try depTokenizer("\t$", + \\ERROR: illegal char '$' at position 1: incomplete escape + ); + try depTokenizer("\n$", + \\ERROR: illegal char '$' at position 1: incomplete escape + ); + try depTokenizer("\r$", + \\ERROR: illegal char '$' at position 1: incomplete escape + ); + try depTokenizer("\r\n$", + \\ERROR: illegal char '$' at position 2: incomplete escape + ); + try depTokenizer(" $", + \\ERROR: illegal char '$' at position 1: incomplete escape + ); +} + +test "error incomplete target" { + try depTokenizer("foo.o", + \\ERROR: incomplete target 'foo.o' at position 0 + ); + try depTokenizer("\tfoo.o", + \\ERROR: incomplete target 'foo.o' at position 1 + ); + try depTokenizer("\nfoo.o", + \\ERROR: incomplete target 'foo.o' at position 1 + ); + try depTokenizer("\rfoo.o", + \\ERROR: incomplete target 'foo.o' at position 1 + ); + try depTokenizer("\r\nfoo.o", + \\ERROR: incomplete target 'foo.o' at position 2 + ); + try depTokenizer(" foo.o", + \\ERROR: incomplete target 'foo.o' at position 1 + ); + + try depTokenizer("\\ foo.o", + \\ERROR: incomplete target ' foo.o' at position 0 + ); + try depTokenizer("\\#foo.o", + \\ERROR: incomplete target '#foo.o' at position 0 + ); + try depTokenizer("\\\\foo.o", + \\ERROR: incomplete target '\foo.o' at position 0 + ); + try depTokenizer("$$foo.o", + \\ERROR: incomplete target '$foo.o' at position 0 + ); +} + +test "error illegal char at position - bad target escape" { + try depTokenizer("\\\t", + \\ERROR: illegal char \x09 at position 1: bad target escape + ); + try depTokenizer("\\\n", + \\ERROR: illegal char \x0A at position 1: bad target escape + ); + try depTokenizer("\\\r", + \\ERROR: illegal char \x0D at position 1: bad target escape + ); + try depTokenizer("\\\r\n", + \\ERROR: illegal char \x0D at position 1: bad target escape + ); +} + +test "error illegal char at position - execting dollar_sign" { + try depTokenizer("$\t", + \\ERROR: illegal char \x09 at position 1: expecting '$' + ); + try depTokenizer("$\n", + \\ERROR: illegal char \x0A at position 1: expecting '$' + ); + try depTokenizer("$\r", + \\ERROR: illegal char \x0D at position 1: expecting '$' + ); + try depTokenizer("$\r\n", + \\ERROR: illegal char \x0D at position 1: expecting '$' + ); +} + +test "error illegal char at position - invalid target" { + try depTokenizer("foo\t.o", + \\ERROR: illegal char \x09 at position 3: invalid target + ); + try depTokenizer("foo\n.o", + \\ERROR: illegal char \x0A at position 3: invalid target + ); + try depTokenizer("foo\r.o", + \\ERROR: illegal char \x0D at position 3: invalid target + ); + try depTokenizer("foo\r\n.o", + \\ERROR: illegal char \x0D at position 3: invalid target + ); +} + +test "error target - continuation expecting end-of-line" { + try depTokenizer("foo.o: \\\t", + \\target = {foo.o} + \\ERROR: illegal char \x09 at position 8: continuation expecting end-of-line + ); + try depTokenizer("foo.o: \\ ", + \\target = {foo.o} + \\ERROR: illegal char ' ' at position 8: continuation expecting end-of-line + ); + try depTokenizer("foo.o: \\x", + \\target = {foo.o} + \\ERROR: illegal char 'x' at position 8: continuation expecting end-of-line + ); + try depTokenizer("foo.o: \\\x0dx", + \\target = {foo.o} + \\ERROR: illegal char 'x' at position 9: continuation expecting end-of-line + ); +} + +test "error prereq - continuation expecting end-of-line" { + try depTokenizer("foo.o: foo.h\\\x0dx", + \\target = {foo.o} + \\ERROR: illegal char 'x' at position 14: continuation expecting end-of-line + ); +} + +// - tokenize input, emit textual representation, and compare to expect +fn depTokenizer(input: []const u8, expect: []const u8) !void { + var arena_allocator = std.heap.ArenaAllocator.init(std.testing.allocator); + const arena = arena_allocator.allocator(); + defer arena_allocator.deinit(); + + var it: Tokenizer = .{ .bytes = input }; + var buffer = std.ArrayList(u8).init(arena); + var resolve_buf = std.ArrayList(u8).init(arena); + var i: usize = 0; + while (it.next()) |token| { + if (i != 0) try buffer.appendSlice("\n"); + switch (token) { + .target, .prereq => |bytes| { + try buffer.appendSlice(@tagName(token)); + try buffer.appendSlice(" = {"); + for (bytes) |b| { + try buffer.append(printable_char_tab[b]); + } + try buffer.appendSlice("}"); + }, + .target_must_resolve => { + try buffer.appendSlice("target = {"); + try token.resolve(resolve_buf.writer()); + for (resolve_buf.items) |b| { + try buffer.append(printable_char_tab[b]); + } + resolve_buf.items.len = 0; + try buffer.appendSlice("}"); + }, + else => { + try buffer.appendSlice("ERROR: "); + try token.printError(buffer.writer()); + break; + }, + } + i += 1; + } + + if (std.mem.eql(u8, expect, buffer.items)) { + try testing.expect(true); + return; + } + + const out = std.io.getStdErr().writer(); + + try out.writeAll("\n"); + try printSection(out, "<<<< input", input); + try printSection(out, "==== expect", expect); + try printSection(out, ">>>> got", buffer.items); + try printRuler(out); + + try testing.expect(false); +} + +fn printSection(out: anytype, label: []const u8, bytes: []const u8) !void { + try printLabel(out, label, bytes); + try hexDump(out, bytes); + try printRuler(out); + try out.writeAll(bytes); + try out.writeAll("\n"); +} + +fn printLabel(out: anytype, label: []const u8, bytes: []const u8) !void { + var buf: [80]u8 = undefined; + var text = try std.fmt.bufPrint(buf[0..], "{s} {d} bytes ", .{ label, bytes.len }); + try out.writeAll(text); + var i: usize = text.len; + const end = 79; + while (i < end) : (i += 1) { + try out.writeAll(&[_]u8{label[0]}); + } + try out.writeAll("\n"); +} + +fn printRuler(out: anytype) !void { + var i: usize = 0; + const end = 79; + while (i < end) : (i += 1) { + try out.writeAll("-"); + } + try out.writeAll("\n"); +} + +fn hexDump(out: anytype, bytes: []const u8) !void { + const n16 = bytes.len >> 4; + var line: usize = 0; + var offset: usize = 0; + while (line < n16) : (line += 1) { + try hexDump16(out, offset, bytes[offset .. offset + 16]); + offset += 16; + } + + const n = bytes.len & 0x0f; + if (n > 0) { + try printDecValue(out, offset, 8); + try out.writeAll(":"); + try out.writeAll(" "); + var end1 = std.math.min(offset + n, offset + 8); + for (bytes[offset..end1]) |b| { + try out.writeAll(" "); + try printHexValue(out, b, 2); + } + var end2 = offset + n; + if (end2 > end1) { + try out.writeAll(" "); + for (bytes[end1..end2]) |b| { + try out.writeAll(" "); + try printHexValue(out, b, 2); + } + } + const short = 16 - n; + var i: usize = 0; + while (i < short) : (i += 1) { + try out.writeAll(" "); + } + if (end2 > end1) { + try out.writeAll(" |"); + } else { + try out.writeAll(" |"); + } + try printCharValues(out, bytes[offset..end2]); + try out.writeAll("|\n"); + offset += n; + } + + try printDecValue(out, offset, 8); + try out.writeAll(":"); + try out.writeAll("\n"); +} + +fn hexDump16(out: anytype, offset: usize, bytes: []const u8) !void { + try printDecValue(out, offset, 8); + try out.writeAll(":"); + try out.writeAll(" "); + for (bytes[0..8]) |b| { + try out.writeAll(" "); + try printHexValue(out, b, 2); + } + try out.writeAll(" "); + for (bytes[8..16]) |b| { + try out.writeAll(" "); + try printHexValue(out, b, 2); + } + try out.writeAll(" |"); + try printCharValues(out, bytes); + try out.writeAll("|\n"); +} + +fn printDecValue(out: anytype, value: u64, width: u8) !void { + var buffer: [20]u8 = undefined; + const len = std.fmt.formatIntBuf(buffer[0..], value, 10, .lower, .{ .width = width, .fill = '0' }); + try out.writeAll(buffer[0..len]); +} + +fn printHexValue(out: anytype, value: u64, width: u8) !void { + var buffer: [16]u8 = undefined; + const len = std.fmt.formatIntBuf(buffer[0..], value, 16, .lower, .{ .width = width, .fill = '0' }); + try out.writeAll(buffer[0..len]); +} + +fn printCharValues(out: anytype, bytes: []const u8) !void { + for (bytes) |b| { + try out.writeAll(&[_]u8{printable_char_tab[b]}); + } +} + +fn printUnderstandableChar(out: anytype, char: u8) !void { + if (std.ascii.isPrint(char)) { + try out.print("'{c}'", .{char}); + } else { + try out.print("\\x{X:0>2}", .{char}); + } +} + +// zig fmt: off +const printable_char_tab: [256]u8 = ( + "................................ !\"#$%&'()*+,-./0123456789:;<=>?" ++ + "@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~." ++ + "................................................................" ++ + "................................................................" +).*; diff --git a/src/Cache.zig b/src/Cache.zig deleted file mode 100644 index 3020f8e8c6..0000000000 --- a/src/Cache.zig +++ /dev/null @@ -1,1265 +0,0 @@ -//! Manages `zig-cache` directories. -//! This is not a general-purpose cache. It is designed to be fast and simple, -//! not to withstand attacks using specially-crafted input. - -gpa: Allocator, -manifest_dir: fs.Dir, -hash: HashHelper = .{}, -/// This value is accessed from multiple threads, protected by mutex. -recent_problematic_timestamp: i128 = 0, -mutex: std.Thread.Mutex = .{}, - -/// A set of strings such as the zig library directory or project source root, which -/// are stripped from the file paths before putting into the cache. They -/// are replaced with single-character indicators. This is not to save -/// space but to eliminate absolute file paths. This improves portability -/// and usefulness of the cache for advanced use cases. -prefixes_buffer: [3]Compilation.Directory = undefined, -prefixes_len: usize = 0, - -const Cache = @This(); -const std = @import("std"); -const builtin = @import("builtin"); -const crypto = std.crypto; -const fs = std.fs; -const assert = std.debug.assert; -const testing = std.testing; -const mem = std.mem; -const fmt = std.fmt; -const Allocator = std.mem.Allocator; -const Compilation = @import("Compilation.zig"); -const log = std.log.scoped(.cache); - -pub fn addPrefix(cache: *Cache, directory: Compilation.Directory) void { - if (directory.path) |p| { - log.debug("Cache.addPrefix {d} {s}", .{ cache.prefixes_len, p }); - } - cache.prefixes_buffer[cache.prefixes_len] = directory; - cache.prefixes_len += 1; -} - -/// Be sure to call `Manifest.deinit` after successful initialization. -pub fn obtain(cache: *Cache) Manifest { - return Manifest{ - .cache = cache, - .hash = cache.hash, - .manifest_file = null, - .manifest_dirty = false, - .hex_digest = undefined, - }; -} - -pub fn prefixes(cache: *const Cache) []const Compilation.Directory { - return cache.prefixes_buffer[0..cache.prefixes_len]; -} - -const PrefixedPath = struct { - prefix: u8, - sub_path: []u8, -}; - -fn findPrefix(cache: *const Cache, file_path: []const u8) !PrefixedPath { - const gpa = cache.gpa; - const resolved_path = try fs.path.resolve(gpa, &[_][]const u8{file_path}); - errdefer gpa.free(resolved_path); - return findPrefixResolved(cache, resolved_path); -} - -/// Takes ownership of `resolved_path` on success. -fn findPrefixResolved(cache: *const Cache, resolved_path: []u8) !PrefixedPath { - const gpa = cache.gpa; - const prefixes_slice = cache.prefixes(); - var i: u8 = 1; // Start at 1 to skip over checking the null prefix. - while (i < prefixes_slice.len) : (i += 1) { - const p = prefixes_slice[i].path.?; - if (mem.startsWith(u8, resolved_path, p)) { - // +1 to skip over the path separator here - const sub_path = try gpa.dupe(u8, resolved_path[p.len + 1 ..]); - gpa.free(resolved_path); - return PrefixedPath{ - .prefix = @intCast(u8, i), - .sub_path = sub_path, - }; - } else { - log.debug("'{s}' does not start with '{s}'", .{ resolved_path, p }); - } - } - - return PrefixedPath{ - .prefix = 0, - .sub_path = resolved_path, - }; -} - -/// This is 128 bits - Even with 2^54 cache entries, the probably of a collision would be under 10^-6 -pub const bin_digest_len = 16; -pub const hex_digest_len = bin_digest_len * 2; -pub const BinDigest = [bin_digest_len]u8; - -const manifest_file_size_max = 50 * 1024 * 1024; - -/// The type used for hashing file contents. Currently, this is SipHash128(1, 3), because it -/// provides enough collision resistance for the Manifest use cases, while being one of our -/// fastest options right now. -pub const Hasher = crypto.auth.siphash.SipHash128(1, 3); - -/// Initial state, that can be copied. -pub const hasher_init: Hasher = Hasher.init(&[_]u8{0} ** Hasher.key_length); - -pub const File = struct { - prefixed_path: ?PrefixedPath, - max_file_size: ?usize, - stat: Stat, - bin_digest: BinDigest, - contents: ?[]const u8, - - pub const Stat = struct { - inode: fs.File.INode, - size: u64, - mtime: i128, - }; - - pub fn deinit(self: *File, gpa: Allocator) void { - if (self.prefixed_path) |pp| { - gpa.free(pp.sub_path); - self.prefixed_path = null; - } - if (self.contents) |contents| { - gpa.free(contents); - self.contents = null; - } - self.* = undefined; - } -}; - -pub const HashHelper = struct { - hasher: Hasher = hasher_init, - - const EmitLoc = Compilation.EmitLoc; - - /// Record a slice of bytes as an dependency of the process being cached - pub fn addBytes(hh: *HashHelper, bytes: []const u8) void { - hh.hasher.update(mem.asBytes(&bytes.len)); - hh.hasher.update(bytes); - } - - pub fn addOptionalBytes(hh: *HashHelper, optional_bytes: ?[]const u8) void { - hh.add(optional_bytes != null); - hh.addBytes(optional_bytes orelse return); - } - - pub fn addEmitLoc(hh: *HashHelper, emit_loc: EmitLoc) void { - hh.addBytes(emit_loc.basename); - } - - pub fn addOptionalEmitLoc(hh: *HashHelper, optional_emit_loc: ?EmitLoc) void { - hh.add(optional_emit_loc != null); - hh.addEmitLoc(optional_emit_loc orelse return); - } - - pub fn addListOfBytes(hh: *HashHelper, list_of_bytes: []const []const u8) void { - hh.add(list_of_bytes.len); - for (list_of_bytes) |bytes| hh.addBytes(bytes); - } - - /// Convert the input value into bytes and record it as a dependency of the process being cached. - pub fn add(hh: *HashHelper, x: anytype) void { - switch (@TypeOf(x)) { - std.builtin.Version => { - hh.add(x.major); - hh.add(x.minor); - hh.add(x.patch); - }, - std.Target.Os.TaggedVersionRange => { - switch (x) { - .linux => |linux| { - hh.add(linux.range.min); - hh.add(linux.range.max); - hh.add(linux.glibc); - }, - .windows => |windows| { - hh.add(windows.min); - hh.add(windows.max); - }, - .semver => |semver| { - hh.add(semver.min); - hh.add(semver.max); - }, - .none => {}, - } - }, - else => switch (@typeInfo(@TypeOf(x))) { - .Bool, .Int, .Enum, .Array => hh.addBytes(mem.asBytes(&x)), - else => @compileError("unable to hash type " ++ @typeName(@TypeOf(x))), - }, - } - } - - pub fn addOptional(hh: *HashHelper, optional: anytype) void { - hh.add(optional != null); - hh.add(optional orelse return); - } - - /// Returns a hex encoded hash of the inputs, without modifying state. - pub fn peek(hh: HashHelper) [hex_digest_len]u8 { - var copy = hh; - return copy.final(); - } - - pub fn peekBin(hh: HashHelper) BinDigest { - var copy = hh; - var bin_digest: BinDigest = undefined; - copy.hasher.final(&bin_digest); - return bin_digest; - } - - /// Returns a hex encoded hash of the inputs, mutating the state of the hasher. - pub fn final(hh: *HashHelper) [hex_digest_len]u8 { - var bin_digest: BinDigest = undefined; - hh.hasher.final(&bin_digest); - - var out_digest: [hex_digest_len]u8 = undefined; - _ = std.fmt.bufPrint( - &out_digest, - "{s}", - .{std.fmt.fmtSliceHexLower(&bin_digest)}, - ) catch unreachable; - return out_digest; - } -}; - -pub const Lock = struct { - manifest_file: fs.File, - - pub fn release(lock: *Lock) void { - if (builtin.os.tag == .windows) { - // Windows does not guarantee that locks are immediately unlocked when - // the file handle is closed. See LockFileEx documentation. - lock.manifest_file.unlock(); - } - - lock.manifest_file.close(); - lock.* = undefined; - } -}; - -pub const Manifest = struct { - cache: *Cache, - /// Current state for incremental hashing. - hash: HashHelper, - manifest_file: ?fs.File, - manifest_dirty: bool, - /// Set this flag to true before calling hit() in order to indicate that - /// upon a cache hit, the code using the cache will not modify the files - /// within the cache directory. This allows multiple processes to utilize - /// the same cache directory at the same time. - want_shared_lock: bool = true, - have_exclusive_lock: bool = false, - // Indicate that we want isProblematicTimestamp to perform a filesystem write in - // order to obtain a problematic timestamp for the next call. Calls after that - // will then use the same timestamp, to avoid unnecessary filesystem writes. - want_refresh_timestamp: bool = true, - files: std.ArrayListUnmanaged(File) = .{}, - hex_digest: [hex_digest_len]u8, - /// Populated when hit() returns an error because of one - /// of the files listed in the manifest. - failed_file_index: ?usize = null, - /// Keeps track of the last time we performed a file system write to observe - /// what time the file system thinks it is, according to its own granularity. - recent_problematic_timestamp: i128 = 0, - - /// Add a file as a dependency of process being cached. When `hit` is - /// called, the file's contents will be checked to ensure that it matches - /// the contents from previous times. - /// - /// Max file size will be used to determine the amount of space the file contents - /// are allowed to take up in memory. If max_file_size is null, then the contents - /// will not be loaded into memory. - /// - /// Returns the index of the entry in the `files` array list. You can use it - /// to access the contents of the file after calling `hit()` like so: - /// - /// ``` - /// var file_contents = cache_hash.files.items[file_index].contents.?; - /// ``` - pub fn addFile(self: *Manifest, file_path: []const u8, max_file_size: ?usize) !usize { - assert(self.manifest_file == null); - - const gpa = self.cache.gpa; - try self.files.ensureUnusedCapacity(gpa, 1); - const prefixed_path = try self.cache.findPrefix(file_path); - errdefer gpa.free(prefixed_path.sub_path); - - log.debug("Manifest.addFile {s} -> {d} {s}", .{ - file_path, prefixed_path.prefix, prefixed_path.sub_path, - }); - - self.files.addOneAssumeCapacity().* = .{ - .prefixed_path = prefixed_path, - .contents = null, - .max_file_size = max_file_size, - .stat = undefined, - .bin_digest = undefined, - }; - - self.hash.add(prefixed_path.prefix); - self.hash.addBytes(prefixed_path.sub_path); - - return self.files.items.len - 1; - } - - pub fn hashCSource(self: *Manifest, c_source: Compilation.CSourceFile) !void { - _ = try self.addFile(c_source.src_path, null); - // Hash the extra flags, with special care to call addFile for file parameters. - // TODO this logic can likely be improved by utilizing clang_options_data.zig. - const file_args = [_][]const u8{"-include"}; - var arg_i: usize = 0; - while (arg_i < c_source.extra_flags.len) : (arg_i += 1) { - const arg = c_source.extra_flags[arg_i]; - self.hash.addBytes(arg); - for (file_args) |file_arg| { - if (mem.eql(u8, file_arg, arg) and arg_i + 1 < c_source.extra_flags.len) { - arg_i += 1; - _ = try self.addFile(c_source.extra_flags[arg_i], null); - } - } - } - } - - pub fn addOptionalFile(self: *Manifest, optional_file_path: ?[]const u8) !void { - self.hash.add(optional_file_path != null); - const file_path = optional_file_path orelse return; - _ = try self.addFile(file_path, null); - } - - pub fn addListOfFiles(self: *Manifest, list_of_files: []const []const u8) !void { - self.hash.add(list_of_files.len); - for (list_of_files) |file_path| { - _ = try self.addFile(file_path, null); - } - } - - /// Check the cache to see if the input exists in it. If it exists, returns `true`. - /// A hex encoding of its hash is available by calling `final`. - /// - /// This function will also acquire an exclusive lock to the manifest file. This means - /// that a process holding a Manifest will block any other process attempting to - /// acquire the lock. If `want_shared_lock` is `true`, a cache hit guarantees the - /// manifest file to be locked in shared mode, and a cache miss guarantees the manifest - /// file to be locked in exclusive mode. - /// - /// The lock on the manifest file is released when `deinit` is called. As another - /// option, one may call `toOwnedLock` to obtain a smaller object which can represent - /// the lock. `deinit` is safe to call whether or not `toOwnedLock` has been called. - pub fn hit(self: *Manifest) !bool { - const gpa = self.cache.gpa; - assert(self.manifest_file == null); - - self.failed_file_index = null; - - const ext = ".txt"; - var manifest_file_path: [self.hex_digest.len + ext.len]u8 = undefined; - - var bin_digest: BinDigest = undefined; - self.hash.hasher.final(&bin_digest); - - _ = std.fmt.bufPrint( - &self.hex_digest, - "{s}", - .{std.fmt.fmtSliceHexLower(&bin_digest)}, - ) catch unreachable; - - self.hash.hasher = hasher_init; - self.hash.hasher.update(&bin_digest); - - mem.copy(u8, &manifest_file_path, &self.hex_digest); - manifest_file_path[self.hex_digest.len..][0..ext.len].* = ext.*; - - if (self.files.items.len == 0) { - // If there are no file inputs, we check if the manifest file exists instead of - // comparing the hashes on the files used for the cached item - while (true) { - if (self.cache.manifest_dir.openFile(&manifest_file_path, .{ - .mode = .read_write, - .lock = .Exclusive, - .lock_nonblocking = self.want_shared_lock, - })) |manifest_file| { - self.manifest_file = manifest_file; - self.have_exclusive_lock = true; - break; - } else |open_err| switch (open_err) { - error.WouldBlock => { - self.manifest_file = try self.cache.manifest_dir.openFile(&manifest_file_path, .{ - .lock = .Shared, - }); - break; - }, - error.FileNotFound => { - if (self.cache.manifest_dir.createFile(&manifest_file_path, .{ - .read = true, - .truncate = false, - .lock = .Exclusive, - .lock_nonblocking = self.want_shared_lock, - })) |manifest_file| { - self.manifest_file = manifest_file; - self.manifest_dirty = true; - self.have_exclusive_lock = true; - return false; // cache miss; exclusive lock already held - } else |err| switch (err) { - error.WouldBlock => continue, - else => |e| return e, - } - }, - else => |e| return e, - } - } - } else { - if (self.cache.manifest_dir.createFile(&manifest_file_path, .{ - .read = true, - .truncate = false, - .lock = .Exclusive, - .lock_nonblocking = self.want_shared_lock, - })) |manifest_file| { - self.manifest_file = manifest_file; - self.have_exclusive_lock = true; - } else |err| switch (err) { - error.WouldBlock => { - self.manifest_file = try self.cache.manifest_dir.openFile(&manifest_file_path, .{ - .lock = .Shared, - }); - }, - else => |e| return e, - } - } - - self.want_refresh_timestamp = true; - - const file_contents = try self.manifest_file.?.reader().readAllAlloc(gpa, manifest_file_size_max); - defer gpa.free(file_contents); - - const input_file_count = self.files.items.len; - var any_file_changed = false; - var line_iter = mem.tokenize(u8, file_contents, "\n"); - var idx: usize = 0; - while (line_iter.next()) |line| { - defer idx += 1; - - const cache_hash_file = if (idx < input_file_count) &self.files.items[idx] else blk: { - const new = try self.files.addOne(gpa); - new.* = .{ - .prefixed_path = null, - .contents = null, - .max_file_size = null, - .stat = undefined, - .bin_digest = undefined, - }; - break :blk new; - }; - - var iter = mem.tokenize(u8, line, " "); - const size = iter.next() orelse return error.InvalidFormat; - const inode = iter.next() orelse return error.InvalidFormat; - const mtime_nsec_str = iter.next() orelse return error.InvalidFormat; - const digest_str = iter.next() orelse return error.InvalidFormat; - const prefix_str = iter.next() orelse return error.InvalidFormat; - const file_path = iter.rest(); - - cache_hash_file.stat.size = fmt.parseInt(u64, size, 10) catch return error.InvalidFormat; - cache_hash_file.stat.inode = fmt.parseInt(fs.File.INode, inode, 10) catch return error.InvalidFormat; - cache_hash_file.stat.mtime = fmt.parseInt(i64, mtime_nsec_str, 10) catch return error.InvalidFormat; - _ = std.fmt.hexToBytes(&cache_hash_file.bin_digest, digest_str) catch return error.InvalidFormat; - const prefix = fmt.parseInt(u8, prefix_str, 10) catch return error.InvalidFormat; - if (prefix >= self.cache.prefixes_len) return error.InvalidFormat; - - if (file_path.len == 0) { - return error.InvalidFormat; - } - if (cache_hash_file.prefixed_path) |pp| { - if (pp.prefix != prefix or !mem.eql(u8, file_path, pp.sub_path)) { - return error.InvalidFormat; - } - } - - if (cache_hash_file.prefixed_path == null) { - cache_hash_file.prefixed_path = .{ - .prefix = prefix, - .sub_path = try gpa.dupe(u8, file_path), - }; - } - - const pp = cache_hash_file.prefixed_path.?; - const dir = self.cache.prefixes()[pp.prefix].handle; - const this_file = dir.openFile(pp.sub_path, .{ .mode = .read_only }) catch |err| switch (err) { - error.FileNotFound => { - try self.upgradeToExclusiveLock(); - return false; - }, - else => return error.CacheUnavailable, - }; - defer this_file.close(); - - const actual_stat = this_file.stat() catch |err| { - self.failed_file_index = idx; - return err; - }; - const size_match = actual_stat.size == cache_hash_file.stat.size; - const mtime_match = actual_stat.mtime == cache_hash_file.stat.mtime; - const inode_match = actual_stat.inode == cache_hash_file.stat.inode; - - if (!size_match or !mtime_match or !inode_match) { - self.manifest_dirty = true; - - cache_hash_file.stat = .{ - .size = actual_stat.size, - .mtime = actual_stat.mtime, - .inode = actual_stat.inode, - }; - - if (self.isProblematicTimestamp(cache_hash_file.stat.mtime)) { - // The actual file has an unreliable timestamp, force it to be hashed - cache_hash_file.stat.mtime = 0; - cache_hash_file.stat.inode = 0; - } - - var actual_digest: BinDigest = undefined; - hashFile(this_file, &actual_digest) catch |err| { - self.failed_file_index = idx; - return err; - }; - - if (!mem.eql(u8, &cache_hash_file.bin_digest, &actual_digest)) { - cache_hash_file.bin_digest = actual_digest; - // keep going until we have the input file digests - any_file_changed = true; - } - } - - if (!any_file_changed) { - self.hash.hasher.update(&cache_hash_file.bin_digest); - } - } - - if (any_file_changed) { - // cache miss - // keep the manifest file open - self.unhit(bin_digest, input_file_count); - try self.upgradeToExclusiveLock(); - return false; - } - - if (idx < input_file_count) { - self.manifest_dirty = true; - while (idx < input_file_count) : (idx += 1) { - const ch_file = &self.files.items[idx]; - self.populateFileHash(ch_file) catch |err| { - self.failed_file_index = idx; - return err; - }; - } - try self.upgradeToExclusiveLock(); - return false; - } - - if (self.want_shared_lock) { - try self.downgradeToSharedLock(); - } - - return true; - } - - pub fn unhit(self: *Manifest, bin_digest: BinDigest, input_file_count: usize) void { - // Reset the hash. - self.hash.hasher = hasher_init; - self.hash.hasher.update(&bin_digest); - - // Remove files not in the initial hash. - for (self.files.items[input_file_count..]) |*file| { - file.deinit(self.cache.gpa); - } - self.files.shrinkRetainingCapacity(input_file_count); - - for (self.files.items) |file| { - self.hash.hasher.update(&file.bin_digest); - } - } - - fn isProblematicTimestamp(man: *Manifest, file_time: i128) bool { - // If the file_time is prior to the most recent problematic timestamp - // then we don't need to access the filesystem. - if (file_time < man.recent_problematic_timestamp) - return false; - - // Next we will check the globally shared Cache timestamp, which is accessed - // from multiple threads. - man.cache.mutex.lock(); - defer man.cache.mutex.unlock(); - - // Save the global one to our local one to avoid locking next time. - man.recent_problematic_timestamp = man.cache.recent_problematic_timestamp; - if (file_time < man.recent_problematic_timestamp) - return false; - - // This flag prevents multiple filesystem writes for the same hit() call. - if (man.want_refresh_timestamp) { - man.want_refresh_timestamp = false; - - var file = man.cache.manifest_dir.createFile("timestamp", .{ - .read = true, - .truncate = true, - }) catch return true; - defer file.close(); - - // Save locally and also save globally (we still hold the global lock). - man.recent_problematic_timestamp = (file.stat() catch return true).mtime; - man.cache.recent_problematic_timestamp = man.recent_problematic_timestamp; - } - - return file_time >= man.recent_problematic_timestamp; - } - - fn populateFileHash(self: *Manifest, ch_file: *File) !void { - const pp = ch_file.prefixed_path.?; - const dir = self.cache.prefixes()[pp.prefix].handle; - const file = try dir.openFile(pp.sub_path, .{}); - defer file.close(); - - const actual_stat = try file.stat(); - ch_file.stat = .{ - .size = actual_stat.size, - .mtime = actual_stat.mtime, - .inode = actual_stat.inode, - }; - - if (self.isProblematicTimestamp(ch_file.stat.mtime)) { - // The actual file has an unreliable timestamp, force it to be hashed - ch_file.stat.mtime = 0; - ch_file.stat.inode = 0; - } - - if (ch_file.max_file_size) |max_file_size| { - if (ch_file.stat.size > max_file_size) { - return error.FileTooBig; - } - - const contents = try self.cache.gpa.alloc(u8, @intCast(usize, ch_file.stat.size)); - errdefer self.cache.gpa.free(contents); - - // Hash while reading from disk, to keep the contents in the cpu cache while - // doing hashing. - var hasher = hasher_init; - var off: usize = 0; - while (true) { - // give me everything you've got, captain - const bytes_read = try file.read(contents[off..]); - if (bytes_read == 0) break; - hasher.update(contents[off..][0..bytes_read]); - off += bytes_read; - } - hasher.final(&ch_file.bin_digest); - - ch_file.contents = contents; - } else { - try hashFile(file, &ch_file.bin_digest); - } - - self.hash.hasher.update(&ch_file.bin_digest); - } - - /// Add a file as a dependency of process being cached, after the initial hash has been - /// calculated. This is useful for processes that don't know all the files that - /// are depended on ahead of time. For example, a source file that can import other files - /// will need to be recompiled if the imported file is changed. - pub fn addFilePostFetch(self: *Manifest, file_path: []const u8, max_file_size: usize) ![]const u8 { - assert(self.manifest_file != null); - - const gpa = self.cache.gpa; - const prefixed_path = try self.cache.findPrefix(file_path); - errdefer gpa.free(prefixed_path.sub_path); - - log.debug("Manifest.addFilePostFetch {s} -> {d} {s}", .{ - file_path, prefixed_path.prefix, prefixed_path.sub_path, - }); - - const new_ch_file = try self.files.addOne(gpa); - new_ch_file.* = .{ - .prefixed_path = prefixed_path, - .max_file_size = max_file_size, - .stat = undefined, - .bin_digest = undefined, - .contents = null, - }; - errdefer self.files.shrinkRetainingCapacity(self.files.items.len - 1); - - try self.populateFileHash(new_ch_file); - - return new_ch_file.contents.?; - } - - /// Add a file as a dependency of process being cached, after the initial hash has been - /// calculated. This is useful for processes that don't know the all the files that - /// are depended on ahead of time. For example, a source file that can import other files - /// will need to be recompiled if the imported file is changed. - pub fn addFilePost(self: *Manifest, file_path: []const u8) !void { - assert(self.manifest_file != null); - - const gpa = self.cache.gpa; - const prefixed_path = try self.cache.findPrefix(file_path); - errdefer gpa.free(prefixed_path.sub_path); - - log.debug("Manifest.addFilePost {s} -> {d} {s}", .{ - file_path, prefixed_path.prefix, prefixed_path.sub_path, - }); - - const new_ch_file = try self.files.addOne(gpa); - new_ch_file.* = .{ - .prefixed_path = prefixed_path, - .max_file_size = null, - .stat = undefined, - .bin_digest = undefined, - .contents = null, - }; - errdefer self.files.shrinkRetainingCapacity(self.files.items.len - 1); - - try self.populateFileHash(new_ch_file); - } - - /// Like `addFilePost` but when the file contents have already been loaded from disk. - /// On success, cache takes ownership of `resolved_path`. - pub fn addFilePostContents( - self: *Manifest, - resolved_path: []u8, - bytes: []const u8, - stat: File.Stat, - ) error{OutOfMemory}!void { - assert(self.manifest_file != null); - const gpa = self.cache.gpa; - - const ch_file = try self.files.addOne(gpa); - errdefer self.files.shrinkRetainingCapacity(self.files.items.len - 1); - - log.debug("Manifest.addFilePostContents resolved_path={s}", .{resolved_path}); - - const prefixed_path = try self.cache.findPrefixResolved(resolved_path); - errdefer gpa.free(prefixed_path.sub_path); - - log.debug("Manifest.addFilePostContents -> {d} {s}", .{ - prefixed_path.prefix, prefixed_path.sub_path, - }); - - ch_file.* = .{ - .prefixed_path = prefixed_path, - .max_file_size = null, - .stat = stat, - .bin_digest = undefined, - .contents = null, - }; - - if (self.isProblematicTimestamp(ch_file.stat.mtime)) { - // The actual file has an unreliable timestamp, force it to be hashed - ch_file.stat.mtime = 0; - ch_file.stat.inode = 0; - } - - { - var hasher = hasher_init; - hasher.update(bytes); - hasher.final(&ch_file.bin_digest); - } - - self.hash.hasher.update(&ch_file.bin_digest); - } - - pub fn addDepFilePost(self: *Manifest, dir: fs.Dir, dep_file_basename: []const u8) !void { - assert(self.manifest_file != null); - - const dep_file_contents = try dir.readFileAlloc(self.cache.gpa, dep_file_basename, manifest_file_size_max); - defer self.cache.gpa.free(dep_file_contents); - - var error_buf = std.ArrayList(u8).init(self.cache.gpa); - defer error_buf.deinit(); - - var it: @import("DepTokenizer.zig") = .{ .bytes = dep_file_contents }; - - // Skip first token: target. - switch (it.next() orelse return) { // Empty dep file OK. - .target, .target_must_resolve, .prereq => {}, - else => |err| { - try err.printError(error_buf.writer()); - log.err("failed parsing {s}: {s}", .{ dep_file_basename, error_buf.items }); - return error.InvalidDepFile; - }, - } - // Process 0+ preqreqs. - // Clang is invoked in single-source mode so we never get more targets. - while (true) { - switch (it.next() orelse return) { - .target, .target_must_resolve => return, - .prereq => |file_path| try self.addFilePost(file_path), - else => |err| { - try err.printError(error_buf.writer()); - log.err("failed parsing {s}: {s}", .{ dep_file_basename, error_buf.items }); - return error.InvalidDepFile; - }, - } - } - } - - /// Returns a hex encoded hash of the inputs. - pub fn final(self: *Manifest) [hex_digest_len]u8 { - assert(self.manifest_file != null); - - // We don't close the manifest file yet, because we want to - // keep it locked until the API user is done using it. - // We also don't write out the manifest yet, because until - // cache_release is called we still might be working on creating - // the artifacts to cache. - - var bin_digest: BinDigest = undefined; - self.hash.hasher.final(&bin_digest); - - var out_digest: [hex_digest_len]u8 = undefined; - _ = std.fmt.bufPrint( - &out_digest, - "{s}", - .{std.fmt.fmtSliceHexLower(&bin_digest)}, - ) catch unreachable; - - return out_digest; - } - - /// If `want_shared_lock` is true, this function automatically downgrades the - /// lock from exclusive to shared. - pub fn writeManifest(self: *Manifest) !void { - assert(self.have_exclusive_lock); - - const manifest_file = self.manifest_file.?; - if (self.manifest_dirty) { - self.manifest_dirty = false; - - var contents = std.ArrayList(u8).init(self.cache.gpa); - defer contents.deinit(); - - const writer = contents.writer(); - var encoded_digest: [hex_digest_len]u8 = undefined; - - for (self.files.items) |file| { - _ = std.fmt.bufPrint( - &encoded_digest, - "{s}", - .{std.fmt.fmtSliceHexLower(&file.bin_digest)}, - ) catch unreachable; - try writer.print("{d} {d} {d} {s} {d} {s}\n", .{ - file.stat.size, - file.stat.inode, - file.stat.mtime, - &encoded_digest, - file.prefixed_path.?.prefix, - file.prefixed_path.?.sub_path, - }); - } - - try manifest_file.setEndPos(contents.items.len); - try manifest_file.pwriteAll(contents.items, 0); - } - - if (self.want_shared_lock) { - try self.downgradeToSharedLock(); - } - } - - fn downgradeToSharedLock(self: *Manifest) !void { - if (!self.have_exclusive_lock) return; - - // WASI does not currently support flock, so we bypass it here. - // TODO: If/when flock is supported on WASI, this check should be removed. - // See https://github.com/WebAssembly/wasi-filesystem/issues/2 - if (builtin.os.tag != .wasi or std.process.can_spawn or !builtin.single_threaded) { - const manifest_file = self.manifest_file.?; - try manifest_file.downgradeLock(); - } - - self.have_exclusive_lock = false; - } - - fn upgradeToExclusiveLock(self: *Manifest) !void { - if (self.have_exclusive_lock) return; - assert(self.manifest_file != null); - - // WASI does not currently support flock, so we bypass it here. - // TODO: If/when flock is supported on WASI, this check should be removed. - // See https://github.com/WebAssembly/wasi-filesystem/issues/2 - if (builtin.os.tag != .wasi or std.process.can_spawn or !builtin.single_threaded) { - const manifest_file = self.manifest_file.?; - // Here we intentionally have a period where the lock is released, in case there are - // other processes holding a shared lock. - manifest_file.unlock(); - try manifest_file.lock(.Exclusive); - } - self.have_exclusive_lock = true; - } - - /// Obtain only the data needed to maintain a lock on the manifest file. - /// The `Manifest` remains safe to deinit. - /// Don't forget to call `writeManifest` before this! - pub fn toOwnedLock(self: *Manifest) Lock { - const lock: Lock = .{ - .manifest_file = self.manifest_file.?, - }; - - self.manifest_file = null; - return lock; - } - - /// Releases the manifest file and frees any memory the Manifest was using. - /// `Manifest.hit` must be called first. - /// Don't forget to call `writeManifest` before this! - pub fn deinit(self: *Manifest) void { - if (self.manifest_file) |file| { - if (builtin.os.tag == .windows) { - // See Lock.release for why this is required on Windows - file.unlock(); - } - - file.close(); - } - for (self.files.items) |*file| { - file.deinit(self.cache.gpa); - } - self.files.deinit(self.cache.gpa); - } -}; - -/// On operating systems that support symlinks, does a readlink. On other operating systems, -/// uses the file contents. Windows supports symlinks but only with elevated privileges, so -/// it is treated as not supporting symlinks. -pub fn readSmallFile(dir: fs.Dir, sub_path: []const u8, buffer: []u8) ![]u8 { - if (builtin.os.tag == .windows) { - return dir.readFile(sub_path, buffer); - } else { - return dir.readLink(sub_path, buffer); - } -} - -/// On operating systems that support symlinks, does a symlink. On other operating systems, -/// uses the file contents. Windows supports symlinks but only with elevated privileges, so -/// it is treated as not supporting symlinks. -/// `data` must be a valid UTF-8 encoded file path and 255 bytes or fewer. -pub fn writeSmallFile(dir: fs.Dir, sub_path: []const u8, data: []const u8) !void { - assert(data.len <= 255); - if (builtin.os.tag == .windows) { - return dir.writeFile(sub_path, data); - } else { - return dir.symLink(data, sub_path, .{}); - } -} - -fn hashFile(file: fs.File, bin_digest: *[Hasher.mac_length]u8) !void { - var buf: [1024]u8 = undefined; - - var hasher = hasher_init; - while (true) { - const bytes_read = try file.read(&buf); - if (bytes_read == 0) break; - hasher.update(buf[0..bytes_read]); - } - - hasher.final(bin_digest); -} - -// Create/Write a file, close it, then grab its stat.mtime timestamp. -fn testGetCurrentFileTimestamp() !i128 { - var file = try fs.cwd().createFile("test-filetimestamp.tmp", .{ - .read = true, - .truncate = true, - }); - defer file.close(); - - return (try file.stat()).mtime; -} - -test "cache file and then recall it" { - if (builtin.os.tag == .wasi) { - // https://github.com/ziglang/zig/issues/5437 - return error.SkipZigTest; - } - - const cwd = fs.cwd(); - - const temp_file = "test.txt"; - const temp_manifest_dir = "temp_manifest_dir"; - - try cwd.writeFile(temp_file, "Hello, world!\n"); - - // Wait for file timestamps to tick - const initial_time = try testGetCurrentFileTimestamp(); - while ((try testGetCurrentFileTimestamp()) == initial_time) { - std.time.sleep(1); - } - - var digest1: [hex_digest_len]u8 = undefined; - var digest2: [hex_digest_len]u8 = undefined; - - { - var cache = Cache{ - .gpa = testing.allocator, - .manifest_dir = try cwd.makeOpenPath(temp_manifest_dir, .{}), - }; - cache.addPrefix(.{ .path = null, .handle = fs.cwd() }); - defer cache.manifest_dir.close(); - - { - var ch = cache.obtain(); - defer ch.deinit(); - - ch.hash.add(true); - ch.hash.add(@as(u16, 1234)); - ch.hash.addBytes("1234"); - _ = try ch.addFile(temp_file, null); - - // There should be nothing in the cache - try testing.expectEqual(false, try ch.hit()); - - digest1 = ch.final(); - try ch.writeManifest(); - } - { - var ch = cache.obtain(); - defer ch.deinit(); - - ch.hash.add(true); - ch.hash.add(@as(u16, 1234)); - ch.hash.addBytes("1234"); - _ = try ch.addFile(temp_file, null); - - // Cache hit! We just "built" the same file - try testing.expect(try ch.hit()); - digest2 = ch.final(); - - try testing.expectEqual(false, ch.have_exclusive_lock); - } - - try testing.expectEqual(digest1, digest2); - } - - try cwd.deleteTree(temp_manifest_dir); - try cwd.deleteFile(temp_file); -} - -test "check that changing a file makes cache fail" { - if (builtin.os.tag == .wasi) { - // https://github.com/ziglang/zig/issues/5437 - return error.SkipZigTest; - } - const cwd = fs.cwd(); - - const temp_file = "cache_hash_change_file_test.txt"; - const temp_manifest_dir = "cache_hash_change_file_manifest_dir"; - const original_temp_file_contents = "Hello, world!\n"; - const updated_temp_file_contents = "Hello, world; but updated!\n"; - - try cwd.deleteTree(temp_manifest_dir); - try cwd.deleteTree(temp_file); - - try cwd.writeFile(temp_file, original_temp_file_contents); - - // Wait for file timestamps to tick - const initial_time = try testGetCurrentFileTimestamp(); - while ((try testGetCurrentFileTimestamp()) == initial_time) { - std.time.sleep(1); - } - - var digest1: [hex_digest_len]u8 = undefined; - var digest2: [hex_digest_len]u8 = undefined; - - { - var cache = Cache{ - .gpa = testing.allocator, - .manifest_dir = try cwd.makeOpenPath(temp_manifest_dir, .{}), - }; - cache.addPrefix(.{ .path = null, .handle = fs.cwd() }); - defer cache.manifest_dir.close(); - - { - var ch = cache.obtain(); - defer ch.deinit(); - - ch.hash.addBytes("1234"); - const temp_file_idx = try ch.addFile(temp_file, 100); - - // There should be nothing in the cache - try testing.expectEqual(false, try ch.hit()); - - try testing.expect(mem.eql(u8, original_temp_file_contents, ch.files.items[temp_file_idx].contents.?)); - - digest1 = ch.final(); - - try ch.writeManifest(); - } - - try cwd.writeFile(temp_file, updated_temp_file_contents); - - { - var ch = cache.obtain(); - defer ch.deinit(); - - ch.hash.addBytes("1234"); - const temp_file_idx = try ch.addFile(temp_file, 100); - - // A file that we depend on has been updated, so the cache should not contain an entry for it - try testing.expectEqual(false, try ch.hit()); - - // The cache system does not keep the contents of re-hashed input files. - try testing.expect(ch.files.items[temp_file_idx].contents == null); - - digest2 = ch.final(); - - try ch.writeManifest(); - } - - try testing.expect(!mem.eql(u8, digest1[0..], digest2[0..])); - } - - try cwd.deleteTree(temp_manifest_dir); - try cwd.deleteTree(temp_file); -} - -test "no file inputs" { - if (builtin.os.tag == .wasi) { - // https://github.com/ziglang/zig/issues/5437 - return error.SkipZigTest; - } - const cwd = fs.cwd(); - const temp_manifest_dir = "no_file_inputs_manifest_dir"; - defer cwd.deleteTree(temp_manifest_dir) catch {}; - - var digest1: [hex_digest_len]u8 = undefined; - var digest2: [hex_digest_len]u8 = undefined; - - var cache = Cache{ - .gpa = testing.allocator, - .manifest_dir = try cwd.makeOpenPath(temp_manifest_dir, .{}), - }; - cache.addPrefix(.{ .path = null, .handle = fs.cwd() }); - defer cache.manifest_dir.close(); - - { - var man = cache.obtain(); - defer man.deinit(); - - man.hash.addBytes("1234"); - - // There should be nothing in the cache - try testing.expectEqual(false, try man.hit()); - - digest1 = man.final(); - - try man.writeManifest(); - } - { - var man = cache.obtain(); - defer man.deinit(); - - man.hash.addBytes("1234"); - - try testing.expect(try man.hit()); - digest2 = man.final(); - try testing.expectEqual(false, man.have_exclusive_lock); - } - - try testing.expectEqual(digest1, digest2); -} - -test "Manifest with files added after initial hash work" { - if (builtin.os.tag == .wasi) { - // https://github.com/ziglang/zig/issues/5437 - return error.SkipZigTest; - } - const cwd = fs.cwd(); - - const temp_file1 = "cache_hash_post_file_test1.txt"; - const temp_file2 = "cache_hash_post_file_test2.txt"; - const temp_manifest_dir = "cache_hash_post_file_manifest_dir"; - - try cwd.writeFile(temp_file1, "Hello, world!\n"); - try cwd.writeFile(temp_file2, "Hello world the second!\n"); - - // Wait for file timestamps to tick - const initial_time = try testGetCurrentFileTimestamp(); - while ((try testGetCurrentFileTimestamp()) == initial_time) { - std.time.sleep(1); - } - - var digest1: [hex_digest_len]u8 = undefined; - var digest2: [hex_digest_len]u8 = undefined; - var digest3: [hex_digest_len]u8 = undefined; - - { - var cache = Cache{ - .gpa = testing.allocator, - .manifest_dir = try cwd.makeOpenPath(temp_manifest_dir, .{}), - }; - cache.addPrefix(.{ .path = null, .handle = fs.cwd() }); - defer cache.manifest_dir.close(); - - { - var ch = cache.obtain(); - defer ch.deinit(); - - ch.hash.addBytes("1234"); - _ = try ch.addFile(temp_file1, null); - - // There should be nothing in the cache - try testing.expectEqual(false, try ch.hit()); - - _ = try ch.addFilePost(temp_file2); - - digest1 = ch.final(); - try ch.writeManifest(); - } - { - var ch = cache.obtain(); - defer ch.deinit(); - - ch.hash.addBytes("1234"); - _ = try ch.addFile(temp_file1, null); - - try testing.expect(try ch.hit()); - digest2 = ch.final(); - - try testing.expectEqual(false, ch.have_exclusive_lock); - } - try testing.expect(mem.eql(u8, &digest1, &digest2)); - - // Modify the file added after initial hash - try cwd.writeFile(temp_file2, "Hello world the second, updated\n"); - - // Wait for file timestamps to tick - const initial_time2 = try testGetCurrentFileTimestamp(); - while ((try testGetCurrentFileTimestamp()) == initial_time2) { - std.time.sleep(1); - } - - { - var ch = cache.obtain(); - defer ch.deinit(); - - ch.hash.addBytes("1234"); - _ = try ch.addFile(temp_file1, null); - - // A file that we depend on has been updated, so the cache should not contain an entry for it - try testing.expectEqual(false, try ch.hit()); - - _ = try ch.addFilePost(temp_file2); - - digest3 = ch.final(); - - try ch.writeManifest(); - } - - try testing.expect(!mem.eql(u8, &digest1, &digest3)); - } - - try cwd.deleteTree(temp_manifest_dir); - try cwd.deleteFile(temp_file1); - try cwd.deleteFile(temp_file2); -} diff --git a/src/Compilation.zig b/src/Compilation.zig index 18d0e46892..ea83d82109 100644 --- a/src/Compilation.zig +++ b/src/Compilation.zig @@ -26,7 +26,7 @@ const wasi_libc = @import("wasi_libc.zig"); const fatal = @import("main.zig").fatal; const clangMain = @import("main.zig").clangMain; const Module = @import("Module.zig"); -const Cache = @import("Cache.zig"); +const Cache = std.Build.Cache; const translate_c = @import("translate_c.zig"); const clang = @import("clang.zig"); const c_codegen = @import("codegen/c.zig"); @@ -807,44 +807,7 @@ pub const AllErrors = struct { } }; -pub const Directory = struct { - /// This field is redundant for operations that can act on the open directory handle - /// directly, but it is needed when passing the directory to a child process. - /// `null` means cwd. - path: ?[]const u8, - handle: std.fs.Dir, - - pub fn join(self: Directory, allocator: Allocator, paths: []const []const u8) ![]u8 { - if (self.path) |p| { - // TODO clean way to do this with only 1 allocation - const part2 = try std.fs.path.join(allocator, paths); - defer allocator.free(part2); - return std.fs.path.join(allocator, &[_][]const u8{ p, part2 }); - } else { - return std.fs.path.join(allocator, paths); - } - } - - pub fn joinZ(self: Directory, allocator: Allocator, paths: []const []const u8) ![:0]u8 { - if (self.path) |p| { - // TODO clean way to do this with only 1 allocation - const part2 = try std.fs.path.join(allocator, paths); - defer allocator.free(part2); - return std.fs.path.joinZ(allocator, &[_][]const u8{ p, part2 }); - } else { - return std.fs.path.joinZ(allocator, paths); - } - } - - /// Whether or not the handle should be closed, or the path should be freed - /// is determined by usage, however this function is provided for convenience - /// if it happens to be what the caller needs. - pub fn closeAndFree(self: *Directory, gpa: Allocator) void { - self.handle.close(); - if (self.path) |p| gpa.free(p); - self.* = undefined; - } -}; +pub const Directory = Cache.Directory; pub const EmitLoc = struct { /// If this is `null` it means the file will be output to the cache directory. @@ -854,6 +817,35 @@ pub const EmitLoc = struct { basename: []const u8, }; +pub const cache_helpers = struct { + pub fn addEmitLoc(hh: *Cache.HashHelper, emit_loc: EmitLoc) void { + hh.addBytes(emit_loc.basename); + } + + pub fn addOptionalEmitLoc(hh: *Cache.HashHelper, optional_emit_loc: ?EmitLoc) void { + hh.add(optional_emit_loc != null); + addEmitLoc(hh, optional_emit_loc orelse return); + } + + pub fn hashCSource(self: *Cache.Manifest, c_source: Compilation.CSourceFile) !void { + _ = try self.addFile(c_source.src_path, null); + // Hash the extra flags, with special care to call addFile for file parameters. + // TODO this logic can likely be improved by utilizing clang_options_data.zig. + const file_args = [_][]const u8{"-include"}; + var arg_i: usize = 0; + while (arg_i < c_source.extra_flags.len) : (arg_i += 1) { + const arg = c_source.extra_flags[arg_i]; + self.hash.addBytes(arg); + for (file_args) |file_arg| { + if (mem.eql(u8, file_arg, arg) and arg_i + 1 < c_source.extra_flags.len) { + arg_i += 1; + _ = try self.addFile(c_source.extra_flags[arg_i], null); + } + } + } + } +}; + pub const ClangPreprocessorMode = enum { no, /// This means we are doing `zig cc -E -o `. @@ -1522,8 +1514,8 @@ pub fn create(gpa: Allocator, options: InitOptions) !*Compilation { cache.hash.add(link_libunwind); cache.hash.add(options.output_mode); cache.hash.add(options.machine_code_model); - cache.hash.addOptionalEmitLoc(options.emit_bin); - cache.hash.addOptionalEmitLoc(options.emit_implib); + cache_helpers.addOptionalEmitLoc(&cache.hash, options.emit_bin); + cache_helpers.addOptionalEmitLoc(&cache.hash, options.emit_implib); cache.hash.addBytes(options.root_name); if (options.target.os.tag == .wasi) cache.hash.add(wasi_exec_model); // TODO audit this and make sure everything is in it @@ -2636,11 +2628,11 @@ fn addNonIncrementalStuffToCacheManifest(comp: *Compilation, man: *Cache.Manifes man.hash.addListOfBytes(key.src.extra_flags); } - man.hash.addOptionalEmitLoc(comp.emit_asm); - man.hash.addOptionalEmitLoc(comp.emit_llvm_ir); - man.hash.addOptionalEmitLoc(comp.emit_llvm_bc); - man.hash.addOptionalEmitLoc(comp.emit_analysis); - man.hash.addOptionalEmitLoc(comp.emit_docs); + cache_helpers.addOptionalEmitLoc(&man.hash, comp.emit_asm); + cache_helpers.addOptionalEmitLoc(&man.hash, comp.emit_llvm_ir); + cache_helpers.addOptionalEmitLoc(&man.hash, comp.emit_llvm_bc); + cache_helpers.addOptionalEmitLoc(&man.hash, comp.emit_analysis); + cache_helpers.addOptionalEmitLoc(&man.hash, comp.emit_docs); man.hash.addListOfBytes(comp.clang_argv); @@ -3959,11 +3951,11 @@ fn updateCObject(comp: *Compilation, c_object: *CObject, c_obj_prog_node: *std.P defer man.deinit(); man.hash.add(comp.clang_preprocessor_mode); - man.hash.addOptionalEmitLoc(comp.emit_asm); - man.hash.addOptionalEmitLoc(comp.emit_llvm_ir); - man.hash.addOptionalEmitLoc(comp.emit_llvm_bc); + cache_helpers.addOptionalEmitLoc(&man.hash, comp.emit_asm); + cache_helpers.addOptionalEmitLoc(&man.hash, comp.emit_llvm_ir); + cache_helpers.addOptionalEmitLoc(&man.hash, comp.emit_llvm_bc); - try man.hashCSource(c_object.src); + try cache_helpers.hashCSource(&man, c_object.src); var arena_allocator = std.heap.ArenaAllocator.init(comp.gpa); defer arena_allocator.deinit(); diff --git a/src/DepTokenizer.zig b/src/DepTokenizer.zig deleted file mode 100644 index 8f9f2f81cd..0000000000 --- a/src/DepTokenizer.zig +++ /dev/null @@ -1,1069 +0,0 @@ -const Tokenizer = @This(); - -index: usize = 0, -bytes: []const u8, -state: State = .lhs, - -const std = @import("std"); -const testing = std.testing; -const assert = std.debug.assert; - -pub fn next(self: *Tokenizer) ?Token { - var start = self.index; - var must_resolve = false; - while (self.index < self.bytes.len) { - const char = self.bytes[self.index]; - switch (self.state) { - .lhs => switch (char) { - '\t', '\n', '\r', ' ' => { - // silently ignore whitespace - self.index += 1; - }, - else => { - start = self.index; - self.state = .target; - }, - }, - .target => switch (char) { - '\t', '\n', '\r', ' ' => { - return errorIllegalChar(.invalid_target, self.index, char); - }, - '$' => { - self.state = .target_dollar_sign; - self.index += 1; - }, - '\\' => { - self.state = .target_reverse_solidus; - self.index += 1; - }, - ':' => { - self.state = .target_colon; - self.index += 1; - }, - else => { - self.index += 1; - }, - }, - .target_reverse_solidus => switch (char) { - '\t', '\n', '\r' => { - return errorIllegalChar(.bad_target_escape, self.index, char); - }, - ' ', '#', '\\' => { - must_resolve = true; - self.state = .target; - self.index += 1; - }, - '$' => { - self.state = .target_dollar_sign; - self.index += 1; - }, - else => { - self.state = .target; - self.index += 1; - }, - }, - .target_dollar_sign => switch (char) { - '$' => { - must_resolve = true; - self.state = .target; - self.index += 1; - }, - else => { - return errorIllegalChar(.expected_dollar_sign, self.index, char); - }, - }, - .target_colon => switch (char) { - '\n', '\r' => { - const bytes = self.bytes[start .. self.index - 1]; - if (bytes.len != 0) { - self.state = .lhs; - return finishTarget(must_resolve, bytes); - } - // silently ignore null target - self.state = .lhs; - }, - '/', '\\' => { - self.state = .target_colon_reverse_solidus; - self.index += 1; - }, - else => { - const bytes = self.bytes[start .. self.index - 1]; - if (bytes.len != 0) { - self.state = .rhs; - return finishTarget(must_resolve, bytes); - } - // silently ignore null target - self.state = .lhs; - }, - }, - .target_colon_reverse_solidus => switch (char) { - '\n', '\r' => { - const bytes = self.bytes[start .. self.index - 2]; - if (bytes.len != 0) { - self.state = .lhs; - return finishTarget(must_resolve, bytes); - } - // silently ignore null target - self.state = .lhs; - }, - else => { - self.state = .target; - }, - }, - .rhs => switch (char) { - '\t', ' ' => { - // silently ignore horizontal whitespace - self.index += 1; - }, - '\n', '\r' => { - self.state = .lhs; - }, - '\\' => { - self.state = .rhs_continuation; - self.index += 1; - }, - '"' => { - self.state = .prereq_quote; - self.index += 1; - start = self.index; - }, - else => { - start = self.index; - self.state = .prereq; - }, - }, - .rhs_continuation => switch (char) { - '\n' => { - self.state = .rhs; - self.index += 1; - }, - '\r' => { - self.state = .rhs_continuation_linefeed; - self.index += 1; - }, - else => { - return errorIllegalChar(.continuation_eol, self.index, char); - }, - }, - .rhs_continuation_linefeed => switch (char) { - '\n' => { - self.state = .rhs; - self.index += 1; - }, - else => { - return errorIllegalChar(.continuation_eol, self.index, char); - }, - }, - .prereq_quote => switch (char) { - '"' => { - self.index += 1; - self.state = .rhs; - return Token{ .prereq = self.bytes[start .. self.index - 1] }; - }, - else => { - self.index += 1; - }, - }, - .prereq => switch (char) { - '\t', ' ' => { - self.state = .rhs; - return Token{ .prereq = self.bytes[start..self.index] }; - }, - '\n', '\r' => { - self.state = .lhs; - return Token{ .prereq = self.bytes[start..self.index] }; - }, - '\\' => { - self.state = .prereq_continuation; - self.index += 1; - }, - else => { - self.index += 1; - }, - }, - .prereq_continuation => switch (char) { - '\n' => { - self.index += 1; - self.state = .rhs; - return Token{ .prereq = self.bytes[start .. self.index - 2] }; - }, - '\r' => { - self.state = .prereq_continuation_linefeed; - self.index += 1; - }, - else => { - // not continuation - self.state = .prereq; - self.index += 1; - }, - }, - .prereq_continuation_linefeed => switch (char) { - '\n' => { - self.index += 1; - self.state = .rhs; - return Token{ .prereq = self.bytes[start .. self.index - 1] }; - }, - else => { - return errorIllegalChar(.continuation_eol, self.index, char); - }, - }, - } - } else { - switch (self.state) { - .lhs, - .rhs, - .rhs_continuation, - .rhs_continuation_linefeed, - => return null, - .target => { - return errorPosition(.incomplete_target, start, self.bytes[start..]); - }, - .target_reverse_solidus, - .target_dollar_sign, - => { - const idx = self.index - 1; - return errorIllegalChar(.incomplete_escape, idx, self.bytes[idx]); - }, - .target_colon => { - const bytes = self.bytes[start .. self.index - 1]; - if (bytes.len != 0) { - self.index += 1; - self.state = .rhs; - return finishTarget(must_resolve, bytes); - } - // silently ignore null target - self.state = .lhs; - return null; - }, - .target_colon_reverse_solidus => { - const bytes = self.bytes[start .. self.index - 2]; - if (bytes.len != 0) { - self.index += 1; - self.state = .rhs; - return finishTarget(must_resolve, bytes); - } - // silently ignore null target - self.state = .lhs; - return null; - }, - .prereq_quote => { - return errorPosition(.incomplete_quoted_prerequisite, start, self.bytes[start..]); - }, - .prereq => { - self.state = .lhs; - return Token{ .prereq = self.bytes[start..] }; - }, - .prereq_continuation => { - self.state = .lhs; - return Token{ .prereq = self.bytes[start .. self.index - 1] }; - }, - .prereq_continuation_linefeed => { - self.state = .lhs; - return Token{ .prereq = self.bytes[start .. self.index - 2] }; - }, - } - } - unreachable; -} - -fn errorPosition(comptime id: std.meta.Tag(Token), index: usize, bytes: []const u8) Token { - return @unionInit(Token, @tagName(id), .{ .index = index, .bytes = bytes }); -} - -fn errorIllegalChar(comptime id: std.meta.Tag(Token), index: usize, char: u8) Token { - return @unionInit(Token, @tagName(id), .{ .index = index, .char = char }); -} - -fn finishTarget(must_resolve: bool, bytes: []const u8) Token { - return if (must_resolve) .{ .target_must_resolve = bytes } else .{ .target = bytes }; -} - -const State = enum { - lhs, - target, - target_reverse_solidus, - target_dollar_sign, - target_colon, - target_colon_reverse_solidus, - rhs, - rhs_continuation, - rhs_continuation_linefeed, - prereq_quote, - prereq, - prereq_continuation, - prereq_continuation_linefeed, -}; - -pub const Token = union(enum) { - target: []const u8, - target_must_resolve: []const u8, - prereq: []const u8, - - incomplete_quoted_prerequisite: IndexAndBytes, - incomplete_target: IndexAndBytes, - - invalid_target: IndexAndChar, - bad_target_escape: IndexAndChar, - expected_dollar_sign: IndexAndChar, - continuation_eol: IndexAndChar, - incomplete_escape: IndexAndChar, - - pub const IndexAndChar = struct { - index: usize, - char: u8, - }; - - pub const IndexAndBytes = struct { - index: usize, - bytes: []const u8, - }; - - /// Resolve escapes in target. Only valid with .target_must_resolve. - pub fn resolve(self: Token, writer: anytype) @TypeOf(writer).Error!void { - const bytes = self.target_must_resolve; // resolve called on incorrect token - - var state: enum { start, escape, dollar } = .start; - for (bytes) |c| { - switch (state) { - .start => { - switch (c) { - '\\' => state = .escape, - '$' => state = .dollar, - else => try writer.writeByte(c), - } - }, - .escape => { - switch (c) { - ' ', '#', '\\' => {}, - '$' => { - try writer.writeByte('\\'); - state = .dollar; - continue; - }, - else => try writer.writeByte('\\'), - } - try writer.writeByte(c); - state = .start; - }, - .dollar => { - try writer.writeByte('$'); - switch (c) { - '$' => {}, - else => try writer.writeByte(c), - } - state = .start; - }, - } - } - } - - pub fn printError(self: Token, writer: anytype) @TypeOf(writer).Error!void { - switch (self) { - .target, .target_must_resolve, .prereq => unreachable, // not an error - .incomplete_quoted_prerequisite, - .incomplete_target, - => |index_and_bytes| { - try writer.print("{s} '", .{self.errStr()}); - if (self == .incomplete_target) { - const tmp = Token{ .target_must_resolve = index_and_bytes.bytes }; - try tmp.resolve(writer); - } else { - try printCharValues(writer, index_and_bytes.bytes); - } - try writer.print("' at position {d}", .{index_and_bytes.index}); - }, - .invalid_target, - .bad_target_escape, - .expected_dollar_sign, - .continuation_eol, - .incomplete_escape, - => |index_and_char| { - try writer.writeAll("illegal char "); - try printUnderstandableChar(writer, index_and_char.char); - try writer.print(" at position {d}: {s}", .{ index_and_char.index, self.errStr() }); - }, - } - } - - fn errStr(self: Token) []const u8 { - return switch (self) { - .target, .target_must_resolve, .prereq => unreachable, // not an error - .incomplete_quoted_prerequisite => "incomplete quoted prerequisite", - .incomplete_target => "incomplete target", - .invalid_target => "invalid target", - .bad_target_escape => "bad target escape", - .expected_dollar_sign => "expecting '$'", - .continuation_eol => "continuation expecting end-of-line", - .incomplete_escape => "incomplete escape", - }; - } -}; - -test "empty file" { - try depTokenizer("", ""); -} - -test "empty whitespace" { - try depTokenizer("\n", ""); - try depTokenizer("\r", ""); - try depTokenizer("\r\n", ""); - try depTokenizer(" ", ""); -} - -test "empty colon" { - try depTokenizer(":", ""); - try depTokenizer("\n:", ""); - try depTokenizer("\r:", ""); - try depTokenizer("\r\n:", ""); - try depTokenizer(" :", ""); -} - -test "empty target" { - try depTokenizer("foo.o:", "target = {foo.o}"); - try depTokenizer( - \\foo.o: - \\bar.o: - \\abcd.o: - , - \\target = {foo.o} - \\target = {bar.o} - \\target = {abcd.o} - ); -} - -test "whitespace empty target" { - try depTokenizer("\nfoo.o:", "target = {foo.o}"); - try depTokenizer("\rfoo.o:", "target = {foo.o}"); - try depTokenizer("\r\nfoo.o:", "target = {foo.o}"); - try depTokenizer(" foo.o:", "target = {foo.o}"); -} - -test "escape empty target" { - try depTokenizer("\\ foo.o:", "target = { foo.o}"); - try depTokenizer("\\#foo.o:", "target = {#foo.o}"); - try depTokenizer("\\\\foo.o:", "target = {\\foo.o}"); - try depTokenizer("$$foo.o:", "target = {$foo.o}"); -} - -test "empty target linefeeds" { - try depTokenizer("\n", ""); - try depTokenizer("\r\n", ""); - - const expect = "target = {foo.o}"; - try depTokenizer( - \\foo.o: - , expect); - try depTokenizer( - \\foo.o: - \\ - , expect); - try depTokenizer( - \\foo.o: - , expect); - try depTokenizer( - \\foo.o: - \\ - , expect); -} - -test "empty target linefeeds + continuations" { - const expect = "target = {foo.o}"; - try depTokenizer( - \\foo.o:\ - , expect); - try depTokenizer( - \\foo.o:\ - \\ - , expect); - try depTokenizer( - \\foo.o:\ - , expect); - try depTokenizer( - \\foo.o:\ - \\ - , expect); -} - -test "empty target linefeeds + hspace + continuations" { - const expect = "target = {foo.o}"; - try depTokenizer( - \\foo.o: \ - , expect); - try depTokenizer( - \\foo.o: \ - \\ - , expect); - try depTokenizer( - \\foo.o: \ - , expect); - try depTokenizer( - \\foo.o: \ - \\ - , expect); -} - -test "prereq" { - const expect = - \\target = {foo.o} - \\prereq = {foo.c} - ; - try depTokenizer("foo.o: foo.c", expect); - try depTokenizer( - \\foo.o: \ - \\foo.c - , expect); - try depTokenizer( - \\foo.o: \ - \\ foo.c - , expect); - try depTokenizer( - \\foo.o: \ - \\ foo.c - , expect); -} - -test "prereq continuation" { - const expect = - \\target = {foo.o} - \\prereq = {foo.h} - \\prereq = {bar.h} - ; - try depTokenizer( - \\foo.o: foo.h\ - \\bar.h - , expect); - try depTokenizer( - \\foo.o: foo.h\ - \\bar.h - , expect); -} - -test "multiple prereqs" { - const expect = - \\target = {foo.o} - \\prereq = {foo.c} - \\prereq = {foo.h} - \\prereq = {bar.h} - ; - try depTokenizer("foo.o: foo.c foo.h bar.h", expect); - try depTokenizer( - \\foo.o: \ - \\foo.c foo.h bar.h - , expect); - try depTokenizer( - \\foo.o: foo.c foo.h bar.h\ - , expect); - try depTokenizer( - \\foo.o: foo.c foo.h bar.h\ - \\ - , expect); - try depTokenizer( - \\foo.o: \ - \\foo.c \ - \\ foo.h\ - \\bar.h - \\ - , expect); - try depTokenizer( - \\foo.o: \ - \\foo.c \ - \\ foo.h\ - \\bar.h\ - \\ - , expect); - try depTokenizer( - \\foo.o: \ - \\foo.c \ - \\ foo.h\ - \\bar.h\ - , expect); -} - -test "multiple targets and prereqs" { - try depTokenizer( - \\foo.o: foo.c - \\bar.o: bar.c a.h b.h c.h - \\abc.o: abc.c \ - \\ one.h two.h \ - \\ three.h four.h - , - \\target = {foo.o} - \\prereq = {foo.c} - \\target = {bar.o} - \\prereq = {bar.c} - \\prereq = {a.h} - \\prereq = {b.h} - \\prereq = {c.h} - \\target = {abc.o} - \\prereq = {abc.c} - \\prereq = {one.h} - \\prereq = {two.h} - \\prereq = {three.h} - \\prereq = {four.h} - ); - try depTokenizer( - \\ascii.o: ascii.c - \\base64.o: base64.c stdio.h - \\elf.o: elf.c a.h b.h c.h - \\macho.o: \ - \\ macho.c\ - \\ a.h b.h c.h - , - \\target = {ascii.o} - \\prereq = {ascii.c} - \\target = {base64.o} - \\prereq = {base64.c} - \\prereq = {stdio.h} - \\target = {elf.o} - \\prereq = {elf.c} - \\prereq = {a.h} - \\prereq = {b.h} - \\prereq = {c.h} - \\target = {macho.o} - \\prereq = {macho.c} - \\prereq = {a.h} - \\prereq = {b.h} - \\prereq = {c.h} - ); - try depTokenizer( - \\a$$scii.o: ascii.c - \\\\base64.o: "\base64.c" "s t#dio.h" - \\e\\lf.o: "e\lf.c" "a.h$$" "$$b.h c.h$$" - \\macho.o: \ - \\ "macho!.c" \ - \\ a.h b.h c.h - , - \\target = {a$scii.o} - \\prereq = {ascii.c} - \\target = {\base64.o} - \\prereq = {\base64.c} - \\prereq = {s t#dio.h} - \\target = {e\lf.o} - \\prereq = {e\lf.c} - \\prereq = {a.h$$} - \\prereq = {$$b.h c.h$$} - \\target = {macho.o} - \\prereq = {macho!.c} - \\prereq = {a.h} - \\prereq = {b.h} - \\prereq = {c.h} - ); -} - -test "windows quoted prereqs" { - try depTokenizer( - \\c:\foo.o: "C:\Program Files (x86)\Microsoft Visual Studio\foo.c" - \\c:\foo2.o: "C:\Program Files (x86)\Microsoft Visual Studio\foo2.c" \ - \\ "C:\Program Files (x86)\Microsoft Visual Studio\foo1.h" \ - \\ "C:\Program Files (x86)\Microsoft Visual Studio\foo2.h" - , - \\target = {c:\foo.o} - \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\foo.c} - \\target = {c:\foo2.o} - \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\foo2.c} - \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\foo1.h} - \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\foo2.h} - ); -} - -test "windows mixed prereqs" { - try depTokenizer( - \\cimport.o: \ - \\ C:\msys64\home\anon\project\zig\master\zig-cache\o\qhvhbUo7GU5iKyQ5mpA8TcQpncCYaQu0wwvr3ybiSTj_Dtqi1Nmcb70kfODJ2Qlg\cimport.h \ - \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\stdio.h" \ - \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt.h" \ - \\ "C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vcruntime.h" \ - \\ "C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\sal.h" \ - \\ "C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\concurrencysal.h" \ - \\ C:\msys64\opt\zig\lib\zig\include\vadefs.h \ - \\ "C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vadefs.h" \ - \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_wstdio.h" \ - \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_stdio_config.h" \ - \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\string.h" \ - \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_memory.h" \ - \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_memcpy_s.h" \ - \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\errno.h" \ - \\ "C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vcruntime_string.h" \ - \\ "C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_wstring.h" - , - \\target = {cimport.o} - \\prereq = {C:\msys64\home\anon\project\zig\master\zig-cache\o\qhvhbUo7GU5iKyQ5mpA8TcQpncCYaQu0wwvr3ybiSTj_Dtqi1Nmcb70kfODJ2Qlg\cimport.h} - \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\stdio.h} - \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt.h} - \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vcruntime.h} - \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\sal.h} - \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\concurrencysal.h} - \\prereq = {C:\msys64\opt\zig\lib\zig\include\vadefs.h} - \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vadefs.h} - \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_wstdio.h} - \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_stdio_config.h} - \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\string.h} - \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_memory.h} - \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_memcpy_s.h} - \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\errno.h} - \\prereq = {C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.21.27702\lib\x64\\..\..\include\vcruntime_string.h} - \\prereq = {C:\Program Files (x86)\Windows Kits\10\\Include\10.0.17763.0\ucrt\corecrt_wstring.h} - ); -} - -test "windows funky targets" { - try depTokenizer( - \\C:\Users\anon\foo.o: - \\C:\Users\anon\foo\ .o: - \\C:\Users\anon\foo\#.o: - \\C:\Users\anon\foo$$.o: - \\C:\Users\anon\\\ foo.o: - \\C:\Users\anon\\#foo.o: - \\C:\Users\anon\$$foo.o: - \\C:\Users\anon\\\ \ \ \ \ foo.o: - , - \\target = {C:\Users\anon\foo.o} - \\target = {C:\Users\anon\foo .o} - \\target = {C:\Users\anon\foo#.o} - \\target = {C:\Users\anon\foo$.o} - \\target = {C:\Users\anon\ foo.o} - \\target = {C:\Users\anon\#foo.o} - \\target = {C:\Users\anon\$foo.o} - \\target = {C:\Users\anon\ foo.o} - ); -} - -test "windows drive and forward slashes" { - try depTokenizer( - \\C:/msys64/what/zig-cache\tmp\48ac4d78dd531abd-cxa_thread_atexit.obj: \ - \\ C:/msys64/opt/zig3/lib/zig/libc/mingw/crt/cxa_thread_atexit.c - , - \\target = {C:/msys64/what/zig-cache\tmp\48ac4d78dd531abd-cxa_thread_atexit.obj} - \\prereq = {C:/msys64/opt/zig3/lib/zig/libc/mingw/crt/cxa_thread_atexit.c} - ); -} - -test "error incomplete escape - reverse_solidus" { - try depTokenizer("\\", - \\ERROR: illegal char '\' at position 0: incomplete escape - ); - try depTokenizer("\t\\", - \\ERROR: illegal char '\' at position 1: incomplete escape - ); - try depTokenizer("\n\\", - \\ERROR: illegal char '\' at position 1: incomplete escape - ); - try depTokenizer("\r\\", - \\ERROR: illegal char '\' at position 1: incomplete escape - ); - try depTokenizer("\r\n\\", - \\ERROR: illegal char '\' at position 2: incomplete escape - ); - try depTokenizer(" \\", - \\ERROR: illegal char '\' at position 1: incomplete escape - ); -} - -test "error incomplete escape - dollar_sign" { - try depTokenizer("$", - \\ERROR: illegal char '$' at position 0: incomplete escape - ); - try depTokenizer("\t$", - \\ERROR: illegal char '$' at position 1: incomplete escape - ); - try depTokenizer("\n$", - \\ERROR: illegal char '$' at position 1: incomplete escape - ); - try depTokenizer("\r$", - \\ERROR: illegal char '$' at position 1: incomplete escape - ); - try depTokenizer("\r\n$", - \\ERROR: illegal char '$' at position 2: incomplete escape - ); - try depTokenizer(" $", - \\ERROR: illegal char '$' at position 1: incomplete escape - ); -} - -test "error incomplete target" { - try depTokenizer("foo.o", - \\ERROR: incomplete target 'foo.o' at position 0 - ); - try depTokenizer("\tfoo.o", - \\ERROR: incomplete target 'foo.o' at position 1 - ); - try depTokenizer("\nfoo.o", - \\ERROR: incomplete target 'foo.o' at position 1 - ); - try depTokenizer("\rfoo.o", - \\ERROR: incomplete target 'foo.o' at position 1 - ); - try depTokenizer("\r\nfoo.o", - \\ERROR: incomplete target 'foo.o' at position 2 - ); - try depTokenizer(" foo.o", - \\ERROR: incomplete target 'foo.o' at position 1 - ); - - try depTokenizer("\\ foo.o", - \\ERROR: incomplete target ' foo.o' at position 0 - ); - try depTokenizer("\\#foo.o", - \\ERROR: incomplete target '#foo.o' at position 0 - ); - try depTokenizer("\\\\foo.o", - \\ERROR: incomplete target '\foo.o' at position 0 - ); - try depTokenizer("$$foo.o", - \\ERROR: incomplete target '$foo.o' at position 0 - ); -} - -test "error illegal char at position - bad target escape" { - try depTokenizer("\\\t", - \\ERROR: illegal char \x09 at position 1: bad target escape - ); - try depTokenizer("\\\n", - \\ERROR: illegal char \x0A at position 1: bad target escape - ); - try depTokenizer("\\\r", - \\ERROR: illegal char \x0D at position 1: bad target escape - ); - try depTokenizer("\\\r\n", - \\ERROR: illegal char \x0D at position 1: bad target escape - ); -} - -test "error illegal char at position - execting dollar_sign" { - try depTokenizer("$\t", - \\ERROR: illegal char \x09 at position 1: expecting '$' - ); - try depTokenizer("$\n", - \\ERROR: illegal char \x0A at position 1: expecting '$' - ); - try depTokenizer("$\r", - \\ERROR: illegal char \x0D at position 1: expecting '$' - ); - try depTokenizer("$\r\n", - \\ERROR: illegal char \x0D at position 1: expecting '$' - ); -} - -test "error illegal char at position - invalid target" { - try depTokenizer("foo\t.o", - \\ERROR: illegal char \x09 at position 3: invalid target - ); - try depTokenizer("foo\n.o", - \\ERROR: illegal char \x0A at position 3: invalid target - ); - try depTokenizer("foo\r.o", - \\ERROR: illegal char \x0D at position 3: invalid target - ); - try depTokenizer("foo\r\n.o", - \\ERROR: illegal char \x0D at position 3: invalid target - ); -} - -test "error target - continuation expecting end-of-line" { - try depTokenizer("foo.o: \\\t", - \\target = {foo.o} - \\ERROR: illegal char \x09 at position 8: continuation expecting end-of-line - ); - try depTokenizer("foo.o: \\ ", - \\target = {foo.o} - \\ERROR: illegal char ' ' at position 8: continuation expecting end-of-line - ); - try depTokenizer("foo.o: \\x", - \\target = {foo.o} - \\ERROR: illegal char 'x' at position 8: continuation expecting end-of-line - ); - try depTokenizer("foo.o: \\\x0dx", - \\target = {foo.o} - \\ERROR: illegal char 'x' at position 9: continuation expecting end-of-line - ); -} - -test "error prereq - continuation expecting end-of-line" { - try depTokenizer("foo.o: foo.h\\\x0dx", - \\target = {foo.o} - \\ERROR: illegal char 'x' at position 14: continuation expecting end-of-line - ); -} - -// - tokenize input, emit textual representation, and compare to expect -fn depTokenizer(input: []const u8, expect: []const u8) !void { - var arena_allocator = std.heap.ArenaAllocator.init(std.testing.allocator); - const arena = arena_allocator.allocator(); - defer arena_allocator.deinit(); - - var it: Tokenizer = .{ .bytes = input }; - var buffer = std.ArrayList(u8).init(arena); - var resolve_buf = std.ArrayList(u8).init(arena); - var i: usize = 0; - while (it.next()) |token| { - if (i != 0) try buffer.appendSlice("\n"); - switch (token) { - .target, .prereq => |bytes| { - try buffer.appendSlice(@tagName(token)); - try buffer.appendSlice(" = {"); - for (bytes) |b| { - try buffer.append(printable_char_tab[b]); - } - try buffer.appendSlice("}"); - }, - .target_must_resolve => { - try buffer.appendSlice("target = {"); - try token.resolve(resolve_buf.writer()); - for (resolve_buf.items) |b| { - try buffer.append(printable_char_tab[b]); - } - resolve_buf.items.len = 0; - try buffer.appendSlice("}"); - }, - else => { - try buffer.appendSlice("ERROR: "); - try token.printError(buffer.writer()); - break; - }, - } - i += 1; - } - - if (std.mem.eql(u8, expect, buffer.items)) { - try testing.expect(true); - return; - } - - const out = std.io.getStdErr().writer(); - - try out.writeAll("\n"); - try printSection(out, "<<<< input", input); - try printSection(out, "==== expect", expect); - try printSection(out, ">>>> got", buffer.items); - try printRuler(out); - - try testing.expect(false); -} - -fn printSection(out: anytype, label: []const u8, bytes: []const u8) !void { - try printLabel(out, label, bytes); - try hexDump(out, bytes); - try printRuler(out); - try out.writeAll(bytes); - try out.writeAll("\n"); -} - -fn printLabel(out: anytype, label: []const u8, bytes: []const u8) !void { - var buf: [80]u8 = undefined; - var text = try std.fmt.bufPrint(buf[0..], "{s} {d} bytes ", .{ label, bytes.len }); - try out.writeAll(text); - var i: usize = text.len; - const end = 79; - while (i < end) : (i += 1) { - try out.writeAll(&[_]u8{label[0]}); - } - try out.writeAll("\n"); -} - -fn printRuler(out: anytype) !void { - var i: usize = 0; - const end = 79; - while (i < end) : (i += 1) { - try out.writeAll("-"); - } - try out.writeAll("\n"); -} - -fn hexDump(out: anytype, bytes: []const u8) !void { - const n16 = bytes.len >> 4; - var line: usize = 0; - var offset: usize = 0; - while (line < n16) : (line += 1) { - try hexDump16(out, offset, bytes[offset .. offset + 16]); - offset += 16; - } - - const n = bytes.len & 0x0f; - if (n > 0) { - try printDecValue(out, offset, 8); - try out.writeAll(":"); - try out.writeAll(" "); - var end1 = std.math.min(offset + n, offset + 8); - for (bytes[offset..end1]) |b| { - try out.writeAll(" "); - try printHexValue(out, b, 2); - } - var end2 = offset + n; - if (end2 > end1) { - try out.writeAll(" "); - for (bytes[end1..end2]) |b| { - try out.writeAll(" "); - try printHexValue(out, b, 2); - } - } - const short = 16 - n; - var i: usize = 0; - while (i < short) : (i += 1) { - try out.writeAll(" "); - } - if (end2 > end1) { - try out.writeAll(" |"); - } else { - try out.writeAll(" |"); - } - try printCharValues(out, bytes[offset..end2]); - try out.writeAll("|\n"); - offset += n; - } - - try printDecValue(out, offset, 8); - try out.writeAll(":"); - try out.writeAll("\n"); -} - -fn hexDump16(out: anytype, offset: usize, bytes: []const u8) !void { - try printDecValue(out, offset, 8); - try out.writeAll(":"); - try out.writeAll(" "); - for (bytes[0..8]) |b| { - try out.writeAll(" "); - try printHexValue(out, b, 2); - } - try out.writeAll(" "); - for (bytes[8..16]) |b| { - try out.writeAll(" "); - try printHexValue(out, b, 2); - } - try out.writeAll(" |"); - try printCharValues(out, bytes); - try out.writeAll("|\n"); -} - -fn printDecValue(out: anytype, value: u64, width: u8) !void { - var buffer: [20]u8 = undefined; - const len = std.fmt.formatIntBuf(buffer[0..], value, 10, .lower, .{ .width = width, .fill = '0' }); - try out.writeAll(buffer[0..len]); -} - -fn printHexValue(out: anytype, value: u64, width: u8) !void { - var buffer: [16]u8 = undefined; - const len = std.fmt.formatIntBuf(buffer[0..], value, 16, .lower, .{ .width = width, .fill = '0' }); - try out.writeAll(buffer[0..len]); -} - -fn printCharValues(out: anytype, bytes: []const u8) !void { - for (bytes) |b| { - try out.writeAll(&[_]u8{printable_char_tab[b]}); - } -} - -fn printUnderstandableChar(out: anytype, char: u8) !void { - if (std.ascii.isPrint(char)) { - try out.print("'{c}'", .{char}); - } else { - try out.print("\\x{X:0>2}", .{char}); - } -} - -// zig fmt: off -const printable_char_tab: [256]u8 = ( - "................................ !\"#$%&'()*+,-./0123456789:;<=>?" ++ - "@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~." ++ - "................................................................" ++ - "................................................................" -).*; diff --git a/src/Module.zig b/src/Module.zig index e4cf0189cc..a129cb0cb6 100644 --- a/src/Module.zig +++ b/src/Module.zig @@ -16,7 +16,7 @@ const Ast = std.zig.Ast; const Module = @This(); const Compilation = @import("Compilation.zig"); -const Cache = @import("Cache.zig"); +const Cache = std.Build.Cache; const Value = @import("value.zig").Value; const Type = @import("type.zig").Type; const TypedValue = @import("TypedValue.zig"); diff --git a/src/Package.zig b/src/Package.zig index 401eef2121..a3afe21009 100644 --- a/src/Package.zig +++ b/src/Package.zig @@ -13,7 +13,7 @@ const Compilation = @import("Compilation.zig"); const Module = @import("Module.zig"); const ThreadPool = @import("ThreadPool.zig"); const WaitGroup = @import("WaitGroup.zig"); -const Cache = @import("Cache.zig"); +const Cache = std.Build.Cache; const build_options = @import("build_options"); const Manifest = @import("Manifest.zig"); diff --git a/src/glibc.zig b/src/glibc.zig index 8dce1c5132..2a2887c334 100644 --- a/src/glibc.zig +++ b/src/glibc.zig @@ -11,7 +11,7 @@ const target_util = @import("target.zig"); const Compilation = @import("Compilation.zig"); const build_options = @import("build_options"); const trace = @import("tracy.zig").trace; -const Cache = @import("Cache.zig"); +const Cache = std.Build.Cache; const Package = @import("Package.zig"); pub const Lib = struct { diff --git a/src/link.zig b/src/link.zig index 2b3ce51667..5650e0679a 100644 --- a/src/link.zig +++ b/src/link.zig @@ -10,7 +10,7 @@ const wasi_libc = @import("wasi_libc.zig"); const Air = @import("Air.zig"); const Allocator = std.mem.Allocator; -const Cache = @import("Cache.zig"); +const Cache = std.Build.Cache; const Compilation = @import("Compilation.zig"); const LibCInstallation = @import("libc_installation.zig").LibCInstallation; const Liveness = @import("Liveness.zig"); diff --git a/src/link/Coff/lld.zig b/src/link/Coff/lld.zig index d705f62f5c..c308ff5989 100644 --- a/src/link/Coff/lld.zig +++ b/src/link/Coff/lld.zig @@ -5,6 +5,7 @@ const assert = std.debug.assert; const fs = std.fs; const log = std.log.scoped(.link); const mem = std.mem; +const Cache = std.Build.Cache; const mingw = @import("../../mingw.zig"); const link = @import("../../link.zig"); @@ -13,7 +14,6 @@ const trace = @import("../../tracy.zig").trace; const Allocator = mem.Allocator; -const Cache = @import("../../Cache.zig"); const Coff = @import("../Coff.zig"); const Compilation = @import("../../Compilation.zig"); diff --git a/src/link/Elf.zig b/src/link/Elf.zig index 45952da6c0..37ebfdc0dc 100644 --- a/src/link/Elf.zig +++ b/src/link/Elf.zig @@ -21,7 +21,7 @@ const trace = @import("../tracy.zig").trace; const Air = @import("../Air.zig"); const Allocator = std.mem.Allocator; pub const Atom = @import("Elf/Atom.zig"); -const Cache = @import("../Cache.zig"); +const Cache = std.Build.Cache; const Compilation = @import("../Compilation.zig"); const Dwarf = @import("Dwarf.zig"); const File = link.File; diff --git a/src/link/MachO.zig b/src/link/MachO.zig index 24ef275c5b..35f5f1b562 100644 --- a/src/link/MachO.zig +++ b/src/link/MachO.zig @@ -28,7 +28,7 @@ const Air = @import("../Air.zig"); const Allocator = mem.Allocator; const Archive = @import("MachO/Archive.zig"); pub const Atom = @import("MachO/Atom.zig"); -const Cache = @import("../Cache.zig"); +const Cache = std.Build.Cache; const CodeSignature = @import("MachO/CodeSignature.zig"); const Compilation = @import("../Compilation.zig"); const Dwarf = File.Dwarf; diff --git a/src/link/MachO/zld.zig b/src/link/MachO/zld.zig index 095ac9b5ce..785fa71445 100644 --- a/src/link/MachO/zld.zig +++ b/src/link/MachO/zld.zig @@ -20,7 +20,7 @@ const trace = @import("../../tracy.zig").trace; const Allocator = mem.Allocator; const Archive = @import("Archive.zig"); const Atom = @import("ZldAtom.zig"); -const Cache = @import("../../Cache.zig"); +const Cache = std.Build.Cache; const CodeSignature = @import("CodeSignature.zig"); const Compilation = @import("../../Compilation.zig"); const DwarfInfo = @import("DwarfInfo.zig"); diff --git a/src/link/Wasm.zig b/src/link/Wasm.zig index 9d20412788..e62a2050d7 100644 --- a/src/link/Wasm.zig +++ b/src/link/Wasm.zig @@ -20,7 +20,7 @@ const lldMain = @import("../main.zig").lldMain; const trace = @import("../tracy.zig").trace; const build_options = @import("build_options"); const wasi_libc = @import("../wasi_libc.zig"); -const Cache = @import("../Cache.zig"); +const Cache = std.Build.Cache; const Type = @import("../type.zig").Type; const TypedValue = @import("../TypedValue.zig"); const LlvmObject = @import("../codegen/llvm.zig").Object; diff --git a/src/main.zig b/src/main.zig index 00a7b126c8..2add2f9165 100644 --- a/src/main.zig +++ b/src/main.zig @@ -20,7 +20,7 @@ const LibCInstallation = @import("libc_installation.zig").LibCInstallation; const wasi_libc = @import("wasi_libc.zig"); const translate_c = @import("translate_c.zig"); const clang = @import("clang.zig"); -const Cache = @import("Cache.zig"); +const Cache = std.Build.Cache; const target_util = @import("target.zig"); const ThreadPool = @import("ThreadPool.zig"); const crash_report = @import("crash_report.zig"); @@ -3607,7 +3607,7 @@ fn cmdTranslateC(comp: *Compilation, arena: Allocator, enable_cache: bool) !void defer if (enable_cache) man.deinit(); man.hash.add(@as(u16, 0xb945)); // Random number to distinguish translate-c from compiling C objects - man.hashCSource(c_source_file) catch |err| { + Compilation.cache_helpers.hashCSource(&man, c_source_file) catch |err| { fatal("unable to process '{s}': {s}", .{ c_source_file.src_path, @errorName(err) }); }; diff --git a/src/mingw.zig b/src/mingw.zig index 06880743c6..4f94e26a98 100644 --- a/src/mingw.zig +++ b/src/mingw.zig @@ -8,7 +8,7 @@ const log = std.log.scoped(.mingw); const builtin = @import("builtin"); const Compilation = @import("Compilation.zig"); const build_options = @import("build_options"); -const Cache = @import("Cache.zig"); +const Cache = std.Build.Cache; pub const CRTFile = enum { crt2_o, -- cgit v1.2.3