// Copyright (c) 2019-2022 Alexander Medvednikov. All rights reserved. // Use of this source code is governed by an MIT license // that can be found in the LICENSE file. [has_globals] module parser import v.scanner import v.ast import v.token import v.pref import v.util import v.vet import v.errors import os import hash.fnv1a [minify] pub struct Parser { pref &pref.Preferences mut: file_base string // "hello.v" file_name string // "/home/user/hello.v" file_name_dir string // "/home/user" unique_prefix string // a hash of p.file_name, used for making anon fn generation unique file_backend_mode ast.Language // .c for .c.v|.c.vv|.c.vsh files; .js for .js.v files, .amd64/.rv32/other arches for .amd64.v/.rv32.v/etc. files, .v otherwise. scanner &scanner.Scanner comments_mode scanner.CommentsMode = .skip_comments // see comment in parse_file tok token.Token prev_tok token.Token peek_tok token.Token table &ast.Table language ast.Language fn_language ast.Language // .c for `fn C.abcd()` declarations expr_level int // prevent too deep recursions for pathological programs inside_vlib_file bool // true for all vlib/ files inside_test_file bool // when inside _test.v or _test.vv file inside_if bool inside_if_expr bool inside_if_cond bool inside_ct_if_expr bool inside_or_expr bool inside_for bool inside_fn bool // true even with implicit main inside_fn_return bool inside_unsafe_fn bool inside_str_interp bool inside_array_lit bool inside_in_array bool inside_match bool // to separate `match A { }` from `Struct{}` inside_select bool // to allow `ch <- Struct{} {` inside `select` inside_match_case bool // to separate `match_expr { }` from `Struct{}` inside_match_body bool // to fix eval not used TODO inside_unsafe bool inside_sum_type bool // to prevent parsing inline sum type again inside_asm_template bool inside_asm bool inside_defer bool inside_generic_params bool // indicates if parsing between `<` and `>` of a method/function inside_receiver_param bool // indicates if parsing the receiver parameter inside the first `(` and `)` of a method inside_struct_field_decl bool or_is_handled bool // ignore `or` in this expression builtin_mod bool // are we in the `builtin` module? mod string // current module name is_manualfree bool // true when `[manualfree] module abc`, makes *all* fns in the current .v file, opt out of autofree has_globals bool // `[has_globals] module abc` - allow globals declarations, even without -enable-globals, in that single .v file __only__ is_generated bool // `[generated] module abc` - turn off compiler notices for that single .v file __only__. is_translated bool // `[translated] module abc` - mark a file as translated, to relax some compiler checks for translated code. attrs []ast.Attr // attributes before next decl stmt expr_mod string // for constructing full type names in parse_type() scope &ast.Scope imports map[string]string // alias => mod_name ast_imports []ast.Import // mod_names used_imports []string // alias auto_imports []string // imports, the user does not need to specify imported_symbols map[string]string is_amp bool // for generating the right code for `&Foo{}` returns bool is_stmt_ident bool // true while the beginning of a statement is an ident/selector expecting_type bool // `is Type`, expecting type errors []errors.Error warnings []errors.Warning notices []errors.Notice vet_errors []vet.Error cur_fn_name string label_names []string name_error bool // indicates if the token is not a name or the name is on another line n_asm int // controls assembly labels global_labels []string comptime_if_cond bool defer_vars []ast.Ident should_abort bool // when too many errors/warnings/notices are accumulated, should_abort becomes true, and the parser should stop codegen_text string struct_init_generic_types []ast.Type if_cond_comments []ast.Comment } __global codegen_files = []&ast.File{} // for tests pub fn parse_stmt(text string, table &ast.Table, scope &ast.Scope) ast.Stmt { $if trace_parse_stmt ? { eprintln('> ${@MOD}.${@FN} text: $text') } mut p := Parser{ scanner: scanner.new_scanner(text, .skip_comments, &pref.Preferences{}) inside_test_file: true table: table pref: &pref.Preferences{} scope: scope } p.init_parse_fns() util.timing_start('PARSE stmt') defer { util.timing_measure_cumulative('PARSE stmt') } p.read_first_token() return p.stmt(false) } pub fn parse_comptime(tmpl_path string, text string, table &ast.Table, pref &pref.Preferences, scope &ast.Scope) &ast.File { $if trace_parse_comptime ? { eprintln('> ${@MOD}.${@FN} text: $text') } mut p := Parser{ file_name: tmpl_path scanner: scanner.new_scanner(text, .skip_comments, pref) table: table pref: pref scope: scope errors: []errors.Error{} warnings: []errors.Warning{} } res := p.parse() unsafe { p.free_scanner() } return res } pub fn parse_text(text string, path string, table &ast.Table, comments_mode scanner.CommentsMode, pref &pref.Preferences) &ast.File { $if trace_parse_text ? { eprintln('> ${@MOD}.${@FN} comments_mode: ${comments_mode:-20} | path: ${path:-20} | text: $text') } mut p := Parser{ scanner: scanner.new_scanner(text, comments_mode, pref) comments_mode: comments_mode table: table pref: pref scope: &ast.Scope{ start_pos: 0 parent: table.global_scope } errors: []errors.Error{} warnings: []errors.Warning{} } p.set_path(path) res := p.parse() unsafe { p.free_scanner() } return res } [unsafe] pub fn (mut p Parser) free() { unsafe { p.free_scanner() } } [unsafe] pub fn (mut p Parser) free_scanner() { unsafe { if p.scanner != 0 { p.scanner.free() p.scanner = &scanner.Scanner(0) } } } pub fn (mut p Parser) set_path(path string) { p.file_name = path p.file_base = os.base(path) p.file_name_dir = os.dir(path) p.inside_vlib_file = p.file_name_dir.contains('vlib') p.inside_test_file = p.file_base.ends_with('_test.v') || p.file_base.ends_with('_test.vv') || p.file_base.all_before_last('.v').all_before_last('.').ends_with('_test') hash := fnv1a.sum64_string(path) p.unique_prefix = hash.hex_full() p.file_backend_mode = .v before_dot_v := path.all_before_last('.v') // also works for .vv and .vsh language := before_dot_v.all_after_last('.') language_with_underscore := before_dot_v.all_after_last('_') if language == before_dot_v && language_with_underscore == before_dot_v { return } actual_language := if language == before_dot_v { language_with_underscore } else { language } match actual_language { 'c' { p.file_backend_mode = .c } 'js' { p.file_backend_mode = .js } else { arch := pref.arch_from_string(actual_language) or { pref.Arch._auto } p.file_backend_mode = ast.pref_arch_to_table_language(arch) if arch == ._auto { p.file_backend_mode = .v } } } } pub fn parse_file(path string, table &ast.Table, comments_mode scanner.CommentsMode, pref &pref.Preferences) &ast.File { // Note: when comments_mode == .toplevel_comments, // the parser gives feedback to the scanner about toplevel statements, so that the scanner can skip // all the tricky inner comments. This is needed because we do not have a good general solution // for handling them, and should be removed when we do (the general solution is also needed for vfmt) $if trace_parse_file ? { eprintln('> ${@MOD}.${@FN} comments_mode: ${comments_mode:-20} | path: $path') } mut p := Parser{ scanner: scanner.new_scanner_file(path, comments_mode, pref) or { panic(err) } comments_mode: comments_mode table: table pref: pref scope: &ast.Scope{ start_pos: 0 parent: table.global_scope } errors: []errors.Error{} warnings: []errors.Warning{} } p.set_path(path) res := p.parse() unsafe { p.free_scanner() } return res } pub fn parse_vet_file(path string, table_ &ast.Table, pref &pref.Preferences) (&ast.File, []vet.Error) { $if trace_parse_vet_file ? { eprintln('> ${@MOD}.${@FN} path: $path') } global_scope := &ast.Scope{ parent: 0 } mut p := Parser{ scanner: scanner.new_scanner_file(path, .parse_comments, pref) or { panic(err) } comments_mode: .parse_comments table: table_ pref: pref scope: &ast.Scope{ start_pos: 0 parent: global_scope } errors: []errors.Error{} warnings: []errors.Warning{} } p.set_path(path) if p.scanner.text.contains_any_substr(['\n ', ' \n']) { source_lines := os.read_lines(path) or { []string{} } for lnumber, line in source_lines { if line.starts_with(' ') { p.vet_error('Looks like you are using spaces for indentation.', lnumber, .vfmt, .space_indent) } if line.ends_with(' ') { p.vet_error('Looks like you have trailing whitespace.', lnumber, .unknown, .trailing_space) } } } p.vet_errors << p.scanner.vet_errors file := p.parse() unsafe { p.free_scanner() } return file, p.vet_errors } pub fn (mut p Parser) parse() &ast.File { util.timing_start('PARSE') defer { util.timing_measure_cumulative('PARSE') } // comments_mode: comments_mode p.init_parse_fns() p.read_first_token() mut stmts := []ast.Stmt{} for p.tok.kind == .comment { stmts << p.comment_stmt() } // module module_decl := p.module_decl() if module_decl.is_skipped { stmts.insert(0, ast.Stmt(module_decl)) } else { stmts << module_decl } // imports for { if p.tok.kind == .key_import { stmts << p.import_stmt() continue } if p.tok.kind == .comment { stmts << p.comment_stmt() continue } break } for { if p.tok.kind == .eof { p.check_unused_imports() break } stmt := p.top_stmt() // clear the attributes after each statement if !(stmt is ast.ExprStmt && (stmt as ast.ExprStmt).expr is ast.Comment) { p.attrs = [] } stmts << stmt if p.should_abort { break } } p.scope.end_pos = p.tok.pos mut errors := p.errors mut warnings := p.warnings mut notices := p.notices if p.pref.check_only { errors << p.scanner.errors warnings << p.scanner.warnings notices << p.scanner.notices } // codegen if p.codegen_text.len > 0 && !p.pref.is_fmt { ptext := 'module ' + p.mod.all_after_last('.') + p.codegen_text codegen_files << parse_text(ptext, p.file_name, p.table, p.comments_mode, p.pref) } return &ast.File{ path: p.file_name path_base: p.file_base is_test: p.inside_test_file is_generated: p.is_generated is_translated: p.is_translated nr_lines: p.scanner.line_nr nr_bytes: p.scanner.text.len mod: module_decl imports: p.ast_imports imported_symbols: p.imported_symbols auto_imports: p.auto_imports stmts: stmts scope: p.scope global_scope: p.table.global_scope errors: errors warnings: warnings notices: notices global_labels: p.global_labels } } /* struct Queue { mut: idx int mu &sync.Mutex mu2 &sync.Mutex paths []string table &ast.Table parsed_ast_files []&ast.File pref &pref.Preferences global_scope &ast.Scope } fn (mut q Queue) run() { for { q.mu.lock() idx := q.idx if idx >= q.paths.len { q.mu.unlock() return } q.idx++ q.mu.unlock() println('run(idx=$idx)') path := q.paths[idx] file := parse_file(path, q.table, .skip_comments, q.pref, q.global_scope) q.mu2.lock() q.parsed_ast_files << file q.mu2.unlock() println('run done(idx=$idx)') } } */ pub fn parse_files(paths []string, table &ast.Table, pref &pref.Preferences) []&ast.File { mut timers := util.new_timers(should_print: false, label: 'parse_files: $paths') $if time_parsing ? { timers.should_print = true } $if macos { /* if !pref.no_parallel && paths[0].contains('/array.v') { println('\n\n\nparse_files() nr_files=$paths.len') println(paths) nr_cpus := runtime.nr_cpus() mut q := &Queue{ paths: paths table: table pref: pref global_scope: global_scope mu: sync.new_mutex() mu2: sync.new_mutex() } for _ in 0 .. nr_cpus - 1 { go q.run() } time.sleep(time.second) println('all done') return q.parsed_ast_files } */ } mut files := []&ast.File{cap: paths.len} for path in paths { timers.start('parse_file $path') files << parse_file(path, table, .skip_comments, pref) timers.show('parse_file $path') } if codegen_files.len > 0 { files << codegen_files codegen_files.clear() } return files } // codegen allows you to generate V code, so that it can be parsed, // checked, markused, cgen-ed etc further, just like user's V code. pub fn (mut p Parser) codegen(code string) { $if debug_codegen ? { eprintln('parser.codegen:\n $code') } p.codegen_text += '\n' + code } pub fn (mut p Parser) init_parse_fns() { // p.prefix_parse_fns = make(100, 100, sizeof(PrefixParseFn)) // p.prefix_parse_fns[token.Kind.name] = parse_name } pub fn (mut p Parser) read_first_token() { // need to call next() 2 times to get peek token and current token p.next() p.next() } [inline] pub fn (p &Parser) peek_token(n int) token.Token { return p.scanner.peek_token(n - 2) } // peek token in if guard `if x,y := opt()` after var_list `x,y` pub fn (p &Parser) peek_token_after_var_list() token.Token { mut n := 0 mut tok := p.tok for { if tok.kind == .key_mut { n += 2 } else { n++ } tok = p.scanner.peek_token(n - 2) if tok.kind != .comma { break } else { n++ tok = p.scanner.peek_token(n - 2) } } return tok } fn (p &Parser) is_array_type() bool { mut i := 1 mut tok := p.tok line_nr := p.tok.line_nr for { tok = p.peek_token(i) if tok.line_nr != line_nr { return false } if tok.kind in [.name, .amp] { return true } i++ if tok.kind == .lsbr || tok.kind != .rsbr { continue } } return false } pub fn (mut p Parser) open_scope() { p.scope = &ast.Scope{ parent: p.scope start_pos: p.tok.pos } } pub fn (mut p Parser) close_scope() { // p.scope.end_pos = p.tok.pos // NOTE: since this is usually called after `p.parse_block()` // ie. when `prev_tok` is rcbr `}` we most likely want `prev_tok` // we could do the following, but probably not needed in 99% of cases: // `end_pos = if p.prev_tok.kind == .rcbr { p.prev_tok.pos } else { p.tok.pos }` p.scope.end_pos = p.prev_tok.pos p.scope.parent.children << p.scope p.scope = p.scope.parent } pub fn (mut p Parser) parse_block() []ast.Stmt { p.open_scope() stmts := p.parse_block_no_scope(false) p.close_scope() return stmts } pub fn (mut p Parser) parse_block_no_scope(is_top_level bool) []ast.Stmt { p.check(.lcbr) mut stmts := []ast.Stmt{cap: 20} if p.tok.kind != .rcbr { mut count := 0 for p.tok.kind !in [.eof, .rcbr] { stmts << p.stmt(is_top_level) count++ if count % 100000 == 0 { eprintln('parsed $count statements so far from fn $p.cur_fn_name ...') } if count > 1000000 { p.error_with_pos('parsed over $count statements from fn $p.cur_fn_name, the parser is probably stuck', p.tok.pos()) return [] } } } if is_top_level { p.top_level_statement_end() } p.check(.rcbr) return stmts } fn (mut p Parser) next() { p.prev_tok = p.tok p.tok = p.peek_tok p.peek_tok = p.scanner.scan() } fn (mut p Parser) check(expected token.Kind) { p.name_error = false if _likely_(p.tok.kind == expected) { p.next() } else { if expected == .name { p.name_error = true } mut s := expected.str() // quote keywords, punctuation, operators if token.is_key(s) || (s.len > 0 && !s[0].is_letter()) { s = '`$s`' } p.error('unexpected $p.tok, expecting $s') } } // JS functions can have multiple dots in their name: // JS.foo.bar.and.a.lot.more.dots() fn (mut p Parser) check_js_name() string { mut name := '' for p.peek_tok.kind == .dot { name += '${p.tok.lit}.' p.next() // .name p.next() // .dot } // last .name name += p.tok.lit p.next() return name } fn (mut p Parser) check_name() string { name := p.tok.lit if p.peek_tok.kind == .dot && name in p.imports { p.register_used_import(name) } p.check(.name) return name } [if trace_parser ?] fn (p &Parser) trace_parser(label string) { eprintln('parsing: ${p.file_name:-30}|tok.pos: ${p.tok.pos().line_str():-39}|tok.kind: ${p.tok.kind:-10}|tok.lit: ${p.tok.lit:-10}|$label') } pub fn (mut p Parser) top_stmt() ast.Stmt { p.trace_parser('top_stmt') for { match p.tok.kind { .key_pub { match p.peek_tok.kind { .key_const { return p.const_decl() } .key_fn { return p.fn_decl() } .key_struct, .key_union { return p.struct_decl() } .key_interface { return p.interface_decl() } .key_enum { return p.enum_decl() } .key_type { return p.type_decl() } else { return p.error('wrong pub keyword usage') } } } .lsbr { // attrs are stored in `p.attrs` p.attributes() continue } .key_asm { return p.asm_stmt(true) } .key_interface { return p.interface_decl() } .key_import { p.error_with_pos('`import x` can only be declared at the beginning of the file', p.tok.pos()) return p.import_stmt() } .key_global { return p.global_decl() } .key_const { return p.const_decl() } .key_fn { return p.fn_decl() } .key_struct { return p.struct_decl() } .dollar { if_expr := p.if_expr(true) return ast.ExprStmt{ expr: if_expr pos: if_expr.pos } } .hash { return p.hash() } .key_type { return p.type_decl() } .key_enum { return p.enum_decl() } .key_union { return p.struct_decl() } .comment { return p.comment_stmt() } else { p.inside_fn = true if p.pref.is_script && !p.pref.is_test { p.open_scope() mut stmts := []ast.Stmt{} for p.tok.kind != .eof { stmts << p.stmt(false) } p.close_scope() return ast.FnDecl{ name: 'main.main' short_name: 'main' mod: 'main' is_main: true stmts: stmts file: p.file_name return_type: ast.void_type scope: p.scope label_names: p.label_names } } else if p.pref.is_fmt { return p.stmt(false) } else { return p.error('bad top level statement ' + p.tok.str()) } } } if p.should_abort { break } } // TODO remove dummy return statement // the compiler complains if it's not there return ast.empty_stmt() } // TODO [if vfmt] pub fn (mut p Parser) check_comment() ast.Comment { if p.tok.kind == .comment { return p.comment() } return ast.Comment{} } pub fn (mut p Parser) comment() ast.Comment { mut pos := p.tok.pos() text := p.tok.lit num_newlines := text.count('\n') is_multi := num_newlines > 0 is_inline := text.len + 4 == p.tok.len // 4: `/` `*` `*` `/` pos.last_line = pos.line_nr + num_newlines p.next() // Filter out false positive space indent vet errors inside comments if p.vet_errors.len > 0 && is_multi { p.vet_errors = p.vet_errors.filter(it.typ != .space_indent || it.pos.line_nr - 1 > pos.last_line || it.pos.line_nr - 1 <= pos.line_nr) } return ast.Comment{ text: text is_multi: is_multi is_inline: is_inline pos: pos } } pub fn (mut p Parser) comment_stmt() ast.ExprStmt { comment := p.comment() return ast.ExprStmt{ expr: comment pos: comment.pos } } [params] struct EatCommentsConfig { same_line bool // Only eat comments on the same line as the previous token follow_up bool // Comments directly below the previous token as long as there is no empty line } pub fn (mut p Parser) eat_comments(cfg EatCommentsConfig) []ast.Comment { mut line := p.prev_tok.line_nr mut comments := []ast.Comment{} for { if p.tok.kind != .comment || (cfg.same_line && p.tok.line_nr > line) || (cfg.follow_up && (p.tok.line_nr > line + 1 || p.tok.lit.contains('\n'))) { break } comments << p.comment() if cfg.follow_up { line = p.prev_tok.line_nr } } return comments } pub fn (mut p Parser) stmt(is_top_level bool) ast.Stmt { p.trace_parser('stmt($is_top_level)') p.is_stmt_ident = p.tok.kind == .name match p.tok.kind { .lcbr { mut pos := p.tok.pos() stmts := p.parse_block() pos.last_line = p.prev_tok.line_nr return ast.Block{ stmts: stmts pos: pos } } .key_assert { p.next() mut pos := p.tok.pos() expr := p.expr(0) pos.update_last_line(p.prev_tok.line_nr) return ast.AssertStmt{ expr: expr pos: pos.extend(p.tok.pos()) is_used: p.inside_test_file || !p.pref.is_prod } } .key_for { return p.for_stmt() } .name { if p.tok.lit == 'sql' && p.peek_tok.kind == .name { return p.sql_stmt() } if p.peek_tok.kind == .colon { // `label:` spos := p.tok.pos() name := p.check_name() if name in p.label_names { return p.error_with_pos('duplicate label `$name`', spos) } p.label_names << name p.next() if p.tok.kind == .key_for { for_pos := p.tok.pos() mut stmt := p.stmt(is_top_level) match mut stmt { ast.ForStmt { stmt.label = name return stmt } ast.ForInStmt { stmt.label = name return stmt } ast.ForCStmt { stmt.label = name return stmt } else { return p.error_with_pos('unknown kind of For statement', for_pos) } } } return ast.GotoLabel{ name: name pos: spos.extend(p.tok.pos()) } } else if p.peek_tok.kind == .name { return p.error_with_pos('unexpected name `$p.tok.lit`', p.tok.pos()) } else if !p.inside_if_expr && !p.inside_match_body && !p.inside_or_expr && p.peek_tok.kind in [.rcbr, .eof] && !p.mark_var_as_used(p.tok.lit) { return p.error_with_pos('`$p.tok.lit` evaluated but not used', p.tok.pos()) } return p.parse_multi_expr(is_top_level) } .comment { return p.comment_stmt() } .key_return { if p.inside_defer { return p.error_with_pos('`return` not allowed inside `defer` block', p.tok.pos()) } else { return p.return_stmt() } } .dollar { match p.peek_tok.kind { .key_if { mut pos := p.tok.pos() expr := p.if_expr(true) pos.update_last_line(p.prev_tok.line_nr) return ast.ExprStmt{ expr: expr pos: pos } } .key_for { return p.comptime_for() } .name { mut pos := p.tok.pos() expr := p.comptime_call() pos.update_last_line(p.prev_tok.line_nr) return ast.ExprStmt{ expr: expr pos: pos } } else { return p.error_with_pos('unexpected \$', p.tok.pos()) } } } .key_continue, .key_break { tok := p.tok line := p.tok.line_nr p.next() mut label := '' if p.tok.line_nr == line && p.tok.kind == .name { label = p.check_name() } return ast.BranchStmt{ kind: tok.kind label: label pos: tok.pos() } } .key_unsafe { return p.unsafe_stmt() } .hash { return p.hash() } .key_defer { if p.inside_defer { return p.error_with_pos('`defer` blocks cannot be nested', p.tok.pos()) } else { p.next() spos := p.tok.pos() p.inside_defer = true p.defer_vars = []ast.Ident{} stmts := p.parse_block() p.inside_defer = false return ast.DeferStmt{ stmts: stmts defer_vars: p.defer_vars.clone() pos: spos.extend_with_last_line(p.tok.pos(), p.prev_tok.line_nr) } } } .key_go { go_expr := p.go_expr() return ast.ExprStmt{ expr: go_expr pos: go_expr.pos } } .key_goto { p.next() spos := p.tok.pos() name := p.check_name() return ast.GotoStmt{ name: name pos: spos } } .key_const { return p.error_with_pos('const can only be defined at the top level (outside of functions)', p.tok.pos()) } .key_asm { return p.asm_stmt(false) } // literals, 'if', etc. in here else { return p.parse_multi_expr(is_top_level) } } } fn (mut p Parser) asm_stmt(is_top_level bool) ast.AsmStmt { p.inside_asm = true p.inside_asm_template = true defer { p.inside_asm = false p.inside_asm_template = false } p.n_asm = 0 if is_top_level { p.top_level_statement_start() } mut backup_scope := p.scope pos := p.tok.pos() p.check(.key_asm) mut arch := pref.arch_from_string(p.tok.lit) or { pref.Arch._auto } mut is_volatile := false mut is_goto := false if p.tok.kind == .key_volatile { arch = pref.arch_from_string(p.peek_tok.lit) or { pref.Arch._auto } is_volatile = true p.next() } else if p.tok.kind == .key_goto { arch = pref.arch_from_string(p.peek_tok.lit) or { pref.Arch._auto } is_goto = true p.next() } if arch == ._auto && !p.pref.is_fmt { if p.tok.lit == '' { p.error('missing assembly architecture. Try i386, amd64 or arm64.') } p.error('unknown assembly architecture') } if p.tok.kind != .name { p.error('must specify assembly architecture') } else { p.next() } p.check_for_impure_v(ast.pref_arch_to_table_language(arch), p.prev_tok.pos()) p.check(.lcbr) p.scope = &ast.Scope{ parent: 0 // you shouldn't be able to reference other variables in assembly blocks detached_from_parent: true start_pos: p.tok.pos objects: ast.all_registers(mut p.table, arch) // } mut local_labels := []string{} // riscv: https://github.com/jameslzhu/riscv-card/blob/master/riscv-card.pdf // x86: https://www.felixcloutier.com/x86/ // arm: https://developer.arm.com/documentation/dui0068/b/arm-instruction-reference mut templates := []ast.AsmTemplate{} for p.tok.kind !in [.semicolon, .rcbr] { template_pos := p.tok.pos() mut name := '' if p.tok.kind == .name && arch == .amd64 && p.tok.lit in ['rex', 'vex', 'xop'] { name += p.tok.lit p.next() for p.tok.kind == .dot { p.next() name += '.' + p.tok.lit p.check(.name) } name += ' ' } is_directive := p.tok.kind == .dot if is_directive { p.next() } if p.tok.kind in [.key_in, .key_lock, .key_orelse] { // `in`, `lock`, `or` are v keywords that are also x86/arm/riscv instructions. name += p.tok.kind.str() p.next() } else if p.tok.kind == .number { name += p.tok.lit p.next() } else { name += p.tok.lit p.check(.name) } // dots are part of instructions for some riscv extensions if arch in [.rv32, .rv64] { for p.tok.kind == .dot { name += '.' p.next() name += p.tok.lit p.check(.name) } } mut is_label := false mut args := []ast.AsmArg{} if p.tok.line_nr == p.prev_tok.line_nr { args_loop: for { if p.prev_tok.pos().line_nr < p.tok.pos().line_nr { break } mut segment := '' if p.tok.kind == .name && p.peek_tok.kind == .colon { segment = p.tok.lit p.next() p.next() } match p.tok.kind { .name { args << p.reg_or_alias() } .number { number_lit := p.parse_number_literal() match number_lit { ast.FloatLiteral { args << ast.FloatLiteral{ ...number_lit } } ast.IntegerLiteral { if is_directive { args << ast.AsmDisp{ val: number_lit.val pos: number_lit.pos } } else { args << ast.IntegerLiteral{ ...number_lit } } } else { verror('p.parse_number_literal() invalid output: `$number_lit`') } } } .chartoken { args << ast.CharLiteral{ val: p.tok.lit pos: p.tok.pos() } p.next() } .colon { is_label = true p.next() local_labels << name break } .lsbr { mut addressing := p.asm_addressing() addressing.segment = segment args << addressing } .rcbr { break } .semicolon { break } else { p.error('invalid token in assembly block') } } if p.tok.kind == .comma { p.next() } else { break } } // if p.prev_tok.pos().line_nr < p.tok.pos().line_nr { // break // } } mut comments := []ast.Comment{} for p.tok.kind == .comment { comments << p.comment() } if is_directive && name in ['globl', 'global'] { for arg in args { p.global_labels << (arg as ast.AsmAlias).name } } templates << ast.AsmTemplate{ name: name args: args comments: comments is_label: is_label is_directive: is_directive pos: template_pos.extend(p.tok.pos()) } } mut scope := p.scope p.scope = backup_scope p.inside_asm_template = false mut output, mut input, mut clobbered, mut global_labels := []ast.AsmIO{}, []ast.AsmIO{}, []ast.AsmClobbered{}, []string{} if !is_top_level { if p.tok.kind == .semicolon { output = p.asm_ios(true) if p.tok.kind == .semicolon { input = p.asm_ios(false) } if p.tok.kind == .semicolon { // because p.reg_or_alias() requires the scope with registers to recognize registers. backup_scope = p.scope p.scope = scope p.next() for p.tok.kind == .name { reg := ast.AsmRegister{ name: p.tok.lit typ: 0 size: -1 } p.next() mut comments := []ast.Comment{} for p.tok.kind == .comment { comments << p.comment() } clobbered << ast.AsmClobbered{ reg: reg comments: comments } if p.tok.kind in [.rcbr, .semicolon] { break } } if is_goto && p.tok.kind == .semicolon { p.next() for p.tok.kind == .name { global_labels << p.tok.lit p.next() } } } } } else if p.tok.kind == .semicolon { p.error('extended assembly is not allowed as a top level statement') } p.scope = backup_scope p.check(.rcbr) if is_top_level { p.top_level_statement_end() } scope.end_pos = p.prev_tok.pos return ast.AsmStmt{ arch: arch is_goto: is_goto is_volatile: is_volatile templates: templates output: output input: input clobbered: clobbered pos: pos.extend(p.prev_tok.pos()) is_basic: is_top_level || output.len + input.len + clobbered.len == 0 scope: scope global_labels: global_labels local_labels: local_labels } } fn (mut p Parser) reg_or_alias() ast.AsmArg { p.check(.name) if p.prev_tok.lit in p.scope.objects { x := unsafe { p.scope.objects[p.prev_tok.lit] } if x is ast.AsmRegister { return ast.AsmArg(x as ast.AsmRegister) } else { verror('non-register ast.ScopeObject found in scope') return ast.AsmDisp{} // should not be reached } } else if p.prev_tok.len >= 2 && p.prev_tok.lit[0] in [`b`, `f`] && p.prev_tok.lit[1..].bytes().all(it.is_digit()) { return ast.AsmDisp{ val: p.prev_tok.lit[1..] + p.prev_tok.lit[0].ascii_str() } } else { return ast.AsmAlias{ name: p.prev_tok.lit pos: p.prev_tok.pos() } } } // fn (mut p Parser) asm_addressing() ast.AsmAddressing { // pos := p.tok.pos() // p.check(.lsbr) // unknown_addressing_mode := 'unknown addressing mode. supported ones are [displacement], [base], [base + displacement] [index ∗ scale + displacement], [base + index ∗ scale + displacement], [base + index + displacement] [rip + displacement]' // mut mode := ast.AddressingMode.invalid // if p.peek_tok.kind == .rsbr { // if p.tok.kind == .name { // mode = .base // } else if p.tok.kind == .number { // mode = .displacement // } else { // p.error(unknown_addressing_mode) // } // } else if p.peek_tok.kind == .mul { // mode = .index_times_scale_plus_displacement // } else if p.tok.lit == 'rip' { // mode = .rip_plus_displacement // } else if p.peek_tok3.kind == .mul { // mode = .base_plus_index_times_scale_plus_displacement // } else if p.peek_tok.kind == .plus && p.peek_tok3.kind == .rsbr { // mode = .base_plus_displacement // } else if p.peek_tok.kind == .plus && p.peek_tok3.kind == .plus { // mode = .base_plus_index_plus_displacement // } else { // p.error(unknown_addressing_mode) // } // mut displacement, mut base, mut index, mut scale := u32(0), ast.AsmArg{}, ast.AsmArg{}, -1 // match mode { // .base { // base = p.reg_or_alias() // } // .displacement { // displacement = p.tok.lit.u32() // p.check(.number) // } // .base_plus_displacement { // base = p.reg_or_alias() // p.check(.plus) // displacement = p.tok.lit.u32() // p.check(.number) // } // .index_times_scale_plus_displacement { // index = p.reg_or_alias() // p.check(.mul) // scale = p.tok.lit.int() // p.check(.number) // p.check(.plus) // displacement = p.tok.lit.u32() // p.check(.number) // } // .base_plus_index_times_scale_plus_displacement { // base = p.reg_or_alias() // p.check(.plus) // index = p.reg_or_alias() // p.check(.mul) // scale = p.tok.lit.int() // p.check(.number) // p.check(.plus) // displacement = p.tok.lit.u32() // p.check(.number) // } // .rip_plus_displacement { // base = p.reg_or_alias() // p.check(.plus) // displacement = p.tok.lit.u32() // p.check(.number) // } // .base_plus_index_plus_displacement { // base = p.reg_or_alias() // p.check(.plus) // index = p.reg_or_alias() // p.check(.plus) // displacement = p.tok.lit.u32() // p.check(.number) // } // .invalid {} // there was already an error above // } // p.check(.rsbr) // return ast.AsmAddressing{ // base: base // displacement: displacement // index: index // scale: scale // mode: mode // pos: pos.extend(p.prev_tok.pos()) // } // } fn (mut p Parser) asm_addressing() ast.AsmAddressing { pos := p.tok.pos() p.check(.lsbr) unknown_addressing_mode := 'unknown addressing mode. supported ones are [displacement], [base], [base + displacement], [index ∗ scale + displacement], [base + index ∗ scale + displacement], [base + index + displacement], [rip + displacement]' // this mess used to look much cleaner before the removal of peek_tok2/3, see above code for cleaner version if p.peek_tok.kind == .rsbr { // [displacement] or [base] if p.tok.kind == .name { base := p.reg_or_alias() p.check(.rsbr) return ast.AsmAddressing{ mode: .base base: base pos: pos.extend(p.prev_tok.pos()) } } else if p.tok.kind == .number { displacement := if p.tok.kind == .name { p.reg_or_alias() } else { x := ast.AsmArg(ast.AsmDisp{ val: p.tok.lit pos: p.tok.pos() }) p.check(.number) x } p.check(.rsbr) return ast.AsmAddressing{ mode: .displacement displacement: displacement pos: pos.extend(p.prev_tok.pos()) } } else { p.error(unknown_addressing_mode) } } if p.peek_tok.kind == .plus && p.tok.kind == .name { // [base + displacement], [base + index ∗ scale + displacement], [base + index + displacement] or [rip + displacement] if p.tok.lit == 'rip' { rip := p.reg_or_alias() p.next() displacement := if p.tok.kind == .name { p.reg_or_alias() } else { x := ast.AsmArg(ast.AsmDisp{ val: p.tok.lit pos: p.tok.pos() }) p.check(.number) x } p.check(.rsbr) return ast.AsmAddressing{ mode: .rip_plus_displacement base: rip displacement: displacement pos: pos.extend(p.prev_tok.pos()) } } base := p.reg_or_alias() p.next() if p.peek_tok.kind == .rsbr { if p.tok.kind == .number { displacement := if p.tok.kind == .name { p.reg_or_alias() } else { x := ast.AsmArg(ast.AsmDisp{ val: p.tok.lit pos: p.tok.pos() }) p.check(.number) x } p.check(.rsbr) return ast.AsmAddressing{ mode: .base_plus_displacement base: base displacement: displacement pos: pos.extend(p.prev_tok.pos()) } } else { p.error(unknown_addressing_mode) } } index := p.reg_or_alias() if p.tok.kind == .mul { p.next() scale := p.tok.lit.int() p.check(.number) p.check(.plus) displacement := if p.tok.kind == .name { p.reg_or_alias() } else { x := ast.AsmArg(ast.AsmDisp{ val: p.tok.lit pos: p.tok.pos() }) p.check(.number) x } p.check(.rsbr) return ast.AsmAddressing{ mode: .base_plus_index_times_scale_plus_displacement base: base index: index scale: scale displacement: displacement pos: pos.extend(p.prev_tok.pos()) } } else if p.tok.kind == .plus { p.next() displacement := if p.tok.kind == .name { p.reg_or_alias() } else { x := ast.AsmArg(ast.AsmDisp{ val: p.tok.lit pos: p.tok.pos() }) p.check(.number) x } p.check(.rsbr) return ast.AsmAddressing{ mode: .base_plus_index_plus_displacement base: base index: index displacement: displacement pos: pos.extend(p.prev_tok.pos()) } } } if p.peek_tok.kind == .mul { // [index ∗ scale + displacement] index := p.reg_or_alias() p.next() scale := p.tok.lit.int() p.check(.number) p.check(.plus) displacement := if p.tok.kind == .name { p.reg_or_alias() } else { x := ast.AsmArg(ast.AsmDisp{ val: p.tok.lit pos: p.tok.pos() }) p.check(.number) x } p.check(.rsbr) return ast.AsmAddressing{ mode: .index_times_scale_plus_displacement index: index scale: scale displacement: displacement pos: pos.extend(p.prev_tok.pos()) } } p.error(unknown_addressing_mode) return ast.AsmAddressing{} } fn (mut p Parser) asm_ios(output bool) []ast.AsmIO { mut res := []ast.AsmIO{} p.check(.semicolon) if p.tok.kind in [.rcbr, .semicolon] { return [] } for { pos := p.tok.pos() mut constraint := '' if p.tok.kind == .lpar { constraint = if output { '+r' } else { 'r' } // default constraint, though vfmt fmts to `+r` and `r` } else { constraint += match p.tok.kind { .assign { '=' } .plus { '+' } .mod { '%' } .amp { '&' } else { '' } } if constraint != '' { p.next() } constraint += p.tok.lit if p.tok.kind == .at { p.next() } else { p.check(.name) } } mut expr := p.expr(0) if mut expr is ast.ParExpr { expr = expr.expr } else { p.error('asm in/output must be enclosed in brackets') } mut alias := '' if p.tok.kind == .key_as { p.next() alias = p.tok.lit p.check(.name) } else if mut expr is ast.Ident { alias = expr.name } // for constraints like `a`, no alias is needed, it is reffered to as rcx mut comments := []ast.Comment{} for p.tok.kind == .comment { comments << p.comment() } res << ast.AsmIO{ alias: alias constraint: constraint expr: expr comments: comments pos: pos.extend(p.prev_tok.pos()) } p.n_asm++ if p.tok.kind in [.semicolon, .rcbr] { break } } return res } fn (mut p Parser) expr_list() ([]ast.Expr, []ast.Comment) { mut exprs := []ast.Expr{} mut comments := []ast.Comment{} for { expr := p.expr(0) if expr is ast.Comment { comments << expr } else { exprs << expr if p.tok.kind != .comma { break } p.next() } } return exprs, comments } fn (mut p Parser) is_attributes() bool { if p.tok.kind != .lsbr { return false } mut i := 0 for { tok := p.peek_token(i) if tok.kind == .eof || tok.line_nr != p.tok.line_nr { return false } if tok.kind == .rsbr { break } i++ } peek_rsbr_tok := p.peek_token(i + 1) if peek_rsbr_tok.line_nr == p.tok.line_nr && peek_rsbr_tok.kind != .rcbr { return false } return true } // when is_top_stmt is true attrs are added to p.attrs fn (mut p Parser) attributes() { p.check(.lsbr) mut has_ctdefine := false for p.tok.kind != .rsbr { start_pos := p.tok.pos() attr := p.parse_attr() if p.attrs.contains(attr.name) && attr.name != 'wasm_export' { p.error_with_pos('duplicate attribute `$attr.name`', start_pos.extend(p.prev_tok.pos())) return } if attr.kind == .comptime_define { if has_ctdefine { p.error_with_pos('only one `[if flag]` may be applied at a time `$attr.name`', start_pos.extend(p.prev_tok.pos())) return } else { has_ctdefine = true } } p.attrs << attr if p.tok.kind != .semicolon { if p.tok.kind == .rsbr { p.next() break } p.error('unexpected $p.tok, expecting `;`') return } p.next() } if p.attrs.len == 0 { p.error_with_pos('attributes cannot be empty', p.prev_tok.pos().extend(p.tok.pos())) return } } fn (mut p Parser) parse_attr() ast.Attr { mut kind := ast.AttrKind.plain apos := p.prev_tok.pos() if p.tok.kind == .key_unsafe { p.next() return ast.Attr{ name: 'unsafe' kind: kind pos: apos.extend(p.tok.pos()) } } mut name := '' mut has_arg := false mut arg := '' mut comptime_cond := ast.empty_expr() mut comptime_cond_opt := false if p.tok.kind == .key_if { kind = .comptime_define p.next() p.comptime_if_cond = true p.inside_if_expr = true p.inside_ct_if_expr = true comptime_cond = p.expr(0) p.comptime_if_cond = false p.inside_if_expr = false p.inside_ct_if_expr = false if comptime_cond is ast.PostfixExpr { comptime_cond_opt = true } name = comptime_cond.str() } else if p.tok.kind == .string { name = p.tok.lit kind = .string p.next() } else { name = p.check_name() // support dot prefix `module.name: arg` if p.tok.kind == .dot { p.next() name += '.' name += p.check_name() } if p.tok.kind == .colon { has_arg = true p.next() if p.tok.kind == .name { // `name: arg` kind = .plain arg = p.check_name() } else if p.tok.kind == .number { // `name: 123` kind = .number arg = p.tok.lit p.next() } else if p.tok.kind == .string { // `name: 'arg'` kind = .string arg = p.tok.lit p.next() } else if p.tok.kind == .key_true || p.tok.kind == .key_false { // `name: true` kind = .bool arg = p.tok.kind.str() p.next() } else { p.error('unexpected $p.tok, an argument is expected after `:`') } } } return ast.Attr{ name: name has_arg: has_arg arg: arg kind: kind ct_expr: comptime_cond ct_opt: comptime_cond_opt pos: apos.extend(p.tok.pos()) } } pub fn (mut p Parser) language_not_allowed_error(language ast.Language, pos token.Pos) { upcase_language := language.str().to_upper() p.error_with_pos('$upcase_language code is not allowed in .${p.file_backend_mode}.v files, please move it to a .${language}.v file', pos) } pub fn (mut p Parser) language_not_allowed_warning(language ast.Language, pos token.Pos) { upcase_language := language.str().to_upper() p.warn_with_pos('$upcase_language code will not be allowed in pure .v files, please move it to a .${language}.v file instead', pos) } pub fn (mut p Parser) check_for_impure_v(language ast.Language, pos token.Pos) { if language == .v { // pure V code is always allowed everywhere return } else { match p.file_backend_mode { .c { if language != .c { p.language_not_allowed_error(language, pos) return } } .js { if language != .js { p.language_not_allowed_error(language, pos) return } } else {} } } if !p.pref.warn_impure_v { // the stricter mode is not ON yet => allow everything for now return } if p.file_backend_mode != language { if p.file_backend_mode == .v { p.language_not_allowed_warning(language, pos) return } } } pub fn (mut p Parser) error(s string) ast.NodeError { return p.error_with_pos(s, p.tok.pos()) } pub fn (mut p Parser) warn(s string) { p.warn_with_pos(s, p.tok.pos()) } pub fn (mut p Parser) note(s string) { p.note_with_pos(s, p.tok.pos()) } pub fn (mut p Parser) error_with_pos(s string, pos token.Pos) ast.NodeError { if p.pref.fatal_errors { exit(1) } mut kind := 'error:' if p.pref.output_mode == .stdout && !p.pref.check_only { if p.pref.is_verbose { print_backtrace() kind = 'parser error:' } ferror := util.formatted_error(kind, s, p.file_name, pos) eprintln(ferror) exit(1) } else { p.errors << errors.Error{ file_path: p.file_name pos: pos reporter: .parser message: s } // To avoid getting stuck after an error, the parser // will proceed to the next token. if p.pref.check_only { p.next() } } if p.pref.output_mode == .silent { // Normally, parser errors mean that the parser exits immediately, so there can be only 1 parser error. // In the silent mode however, the parser continues to run, even though it would have stopped. Some // of the parser logic does not expect that, and may loop forever. // The p.next() here is needed, so the parser is more robust, and *always* advances, even in the -silent mode. p.next() } return ast.NodeError{ idx: p.errors.len - 1 pos: pos } } pub fn (mut p Parser) error_with_error(error errors.Error) { if p.pref.fatal_errors { exit(1) } mut kind := 'error:' if p.pref.output_mode == .stdout && !p.pref.check_only { if p.pref.is_verbose { print_backtrace() kind = 'parser error:' } ferror := util.formatted_error(kind, error.message, error.file_path, error.pos) eprintln(ferror) exit(1) } else { if p.pref.message_limit >= 0 && p.errors.len >= p.pref.message_limit { p.should_abort = true return } p.errors << error } if p.pref.output_mode == .silent { // Normally, parser errors mean that the parser exits immediately, so there can be only 1 parser error. // In the silent mode however, the parser continues to run, even though it would have stopped. Some // of the parser logic does not expect that, and may loop forever. // The p.next() here is needed, so the parser is more robust, and *always* advances, even in the -silent mode. p.next() } } pub fn (mut p Parser) warn_with_pos(s string, pos token.Pos) { if p.pref.warns_are_errors { p.error_with_pos(s, pos) return } if p.pref.skip_warnings { return } if p.pref.output_mode == .stdout && !p.pref.check_only { ferror := util.formatted_error('warning:', s, p.file_name, pos) eprintln(ferror) } else { if p.pref.message_limit >= 0 && p.warnings.len >= p.pref.message_limit { p.should_abort = true return } p.warnings << errors.Warning{ file_path: p.file_name pos: pos reporter: .parser message: s } } } pub fn (mut p Parser) note_with_pos(s string, pos token.Pos) { if p.pref.skip_warnings { return } if p.is_generated { return } if p.pref.output_mode == .stdout && !p.pref.check_only { ferror := util.formatted_error('notice:', s, p.file_name, pos) eprintln(ferror) } else { p.notices << errors.Notice{ file_path: p.file_name pos: pos reporter: .parser message: s } } } pub fn (mut p Parser) vet_error(msg string, line int, fix vet.FixKind, typ vet.ErrorType) { pos := token.Pos{ line_nr: line + 1 } p.vet_errors << vet.Error{ message: msg file_path: p.scanner.file_path pos: pos kind: .error fix: fix typ: typ } } fn (mut p Parser) parse_multi_expr(is_top_level bool) ast.Stmt { // in here might be 1) multi-expr 2) multi-assign // 1, a, c ... } // multi-expression // a, mut b ... :=/= // multi-assign // collect things upto hard boundaries tok := p.tok mut pos := tok.pos() mut defer_vars := p.defer_vars p.defer_vars = []ast.Ident{} left, left_comments := p.expr_list() if !(p.inside_defer && p.tok.kind == .decl_assign) { defer_vars << p.defer_vars } p.defer_vars = defer_vars left0 := left[0] if tok.kind == .key_mut && p.tok.kind != .decl_assign { return p.error('expecting `:=` (e.g. `mut x :=`)') } // TODO remove translated if p.tok.kind in [.assign, .decl_assign] || p.tok.kind.is_assign() { return p.partial_assign_stmt(left, left_comments) } else if !p.pref.translated && !p.is_translated && !p.pref.is_fmt && tok.kind !in [.key_if, .key_match, .key_lock, .key_rlock, .key_select] { for node in left { if (is_top_level || p.tok.kind != .rcbr) && node !is ast.CallExpr && node !is ast.PostfixExpr && node !is ast.ComptimeCall && node !is ast.SelectorExpr && node !is ast.DumpExpr { is_complex_infix_expr := node is ast.InfixExpr && (node as ast.InfixExpr).op in [.left_shift, .right_shift, .unsigned_right_shift, .arrow] if !is_complex_infix_expr { return p.error_with_pos('expression evaluated but not used', node.pos()) } } } } pos.update_last_line(p.prev_tok.line_nr) if left.len == 1 { return ast.ExprStmt{ expr: left0 pos: left0.pos() comments: left_comments is_expr: p.inside_for } } return ast.ExprStmt{ expr: ast.ConcatExpr{ vals: left pos: tok.pos() } pos: pos comments: left_comments } } pub fn (mut p Parser) parse_ident(language ast.Language) ast.Ident { // p.warn('name ') is_shared := p.tok.kind == .key_shared is_atomic := p.tok.kind == .key_atomic if is_shared { p.register_auto_import('sync') } mut_pos := p.tok.pos() modifier_kind := p.tok.kind is_mut := p.tok.kind == .key_mut || is_shared || is_atomic if is_mut { p.next() } is_static := p.tok.kind == .key_static if is_static { p.next() } is_volatile := p.tok.kind == .key_volatile if is_volatile { p.next() } if p.tok.kind != .name { if is_mut || is_static || is_volatile { p.error_with_pos('the `$modifier_kind` keyword is invalid here', mut_pos) } else { p.error('unexpected token `$p.tok.lit`') } return ast.Ident{ scope: p.scope } } pos := p.tok.pos() mut name := p.check_name() if name == '_' { return ast.Ident{ tok_kind: p.tok.kind name: '_' comptime: p.comptime_if_cond kind: .blank_ident pos: pos info: ast.IdentVar{ is_mut: false is_static: false is_volatile: false } scope: p.scope } } if p.inside_match_body && name == 'it' { // p.warn('it') } if p.expr_mod.len > 0 { name = '${p.expr_mod}.$name' } return ast.Ident{ tok_kind: p.tok.kind kind: .unresolved name: name comptime: p.comptime_if_cond language: language mod: p.mod pos: pos is_mut: is_mut mut_pos: mut_pos info: ast.IdentVar{ is_mut: is_mut is_static: is_static is_volatile: is_volatile share: ast.sharetype_from_flags(is_shared, is_atomic) } scope: p.scope } } fn (p &Parser) is_typename(t token.Token) bool { return t.kind == .name && (t.lit[0].is_capital() || p.table.known_type(t.lit)) } // heuristics to detect `func()` from `var < expr` // 1. `f<[]` is generic(e.g. `f<[]int>`) because `var < []` is invalid // 2. `f) // 3. `f` is generic because `v1 < foo > v2` is invalid syntax // 4. `f(), in contrast, is not generic!) // 5. `f` is same as case 3 // 7. `f 0 && p.tok.lit[0].is_capital() if lit0_is_capital || p.peek_tok.kind != .lt { return false } mut tok2 := p.peek_token(2) mut tok3 := p.peek_token(3) mut tok4 := p.peek_token(4) mut tok5 := p.peek_token(5) mut kind2, mut kind3, mut kind4, mut kind5 := tok2.kind, tok3.kind, tok4.kind, tok5.kind if kind2 == .amp { // if there is a & in front, shift everything left tok2 = tok3 kind2 = kind3 tok3 = tok4 kind3 = kind4 tok4 = tok5 kind4 = kind5 tok5 = p.peek_token(6) kind5 = tok5.kind } if kind2 == .lsbr { // case 1 return tok3.kind == .rsbr } if kind2 == .name { if tok2.lit == 'map' && kind3 == .lsbr { // case 2 return true } return match kind3 { .gt { true } // case 3 .lt { !(tok4.lit.len == 1 && tok4.lit[0].is_capital()) } // case 4 .comma { p.is_typename(tok2) } // case 5 // case 6 and 7 .dot { kind4 == .name && (kind5 == .gt || (kind5 == .comma && p.is_typename(tok4))) } else { false } } } return false } const valid_tokens_inside_types = [token.Kind.lsbr, .rsbr, .name, .dot, .comma, .key_fn, .lt] fn (mut p Parser) is_generic_cast() bool { if !ast.type_can_start_with_token(p.tok) { return false } mut i := 0 mut level := 0 mut lt_count := 0 for { i++ tok := p.peek_token(i) if tok.kind == .lt { lt_count++ level++ } else if tok.kind == .gt { level-- } if lt_count > 0 && level == 0 { break } if i > 20 || tok.kind !in parser.valid_tokens_inside_types { return false } } next_tok := p.peek_token(i + 1) // `next_tok` is the token following the closing `>` of the generic type: MyType{ // ^ // if `next_tok` is a left paren, then the full expression looks something like // `Foo(` or `Foo(`, which are valid type casts - return true if next_tok.kind == .lpar { return true } // any other token is not a valid generic cast, however return false } pub fn (mut p Parser) name_expr() ast.Expr { prev_tok_kind := p.prev_tok.kind mut node := ast.empty_expr() if p.expecting_type { if p.tok.kind == .dollar { node = p.parse_comptime_type() p.expecting_type = false return node } p.expecting_type = false // get type position before moving to next type_pos := p.tok.pos() typ := p.parse_type() return ast.TypeNode{ typ: typ pos: type_pos } } mut language := ast.Language.v if p.tok.lit == 'C' { language = ast.Language.c p.check_for_impure_v(language, p.tok.pos()) } else if p.tok.lit == 'JS' { language = ast.Language.js p.check_for_impure_v(language, p.tok.pos()) } mut mod := '' // p.warn('resetting') p.expr_mod = '' // `map[string]int` initialization if p.tok.lit == 'map' && p.peek_tok.kind == .lsbr { mut pos := p.tok.pos() map_type := p.parse_map_type() if p.tok.kind == .lcbr { p.next() if p.tok.kind == .rcbr { pos = pos.extend(p.tok.pos()) p.next() } else { if p.pref.is_fmt { map_init := p.map_init() p.check(.rcbr) return map_init } p.error('`}` expected; explicit `map` initialization does not support parameters') } } return ast.MapInit{ typ: map_type pos: pos } } // `chan typ{...}` if p.tok.lit == 'chan' { first_pos := p.tok.pos() mut last_pos := first_pos chan_type := p.parse_chan_type() mut has_cap := false mut cap_expr := ast.empty_expr() p.check(.lcbr) if p.tok.kind == .rcbr { last_pos = p.tok.pos() p.next() } else { key := p.check_name() p.check(.colon) match key { 'cap' { has_cap = true cap_expr = p.expr(0) } 'len', 'init' { return p.error('`$key` cannot be initialized for `chan`. Did you mean `cap`?') } else { return p.error('wrong field `$key`, expecting `cap`') } } last_pos = p.tok.pos() p.check(.rcbr) } return ast.ChanInit{ pos: first_pos.extend(last_pos) has_cap: has_cap cap_expr: cap_expr typ: chan_type } } // Raw string (`s := r'hello \n ') if p.peek_tok.kind == .string && !p.inside_str_interp && p.peek_token(2).kind != .colon { if p.tok.kind == .name && p.tok.lit in ['r', 'c', 'js'] { return p.string_expr() } else { // don't allow any other string prefix except `r`, `js` and `c` return p.error('only `c`, `r`, `js` are recognized string prefixes, but you tried to use `$p.tok.lit`') } } // don't allow r`byte` and c`byte` if p.peek_tok.kind == .chartoken && p.tok.lit.len == 1 && p.tok.lit[0] in [`r`, `c`] { opt := if p.tok.lit == 'r' { '`r` (raw string)' } else { '`c` (c string)' } return p.error('cannot use $opt with `byte` and `rune`') } // Make sure that the var is not marked as used in assignments: `x = 1`, `x += 2` etc // but only when it's actually used (e.g. `println(x)`) known_var := if p.peek_tok.kind.is_assign() { p.scope.known_var(p.tok.lit) } else { p.mark_var_as_used(p.tok.lit) } // Handle modules mut is_mod_cast := false if p.peek_tok.kind == .dot && !known_var && (language != .v || p.known_import(p.tok.lit) || p.mod.all_after_last('.') == p.tok.lit) { // p.tok.lit has been recognized as a module if language == .c { mod = 'C' } else if language == .js { mod = 'JS' } else { if p.tok.lit in p.imports { // mark the imported module as used p.register_used_import(p.tok.lit) if p.peek_tok.kind == .dot && p.peek_token(2).kind != .eof && p.peek_token(2).lit.len > 0 && p.peek_token(2).lit[0].is_capital() { is_mod_cast = true } else if p.peek_tok.kind == .dot && p.peek_token(2).kind != .eof && p.peek_token(2).lit.len == 0 { // incomplete module selector must be handled by dot_expr instead ident := p.parse_ident(language) node = ident if p.inside_defer { if !p.defer_vars.any(it.name == ident.name && it.mod == ident.mod) && ident.name != 'err' { p.defer_vars << ident } } return node } } // prepend the full import mod = p.imports[p.tok.lit] } p.next() p.check(.dot) p.expr_mod = mod } lit0_is_capital := if p.tok.kind != .eof && p.tok.lit.len > 0 { p.tok.lit[0].is_capital() } else { false } is_optional := p.tok.kind == .question is_generic_call := p.is_generic_call() is_generic_cast := p.is_generic_cast() // p.warn('name expr $p.tok.lit $p.peek_tok.str()') same_line := p.tok.line_nr == p.peek_tok.line_nr // `(` must be on same line as name token otherwise it's a ParExpr if !same_line && p.peek_tok.kind == .lpar { ident := p.parse_ident(language) node = ident if p.inside_defer { if !p.defer_vars.any(it.name == ident.name && it.mod == ident.mod) && ident.name != 'err' { p.defer_vars << ident } } } else if p.peek_tok.kind == .lpar || is_generic_call || is_generic_cast || (is_optional && p.peek_token(2).kind == .lpar) { // foo(), foo() or type() cast mut name := if is_optional { p.peek_tok.lit } else { p.tok.lit } if mod.len > 0 { name = '${mod}.$name' } name_w_mod := p.prepend_mod(name) // type cast. TODO: finish // if name in ast.builtin_type_names_to_idx { if (!known_var && (name in p.table.type_idxs || name_w_mod in p.table.type_idxs) && name !in ['C.stat', 'C.sigaction']) || is_mod_cast || is_generic_cast || (language == .v && name.len > 0 && name[0].is_capital()) { // MainLetter(x) is *always* a cast, as long as it is not `C.` // TODO handle C.stat() start_pos := p.tok.pos() mut to_typ := p.parse_type() // this prevents inner casts to also have an `&` // example: &Foo(malloc(int(num))) // without the next line int would result in int* p.is_amp = false p.check(.lpar) mut expr := ast.empty_expr() mut arg := ast.empty_expr() mut has_arg := false expr = p.expr(0) // TODO, string(b, len) if p.tok.kind == .comma && to_typ.idx() == ast.string_type_idx { p.next() arg = p.expr(0) // len has_arg = true } end_pos := p.tok.pos() p.check(.rpar) node = ast.CastExpr{ typ: to_typ typname: p.table.sym(to_typ).name expr: expr arg: arg has_arg: has_arg pos: start_pos.extend(end_pos) } p.expr_mod = '' return node } else { // fn call if is_optional { p.error_with_pos('unexpected $p.prev_tok', p.prev_tok.pos()) } node = p.call_expr(language, mod) if p.tok.kind == .lpar && p.prev_tok.line_nr == p.tok.line_nr { p.next() pos := p.tok.pos() args := p.call_args() p.check(.rpar) node = ast.CallExpr{ left: node args: args pos: pos scope: p.scope } } } } else if (p.peek_tok.kind == .lcbr || (p.peek_tok.kind == .lt && lit0_is_capital)) && (!p.inside_match || (p.inside_select && prev_tok_kind == .arrow && lit0_is_capital)) && !p.inside_match_case && (!p.inside_if || p.inside_select) && (!p.inside_for || p.inside_select) && !known_var { return p.struct_init(p.mod + '.' + p.tok.lit, false) // short_syntax: false } else if p.peek_tok.kind == .lcbr && p.inside_if && lit0_is_capital && !known_var && language == .v { // if a == Foo{} {...} return p.struct_init(p.mod + '.' + p.tok.lit, false) } else if p.peek_tok.kind == .dot && (lit0_is_capital && !known_var && language == .v) { // T.name if p.is_generic_name() { pos := p.tok.pos() name := p.check_name() p.check(.dot) field := p.check_name() fkind := match field { 'name' { ast.GenericKindField.name } 'typ' { ast.GenericKindField.typ } else { ast.GenericKindField.unknown } } pos.extend(p.tok.pos()) return ast.SelectorExpr{ expr: ast.Ident{ name: name scope: p.scope } field_name: field gkind_field: fkind pos: pos scope: p.scope } } if p.peek_token(2).kind == .name && p.peek_token(3).kind == .lpar && !known_var { p.error_with_pos('the receiver of the method call must be an instantiated object, e.g. `foo.bar()`', p.tok.pos()) } // `Color.green` mut enum_name := p.check_name() enum_name_pos := p.prev_tok.pos() if mod != '' { enum_name = mod + '.' + enum_name } else { enum_name = p.imported_symbols[enum_name] or { p.prepend_mod(enum_name) } } p.check(.dot) val := p.check_name() p.expr_mod = '' return ast.EnumVal{ enum_name: enum_name val: val pos: enum_name_pos.extend(p.prev_tok.pos()) mod: mod } } else if language == .js && p.peek_tok.kind == .dot && p.peek_token(2).kind == .name { // JS. function call with more than 1 dot node = p.call_expr(language, mod) } else { if p.inside_in_array && ((lit0_is_capital && !known_var && language == .v) || (p.peek_tok.kind == .dot && p.peek_token(2).lit.len > 0 && p.peek_token(2).lit[0].is_capital()) || p.table.find_type_idx(p.mod + '.' + p.tok.lit) > 0) { type_pos := p.tok.pos() typ := p.parse_type() return ast.TypeNode{ typ: typ pos: type_pos } } ident := p.parse_ident(language) node = ident if p.inside_defer { if !p.defer_vars.any(it.name == ident.name && it.mod == ident.mod) && ident.name != 'err' { p.defer_vars << ident } } } p.expr_mod = '' return node } fn (mut p Parser) index_expr(left ast.Expr, is_gated bool) ast.IndexExpr { // left == `a` in `a[0]` start_pos := p.tok.pos() p.next() // [ mut has_low := true if p.tok.kind == .dotdot { has_low = false // [..end] p.next() mut high := ast.empty_expr() mut has_high := false if p.tok.kind != .rsbr { high = p.expr(0) has_high = true } pos_high := start_pos.extend(p.tok.pos()) p.check(.rsbr) mut or_kind_high := ast.OrKind.absent mut or_stmts_high := []ast.Stmt{} mut or_pos_high := token.Pos{} if !p.or_is_handled { // a[..end] or {...} if p.tok.kind == .key_orelse { was_inside_or_expr := p.inside_or_expr p.inside_or_expr = true or_pos_high = p.tok.pos() p.next() p.open_scope() or_stmts_high = p.parse_block_no_scope(false) or_pos_high = or_pos_high.extend(p.prev_tok.pos()) p.close_scope() p.inside_or_expr = was_inside_or_expr return ast.IndexExpr{ left: left pos: pos_high index: ast.RangeExpr{ low: ast.empty_expr() high: high has_high: has_high pos: pos_high is_gated: is_gated } or_expr: ast.OrExpr{ kind: .block stmts: or_stmts_high pos: or_pos_high } is_gated: is_gated } } // `a[start..end] ?` if p.tok.kind == .question { or_pos_high = p.tok.pos() or_kind_high = .propagate_option p.next() } } return ast.IndexExpr{ left: left pos: pos_high index: ast.RangeExpr{ low: ast.empty_expr() high: high has_high: has_high pos: pos_high is_gated: is_gated } or_expr: ast.OrExpr{ kind: or_kind_high stmts: or_stmts_high pos: or_pos_high } is_gated: is_gated } } expr := p.expr(0) // `[expr]` or `[expr..` mut has_high := false if p.tok.kind == .dotdot { // either [start..end] or [start..] p.next() mut high := ast.empty_expr() if p.tok.kind != .rsbr { has_high = true high = p.expr(0) } pos_low := start_pos.extend(p.tok.pos()) p.check(.rsbr) mut or_kind_low := ast.OrKind.absent mut or_stmts_low := []ast.Stmt{} mut or_pos_low := token.Pos{} if !p.or_is_handled { // a[start..end] or {...} if p.tok.kind == .key_orelse { was_inside_or_expr := p.inside_or_expr p.inside_or_expr = true or_pos_low = p.tok.pos() p.next() p.open_scope() or_stmts_low = p.parse_block_no_scope(false) or_pos_low = or_pos_low.extend(p.prev_tok.pos()) p.close_scope() p.inside_or_expr = was_inside_or_expr return ast.IndexExpr{ left: left pos: pos_low index: ast.RangeExpr{ low: expr high: high has_high: has_high has_low: has_low pos: pos_low is_gated: is_gated } or_expr: ast.OrExpr{ kind: .block stmts: or_stmts_low pos: or_pos_low } is_gated: is_gated } } // `a[start..end] ?` if p.tok.kind == .question { or_pos_low = p.tok.pos() or_kind_low = .propagate_option p.next() } } return ast.IndexExpr{ left: left pos: pos_low index: ast.RangeExpr{ low: expr high: high has_high: has_high has_low: has_low pos: pos_low is_gated: is_gated } or_expr: ast.OrExpr{ kind: or_kind_low stmts: or_stmts_low pos: or_pos_low } is_gated: is_gated } } // [expr] pos := start_pos.extend(p.tok.pos()) p.check(.rsbr) mut or_kind := ast.OrKind.absent mut or_stmts := []ast.Stmt{} mut or_pos := token.Pos{} if !p.or_is_handled { // a[i] or { ... } if p.tok.kind == .key_orelse { was_inside_or_expr := p.inside_or_expr p.inside_or_expr = true or_pos = p.tok.pos() p.next() p.open_scope() or_stmts = p.parse_block_no_scope(false) or_pos = or_pos.extend(p.prev_tok.pos()) p.close_scope() p.inside_or_expr = was_inside_or_expr return ast.IndexExpr{ left: left index: expr pos: pos or_expr: ast.OrExpr{ kind: .block stmts: or_stmts pos: or_pos } is_gated: is_gated } } // `a[i] ?` if p.tok.kind == .question { or_pos = p.tok.pos() or_kind = .propagate_option p.next() } } return ast.IndexExpr{ left: left index: expr pos: pos or_expr: ast.OrExpr{ kind: or_kind stmts: or_stmts pos: or_pos } is_gated: is_gated } } fn (mut p Parser) scope_register_it() { p.scope.register(ast.Var{ name: 'it' pos: p.tok.pos() is_used: true }) } fn (mut p Parser) scope_register_ab() { p.scope.register(ast.Var{ name: 'a' pos: p.tok.pos() is_used: true }) p.scope.register(ast.Var{ name: 'b' pos: p.tok.pos() is_used: true }) } fn (mut p Parser) dot_expr(left ast.Expr) ast.Expr { p.next() if p.tok.kind == .dollar { return p.comptime_selector(left) } is_generic_call := p.is_generic_call() name_pos := p.tok.pos() mut field_name := '' // check if the name is on the same line as the dot if (p.prev_tok.pos().line_nr == name_pos.line_nr) || p.tok.kind != .name { field_name = p.check_name() } else { p.name_error = true } is_filter := field_name in ['filter', 'map', 'any', 'all'] if is_filter || field_name == 'sort' { p.open_scope() } // ! in mutable methods if p.tok.kind == .not && p.peek_tok.kind == .lpar { p.next() } // Method call // TODO move to fn.v call_expr() mut concrete_types := []ast.Type{} mut concrete_list_pos := p.tok.pos() if is_generic_call { // `g.foo(10)` concrete_types = p.parse_concrete_types() concrete_list_pos = concrete_list_pos.extend(p.prev_tok.pos()) // In case of `foo()` // T is unwrapped and registered in the checker. has_generic := concrete_types.any(it.has_flag(.generic)) if !has_generic { // will be added in checker p.table.register_fn_concrete_types(field_name, concrete_types) } } if p.tok.kind == .lpar { p.next() args := p.call_args() p.check(.rpar) mut or_stmts := []ast.Stmt{} mut or_kind := ast.OrKind.absent mut or_pos := p.tok.pos() if p.tok.kind == .key_orelse { was_inside_or_expr := p.inside_or_expr p.inside_or_expr = true p.next() p.open_scope() p.scope.register(ast.Var{ name: 'err' typ: ast.error_type pos: p.tok.pos() is_used: true is_stack_obj: true }) or_kind = .block or_stmts = p.parse_block_no_scope(false) or_pos = or_pos.extend(p.prev_tok.pos()) p.close_scope() p.inside_or_expr = was_inside_or_expr } // `foo()?` if p.tok.kind in [.question, .not] { is_not := p.tok.kind == .not p.next() if p.inside_defer { p.error_with_pos('error propagation not allowed inside `defer` blocks', p.prev_tok.pos()) } or_kind = if is_not { .propagate_result } else { .propagate_option } } end_pos := p.prev_tok.pos() pos := name_pos.extend(end_pos) comments := p.eat_comments(same_line: true) mcall_expr := ast.CallExpr{ left: left name: field_name args: args name_pos: name_pos pos: pos is_method: true concrete_types: concrete_types concrete_list_pos: concrete_list_pos or_block: ast.OrExpr{ stmts: or_stmts kind: or_kind pos: or_pos } scope: p.scope comments: comments } if is_filter || field_name == 'sort' { p.close_scope() } return mcall_expr } mut is_mut := false mut mut_pos := token.Pos{} if p.inside_match || p.inside_if_expr || p.inside_for { match left { ast.Ident, ast.SelectorExpr { is_mut = left.is_mut mut_pos = left.mut_pos } else {} } } pos := if p.name_error { left.pos().extend(name_pos) } else { name_pos } sel_expr := ast.SelectorExpr{ expr: left field_name: field_name pos: pos is_mut: is_mut mut_pos: mut_pos scope: p.scope next_token: p.tok.kind } if is_filter { p.close_scope() } return sel_expr } fn (mut p Parser) parse_generic_types() ([]ast.Type, []string) { mut types := []ast.Type{} mut param_names := []string{} if p.tok.kind != .lt { return types, param_names } p.check(.lt) mut first_done := false mut count := 0 for p.tok.kind !in [.gt, .eof] { if first_done { p.check(.comma) } name := p.tok.lit if name.len > 0 && !name[0].is_capital() { p.error('generic parameter needs to be uppercase') } if name.len > 1 { p.error('generic parameter name needs to be exactly one char') } if !util.is_generic_type_name(p.tok.lit) { p.error('`$p.tok.lit` is a reserved name and cannot be used for generics') } if name in param_names { p.error('duplicated generic parameter `$name`') } if count > 8 { p.error('cannot have more than 9 generic parameters') } p.check(.name) param_names << name mut idx := p.table.find_type_idx(name) if idx == 0 { idx = p.table.register_sym(ast.TypeSymbol{ name: name cname: util.no_dots(name) mod: p.mod kind: .any is_pub: true }) } types << ast.new_type(idx).set_flag(.generic) first_done = true count++ } p.check(.gt) return types, param_names } fn (mut p Parser) parse_concrete_types() []ast.Type { mut types := []ast.Type{} if p.tok.kind != .lt { return types } p.next() // `<` mut first_done := false for p.tok.kind !in [.eof, .gt] { if first_done { p.check(.comma) } types << p.parse_type() first_done = true } p.check(.gt) // `>` return types } // is_generic_name returns true if the current token is a generic name. fn (p Parser) is_generic_name() bool { return p.tok.kind == .name && util.is_generic_type_name(p.tok.lit) } // `.green` // `pref.BuildMode.default_mode` fn (mut p Parser) enum_val() ast.EnumVal { start_pos := p.tok.pos() p.check(.dot) val := p.check_name() return ast.EnumVal{ val: val pos: start_pos.extend(p.prev_tok.pos()) } } fn (mut p Parser) filter_string_vet_errors(pos token.Pos) { if p.vet_errors.len == 0 { return } p.vet_errors = p.vet_errors.filter( (it.typ == .trailing_space && it.pos.line_nr - 1 >= pos.last_line) || (it.typ != .trailing_space && it.pos.line_nr - 1 > pos.last_line) || (it.typ == .space_indent && it.pos.line_nr - 1 <= pos.line_nr) || (it.typ != .space_indent && it.pos.line_nr - 1 < pos.line_nr)) } fn (mut p Parser) string_expr() ast.Expr { is_raw := p.tok.kind == .name && p.tok.lit == 'r' is_cstr := p.tok.kind == .name && p.tok.lit == 'c' if is_raw || is_cstr { p.next() } mut node := ast.empty_expr() val := p.tok.lit mut pos := p.tok.pos() pos.last_line = pos.line_nr + val.count('\n') if p.peek_tok.kind != .str_dollar { p.next() p.filter_string_vet_errors(pos) node = ast.StringLiteral{ val: val is_raw: is_raw language: if is_cstr { ast.Language.c } else { ast.Language.v } pos: pos } return node } mut exprs := []ast.Expr{} mut vals := []string{} mut has_fmts := []bool{} mut fwidths := []int{} mut precisions := []int{} mut visible_pluss := []bool{} mut fills := []bool{} mut fmts := []u8{} mut fposs := []token.Pos{} // Handle $ interpolation p.inside_str_interp = true for p.tok.kind == .string { vals << p.tok.lit p.next() if p.tok.kind != .str_dollar { break } p.next() exprs << p.expr(0) mut has_fmt := false mut fwidth := 0 mut fwidthneg := false // 987698 is a magic default value, unlikely to be present in user input. Note: 0 is valid precision mut precision := 987698 mut visible_plus := false mut fill := false mut fmt := `_` // placeholder if p.tok.kind == .colon { p.next() // ${num:-2d} if p.tok.kind == .minus { fwidthneg = true p.next() } else if p.tok.kind == .plus { visible_plus = true p.next() } // ${num:2d} if p.tok.kind == .number { fields := p.tok.lit.split('.') if fields[0].len > 0 && fields[0][0] == `0` { fill = true } fwidth = fields[0].int() if fwidthneg { fwidth = -fwidth } if fields.len > 1 { precision = fields[1].int() } p.next() } if p.tok.kind == .name { if p.tok.lit.len == 1 { fmt = p.tok.lit[0] has_fmt = true p.next() } else { return p.error('format specifier may only be one letter') } } } fwidths << fwidth has_fmts << has_fmt precisions << precision visible_pluss << visible_plus fmts << fmt fills << fill fposs << p.prev_tok.pos() } pos = pos.extend(p.prev_tok.pos()) p.filter_string_vet_errors(pos) node = ast.StringInterLiteral{ vals: vals exprs: exprs need_fmts: has_fmts fwidths: fwidths precisions: precisions pluss: visible_pluss fills: fills fmts: fmts fmt_poss: fposs pos: pos } // need_fmts: prelimery - until checker finds out if really needed p.inside_str_interp = false return node } fn (mut p Parser) parse_number_literal() ast.Expr { mut pos := p.tok.pos() is_neg := p.tok.kind == .minus if is_neg { p.next() pos = pos.extend(p.tok.pos()) } lit := p.tok.lit full_lit := if is_neg { '-' + lit } else { lit } mut node := ast.empty_expr() if lit.index_any('.eE') >= 0 && lit[..2] !in ['0x', '0X', '0o', '0O', '0b', '0B'] { node = ast.FloatLiteral{ val: full_lit pos: pos } } else { node = ast.IntegerLiteral{ val: full_lit pos: pos } } p.next() return node } fn (mut p Parser) module_decl() ast.Module { mut module_attrs := []ast.Attr{} mut attrs_pos := p.tok.pos() for p.tok.kind == .lsbr { p.attributes() } module_attrs << p.attrs mut name := 'main' mut module_pos := token.Pos{} mut name_pos := token.Pos{} mut mod_node := ast.Module{} is_skipped := p.tok.kind != .key_module if is_skipped { // the attributes were for something else != module, like a struct/fn/type etc. module_attrs = [] } else { p.attrs = [] module_pos = p.tok.pos() p.next() name_pos = p.tok.pos() name = p.check_name() mod_node = ast.Module{ pos: module_pos } if module_pos.line_nr != name_pos.line_nr { p.error_with_pos('`module` and `$name` must be at same line', name_pos) return mod_node } // Note: this shouldn't be reassigned into name_pos // as it creates a wrong position when extended // to module_pos n_pos := p.tok.pos() if module_pos.line_nr == n_pos.line_nr && p.tok.kind != .comment && p.tok.kind != .eof { if p.tok.kind == .name { p.error_with_pos('`module $name`, you can only declare one module, unexpected `$p.tok.lit`', n_pos) return mod_node } else { p.error_with_pos('`module $name`, unexpected `$p.tok.kind` after module name', n_pos) return mod_node } } module_pos = attrs_pos.extend(name_pos) } full_name := util.qualify_module(p.pref, name, p.file_name) p.mod = full_name p.builtin_mod = p.mod == 'builtin' mod_node = ast.Module{ name: full_name short_name: name attrs: module_attrs is_skipped: is_skipped pos: module_pos name_pos: name_pos } if !is_skipped { for ma in module_attrs { match ma.name { 'deprecated' { // [deprecated: 'use a replacement'] p.table.mark_module_as_deprecated(p.mod, ma.arg) } 'deprecated_after' { // [deprecated_after: '2027-12-30'] p.table.mark_module_as_deprecated_after(p.mod, ma.arg) } 'manualfree' { p.is_manualfree = true } 'generated' { p.is_generated = true } 'has_globals' { if p.inside_vlib_file { p.has_globals = true } else { p.error_with_pos('[has_globals] is allowed only in .v files of `vlib` modules', ma.pos) } } 'translated' { p.is_translated = true } else { p.error_with_pos('unknown module attribute `[$ma.name]`', ma.pos) return mod_node } } } } return mod_node } fn (mut p Parser) import_stmt() ast.Import { import_pos := p.tok.pos() p.check(.key_import) mut pos := p.tok.pos() mut import_node := ast.Import{ pos: import_pos.extend(pos) } if p.tok.kind == .lpar { p.error_with_pos('`import()` has been deprecated, use `import x` instead', pos) return import_node } mut mod_name_arr := []string{} mod_name_arr << p.check_name() if import_pos.line_nr != pos.line_nr { p.error_with_pos('`import` statements must be a single line', pos) return import_node } mut mod_alias := mod_name_arr[0] import_node = ast.Import{ pos: import_pos.extend(pos) mod_pos: pos alias_pos: pos } for p.tok.kind == .dot { p.next() submod_pos := p.tok.pos() if p.tok.kind != .name { p.error_with_pos('module syntax error, please use `x.y.z`', submod_pos) return import_node } if import_pos.line_nr != submod_pos.line_nr { p.error_with_pos('`import` and `submodule` must be at same line', submod_pos) return import_node } submod_name := p.check_name() mod_name_arr << submod_name mod_alias = submod_name pos = pos.extend(submod_pos) import_node = ast.Import{ pos: import_pos.extend(pos) mod_pos: pos alias_pos: submod_pos mod: util.qualify_import(p.pref, mod_name_arr.join('.'), p.file_name) alias: mod_alias } } if mod_name_arr.len == 1 { import_node = ast.Import{ pos: import_node.pos mod_pos: import_node.mod_pos alias_pos: import_node.alias_pos mod: util.qualify_import(p.pref, mod_name_arr[0], p.file_name) alias: mod_alias } } mod_name := import_node.mod if p.tok.kind == .key_as { p.next() alias_pos := p.tok.pos() mod_alias = p.check_name() if mod_alias == mod_name_arr.last() { p.error_with_pos('import alias `$mod_name as $mod_alias` is redundant', p.prev_tok.pos()) return import_node } import_node = ast.Import{ pos: import_node.pos.extend(alias_pos) mod_pos: import_node.mod_pos alias_pos: alias_pos mod: import_node.mod alias: mod_alias } } if p.tok.kind == .lcbr { // import module { fn1, Type2 } syntax mut initial_syms_pos := p.tok.pos() p.import_syms(mut import_node) initial_syms_pos = initial_syms_pos.extend(p.tok.pos()) import_node = ast.Import{ ...import_node syms_pos: initial_syms_pos pos: import_node.pos.extend(initial_syms_pos) } p.register_used_import(mod_alias) // no `unused import` msg for parent } pos_t := p.tok.pos() if import_pos.line_nr == pos_t.line_nr { if p.tok.kind !in [.lcbr, .eof, .comment] { p.error_with_pos('cannot import multiple modules at a time', pos_t) return import_node } } import_node.comments = p.eat_comments(same_line: true) import_node.next_comments = p.eat_comments(follow_up: true) p.imports[mod_alias] = mod_name // if mod_name !in p.table.imports { p.table.imports << mod_name p.ast_imports << import_node // } return import_node } // import_syms parses the inner part of `import module { submod1, submod2 }` fn (mut p Parser) import_syms(mut parent ast.Import) { p.next() pos_t := p.tok.pos() if p.tok.kind == .rcbr { // closed too early p.error_with_pos('empty `$parent.mod` import set, remove `{}`', pos_t) return } if p.tok.kind != .name { // not a valid inner name p.error_with_pos('import syntax error, please specify a valid fn or type name', pos_t) return } for p.tok.kind == .name { pos := p.tok.pos() alias := p.check_name() p.imported_symbols[alias] = parent.mod + '.' + alias // so we can work with this in fmt+checker parent.syms << ast.ImportSymbol{ pos: pos name: alias } if p.tok.kind == .comma { // go again if more than one p.next() continue } if p.tok.kind == .rcbr { // finish if closing `}` is seen break } } if p.tok.kind != .rcbr { p.error_with_pos('import syntax error, no closing `}`', p.tok.pos()) return } p.next() } fn (mut p Parser) const_decl() ast.ConstDecl { p.top_level_statement_start() mut attrs := []ast.Attr{} if p.attrs.len > 0 { attrs = p.attrs p.attrs = [] } mut is_markused := false for ga in attrs { match ga.name { 'markused' { is_markused = true } else {} } } start_pos := p.tok.pos() is_pub := p.tok.kind == .key_pub if is_pub { p.next() } const_pos := p.tok.pos() p.check(.key_const) is_block := p.tok.kind == .lpar if is_block { p.next() // ( } mut fields := []ast.ConstField{} mut comments := []ast.Comment{} mut end_comments := []ast.Comment{} for { comments = p.eat_comments() if is_block && p.tok.kind == .eof { p.error('unexpected eof, expecting ´)´') return ast.ConstDecl{} } if p.tok.kind == .rpar { break } pos := p.tok.pos() name := p.check_name() end_comments << p.eat_comments() if util.contains_capital(name) { p.warn_with_pos('const names cannot contain uppercase letters, use snake_case instead', pos) } full_name := p.prepend_mod(name) p.check(.assign) end_comments << p.eat_comments() if p.tok.kind == .key_fn { p.error('const initializer fn literal is not a constant') return ast.ConstDecl{} } if p.tok.kind == .eof { p.error('unexpected eof, expecting an expression') return ast.ConstDecl{} } expr := p.expr(0) field := ast.ConstField{ name: full_name mod: p.mod is_pub: is_pub expr: expr pos: pos.extend(expr.pos()) comments: comments end_comments: end_comments is_markused: is_markused } fields << field p.table.global_scope.register(field) comments = [] if !is_block { break } } p.top_level_statement_end() if is_block { p.check(.rpar) } else { comments << p.eat_comments(same_line: true) } return ast.ConstDecl{ pos: start_pos.extend_with_last_line(const_pos, p.prev_tok.line_nr) fields: fields is_pub: is_pub end_comments: comments is_block: is_block attrs: attrs } } fn (mut p Parser) return_stmt() ast.Return { first_pos := p.tok.pos() p.next() // no return mut comments := p.eat_comments() if p.tok.kind == .rcbr || (p.tok.kind == .name && p.peek_tok.kind == .colon) { return ast.Return{ comments: comments pos: first_pos } } // return exprs exprs, comments2 := p.expr_list() comments << comments2 end_pos := exprs.last().pos() return ast.Return{ exprs: exprs comments: comments pos: first_pos.extend(end_pos) } } // left hand side of `=` or `:=` in `a,b,c := 1,2,3` fn (mut p Parser) global_decl() ast.GlobalDecl { mut attrs := []ast.Attr{} if p.attrs.len > 0 { attrs = p.attrs p.attrs = [] } mut is_markused := false for ga in attrs { match ga.name { 'markused' { is_markused = true } else {} } } if !p.has_globals && !p.pref.enable_globals && !p.pref.is_fmt && !p.pref.translated && !p.is_translated && !p.pref.is_livemain && !p.pref.building_v && !p.builtin_mod { p.error('use `v -enable-globals ...` to enable globals') return ast.GlobalDecl{} } start_pos := p.tok.pos() p.check(.key_global) is_block := p.tok.kind == .lpar if is_block { p.next() // ( } mut fields := []ast.GlobalField{} mut comments := []ast.Comment{} for { comments = p.eat_comments() is_volatile := p.tok.kind == .key_volatile if is_volatile { p.next() } if is_block && p.tok.kind == .eof { p.error('unexpected eof, expecting `)`') return ast.GlobalDecl{} } if p.tok.kind == .rpar { break } pos := p.tok.pos() name := p.check_name() has_expr := p.tok.kind == .assign mut expr := ast.empty_expr() mut typ := ast.void_type mut typ_pos := token.Pos{} if has_expr { p.next() // = expr = p.expr(0) match mut expr { ast.CastExpr, ast.StructInit, ast.ArrayInit, ast.ChanInit { typ = expr.typ } ast.BoolLiteral, ast.IsRefType { typ = ast.bool_type } ast.CharLiteral { typ = ast.char_type } ast.FloatLiteral { typ = ast.f64_type } ast.IntegerLiteral, ast.SizeOf { typ = ast.int_type } ast.StringLiteral, ast.StringInterLiteral { typ = ast.string_type } else { // type will be deduced by checker } } } else { typ_pos = p.tok.pos() typ = p.parse_type() } field := ast.GlobalField{ name: name has_expr: has_expr expr: expr pos: pos typ_pos: typ_pos typ: typ comments: comments is_markused: is_markused is_volatile: is_volatile } fields << field p.table.global_scope.register(field) comments = [] if !is_block { break } } if is_block { p.check(.rpar) } return ast.GlobalDecl{ pos: start_pos.extend(p.prev_tok.pos()) mod: p.mod fields: fields end_comments: comments is_block: is_block attrs: attrs } } fn (mut p Parser) enum_decl() ast.EnumDecl { p.top_level_statement_start() is_pub := p.tok.kind == .key_pub start_pos := p.tok.pos() if is_pub { p.next() } p.check(.key_enum) end_pos := p.tok.pos() enum_name := p.check_name() if enum_name.len == 1 { p.error_with_pos('single letter capital names are reserved for generic template types.', end_pos) return ast.EnumDecl{} } if enum_name in p.imported_symbols { p.error_with_pos('cannot register enum `$enum_name`, this type was already imported', end_pos) return ast.EnumDecl{} } name := p.prepend_mod(enum_name) p.check(.lcbr) enum_decl_comments := p.eat_comments() mut vals := []string{} // mut default_exprs := []ast.Expr{} mut fields := []ast.EnumField{} mut uses_exprs := false for p.tok.kind != .eof && p.tok.kind != .rcbr { pos := p.tok.pos() val := p.check_name() vals << val mut expr := ast.empty_expr() mut has_expr := false // p.warn('enum val $val') if p.tok.kind == .assign { p.next() expr = p.expr(0) has_expr = true uses_exprs = true } fields << ast.EnumField{ name: val pos: pos expr: expr has_expr: has_expr comments: p.eat_comments(same_line: true) next_comments: p.eat_comments() } } p.top_level_statement_end() p.check(.rcbr) is_flag := p.attrs.contains('flag') is_multi_allowed := p.attrs.contains('_allow_multiple_values') if is_flag { if fields.len > 32 { p.error('when an enum is used as bit field, it must have a max of 32 fields') return ast.EnumDecl{} } for f in fields { if f.has_expr { p.error_with_pos('when an enum is used as a bit field, you can not assign custom values', f.pos) return ast.EnumDecl{} } } pubfn := if p.mod == 'main' { 'fn' } else { 'pub fn' } p.codegen(' // [inline] $pubfn ( e &$enum_name) is_empty() bool { return int(*e) == 0 } [inline] $pubfn ( e &$enum_name) has(flag $enum_name) bool { return (int(*e) & (int(flag))) != 0 } [inline] $pubfn ( e &$enum_name) all(flag $enum_name) bool { return (int(*e) & (int(flag))) == int(flag) } [inline] $pubfn (mut e $enum_name) set(flag $enum_name) { unsafe{ *e = ${enum_name}(int(*e) | (int(flag))) } } [inline] $pubfn (mut e $enum_name) clear(flag $enum_name) { unsafe{ *e = ${enum_name}(int(*e) & ~(int(flag))) } } [inline] $pubfn (mut e $enum_name) toggle(flag $enum_name) { unsafe{ *e = ${enum_name}(int(*e) ^ (int(flag))) } } // ') } idx := p.table.register_sym(ast.TypeSymbol{ kind: .enum_ name: name cname: util.no_dots(name) mod: p.mod info: ast.Enum{ vals: vals is_flag: is_flag is_multi_allowed: is_multi_allowed uses_exprs: uses_exprs } is_pub: is_pub }) if idx == -1 { p.error_with_pos('cannot register enum `$name`, another type with this name exists', end_pos) } enum_decl := ast.EnumDecl{ name: name is_pub: is_pub is_flag: is_flag is_multi_allowed: is_multi_allowed fields: fields pos: start_pos.extend_with_last_line(end_pos, p.prev_tok.line_nr) attrs: p.attrs comments: enum_decl_comments } p.table.register_enum_decl(enum_decl) return enum_decl } fn (mut p Parser) type_decl() ast.TypeDecl { start_pos := p.tok.pos() is_pub := p.tok.kind == .key_pub if is_pub { p.next() } p.check(.key_type) end_pos := p.tok.pos() decl_pos := start_pos.extend(end_pos) name_pos := p.tok.pos() name := p.check_name() if name.len == 1 && name[0].is_capital() { p.error_with_pos('single letter capital names are reserved for generic template types.', decl_pos) return ast.FnTypeDecl{} } if name in p.imported_symbols { p.error_with_pos('cannot register alias `$name`, this type was already imported', end_pos) return ast.AliasTypeDecl{} } mut sum_variants := []ast.TypeNode{} generic_types, _ := p.parse_generic_types() decl_pos_with_generics := decl_pos.extend(p.prev_tok.pos()) p.check(.assign) mut type_pos := p.tok.pos() mut comments := []ast.Comment{} if p.tok.kind == .key_fn { // function type: `type mycallback = fn(string, int)` fn_name := p.prepend_mod(name) fn_type := p.parse_fn_type(fn_name) p.table.sym(fn_type).is_pub = is_pub type_pos = type_pos.extend(p.tok.pos()) comments = p.eat_comments(same_line: true) attrs := p.attrs p.attrs = [] return ast.FnTypeDecl{ name: fn_name is_pub: is_pub typ: fn_type pos: decl_pos type_pos: type_pos comments: comments attrs: attrs } } sum_variants << p.parse_sum_type_variants() // type SumType = A | B | c if sum_variants.len > 1 { for variant in sum_variants { variant_sym := p.table.sym(variant.typ) // TODO: implement this check for error too if variant_sym.kind == .none_ { p.error_with_pos('named sum type cannot have none as its variant', variant.pos) return ast.AliasTypeDecl{} } } variant_types := sum_variants.map(it.typ) prepend_mod_name := p.prepend_mod(name) typ := p.table.register_sym(ast.TypeSymbol{ kind: .sum_type name: prepend_mod_name cname: util.no_dots(prepend_mod_name) mod: p.mod info: ast.SumType{ variants: variant_types is_generic: generic_types.len > 0 generic_types: generic_types } is_pub: is_pub }) if typ == ast.invalid_type_idx { p.error_with_pos('cannot register sum type `$name`, another type with this name exists', name_pos) return ast.SumTypeDecl{} } comments = p.eat_comments(same_line: true) return ast.SumTypeDecl{ name: name typ: typ is_pub: is_pub variants: sum_variants generic_types: generic_types attrs: p.attrs pos: decl_pos comments: comments } } // type MyType = int if generic_types.len > 0 { p.error_with_pos('generic type aliases are not yet implemented', decl_pos_with_generics) return ast.AliasTypeDecl{} } // sum_variants will have only one element parent_type := sum_variants[0].typ parent_sym := p.table.sym(parent_type) pidx := parent_type.idx() p.check_for_impure_v(parent_sym.language, decl_pos) prepend_mod_name := p.prepend_mod(name) idx := p.table.register_sym(ast.TypeSymbol{ kind: .alias name: prepend_mod_name cname: util.no_dots(prepend_mod_name) mod: p.mod parent_idx: pidx info: ast.Alias{ parent_type: parent_type language: parent_sym.language } is_pub: is_pub }) type_end_pos := p.prev_tok.pos() if idx == ast.invalid_type_idx { p.error_with_pos('cannot register alias `$name`, another type with this name exists', name_pos) return ast.AliasTypeDecl{} } if idx == pidx { type_alias_pos := sum_variants[0].pos p.error_with_pos('a type alias can not refer to itself: $name', decl_pos.extend(type_alias_pos)) return ast.AliasTypeDecl{} } comments = p.eat_comments(same_line: true) return ast.AliasTypeDecl{ name: name is_pub: is_pub parent_type: parent_type type_pos: type_pos.extend(type_end_pos) pos: decl_pos comments: comments } } fn (mut p Parser) assoc() ast.Assoc { var_name := p.check_name() pos := p.tok.pos() mut v := p.scope.find_var(var_name) or { p.error('unknown variable `$var_name`') return ast.Assoc{ scope: 0 } } v.is_used = true mut fields := []string{} mut vals := []ast.Expr{} p.check(.pipe) for p.tok.kind != .eof { fields << p.check_name() p.check(.colon) expr := p.expr(0) vals << expr if p.tok.kind == .comma { p.next() } if p.tok.kind == .rcbr { break } } return ast.Assoc{ var_name: var_name fields: fields exprs: vals pos: pos scope: p.scope } } fn (p &Parser) new_true_expr() ast.Expr { return ast.BoolLiteral{ val: true pos: p.tok.pos() } } [noreturn] fn verror(s string) { util.verror('parser error', s) } fn (mut p Parser) top_level_statement_start() { if p.comments_mode == .toplevel_comments { p.scanner.set_is_inside_toplevel_statement(true) p.rewind_scanner_to_current_token_in_new_mode() $if debugscanner ? { eprintln('>> p.top_level_statement_start | tidx:${p.tok.tidx:-5} | p.tok.kind: ${p.tok.kind:-10} | p.tok.lit: $p.tok.lit $p.peek_tok.lit ${p.peek_token(2).lit} ${p.peek_token(3).lit} ...') } } } fn (mut p Parser) top_level_statement_end() { if p.comments_mode == .toplevel_comments { p.scanner.set_is_inside_toplevel_statement(false) p.rewind_scanner_to_current_token_in_new_mode() $if debugscanner ? { eprintln('>> p.top_level_statement_end | tidx:${p.tok.tidx:-5} | p.tok.kind: ${p.tok.kind:-10} | p.tok.lit: $p.tok.lit $p.peek_tok.lit ${p.peek_token(2).lit} ${p.peek_token(3).lit} ...') } } } fn (mut p Parser) rewind_scanner_to_current_token_in_new_mode() { // Go back and rescan some tokens, ensuring that the parser's // lookahead buffer p.peek_tok .. p.peek_token(3), will now contain // the correct tokens (possible comments), for the new mode // This refilling of the lookahead buffer is needed for the // .toplevel_comments parsing mode. tidx := p.tok.tidx p.scanner.set_current_tidx(tidx - 5) no_token := token.Token{} p.prev_tok = no_token p.tok = no_token p.peek_tok = no_token // requires 2 calls p.next() or check p.tok.kind != token.Kind.unknown p.next() for { p.next() // eprintln('rewinding to ${p.tok.tidx:5} | goal: ${tidx:5}') if tidx == p.tok.tidx { break } } } // returns true if `varname` is known pub fn (mut p Parser) mark_var_as_used(varname string) bool { if mut obj := p.scope.find(varname) { match mut obj { ast.Var { obj.is_used = true return true } ast.GlobalField { // obj.is_used = true return true } // ast.ConstField { // return true //} else {} } } return false } fn (mut p Parser) unsafe_stmt() ast.Stmt { mut pos := p.tok.pos() p.next() if p.tok.kind != .lcbr { return p.error_with_pos('please use `unsafe {`', p.tok.pos()) } p.next() if p.inside_unsafe { return p.error_with_pos('already inside `unsafe` block', pos) } if p.tok.kind == .rcbr { // `unsafe {}` pos.update_last_line(p.tok.line_nr) p.next() return ast.Block{ is_unsafe: true pos: pos } } p.inside_unsafe = true p.open_scope() // needed in case of `unsafe {stmt}` defer { p.inside_unsafe = false p.close_scope() } stmt := p.stmt(false) if p.tok.kind == .rcbr { if stmt is ast.ExprStmt { // `unsafe {expr}` if stmt.expr.is_expr() { p.next() pos.update_last_line(p.prev_tok.line_nr) ue := ast.UnsafeExpr{ expr: stmt.expr pos: pos } // parse e.g. `unsafe {expr}.foo()` expr := p.expr_with_left(ue, 0, p.is_stmt_ident) return ast.ExprStmt{ expr: expr pos: pos } } } } // unsafe {stmts} mut stmts := [stmt] for p.tok.kind != .rcbr { stmts << p.stmt(false) } p.next() pos.update_last_line(p.tok.line_nr) return ast.Block{ stmts: stmts is_unsafe: true pos: pos } } fn (mut p Parser) trace(fbase string, message string) { if p.file_base == fbase { println('> p.trace | ${fbase:-10s} | $message') } }