// Copyright 2019 The CC Authors. All rights reserved. // Use of this source code is governed by a BSD-style // license that can be found in the LICENSE file. package cc // import "modernc.org/cc/v3" import ( "bytes" "fmt" gotoken "go/token" "math" "os" "path/filepath" "strconv" "strings" "sync" "time" "unicode/utf8" "modernc.org/token" ) const ( maxIncludeLevel = 200 // gcc, std is at least 15. ) var ( _ tokenReader = (*cpp)(nil) _ tokenWriter = (*cpp)(nil) idCOUNTER = dict.sid("__COUNTER__") idCxLimitedRange = dict.sid("CX_LIMITED_RANGE") idDATE = dict.sid("__DATE__") idDefault = dict.sid("DEFAULT") idDefined = dict.sid("defined") idEmptyString = dict.sid(`""`) idFILE = dict.sid("__FILE__") idFPContract = dict.sid("FP_CONTRACT") idFdZero = dict.sid("FD_ZERO") idFenvAccess = dict.sid("FENV_ACCESS") idGNUC = dict.sid("__GNUC__") idHasIncludeImpl = dict.sid("__has_include_impl") idIntMaxWidth = dict.sid("__INTMAX_WIDTH__") idL = dict.sid("L") idLINE = dict.sid("__LINE__") idNL = dict.sid("\n") idOff = dict.sid("OFF") idOn = dict.sid("ON") idOne = dict.sid("1") idPragmaSTDC = dict.sid("__pragma_stdc") idSTDC = dict.sid("STDC") idTIME = dict.sid("__TIME__") idTclDefaultDoubleRounding = dict.sid("TCL_DEFAULT_DOUBLE_ROUNDING") idTclIeeeDoubleRounding = dict.sid("TCL_IEEE_DOUBLE_ROUNDING") idVaArgs = dict.sid("__VA_ARGS__") idZero = dict.sid("0") cppTokensPool = sync.Pool{New: func() interface{} { r := []cppToken{}; return &r }} protectedMacros = hideSet{ // [0], 6.10.8, 4 dict.sid("__STDC_HOSTED__"): {}, dict.sid("__STDC_IEC_559_COMPLEX__"): {}, dict.sid("__STDC_IEC_559__"): {}, dict.sid("__STDC_ISO_10646__"): {}, dict.sid("__STDC_MB_MIGHT_NEQ_WC__"): {}, dict.sid("__STDC_VERSION__"): {}, dict.sid("__STDC__"): {}, idCOUNTER: {}, idDATE: {}, idFILE: {}, idLINE: {}, idTIME: {}, } ) type tokenReader interface { read() (cppToken, bool) unget(cppToken) ungets([]cppToken) } type tokenWriter interface { write(cppToken) writes([]cppToken) } // token4 is produced by translation phase 4. type token4 struct { file *tokenFile //TODO sort fields token3 } func (t *token4) Position() (r token.Position) { if t.pos != 0 && t.file != nil { r = t.file.PositionFor(token.Pos(t.pos), true) } return r } type hideSet map[StringID]struct{} type cppToken struct { token4 hs hideSet } func (t *cppToken) has(nm StringID) bool { _, ok := t.hs[nm]; return ok } type cppWriter struct { toks []cppToken } func (w *cppWriter) write(tok cppToken) { w.toks = append(w.toks, tok) } func (w *cppWriter) writes(toks []cppToken) { w.toks = append(w.toks, toks...) } type ungetBuf []cppToken func (u *ungetBuf) unget(t cppToken) { *u = append(*u, t) } func (u *ungetBuf) read() (t cppToken) { s := *u n := len(s) - 1 t = s[n] *u = s[:n] return t } func (u *ungetBuf) ungets(toks []cppToken) { s := *u for i := len(toks) - 1; i >= 0; i-- { s = append(s, toks[i]) } *u = s } func cppToksStr(toks []cppToken, sep string) string { var b strings.Builder for i, v := range toks { if i != 0 { b.WriteString(sep) } b.WriteString(v.String()) } return b.String() } func cppToksStr2(toks [][]cppToken) string { panic(todo("")) var a []string for _, v := range toks { a = append(a, fmt.Sprintf("%q", cppToksStr(v, "|"))) } return fmt.Sprint(a) } type cppReader struct { buf []cppToken ungetBuf } func (r *cppReader) read() (tok cppToken, ok bool) { if len(r.ungetBuf) != 0 { return r.ungetBuf.read(), true } if len(r.buf) == 0 { return tok, false } tok = r.buf[0] r.buf = r.buf[1:] return tok, true } type cppScanner []cppToken func (s *cppScanner) peek() (r cppToken) { r.char = -1 if len(*s) == 0 { return r } return (*s)[0] } func (s *cppScanner) next() (r cppToken) { r.char = -1 if len(*s) == 0 { return r } *s = (*s)[1:] return s.peek() } func (s *cppScanner) Pos() token.Pos { if len(*s) == 0 { return 0 } return (*s)[0].Pos() } // Macro represents a preprocessor macro definition. type Macro struct { fp []StringID repl []token3 repl2 []Token name token4 pos int32 isFnLike bool namedVariadic bool // foo..., note no comma before ellipsis. variadic bool } // Position reports the position of the macro definition. func (m *Macro) Position() token.Position { if m.pos != 0 && m.name.file != nil { return m.name.file.PositionFor(token.Pos(m.pos), true) } return token.Position{} } // Parameters return the list of function-like macro parameters. func (m *Macro) Parameters() []StringID { return m.fp } // ReplacementTokens return the list of tokens m is replaced with. Tokens in // the returned list have only the Rune and Value fields valid. func (m *Macro) ReplacementTokens() []Token { if m.repl2 != nil { return m.repl2 } m.repl2 = make([]Token, len(m.repl)) for i, v := range m.repl { m.repl2[i] = Token{Rune: v.char, Value: v.value, Src: v.src} } return m.repl2 } // IsFnLike reports whether m is a function-like macro. func (m *Macro) IsFnLike() bool { return m.isFnLike } func (m *Macro) isNamedVariadicParam(nm StringID) bool { return m.namedVariadic && nm == m.fp[len(m.fp)-1] } func (m *Macro) param2(varArgs []cppToken, ap [][]cppToken, nm StringID, out *[]cppToken, argIndex *int) bool { *out = nil if nm == idVaArgs || m.isNamedVariadicParam(nm) { if !m.variadic { return false } *out = append([]cppToken(nil), varArgs...) return true } for i, v := range m.fp { if v == nm { if i < len(ap) { a := ap[i] for len(a) != 0 && a[0].char == ' ' { a = a[1:] } *out = a } if argIndex != nil { *argIndex = i } return true } } return false } func (m *Macro) param(varArgs []cppToken, ap [][]cppToken, nm StringID, out *[]cppToken) bool { // trc("select (A) varArgs %q, ap %v, nm %q, out %q", cppToksStr(varArgs, "|"), cppToksStr2(ap), nm, cppToksStr(*out, "|")) // defer func() { // trc("select (A) varArgs %q, ap %v, nm %q, out %q", cppToksStr(varArgs, "|"), cppToksStr2(ap), nm, cppToksStr(*out, "|")) // }() return m.param2(varArgs, ap, nm, out, nil) } // --------------------------------------------------------------- Preprocessor type cpp struct { counter int counterMacro Macro ctx *context dateMacro Macro file *tokenFile fileMacro Macro in chan []token3 inBuf []token3 includeLevel int lineMacro Macro macroStack map[StringID][]*Macro macros map[StringID]*Macro out chan *[]token4 outBuf *[]token4 pragmaOpBuf []token4 rq chan struct{} timeMacro Macro ungetBuf last rune intmaxChecked bool nonFirstRead bool seenEOF bool inPragmaOp bool } func newCPP(ctx *context) *cpp { b := token4Pool.Get().(*[]token4) *b = (*b)[:0] r := &cpp{ ctx: ctx, macroStack: map[StringID][]*Macro{}, macros: map[StringID]*Macro{}, outBuf: b, } r.counterMacro = Macro{repl: []token3{{char: PPNUMBER}}} r.dateMacro = Macro{repl: []token3{{char: STRINGLITERAL}}} r.timeMacro = Macro{repl: []token3{{char: STRINGLITERAL}}} r.fileMacro = Macro{repl: []token3{{char: STRINGLITERAL}}} r.lineMacro = Macro{repl: []token3{{char: PPNUMBER}}} r.macros = map[StringID]*Macro{ idCOUNTER: &r.counterMacro, idDATE: &r.dateMacro, idFILE: &r.fileMacro, idLINE: &r.lineMacro, idTIME: &r.timeMacro, } t := time.Now() // This macro expands to a string constant that describes the date on which the // preprocessor is being run. The string constant contains eleven characters // and looks like "Feb 12 1996". If the day of the month is less than 10, it is // padded with a space on the left. r.dateMacro.repl[0].value = dict.sid(t.Format("\"Jan _2 2006\"")) // This macro expands to a string constant that describes the time at which the // preprocessor is being run. The string constant contains eight characters and // looks like "23:59:01". r.timeMacro.repl[0].value = dict.sid(t.Format("\"15:04:05\"")) return r } func (c *cpp) cppToks(toks []token3) (r []cppToken) { r = make([]cppToken, len(toks)) for i, v := range toks { r[i].token4.token3 = v r[i].token4.file = c.file } return r } func (c *cpp) err(n node, msg string, args ...interface{}) (stop bool) { var position token.Position switch x := n.(type) { case nil: case token4: position = x.Position() default: if p := n.Pos(); p.IsValid() { position = c.file.PositionFor(p, true) } } return c.ctx.err(position, msg, args...) } func (c *cpp) read() (cppToken, bool) { if len(c.ungetBuf) != 0 { return c.ungetBuf.read(), true } if len(c.inBuf) == 0 { if c.seenEOF { return cppToken{}, false } if c.nonFirstRead { c.rq <- struct{}{} } c.nonFirstRead = true var ok bool if c.inBuf, ok = <-c.in; !ok { c.seenEOF = true return cppToken{}, false } } tok := c.inBuf[0] c.inBuf = c.inBuf[1:] return cppToken{token4{token3: tok, file: c.file}, nil}, true } func (c *cpp) write(tok cppToken) { if tok.char == ' ' && c.last == ' ' { return } if c.ctx.cfg.PreprocessOnly { switch { case //TODO cover ALL the bad combinations c.last == '+' && tok.char == '+', c.last == '+' && tok.char == INC, c.last == '-' && tok.char == '-', c.last == '-' && tok.char == DEC, c.last == IDENTIFIER && tok.char == IDENTIFIER, c.last == PPNUMBER && tok.char == '+', //TODO not when ends in a digit c.last == PPNUMBER && tok.char == '-': //TODO not when ends in a digit sp := tok sp.char = ' ' sp.value = idSpace *c.outBuf = append(*c.outBuf, sp.token4) } } //dbg("%T.write %q", c, tok) c.last = tok.char switch { case c.inPragmaOp: out: switch tok.char { case ')': c.inPragmaOp = false b := c.pragmaOpBuf if len(b) == 0 || b[0].char != '(' { c.err(b[0], "expected (") break } var a []string for _, v := range b[1:] { if v.char != STRINGLITERAL { c.err(v, "expected string literal") break out } a = append(a, v.String()) } if len(a) == 0 { break } for i, v := range a { // [0], 6.10.9, 1 if v[0] == 'L' { v = v[1:] } v = v[1 : len(v)-1] v = strings.ReplaceAll(v, `\"`, `"`) a[i] = "#pragma " + strings.ReplaceAll(v, `\\`, `\`) + "\n" } src := strings.Join(a, "") s := newScanner0(c.ctx, strings.NewReader(src), tokenNewFile("", len(src)), 4096) if ppf := s.translationPhase3(); ppf != nil { ppf.translationPhase4(c) } default: c.pragmaOpBuf = append(c.pragmaOpBuf, tok.token4) } default: switch { case tok.char == '\n': *c.outBuf = append(*c.outBuf, tok.token4) c.out <- c.outBuf b := token4Pool.Get().(*[]token4) *b = (*b)[:0] c.outBuf = b case tok.char == IDENTIFIER && tok.value == idPragmaOp: if len(*c.outBuf) != 0 { tok.char = '\n' tok.value = 0 *c.outBuf = append(*c.outBuf, tok.token4) c.out <- c.outBuf b := token4Pool.Get().(*[]token4) *b = (*b)[:0] c.outBuf = b } c.inPragmaOp = true c.pragmaOpBuf = c.pragmaOpBuf[:0] default: *c.outBuf = append(*c.outBuf, tok.token4) } } } func ltrim4(toks []token4) []token4 { for len(toks) != 0 && toks[0].char == ' ' { toks = toks[1:] } return toks } func (c *cpp) writes(toks []cppToken) { for _, v := range toks { c.write(v) } } // [1]pg 1. // // expand(TS) /* recur, substitute, pushback, rescan */ // { // if TS is {} then // // ---------------------------------------------------------- A // return {}; // // else if TS is T^HS • TS’ and T is in HS then // //----------------------------------------------------------- B // return T^HS • expand(TS’); // // else if TS is T^HS • TS’ and T is a "()-less macro" then // // ---------------------------------------------------------- C // return expand(subst(ts(T), {}, {}, HS \cup {T}, {}) • TS’ ); // // else if TS is T^HS •(•TS’ and T is a "()’d macro" then // // ---------------------------------------------------------- D // check TS’ is actuals • )^HS’ • TS’’ and actuals are "correct for T" // return expand(subst(ts(T), fp(T), actuals,(HS \cap HS’) \cup {T }, {}) • TS’’); // // // ------------------------------------------------------------------ E // note TS must be T^HS • TS’ // return T^HS • expand(TS’); // } func (c *cpp) expand(ts tokenReader, w tokenWriter, expandDefined bool) { // trc("==== expand enter") start: tok, ok := ts.read() tok.file = c.file // First, if TS is the empty set, the result is the empty set. if !ok { // ---------------------------------------------------------- A // return {}; // trc("---- expand A") return } // dbg("expand start %q", tok) if tok.char == IDENTIFIER { nm := tok.value if nm == idDefined && expandDefined { c.parseDefined(tok, ts, w) goto start } // Otherwise, if the token sequence begins with a token whose // hide set contains that token, then the result is the token // sequence beginning with that token (including its hide set) // followed by the result of expand on the rest of the token // sequence. if tok.has(nm) { // -------------------------------------------------- B // return T^HS • expand(TS’); // trc("---- expand B") // trc("expand write %q", tok) w.write(tok) goto start } m := c.macros[nm] if m != nil && !m.isFnLike { // Otherwise, if the token sequence begins with an // object-like macro, the result is the expansion of // the rest of the token sequence beginning with the // sequence returned by subst invoked with the // replacement token sequence for the macro, two empty // sets, the union of the macro’s hide set and the // macro itself, and an empty set. switch nm { case idLINE: c.lineMacro.repl[0].value = dict.sid(fmt.Sprint(tok.Position().Line)) case idCOUNTER: c.counterMacro.repl[0].value = dict.sid(fmt.Sprint(c.counter)) c.counter++ case idTclDefaultDoubleRounding: if c.ctx.cfg.ReplaceMacroTclDefaultDoubleRounding != "" { m = c.macros[dict.sid(c.ctx.cfg.ReplaceMacroTclDefaultDoubleRounding)] } case idTclIeeeDoubleRounding: if c.ctx.cfg.ReplaceMacroTclIeeeDoubleRounding != "" { m = c.macros[dict.sid(c.ctx.cfg.ReplaceMacroTclIeeeDoubleRounding)] } } if m != nil { // -------------------------------------------------- C // return expand(subst(ts(T), {}, {}, HS \cup {T}, {}) • TS’ ); // trc("---- expand C") hs := hideSet{nm: {}} for k, v := range tok.hs { hs[k] = v } os := cppTokensPool.Get().(*[]cppToken) toks := c.subst(m, c.cppToks(m.repl), nil, nil, nil, hs, os, expandDefined) for i := range toks { toks[i].pos = tok.pos } if len(toks) == 1 { toks[0].macro = nm } ts.ungets(toks) (*os) = (*os)[:0] cppTokensPool.Put(os) goto start } } if m != nil && m.isFnLike { switch nm { case idFdZero: if c.ctx.cfg.ReplaceMacroFdZero != "" { m = c.macros[dict.sid(c.ctx.cfg.ReplaceMacroFdZero)] } } if m != nil { // -------------------------------------------------- D // check TS’ is actuals • )^HS’ • TS’’ and actuals are "correct for T" // return expand(subst(ts(T), fp(T), actuals,(HS \cap HS’) \cup {T }, {}) • TS’’); // trc("---- expand D") hs := tok.hs var skip []cppToken again: t2, ok := ts.read() if !ok { // dbg("expand write %q", tok) w.write(tok) ts.ungets(skip) goto start } skip = append(skip, t2) switch t2.char { case '\n', ' ': goto again case '(': // ok default: w.write(tok) ts.ungets(skip) goto start } varArgs, ap, hs2 := c.actuals(m, ts) if nm == idHasIncludeImpl { //TODO- if len(ap) != 1 || len(ap[0]) != 1 { panic(todo("internal error")) } arg := ap[0][0].value.String() switch { case strings.HasPrefix(arg, `"\"`): // `"\"stdio.h\""` arg = arg[2:len(arg)-3] + `"` // -> `"stdio.h"` case strings.HasPrefix(arg, `"<`): // `""` arg = arg[1 : len(arg)-1] // -> `` default: arg = "" } var tok3 token3 tok3.char = PPNUMBER tok3.value = idZero if arg != "" { if _, err := c.hasInclude(&tok, arg); err == nil { tok3.value = idOne } } tok := cppToken{token4{token3: tok3, file: c.file}, nil} ts.ungets([]cppToken{tok}) goto start } switch { case len(hs2) == 0: hs2 = hideSet{nm: {}} default: nhs := hideSet{} for k := range hs { if _, ok := hs2[k]; ok { nhs[k] = struct{}{} } } nhs[nm] = struct{}{} hs2 = nhs } os := cppTokensPool.Get().(*[]cppToken) toks := c.subst(m, c.cppToks(m.repl), m.fp, varArgs, ap, hs2, os, expandDefined) for i := range toks { toks[i].pos = tok.pos } ts.ungets(toks) (*os) = (*os)[:0] cppTokensPool.Put(os) goto start } } } // ------------------------------------------------------------------ E // note TS must be T^HS • TS’ // return T^HS • expand(TS’); // trc("---- expand E") // trc("expand write %q", tok) w.write(tok) goto start } func (c *cpp) hasInclude(n Node, nm string) (rs string, err error) { // nm0 := nm // defer func() { //TODO- // trc("nm0 %q nm %q rs %q err %v", nm0, nm, rs, err) // }() var ( b byte paths []string sys bool ) switch { case nm != "" && nm[0] == '"': paths = c.ctx.includePaths b = '"' case nm != "" && nm[0] == '<': paths = c.ctx.sysIncludePaths sys = true b = '>' case nm == "": return "", fmt.Errorf("%v: invalid empty include argument", n.Position()) default: return "", fmt.Errorf("%v: invalid include argument %s", n.Position(), nm) } x := strings.IndexByte(nm[1:], b) if x < 0 { return "", fmt.Errorf("%v: invalid include argument %s", n.Position(), nm) } nm = filepath.FromSlash(nm[1 : x+1]) switch { case filepath.IsAbs(nm): fi, err := c.ctx.statFile(nm, sys) if err != nil { return "", fmt.Errorf("%v: %s", n.Position(), err) } if fi.IsDir() { return "", fmt.Errorf("%v: %s is a directory, not a file", n.Position(), nm) } return nm, nil default: dir := filepath.Dir(c.file.Name()) for _, v := range paths { if v == "@" { v = dir } var p string switch { case strings.HasPrefix(nm, "./"): wd := c.ctx.cfg.WorkingDir if wd == "" { var err error if wd, err = os.Getwd(); err != nil { return "", fmt.Errorf("%v: cannot determine working dir: %v", n.Position(), err) } } p = filepath.Join(wd, nm) default: p = filepath.Join(v, nm) } fi, err := c.ctx.statFile(p, sys) if err != nil || fi.IsDir() { continue } return p, nil } wd, _ := os.Getwd() return "", fmt.Errorf("include file not found: %s (wd %s)\nsearch paths:\n\t%s", nm, wd, strings.Join(paths, "\n\t")) } } func (c *cpp) actuals(m *Macro, r tokenReader) (varArgs []cppToken, ap [][]cppToken, hs hideSet) { var lvl, n int varx := len(m.fp) if m.namedVariadic { varx-- } var last rune for { t, ok := r.read() if !ok { c.err(t, "unexpected EOF") return nil, nil, nil } // 6.10.3, 10 // // Within the sequence of preprocessing tokens making up an // invocation of a function-like macro, new-line is considered // a normal white-space character. if t.char == '\n' { t.char = ' ' t.value = idSpace } if t.char == ' ' && last == ' ' { continue } last = t.char switch t.char { case ',': if lvl == 0 { if n >= varx && (len(varArgs) != 0 || !isWhite(t.char)) { varArgs = append(varArgs, t) } n++ continue } case ')': if lvl == 0 { for len(ap) < len(m.fp) { ap = append(ap, nil) } for i, v := range ap { ap[i] = c.trim(v) } // for i, v := range ap { // dbg("%T.actuals %v/%v %q", c, i, len(ap), tokStr(v, "|")) // } return c.trim(varArgs), ap, t.hs } lvl-- case '(': lvl++ } if n >= varx && (len(varArgs) != 0 || !isWhite(t.char)) { varArgs = append(varArgs, t) } for len(ap) <= n { ap = append(ap, []cppToken{}) } ap[n] = append(ap[n], t) } } // [1]pg 2. // // subst(IS, FP, AP, HS, OS) /* substitute args, handle stringize and paste */ // { // if IS is {} then // // ---------------------------------------------------------- A // return hsadd(HS, OS); // // else if IS is # • T • IS’ and T is FP[i] then // // ---------------------------------------------------------- B // return subst(IS’, FP, AP, HS, OS • stringize(select(i, AP))); // // else if IS is ## • T • IS’ and T is FP[i] then // { // // ---------------------------------------------------------- C // if select(i, AP) is {} then /* only if actuals can be empty */ // // -------------------------------------------------- D // return subst(IS’, FP, AP, HS, OS); // else // // -------------------------------------------------- E // return subst(IS’, FP, AP, HS, glue(OS, select(i, AP))); // } // // else if IS is ## • T^HS’ • IS’ then // // ---------------------------------------------------------- F // return subst(IS’, FP, AP, HS, glue(OS, T^HS’)); // // else if IS is T • ##^HS’ • IS’ and T is FP[i] then // { // // ---------------------------------------------------------- G // if select(i, AP) is {} then /* only if actuals can be empty */ // { // // -------------------------------------------------- H // if IS’ is T’ • IS’’ and T’ is FP[j] then // // ------------------------------------------ I // return subst(IS’’, FP, AP, HS, OS • select(j, AP)); // else // // ------------------------------------------ J // return subst(IS’, FP, AP, HS, OS); // } // else // // -------------------------------------------------- K // return subst(##^HS’ • IS’, FP, AP, HS, OS • select(i, AP)); // // } // // else if IS is T • IS’ and T is FP[i] then // // ---------------------------------------------------------- L // return subst(IS’, FP, AP, HS, OS • expand(select(i, AP))); // // // ------------------------------------------------------------------ M // note IS must be T^HS’ • IS’ // return subst(IS’, FP, AP, HS, OS • T^HS’); // } // // A quick overview of subst is that it walks through the input sequence, IS, // building up an output sequence, OS, by handling each token from left to // right. (The order that this operation takes is left to the implementation // also, walking from left to right is more natural since the rest of the // algorithm is constrained to this ordering.) Stringizing is easy, pasting // requires trickier handling because the operation has a bunch of // combinations. After the entire input sequence is finished, the updated hide // set is applied to the output sequence, and that is the result of subst. func (c *cpp) subst(m *Macro, is []cppToken, fp []StringID, varArgs []cppToken, ap [][]cppToken, hs hideSet, os *[]cppToken, expandDefined bool) (r []cppToken) { var ap0 [][]cppToken for _, v := range ap { ap0 = append(ap0, append([]cppToken(nil), v...)) } // trc("==== subst: is %q, fp, %v ap@%p %v", cppToksStr(is, "|"), fp, &ap, cppToksStr2(ap)) start: // trc("start: is %q, fp %v, ap@%p %v, os %q", cppToksStr(is, "|"), fp, &ap, cppToksStr2(ap), cppToksStr(*os, "|")) if len(is) == 0 { // ---------------------------------------------------------- A // return hsadd(HS, OS); // trc("---- A") // trc("subst RETURNS %q", cppToksStr(*os, "|")) return c.hsAdd(hs, os) } tok := is[0] var arg []cppToken if tok.char == '#' { if len(is) > 1 && is[1].char == IDENTIFIER && m.param(varArgs, ap0, is[1].value, &arg) { // -------------------------------------------------- B // return subst(IS’, FP, AP, HS, OS • stringize(select(i, AP))); // trc("---- subst B") *os = append(*os, c.stringize(arg)) is = is[2:] goto start } } if tok.char == PPPASTE { if len(is) > 1 && is[1].char == IDENTIFIER && m.param(varArgs, ap0, is[1].value, &arg) { // -------------------------------------------------- C // trc("---- subst C") if len(arg) == 0 { // TODO "only if actuals can be empty" // ------------------------------------------ D // return subst(IS’, FP, AP, HS, OS); // trc("---- D") if c := len(*os); c != 0 && (*os)[c-1].char == ',' { *os = (*os)[:c-1] } is = is[2:] goto start } // -------------------------------------------------- E // return subst(IS’, FP, AP, HS, glue(OS, select(i, AP))); // trc("---- subst E, arg %q", cppToksStr(arg, "|")) *os = c.glue(*os, arg) is = is[2:] goto start } if len(is) > 1 { // -------------------------------------------------- F // return subst(IS’, FP, AP, HS, glue(OS, T^HS’)); // trc("---- subst F") *os = c.glue(*os, is[1:2]) is = is[2:] goto start } } if tok.char == IDENTIFIER && (len(is) > 1 && is[1].char == PPPASTE) && m.param(varArgs, ap0, tok.value, &arg) { // ---------------------------------------------------------- G // trc("---- subst G") if len(arg) == 0 { // TODO "only if actuals can be empty" // -------------------------------------------------- H // trc("---- subst H") is = is[2:] // skip T## if len(is) > 0 && is[0].char == IDENTIFIER && m.param(varArgs, ap, is[0].value, &arg) { // -------------------------------------------------- I // return subst(IS’’, FP, AP, HS, OS • select(j, AP)); // trc("---- subst I") *os = append(*os, arg...) is = is[1:] goto start } else { // -------------------------------------------------- J // return subst(IS’, FP, AP, HS, OS); // trc("---- subst J") goto start } } // ---------------------------------------------------------- K // return subst(##^HS’ • IS’, FP, AP, HS, OS • select(i, AP)); // trc("---- subst K") *os = append(*os, arg...) is = is[1:] goto start } ax := -1 if tok.char == IDENTIFIER && m.param2(varArgs, ap, tok.value, &arg, &ax) { // ------------------------------------------ L // return subst(IS’, FP, AP, HS, OS • expand(select(i, AP))); // trc("---- subst L") // if toks, ok := cache[tok.value]; ok { // os = append(os, toks...) // is = is[1:] // goto start // } sel := cppReader{buf: arg} var w cppWriter // trc("---- L(1) ap@%p %v", &ap, cppToksStr2(ap)) c.expand(&sel, &w, expandDefined) // trc("---- L(2) ap@%p %v", &ap, cppToksStr2(ap)) *os = append(*os, w.toks...) if ax >= 0 { ap[ax] = w.toks } is = is[1:] goto start } // ------------------------------------------------------------------ M // note IS must be T^HS’ • IS’ // return subst(IS’, FP, AP, HS, OS • T^HS’); *os = append(*os, tok) is = is[1:] // trc("---- subst M: is %q, os %q", cppToksStr(is, "|"), cppToksStr(*os, "|")) goto start } // paste last of left side with first of right side // // [1] pg. 3 // //TODO implement properly [0], 6.10.3.3, 2. Must rescan the resulting token(s). // // $ cat main.c // #include // // #define foo(a, b) a ## b // // int main() { // int i = 42; // i foo(+, +); // printf("%i\n", i); // return 0; // } // $ rm -f a.out ; gcc -Wall main.c && ./a.out ; echo $? // 43 // 0 // $ // // ---------------------------------------------------------------------------- // glue(LS,RS ) /* paste last of left side with first of right side */ // { // if LS is L^HS and RS is R^HS’ • RS’ then // return L&R^(HS∩HS’) • RS’; /* undefined if L&R is invalid */ // // note LS must be L HS • LS’ // return L^HS • glue(LS’,RS ); // } func (c *cpp) glue(ls, rs []cppToken) (out []cppToken) { // trc("ls %q, rs %q", cppToksStr(ls, "|"), cppToksStr(rs, "|")) if len(rs) == 0 { return ls } if len(ls) == 0 { return rs } l := ls[len(ls)-1] ls = ls[:len(ls)-1] r := rs[0] rs = rs[1:] if l.char == IDENTIFIER && l.value == idL && r.char == STRINGLITERAL { l.char = LONGSTRINGLITERAL } l.value = dict.sid(l.String() + r.String()) return append(append(ls, l), rs...) } // Given a token sequence, stringize returns a single string literal token // containing the concatenated spellings of the tokens. // // [1] pg. 3 func (c *cpp) stringize(s0 []cppToken) (r cppToken) { // 6.10.3.2 // // Each occurrence of white space between the argument’s preprocessing // tokens becomes a single space character in the character string // literal. s := make([]cppToken, 0, len(s0)) var last rune for i := range s0 { t := s0[i] if isWhite(t.char) { t.char = ' ' t.value = idSpace if last == ' ' { continue } } last = t.char s = append(s, t) } // White space before the first preprocessing token and after the last // preprocessing token composing the argument is deleted. s = c.trim(s) // The character string literal corresponding to an empty argument is // "" if len(s) == 0 { r.hs = nil r.char = STRINGLITERAL r.value = idEmptyString return r } var a []string // Otherwise, the original spelling of each preprocessing token in the // argument is retained in the character string literal, except for // special handling for producing the spelling of string literals and // character constants: a \ character is inserted before each " and \ // character of a character constant or string literal (including the // delimiting " characters), except that it is implementation-defined // whether a \ character is inserted before the \ character beginning a // universal character name. for _, v := range s { s := v.String() switch v.char { case CHARCONST, STRINGLITERAL: s = strings.ReplaceAll(s, `\`, `\\`) s = strings.ReplaceAll(s, `"`, `\"`) case LONGCHARCONST, LONGSTRINGLITERAL: panic("TODO") } a = append(a, s) } r = s[0] r.hs = nil r.char = STRINGLITERAL r.value = dict.sid(`"` + strings.Join(a, "") + `"`) return r } func (c *cpp) trim(toks []cppToken) []cppToken { for len(toks) != 0 && isWhite(toks[0].char) { toks = toks[1:] } for len(toks) != 0 && isWhite(toks[len(toks)-1].char) { toks = toks[:len(toks)-1] } return toks } func (c *cpp) hsAdd(hs hideSet, toks *[]cppToken) []cppToken { for i, v := range *toks { if v.hs == nil { v.hs = hideSet{} } for k, w := range hs { v.hs[k] = w } v.file = c.file (*toks)[i] = v } return *toks } func (c *cpp) parseDefined(tok cppToken, r tokenReader, w tokenWriter) { toks := []cppToken{tok} if tok = c.scanToNonBlankToken(&toks, r, w); tok.char < 0 { return } switch tok.char { case IDENTIFIER: // ok case '(': if tok = c.scanToNonBlankToken(&toks, r, w); tok.char < 0 { return } if tok.char != IDENTIFIER { w.writes(toks) return } tok2 := c.scanToNonBlankToken(&toks, r, w) if tok2.char < 0 { return } if tok2.char != ')' { w.writes(toks) return } } tok.char = PPNUMBER switch _, ok := c.macros[tok.value]; { case ok: tok.value = idOne default: tok.value = idZero } w.write(tok) } func (c *cpp) scanToNonBlankToken(toks *[]cppToken, r tokenReader, w tokenWriter) cppToken { tok, ok := r.read() if !ok { w.writes(*toks) tok.char = -1 return tok } *toks = append(*toks, tok) if tok.char == ' ' || tok.char == '\n' { if tok, ok = r.read(); !ok { w.writes(*toks) tok.char = -1 return tok } *toks = append(*toks, tok) } return (*toks)[len(*toks)-1] } // [0], 6.10.1 func (c *cpp) evalInclusionCondition(expr []token3) (r bool) { if !c.intmaxChecked { if m := c.macros[idIntMaxWidth]; m != nil && len(m.repl) != 0 { if val := c.intMaxWidth(); val != 0 && val != 64 { c.err(m.name, "%s is %v, but only 64 is supported", idIntMaxWidth, val) } } c.intmaxChecked = true } val := c.eval(expr) return val != nil && c.isNonZero(val) } func (c *cpp) intMaxWidth() int64 { if m := c.macros[idIntMaxWidth]; m != nil && len(m.repl) != 0 { switch x := c.eval(m.repl).(type) { case nil: return 0 case int64: return x case uint64: return int64(x) default: panic(internalError()) } } return 0 } func (c *cpp) eval(expr []token3) interface{} { toks := make([]cppToken, len(expr)) for i, v := range expr { toks[i] = cppToken{token4{token3: v}, nil} } var w cppWriter c.expand(&cppReader{buf: toks}, &w, true) toks = w.toks p := 0 for _, v := range toks { switch v.char { case ' ', '\n': // nop default: toks[p] = v p++ } } toks = toks[:p] s := cppScanner(toks) val := c.expression(&s, true) switch s.peek().char { case -1, '#': // ok default: t := s.peek() c.err(t, "unexpected %s", tokName(t.char)) return nil } return val } // [0], 6.5.17 Comma operator // // expression: // assignment-expression // expression , assignment-expression func (c *cpp) expression(s *cppScanner, eval bool) interface{} { for { r := c.assignmentExpression(s, eval) if s.peek().char != ',' { return r } s.next() } } // [0], 6.5.16 Assignment operators // // assignment-expression: // conditional-expression // unary-expression assignment-operator assignment-expression // // assignment-operator: one of // = *= /= %= += -= <<= >>= &= ^= |= func (c *cpp) assignmentExpression(s *cppScanner, eval bool) interface{} { return c.conditionalExpression(s, eval) } // [0], 6.5.15 Conditional operator // // conditional-expression: // logical-OR-expression // logical-OR-expression ? expression : conditional-expression func (c *cpp) conditionalExpression(s *cppScanner, eval bool) interface{} { expr := c.logicalOrExpression(s, eval) if s.peek().char == '?' { s.next() exprIsNonZero := c.isNonZero(expr) expr2 := c.conditionalExpression(s, exprIsNonZero) if tok := s.peek(); tok.char != ':' { c.err(tok, "expected ':'") return expr } s.next() expr3 := c.conditionalExpression(s, !exprIsNonZero) // [0] 6.5.15 // // 5. If both the second and third operands have arithmetic type, the result // type that would be determined by the usual arithmetic conversions, were they // applied to those two operands, is the type of the result. x := c.operand(expr2) y := c.operand(expr3) if x != nil && y != nil { x, y = usualArithmeticConversions(c.ctx, nil, x, y, false) expr2 = c.fromOperand(x) expr3 = c.fromOperand(y) } switch { case exprIsNonZero: expr = expr2 default: expr = expr3 } } return expr } func (c *cpp) operand(v interface{}) Operand { switch x := v.(type) { case int64: return &operand{typ: &typeBase{size: 8, kind: byte(LongLong), flags: fSigned}, value: Int64Value(x)} case uint64: return &operand{typ: &typeBase{size: 8, kind: byte(ULongLong)}, value: Uint64Value(x)} default: return nil } } func (c *cpp) fromOperand(op Operand) interface{} { switch x := op.Value().(type) { case Int64Value: return int64(x) case Uint64Value: return uint64(x) default: return nil } } // [0], 6.5.14 Logical OR operator // // logical-OR-expression: // logical-AND-expression // logical-OR-expression || logical-AND-expression func (c *cpp) logicalOrExpression(s *cppScanner, eval bool) interface{} { lhs := c.logicalAndExpression(s, eval) for s.peek().char == OROR { s.next() if c.isNonZero(lhs) { eval = false } rhs := c.logicalAndExpression(s, eval) if c.isNonZero(lhs) || c.isNonZero(rhs) { lhs = int64(1) } } return lhs } // [0], 6.5.13 Logical AND operator // // logical-AND-expression: // inclusive-OR-expression // logical-AND-expression && inclusive-OR-expression func (c *cpp) logicalAndExpression(s *cppScanner, eval bool) interface{} { lhs := c.inclusiveOrExpression(s, eval) for s.peek().char == ANDAND { s.next() if c.isZero(lhs) { eval = false } rhs := c.inclusiveOrExpression(s, eval) if c.isZero(lhs) || c.isZero(rhs) { lhs = int64(0) } } return lhs } func (c *cpp) isZero(val interface{}) bool { switch x := val.(type) { case int64: return x == 0 case uint64: return x == 0 } panic(internalError()) } // [0], 6.5.12 Bitwise inclusive OR operator // // inclusive-OR-expression: // exclusive-OR-expression // inclusive-OR-expression | exclusive-OR-expression func (c *cpp) inclusiveOrExpression(s *cppScanner, eval bool) interface{} { lhs := c.exclusiveOrExpression(s, eval) for s.peek().char == '|' { s.next() rhs := c.exclusiveOrExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: lhs = x | y case uint64: lhs = uint64(x) | y } case uint64: switch y := rhs.(type) { case int64: lhs = x | uint64(y) case uint64: lhs = x | y } } } } return lhs } // [0], 6.5.11 Bitwise exclusive OR operator // // exclusive-OR-expression: // AND-expression // exclusive-OR-expression ^ AND-expression func (c *cpp) exclusiveOrExpression(s *cppScanner, eval bool) interface{} { lhs := c.andExpression(s, eval) for s.peek().char == '^' { s.next() rhs := c.andExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: lhs = x ^ y case uint64: lhs = uint64(x) ^ y } case uint64: switch y := rhs.(type) { case int64: lhs = x ^ uint64(y) case uint64: lhs = x ^ y } } } } return lhs } // [0], 6.5.10 Bitwise AND operator // // AND-expression: // equality-expression // AND-expression & equality-expression func (c *cpp) andExpression(s *cppScanner, eval bool) interface{} { lhs := c.equalityExpression(s, eval) for s.peek().char == '&' { s.next() rhs := c.equalityExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: lhs = x & y case uint64: lhs = uint64(x) & y } case uint64: switch y := rhs.(type) { case int64: lhs = x & uint64(y) case uint64: lhs = x & y } } } } return lhs } // [0], 6.5.9 Equality operators // // equality-expression: // relational-expression // equality-expression == relational-expression // equality-expression != relational-expression func (c *cpp) equalityExpression(s *cppScanner, eval bool) interface{} { lhs := c.relationalExpression(s, eval) for { var v bool switch s.peek().char { case EQ: s.next() rhs := c.relationalExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: v = x == y case uint64: v = uint64(x) == y } case uint64: switch y := rhs.(type) { case int64: v = x == uint64(y) case uint64: v = x == y } } } case NEQ: s.next() rhs := c.relationalExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: v = x != y case uint64: v = uint64(x) != y } case uint64: switch y := rhs.(type) { case int64: v = x != uint64(y) case uint64: v = x != y } } } default: return lhs } switch { case v: lhs = int64(1) default: lhs = int64(0) } } } // [0], 6.5.8 Relational operators // // relational-expression: // shift-expression // relational-expression < shift-expression // relational-expression > shift-expression // relational-expression <= shift-expression // relational-expression >= shift-expression func (c *cpp) relationalExpression(s *cppScanner, eval bool) interface{} { lhs := c.shiftExpression(s, eval) for { var v bool switch s.peek().char { case '<': s.next() rhs := c.shiftExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: v = x < y case uint64: v = uint64(x) < y } case uint64: switch y := rhs.(type) { case int64: v = x < uint64(y) case uint64: v = x < y } } } case '>': s.next() rhs := c.shiftExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: v = x > y case uint64: v = uint64(x) > y } case uint64: switch y := rhs.(type) { case int64: v = x > uint64(y) case uint64: v = x > y } } } case LEQ: s.next() rhs := c.shiftExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: v = x <= y case uint64: v = uint64(x) <= y } case uint64: switch y := rhs.(type) { case int64: v = x <= uint64(y) case uint64: v = x <= y } } } case GEQ: s.next() rhs := c.shiftExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: v = x >= y case uint64: v = uint64(x) >= y } case uint64: switch y := rhs.(type) { case int64: v = x >= uint64(y) case uint64: v = x >= y } } } default: return lhs } switch { case v: lhs = int64(1) default: lhs = int64(0) } } } // [0], 6.5.7 Bitwise shift operators // // shift-expression: // additive-expression // shift-expression << additive-expression // shift-expression >> additive-expression func (c *cpp) shiftExpression(s *cppScanner, eval bool) interface{} { lhs := c.additiveExpression(s, eval) for { switch s.peek().char { case LSH: s.next() rhs := c.additiveExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: lhs = x << uint(y) case uint64: lhs = uint64(x) << uint(y) } case uint64: switch y := rhs.(type) { case int64: lhs = x << uint(y) case uint64: lhs = x << uint(y) } } } case RSH: s.next() rhs := c.additiveExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: lhs = x >> uint(y) case uint64: lhs = uint64(x) >> uint(y) } case uint64: switch y := rhs.(type) { case int64: lhs = x >> uint(y) case uint64: lhs = x >> uint(y) } } } default: return lhs } } } // [0], 6.5.6 Additive operators // // additive-expression: // multiplicative-expression // additive-expression + multiplicative-expression // additive-expression - multiplicative-expression func (c *cpp) additiveExpression(s *cppScanner, eval bool) interface{} { lhs := c.multiplicativeExpression(s, eval) for { switch s.peek().char { case '+': s.next() rhs := c.multiplicativeExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: lhs = x + y case uint64: lhs = uint64(x) + y } case uint64: switch y := rhs.(type) { case int64: lhs = x + uint64(y) case uint64: lhs = x + y } } } case '-': s.next() rhs := c.multiplicativeExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: lhs = x - y case uint64: lhs = uint64(x) - y } case uint64: switch y := rhs.(type) { case int64: lhs = x - uint64(y) case uint64: lhs = x - y } } } default: return lhs } } } // [0], 6.5.5 Multiplicative operators // // multiplicative-expression: // unary-expression // [0], 6.10.1, 1. // multiplicative-expression * unary-expression // multiplicative-expression / unary-expression // multiplicative-expression % unary-expression func (c *cpp) multiplicativeExpression(s *cppScanner, eval bool) interface{} { lhs := c.unaryExpression(s, eval) for { switch s.peek().char { case '*': s.next() rhs := c.unaryExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: lhs = x * y case uint64: lhs = uint64(x) * y } case uint64: switch y := rhs.(type) { case int64: lhs = x * uint64(y) case uint64: lhs = x * y } } } case '/': tok := s.next() rhs := c.unaryExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: if y == 0 { c.err(tok, "division by zero") break } lhs = x / y case uint64: if y == 0 { c.err(tok, "division by zero") break } lhs = uint64(x) / y } case uint64: switch y := rhs.(type) { case int64: if y == 0 { c.err(tok, "division by zero") break } lhs = x / uint64(y) case uint64: if y == 0 { c.err(tok, "division by zero") break } lhs = x / y } } } case '%': tok := s.next() rhs := c.unaryExpression(s, eval) if eval { switch x := lhs.(type) { case int64: switch y := rhs.(type) { case int64: if y == 0 { c.err(tok, "division by zero") break } lhs = x % y case uint64: if y == 0 { c.err(tok, "division by zero") break } lhs = uint64(x) % y } case uint64: switch y := rhs.(type) { case int64: if y == 0 { c.err(tok, "division by zero") break } lhs = x % uint64(y) case uint64: if y == 0 { c.err(tok, "division by zero") break } lhs = x % y } } } default: return lhs } } } // [0], 6.5.3 Unary operators // // unary-expression: // primary-expression // unary-operator unary-expression // // unary-operator: one of // + - ~ ! func (c *cpp) unaryExpression(s *cppScanner, eval bool) interface{} { switch s.peek().char { case '+': s.next() return c.unaryExpression(s, eval) case '-': s.next() expr := c.unaryExpression(s, eval) if eval { switch x := expr.(type) { case int64: expr = -x case uint64: expr = -x } } return expr case '~': s.next() expr := c.unaryExpression(s, eval) if eval { switch x := expr.(type) { case int64: expr = ^x case uint64: expr = ^x } } return expr case '!': s.next() expr := c.unaryExpression(s, eval) if eval { var v bool switch x := expr.(type) { case int64: v = x == 0 case uint64: v = x == 0 } switch { case v: expr = int64(1) default: expr = int64(0) } } return expr default: return c.primaryExpression(s, eval) } } // [0], 6.5.1 Primary expressions // // primary-expression: // identifier // constant // ( expression ) func (c *cpp) primaryExpression(s *cppScanner, eval bool) interface{} { switch tok := s.peek(); tok.char { case CHARCONST, LONGCHARCONST: s.next() r := charConst(c.ctx, tok) return int64(r) case IDENTIFIER: if c.ctx.evalIdentError { panic("cannot evaluate identifier") } s.next() if s.peek().char == '(' { s.next() n := 1 loop: for n != 0 { switch s.peek().char { case '(': n++ case ')': n-- case -1: c.err(s.peek(), "expected )") break loop } s.next() } } return int64(0) case PPNUMBER: s.next() return c.intConst(tok) case '(': s.next() expr := c.expression(s, eval) if s.peek().char == ')' { s.next() } return expr default: return int64(0) } } // [0], 6.4.4.1 Integer constants // // integer-constant: // decimal-constant integer-suffix_opt // octal-constant integer-suffix_opt // hexadecimal-constant integer-suffix_opt // // decimal-constant: // nonzero-digit // decimal-constant digit // // octal-constant: // 0 // octal-constant octal-digit // // hexadecimal-prefix: one of // 0x 0X // // integer-suffix_opt: one of // u ul ull l lu ll llu func (c *cpp) intConst(tok cppToken) (r interface{}) { var n uint64 s0 := tok.String() s := strings.TrimRight(s0, "uUlL") switch { case strings.HasPrefix(s, "0x") || strings.HasPrefix(s, "0X"): var err error if n, err = strconv.ParseUint(s[2:], 16, 64); err != nil { c.err(tok, "%v", err) return int64(0) } case strings.HasPrefix(s, "0"): var err error if n, err = strconv.ParseUint(s, 8, 64); err != nil { c.err(tok, "%v", err) return int64(0) } default: var err error if n, err = strconv.ParseUint(s, 10, 64); err != nil { c.err(tok, "%v", err) return int64(0) } } suffix := s0[len(s):] if suffix == "" { if n > math.MaxInt64 { return n } return int64(n) } switch suffix = strings.ToLower(suffix); suffix { default: c.err(tok, "invalid suffix: %v", s0) fallthrough case "l", "ll": if n > math.MaxInt64 { return n } return int64(n) case "llu", "lu", "u", "ul", "ull": return n } } func charConst(ctx *context, tok cppToken) rune { s := tok.String() switch tok.char { case LONGCHARCONST: s = s[1:] // Remove leading 'L'. fallthrough case CHARCONST: s = s[1 : len(s)-1] // Remove outer 's. if len(s) == 1 { return rune(s[0]) } var r rune var n int switch s[0] { case '\\': r, n = decodeEscapeSequence(ctx, tok, s) if r < 0 { r = -r } default: r, n = utf8.DecodeRuneInString(s) } if n != len(s) { ctx.errNode(&tok, "invalid character constant") } return r } panic(internalError()) } // escape-sequence {simple-sequence}|{octal-escape-sequence}|{hexadecimal-escape-sequence}|{universal-character-name} // simple-sequence \\['\x22?\\abfnrtv] // octal-escape-sequence \\{octal-digit}{octal-digit}?{octal-digit}? // hexadecimal-escape-sequence \\x{hexadecimal-digit}+ func decodeEscapeSequence(ctx *context, tok cppToken, s string) (rune, int) { if s[0] != '\\' { panic(internalError()) } if len(s) == 1 { return rune(s[0]), 1 } r := rune(s[1]) switch r { case '\'', '"', '?', '\\': return r, 2 case 'a': return 7, 2 case 'b': return 8, 2 case 'e': return 0x1b, 2 case 'f': return 12, 2 case 'n': return 10, 2 case 'r': return 13, 2 case 't': return 9, 2 case 'v': return 11, 2 case 'x': v, n := 0, 2 loop2: for i := 2; i < len(s); i++ { r := s[i] switch { case r >= '0' && r <= '9', r >= 'a' && r <= 'f', r >= 'A' && r <= 'F': v = v<<4 | decodeHex(r) n++ default: break loop2 } } return -rune(v & 0xff), n case 'u', 'U': return decodeUCN(s) } if r < '0' || r > '7' { panic(internalError()) } v, n := 0, 1 ok := false loop: for i := 1; i < len(s); i++ { r := s[i] switch { case i < 4 && r >= '0' && r <= '7': ok = true v = v<<3 | (int(r) - '0') n++ default: break loop } } if !ok { ctx.errNode(&tok, "invalid octal sequence") } return -rune(v), n } // universal-character-name \\u{hex-quad}|\\U{hex-quad}{hex-quad} func decodeUCN(s string) (rune, int) { if s[0] != '\\' { panic(internalError()) } s = s[1:] switch s[0] { case 'u': return rune(decodeHexQuad(s[1:])), 6 case 'U': return rune(decodeHexQuad(s[1:])<<16 | decodeHexQuad(s[5:])), 10 } panic(internalError()) } // hex-quad {hexadecimal-digit}{hexadecimal-digit}{hexadecimal-digit}{hexadecimal-digit} func decodeHexQuad(s string) int { n := 0 for i := 0; i < 4; i++ { n = n<<4 | decodeHex(s[i]) } return n } func decodeHex(r byte) int { switch { case r >= '0' && r <= '9': return int(r) - '0' default: x := int(r) &^ 0x20 return x - 'A' + 10 } } func (c *cpp) isNonZero(val interface{}) bool { switch x := val.(type) { case int64: return x != 0 case uint64: return x != 0 } panic(internalError()) } type ppLine interface { getToks() []token3 } type ppIfGroupDirective interface { evalInclusionCondition(*cpp) bool } type ppElifDirective struct { toks []token3 expr []token3 } func (n *ppElifDirective) getToks() []token3 { return n.toks } type ppElseDirective struct { toks []token3 } func (n *ppElseDirective) getToks() []token3 { return n.toks } type ppEndifDirective struct { toks []token3 } func (n *ppEndifDirective) getToks() []token3 { return n.toks } type ppEmptyDirective struct { toks []token3 } func (n *ppEmptyDirective) getToks() []token3 { return n.toks } func (n *ppEmptyDirective) translationPhase4(c *cpp) { // nop } type ppIncludeDirective struct { arg []token3 toks []token3 includeNext bool // false: #include, true: #include_next } func (n *ppIncludeDirective) getToks() []token3 { return n.toks } func (n *ppIncludeDirective) translationPhase4(c *cpp) { if c.ctx.cfg.ignoreIncludes { return } args := make([]cppToken, 0, len(n.arg)) for _, v := range n.arg { switch v.char { case ' ', '\t', '\v', '\f': // nop default: args = append(args, cppToken{token4{token3: v}, nil}) } } var sb strings.Builder for _, v := range args { sb.WriteString(v.String()) } nm := strings.TrimSpace(sb.String()) if nm == "" { c.err(n.toks[0], "invalid empty include argument") return } switch nm[0] { case '"', '<': // ok default: var w cppWriter c.expand(&cppReader{buf: args}, &w, false) x := 0 for _, v := range w.toks { switch v.char { case ' ', '\t', '\v', '\f': // nop default: w.toks[x] = v x++ } } w.toks = w.toks[:x] nm = strings.TrimSpace(cppToksStr(w.toks, "")) } toks := n.toks if c.ctx.cfg.RejectIncludeNext { c.err(toks[0], "#include_next is a GCC extension") return } if c.ctx.cfg.fakeIncludes { c.send([]token3{{char: STRINGLITERAL, value: dict.sid(nm), src: dict.sid(nm)}, {char: '\n', value: idNL}}) return } if re := c.ctx.cfg.IgnoreInclude; re != nil && re.MatchString(nm) { return } if c.includeLevel == maxIncludeLevel { c.err(toks[0], "too many include levels") return } c.includeLevel++ defer func() { c.includeLevel-- }() var ( b byte paths []string sys bool ) switch { case nm != "" && nm[0] == '"': paths = c.ctx.includePaths b = '"' case nm != "" && nm[0] == '<': paths = c.ctx.sysIncludePaths sys = true b = '>' case nm == "": c.err(toks[0], "invalid empty include argument") return default: c.err(toks[0], "invalid include argument %s", nm) return } x := strings.IndexByte(nm[1:], b) if x < 0 { c.err(toks[0], "invalid include argument %s", nm) return } nm = filepath.FromSlash(nm[1 : x+1]) var path string switch { case filepath.IsAbs(nm): path = nm default: dir := filepath.Dir(c.file.Name()) if n.includeNext { nmDir, _ := filepath.Split(nm) for i, v := range paths { if w, err := filepath.Abs(v); err == nil { v = w } v = filepath.Join(v, nmDir) if v == dir { paths = paths[i+1:] break } } } for _, v := range paths { if v == "@" { v = dir } p := filepath.Join(v, nm) fi, err := c.ctx.statFile(p, sys) if err != nil || fi.IsDir() { continue } path = p break } } if path == "" { wd, _ := os.Getwd() c.err(toks[0], "include file not found: %s (wd %s)\nsearch paths:\n\t%s", nm, wd, strings.Join(paths, "\n\t")) return } if h := c.ctx.cfg.IncludeFileHandler; h != nil { var position gotoken.Position if p := toks[0].Pos(); p.IsValid() { position = gotoken.Position(c.file.PositionFor(p, true)) } apath, err := filepath.Abs(path) if err != nil { c.err(toks[0], "%s: cannot compute absolute path: %v", path, err) } h(position, apath) } cf, err := cache.getFile(c.ctx, path, sys, false) if err != nil { c.err(toks[0], "%s: %v", path, err) return } pf, err := cf.ppFile() if err != nil { c.err(toks[0], "%s: %v", path, err) return } saveFile := c.file saveFileMacro := c.fileMacro.repl[0].value c.file = pf.file c.fileMacro.repl[0].value = dict.sid(fmt.Sprintf("%q", c.file.Name())) defer func() { c.file = saveFile c.fileMacro.repl[0].value = saveFileMacro }() pf.translationPhase4(c) } func (c *cpp) send(toks []token3) { c.in <- toks <-c.rq } func (c *cpp) identicalReplacementLists(a, b []token3) bool { for len(a) != 0 && a[0].char == ' ' { a = a[1:] } for len(b) != 0 && b[0].char == ' ' { b = b[1:] } for len(a) != 0 && a[len(a)-1].char == ' ' { a = a[:len(a)-1] } for len(b) != 0 && b[len(b)-1].char == ' ' { b = b[:len(b)-1] } if len(a) != len(b) { return false } for i, v := range a { w := b[i] if v.char != w.char || v.value != w.value { return false } } return true } func stringConst(ctx *context, t cppToken) string { s := t.String() switch t.char { case LONGSTRINGLITERAL: s = s[1:] // Remove leading 'L'. fallthrough case STRINGLITERAL: var buf bytes.Buffer for i := 1; i < len(s)-1; { switch c := s[i]; c { case '\\': r, n := decodeEscapeSequence(ctx, t, s[i:]) switch { case r < 0: buf.WriteByte(byte(-r)) default: buf.WriteRune(r) } i += n default: buf.WriteByte(c) i++ } } return buf.String() } panic(internalError()) } // -------------------------------------------------------- Translation phase 4 // [0], 5.1.1.2, 4 // // Preprocessing directives are executed, macro invocations are expanded, and // _Pragma unary operator expressions are executed. If a character sequence // that matches the syntax of a universal character name is produced by token // concatenation (6.10.3.3), the behavior is undefined. A #include // preprocessing directive causes the named header or source file to be // processed from phase 1 through phase 4, recursively. All preprocessing // directives are then deleted. func (c *cpp) translationPhase4(in []source) chan *[]token4 { c.rq = make(chan struct{}) // Must be unbufferred c.in = make(chan []token3) // Must be unbufferred c.out = make(chan *[]token4, 10) //DONE benchmark tuned go func() { defer close(c.out) c.expand(c, c, false) }() go func() { defer close(c.in) for _, v := range in { pf, err := v.ppFile() if err != nil { c.err(nil, "%s", err) break } c.file = pf.file c.fileMacro.repl[0].value = dict.sid(fmt.Sprintf("%q", c.file.Name())) pf.translationPhase4(c) } }() return c.out } type ppErrorDirective struct { toks []token3 msg []token3 } func (n *ppErrorDirective) getToks() []token3 { return n.toks } func (n *ppErrorDirective) translationPhase4(c *cpp) { var b strings.Builder for _, v := range n.msg { b.WriteString(v.String()) } c.err(n.toks[0], "%s", strings.TrimSpace(b.String())) } type ppPragmaDirective struct { toks []token3 args []token3 } func (n *ppPragmaDirective) getToks() []token3 { return n.toks } func (n *ppPragmaDirective) translationPhase4(c *cpp) { parsePragma(c, n.args) } func parsePragma(c *cpp, args0 []token3) { if len(args0) == 1 { // \n return } if t := args0[0]; t.char == IDENTIFIER && t.value == idSTDC { p := t p.char = PRAGMASTDC p.value = idPragmaSTDC send := []token3{p, {char: ' ', value: idSpace, src: idSpace, pos: t.pos}} args := ltrim3(args0[1:]) if len(args) == 0 { c.err(args[0], "expected argument of STDC") return } if t = args[0]; t.char != IDENTIFIER { c.err(t, "expected identifier") return } switch t.value { case idFPContract, idFenvAccess, idCxLimitedRange: // ok default: c.err(t, "expected FP_CONTRACT or FENV_ACCESS or CX_LIMITED_RANGE") return } args = ltrim3(args[1:]) if len(args) == 0 { c.err(args[0], "expected ON or OFF or DEFAULT") return } if t = args[0]; t.char != IDENTIFIER { c.err(t, "expected identifier") return } switch t.value { case idOn, idOff, idDefault: c.writes(c.cppToks(append(send, args0...))) default: c.err(t, "expected ON or OFF or DEFAULT") return } } if c.ctx.cfg.PragmaHandler == nil { return } var toks []cppToken for _, v := range args0[:len(args0)-1] { toks = append(toks, cppToken{token4: token4{file: c.file, token3: v}}) } if len(toks) == 0 { return } var toks2 []Token var sep StringID for _, tok := range toks { switch tok.char { case ' ', '\n': if c.ctx.cfg.PreserveOnlyLastNonBlankSeparator { if strings.TrimSpace(tok.value.String()) != "" { sep = tok.value } break } switch { case sep != 0: sep = dict.sid(sep.String() + tok.String()) //TODO quadratic default: sep = tok.value } default: var t Token t.Rune = tok.char t.Sep = sep t.Value = tok.value t.file = tok.file t.pos = tok.pos toks2 = append(toks2, t) sep = 0 } } if len(toks2) == 0 { return } // dbg("%v: %q", c.file.PositionFor(args0[0].Pos(), true), tokStr(toks2, "|")) c.ctx.cfg.PragmaHandler(&pragma{tok: toks[0], c: c}, toks2) } type ppNonDirective struct { toks []token3 } func (n *ppNonDirective) getToks() []token3 { return n.toks } func (n *ppNonDirective) translationPhase4(c *cpp) { // nop } type ppTextLine struct { toks []token3 } func (n *ppTextLine) getToks() []token3 { return n.toks } func (n *ppTextLine) translationPhase4(c *cpp) { c.send(n.toks) } type ppLineDirective struct { toks []token3 args []token3 nextPos int } func (n *ppLineDirective) getToks() []token3 { return n.toks } func (n *ppLineDirective) translationPhase4(c *cpp) { toks := expandArgs(c, n.args) if len(toks) == 0 { return } switch t := toks[0]; t.char { case PPNUMBER: ln, err := strconv.ParseInt(t.String(), 10, 31) if err != nil || ln < 1 { c.err(t, "expected positive integer less or equal 2147483647") return } for len(toks) != 0 && toks[0].char == ' ' { toks = toks[1:] } if len(toks) == 1 { c.file.AddLineInfo(int(n.nextPos)-1, c.file.Name(), int(ln)) return } toks = toks[1:] for len(toks) != 0 && toks[0].char == ' ' { toks = toks[1:] } if len(toks) == 0 { c.file.AddLineInfo(int(n.nextPos)-1, c.file.Name(), int(ln)) return } switch t := toks[0]; t.char { case STRINGLITERAL: s := t.String() s = s[1 : len(s)-1] c.file.AddLineInfo(int(n.nextPos)-1, s, int(ln)) c.fileMacro.repl[0].value = t.value for len(toks) != 0 && toks[0].char == ' ' { toks = toks[1:] } if len(toks) != 0 && c.ctx.cfg.RejectLineExtraTokens { c.err(toks[0], "expected new-line") } default: c.err(t, "expected string literal") return } default: c.err(toks[0], "expected integer literal") return } } func expandArgs(c *cpp, args []token3) []cppToken { var w cppWriter var toks []cppToken for _, v := range args { toks = append(toks, cppToken{token4: token4{file: c.file, token3: v}}) } c.expand(&cppReader{buf: toks}, &w, true) return w.toks } type ppUndefDirective struct { name token3 toks []token3 } func (n *ppUndefDirective) getToks() []token3 { return n.toks } func (n *ppUndefDirective) translationPhase4(c *cpp) { nm := n.name.value if _, ok := protectedMacros[nm]; ok || nm == idDefined { c.err(n.name, "cannot undefine a protected name") return } // dbg("#undef %s", nm) delete(c.macros, nm) } type ppIfdefDirective struct { name StringID toks []token3 } func (n *ppIfdefDirective) evalInclusionCondition(c *cpp) bool { _, ok := c.macros[n.name]; return ok } func (n *ppIfdefDirective) getToks() []token3 { return n.toks } type ppIfndefDirective struct { name StringID toks []token3 } func (n *ppIfndefDirective) evalInclusionCondition(c *cpp) bool { _, ok := c.macros[n.name] return !ok } func (n *ppIfndefDirective) getToks() []token3 { return n.toks } type ppIfDirective struct { toks []token3 expr []token3 } func (n *ppIfDirective) getToks() []token3 { return n.toks } func (n *ppIfDirective) evalInclusionCondition(c *cpp) bool { return c.evalInclusionCondition(n.expr) } type ppDefineObjectMacroDirective struct { name token3 toks []token3 replacementList []token3 } func (n *ppDefineObjectMacroDirective) getToks() []token3 { return n.toks } func (n *ppDefineObjectMacroDirective) translationPhase4(c *cpp) { nm := n.name.value m := c.macros[nm] if m != nil { if _, ok := protectedMacros[nm]; ok || nm == idDefined { c.err(n.name, "cannot define protected name") return } if m.isFnLike { c.err(n.name, "redefinition of a function-like macro with an object-like one") } if !c.identicalReplacementLists(n.replacementList, m.repl) && c.ctx.cfg.RejectIncompatibleMacroRedef { c.err(n.name, "redefinition with different replacement list") return } } // find first non-blank token to claim as our location var pos int32 for _, t := range n.toks { if t.char != ' ' { pos = t.pos break } } // dbg("#define %s %s // %v", n.name, tokStr(n.replacementList, " "), c.file.PositionFor(n.name.Pos(), true)) c.macros[nm] = &Macro{pos: pos, name: token4{token3: n.name, file: c.file}, repl: n.replacementList} if nm != idGNUC { return } c.ctx.keywords = gccKeywords } type ppDefineFunctionMacroDirective struct { identifierList []token3 toks []token3 replacementList []token3 name token3 namedVariadic bool // foo..., note no comma before ellipsis. variadic bool } func (n *ppDefineFunctionMacroDirective) getToks() []token3 { return n.toks } func (n *ppDefineFunctionMacroDirective) translationPhase4(c *cpp) { nm := n.name.value m := c.macros[nm] if m != nil { if _, ok := protectedMacros[nm]; ok || nm == idDefined { c.err(n.name, "cannot define protected name") return } if !m.isFnLike && c.ctx.cfg.RejectIncompatibleMacroRedef { c.err(n.name, "redefinition of an object-like macro with a function-like one") return } ok := len(m.fp) == len(n.identifierList) if ok { for i, v := range m.fp { if v != n.identifierList[i].value { ok = false break } } } if !ok && (len(n.replacementList) != 0 || len(m.repl) != 0) && c.ctx.cfg.RejectIncompatibleMacroRedef { c.err(n.name, "redefinition with different formal parameters") return } if !c.identicalReplacementLists(n.replacementList, m.repl) && c.ctx.cfg.RejectIncompatibleMacroRedef { c.err(n.name, "redefinition with different replacement list") return } if m.variadic != n.variadic && c.ctx.cfg.RejectIncompatibleMacroRedef { c.err(n.name, "redefinition differs in being variadic") return } } nms := map[StringID]struct{}{} for _, v := range n.identifierList { if _, ok := nms[v.value]; ok { c.err(v, "duplicate identifier %s", v.value) } } var fp []StringID for _, v := range n.identifierList { fp = append(fp, v.value) } // dbg("#define %s %s // %v", n.name, tokStr(n.replacementList, " "), c.file.PositionFor(n.name.Pos(), true)) c.macros[nm] = &Macro{fp: fp, isFnLike: true, name: token4{token3: n.name, file: c.file}, repl: n.replacementList, variadic: n.variadic, namedVariadic: n.namedVariadic} } // [0], 6.10.1 // // elif-group: // # elif constant-expression new-line group_opt type ppElifGroup struct { elif *ppElifDirective groups []ppGroup } func (n *ppElifGroup) evalInclusionCondition(c *cpp) bool { if !c.evalInclusionCondition(n.elif.expr) { return false } for _, v := range n.groups { v.translationPhase4(c) } return true } // [0], 6.10.1 // // else-group: // # else new-line group_opt type ppElseGroup struct { elseLine *ppElseDirective groups []ppGroup } func (n *ppElseGroup) translationPhase4(c *cpp) { if n == nil { return } for _, v := range n.groups { v.translationPhase4(c) } } // [0], 6.10.1 // // PreprocessingFile: // GroupOpt type ppFile struct { file *tokenFile groups []ppGroup } func (n *ppFile) translationPhase4(c *cpp) { c.ctx.tuSourcesAdd(1) if f := n.file; f != nil { c.ctx.tuSizeAdd(int64(f.Size())) } for _, v := range n.groups { v.translationPhase4(c) } } // [0], 6.10.1 // // group-part: // if-section // control-line // text-line // # non-directive type ppGroup interface { translationPhase4(*cpp) } // [0], 6.10.1 // // if-group: // # if constant-expression new-line group opt // # ifdef identifier new-line group opt // # ifndef identifier new-line group opt type ppIfGroup struct { directive ppIfGroupDirective groups []ppGroup } func (n *ppIfGroup) evalInclusionCondition(c *cpp) bool { if !n.directive.evalInclusionCondition(c) { return false } for _, v := range n.groups { v.translationPhase4(c) } return true } // [0], 6.10.1 // // if-section: // if-group elif-groups_opt else-group_opt endif-line type ppIfSection struct { ifGroup *ppIfGroup elifGroups []*ppElifGroup elseGroup *ppElseGroup endifLine *ppEndifDirective } func (n *ppIfSection) translationPhase4(c *cpp) { if !n.ifGroup.evalInclusionCondition(c) { for _, v := range n.elifGroups { if v.evalInclusionCondition(c) { return } } n.elseGroup.translationPhase4(c) } }