gc: introduce rune

R=ken, r
CC=golang-dev
https://golang.org/cl/5293046

15 files changed, 186 insertions, 35 deletions

diff --git a/src/cmd/gc/Makefile b/src/cmd/gc/Makefile
index f7e3051783..11f466ae80 100644
--- a/src/cmd/gc/Makefile
+++ b/src/cmd/gc/Makefile
@@ -44,6 +44,8 @@ OFILES=\
 	walk.$O\
 	y1.tab.$O\
 
+HOST_CFLAGS+=-DGOEXPERIMENT='"$(GOEXPERIMENT)"' 
+
 NOINSTALL=1
 include ../../Make.clib
 
diff --git a/src/cmd/gc/builtin.c.boot b/src/cmd/gc/builtin.c.boot
index bf9d96d6a7..f6508b55a3 100644
--- a/src/cmd/gc/builtin.c.boot
+++ b/src/cmd/gc/builtin.c.boot
@@ -31,11 +31,11 @@ char *runtimeimport =
 	"func @\"\".slicestring1 (? string, ? int) string\n"
 	"func @\"\".intstring (? int64) string\n"
 	"func @\"\".slicebytetostring (? []byte) string\n"
-	"func @\"\".sliceinttostring (? []int) string\n"
+	"func @\"\".slicerunetostring (? []rune) string\n"
 	"func @\"\".stringtoslicebyte (? string) []byte\n"
-	"func @\"\".stringtosliceint (? string) []int\n"
+	"func @\"\".stringtoslicerune (? string) []rune\n"
 	"func @\"\".stringiter (? string, ? int) int\n"
-	"func @\"\".stringiter2 (? string, ? int) (retk int, retv int)\n"
+	"func @\"\".stringiter2 (? string, ? int) (retk int, retv rune)\n"
 	"func @\"\".slicecopy (to any, fr any, wid uint32) int\n"
 	"func @\"\".slicestringcopy (to any, fr any) int\n"
 	"func @\"\".convI2E (elem any) any\n"
diff --git a/src/cmd/gc/export.c b/src/cmd/gc/export.c
index e2f8c6f0c7..f79619e8f5 100644
--- a/src/cmd/gc/export.c
+++ b/src/cmd/gc/export.c
@@ -94,7 +94,7 @@ dumpprereq(Type *t)
 	if(t == T)
 		return;
 
-	if(t->printed || t == types[t->etype] || t == bytetype)
+	if(t->printed || t == types[t->etype] || t == bytetype || t == runetype)
 		return;
 	t->printed = 1;
 
diff --git a/src/cmd/gc/go.h b/src/cmd/gc/go.h
index 73cef3ddea..74d38bb789 100644
--- a/src/cmd/gc/go.h
+++ b/src/cmd/gc/go.h
@@ -785,6 +785,7 @@ EXTERN	Type*	types[NTYPE];
 EXTERN	Type*	idealstring;
 EXTERN	Type*	idealbool;
 EXTERN	Type*	bytetype;
+EXTERN	Type*	runetype;
 EXTERN	uchar	simtype[NTYPE];
 EXTERN	uchar	isptr[NTYPE];
 EXTERN	uchar	isforw[NTYPE];
@@ -840,6 +841,7 @@ EXTERN	Node*	nblank;
 
 extern	int	thechar;
 extern	char*	thestring;
+
 EXTERN	char*	hunk;
 EXTERN	int32	nhunk;
 EXTERN	int32	thunk;
@@ -854,6 +856,8 @@ EXTERN	int	packagequotes;
 EXTERN	int	longsymnames;
 EXTERN	int	compiling_runtime;
 
+EXTERN	int	rune32;
+
 /*
  *	y.tab.c
  */
@@ -1009,6 +1013,7 @@ Node*	renameinit(Node *n);
 void	cannedimports(char *file, char *cp);
 void	importfile(Val *f, int line);
 char*	lexname(int lex);
+char*	expstring(void);
 void	mkpackage(char* pkgname);
 void	unimportfile(void);
 int32	yylex(void);
diff --git a/src/cmd/gc/lex.c b/src/cmd/gc/lex.c
index 73a23ee5a4..1d7f0e82e3 100644
--- a/src/cmd/gc/lex.c
+++ b/src/cmd/gc/lex.c
@@ -30,6 +30,60 @@ static void	addidir(char*);
 static int	getlinepragma(void);
 static char *goos, *goarch, *goroot;
 
+// Compiler experiments.
+// These are controlled by the GCEXPERIMENT environment
+// variable recorded when the compiler is built.
+static struct {
+	char *name;
+	int *val;
+} exper[] = {
+	{"rune32", &rune32},
+};
+
+static void
+addexp(char *s)
+{
+	int i;
+	
+	for(i=0; i<nelem(exper); i++) {
+		if(strcmp(exper[i].name, s) == 0) {
+			*exper[i].val = 1;
+			return;
+		}
+	}
+	
+	print("unknown experiment %s\n", s);
+	exits("unknown experiment");
+}
+
+static void
+setexp(void)
+{
+	char *f[20];
+	int i, nf;
+	
+	// The makefile #defines GOEXPERIMENT for us.
+	nf = getfields(GOEXPERIMENT, f, nelem(f), 1, ",");
+	for(i=0; i<nf; i++)
+		addexp(f[i]);
+}
+
+char*
+expstring(void)
+{
+	int i;
+	static char buf[512];
+
+	strcpy(buf, "X");
+	for(i=0; i<nelem(exper); i++)
+		if(*exper[i].val)
+			seprint(buf+strlen(buf), buf+sizeof buf, ",%s", exper[i].name);
+	if(strlen(buf) == 1)
+		strcpy(buf, "X,none");
+	buf[1] = ':';
+	return buf;
+}
+
 // Our own isdigit, isspace, isalpha, isalnum that take care 
 // of EOF and other out of range arguments.
 static int
@@ -94,7 +148,7 @@ usage(void)
 	print("  -u disable package unsafe\n");
 	print("  -w print the parse tree after typing\n");
 	print("  -x print lex tokens\n");
-	exits(0);
+	exits("usage");
 }
 
 void
@@ -144,6 +198,8 @@ main(int argc, char *argv[])
 	goroot = getgoroot();
 	goos = getgoos();
 	goarch = thestring;
+	
+	setexp();
 
 	outfile = nil;
 	ARGBEGIN {
@@ -170,7 +226,10 @@ main(int argc, char *argv[])
 		break;
 
 	case 'V':
-		print("%cg version %s\n", thechar, getgoversion());
+		p = expstring();
+		if(strcmp(p, "X:none") == 0)
+			p = "";
+		print("%cg version %s%s%s%s\n", thechar, getgoversion(), *p ? " " : "", p);
 		exits(0);
 	} ARGEND
 
@@ -540,7 +599,7 @@ importfile(Val *f, int line)
 			yyerror("import %s: not a go object file", file);
 			errorexit();
 		}
-		q = smprint("%s %s %s", getgoos(), thestring, getgoversion());
+		q = smprint("%s %s %s %s", getgoos(), thestring, getgoversion(), expstring());
 		if(strcmp(p+10, q) != 0) {
 			yyerror("import %s: object is [%s] expected [%s]", file, p+10, q);
 			errorexit();
@@ -1720,6 +1779,18 @@ lexinit1(void)
 	s1 = pkglookup("byte", builtinpkg);
 	s1->lexical = LNAME;
 	s1->def = typenod(bytetype);
+
+	// rune alias
+	s = lookup("rune");
+	s->lexical = LNAME;
+	if(rune32)
+		runetype = typ(TINT32);
+	else
+		runetype = typ(TINT);
+	runetype->sym = s;
+	s1 = pkglookup("rune", builtinpkg);
+	s1->lexical = LNAME;
+	s1->def = typenod(runetype);
 }
 
 static void
@@ -1761,6 +1832,10 @@ lexfini(void)
 	if(s->def == N)
 		s->def = typenod(bytetype);
 
+	s = lookup("rune");
+	if(s->def == N)
+		s->def = typenod(runetype);
+
 	types[TNIL] = typ(TNIL);
 	s = lookup("nil");
 	if(s->def == N) {
diff --git a/src/cmd/gc/obj.c b/src/cmd/gc/obj.c
index aba2aafd81..994f71f3f8 100644
--- a/src/cmd/gc/obj.c
+++ b/src/cmd/gc/obj.c
@@ -23,7 +23,7 @@ dumpobj(void)
 		errorexit();
 	}
 
-	Bprint(bout, "go object %s %s %s\n", getgoos(), thestring, getgoversion());
+	Bprint(bout, "go object %s %s %s %s\n", getgoos(), thestring, getgoversion(), expstring());
 	Bprint(bout, "  exports automatically generated from\n");
 	Bprint(bout, "  %s in package \"%s\"\n", curio.infile, localpkg->name);
 	dumpexport();
diff --git a/src/cmd/gc/range.c b/src/cmd/gc/range.c
index 5cbafd895a..062e793be2 100644
--- a/src/cmd/gc/range.c
+++ b/src/cmd/gc/range.c
@@ -54,7 +54,7 @@ typecheckrange(Node *n)
 
 	case TSTRING:
 		t1 = types[TINT];
-		t2 = types[TINT];
+		t2 = runetype;
 		break;
 	}
 
@@ -216,7 +216,7 @@ walkrange(Node *n)
 		if(v2 == N)
 			a = nod(OAS, hv1, mkcall("stringiter", types[TINT], nil, ha, hv1));
 		else {
-			hv2 = temp(types[TINT]);
+			hv2 = temp(runetype);
 			a = nod(OAS2, N, N);
 			a->list = list(list1(hv1), hv2);
 			fn = syslook("stringiter2", 0);
diff --git a/src/cmd/gc/reflect.c b/src/cmd/gc/reflect.c
index 816235bcc2..67ad6bc5a4 100644
--- a/src/cmd/gc/reflect.c
+++ b/src/cmd/gc/reflect.c
@@ -692,7 +692,7 @@ dtypesym(Type *t)
 		tbase = t->type;
 	dupok = tbase->sym == S;
 
-	if(compiling_runtime && (tbase == types[tbase->etype] || tbase == bytetype))	// int, float, etc
+	if(compiling_runtime && (tbase == types[tbase->etype] || tbase == bytetype || tbase == runetype))	// int, float, etc
 		goto ok;
 
 	// named types from other files are defined only by those files
diff --git a/src/cmd/gc/runtime.go b/src/cmd/gc/runtime.go
index b190e50a7a..1bf2ad482e 100644
--- a/src/cmd/gc/runtime.go
+++ b/src/cmd/gc/runtime.go
@@ -47,11 +47,11 @@ func slicestring(string, int, int) string
 func slicestring1(string, int) string
 func intstring(int64) string
 func slicebytetostring([]byte) string
-func sliceinttostring([]int) string
+func slicerunetostring([]rune) string
 func stringtoslicebyte(string) []byte
-func stringtosliceint(string) []int
+func stringtoslicerune(string) []rune
 func stringiter(string, int) int
-func stringiter2(string, int) (retk int, retv int)
+func stringiter2(string, int) (retk int, retv rune)
 func slicecopy(to any, fr any, wid uint32) int
 func slicestringcopy(to any, fr any) int
 
diff --git a/src/cmd/gc/subr.c b/src/cmd/gc/subr.c
index 56537efa02..bd6585518b 100644
--- a/src/cmd/gc/subr.c
+++ b/src/cmd/gc/subr.c
@@ -1285,13 +1285,15 @@ Tpretty(Fmt *fp, Type *t)
 		// called from typesym
 		if(t == bytetype)
 			t = types[bytetype->etype];
+		if(t == runetype)
+			t = types[runetype->etype];
 	}
 
 	if(t->etype != TFIELD
 	&& t->sym != S
 	&& !(fp->flags&FmtLong)) {
 		s = t->sym;
-		if((t == types[t->etype] && t->etype != TUNSAFEPTR) || t == bytetype)
+		if((t == types[t->etype] && t->etype != TUNSAFEPTR) || t == bytetype || t == runetype)
 			return fmtprint(fp, "%s", s->name);
 		if(exporting) {
 			if(fp->flags & FmtShort)
@@ -1875,6 +1877,11 @@ eqtype(Type *t1, Type *t2)
 			if((t1 == types[TUINT8] || t1 == bytetype) && (t2 == types[TUINT8] || t2 == bytetype))
 				return 1;
 			break;
+		case TINT:
+		case TINT32:
+			if((t1 == types[runetype->etype] || t1 == runetype) && (t2 == types[runetype->etype] || t2 == runetype))
+				return 1;
+			break;
 		}
 		return 0;
 	}
@@ -2100,7 +2107,7 @@ convertop(Type *src, Type *dst, char **why)
 		return OCONV;
 	}
 
-	// 6. src is an integer or has type []byte or []int
+	// 6. src is an integer or has type []byte or []rune
 	// and dst is a string type.
 	if(isint[src->etype] && dst->etype == TSTRING)
 		return ORUNESTR;
@@ -2108,16 +2115,16 @@ convertop(Type *src, Type *dst, char **why)
 	if(isslice(src) && src->sym == nil && dst->etype == TSTRING) {
 		if(eqtype(src->type, bytetype))
 			return OARRAYBYTESTR;
-		if(eqtype(src->type, types[TINT]))
+		if(eqtype(src->type, runetype))
 			return OARRAYRUNESTR;
 	}
 	
-	// 7. src is a string and dst is []byte or []int.
+	// 7. src is a string and dst is []byte or []rune.
 	// String to slice.
 	if(src->etype == TSTRING && isslice(dst) && dst->sym == nil) {
 		if(eqtype(dst->type, bytetype))
 			return OSTRARRAYBYTE;
-		if(eqtype(dst->type, types[TINT]))
+		if(eqtype(dst->type, runetype))
 			return OSTRARRAYRUNE;
 	}
 	
diff --git a/src/cmd/gc/typecheck.c b/src/cmd/gc/typecheck.c
index 21cf77e300..d42477fd8b 100644
--- a/src/cmd/gc/typecheck.c
+++ b/src/cmd/gc/typecheck.c
@@ -2469,7 +2469,7 @@ stringtoarraylit(Node **np)
 		while(p < ep)
 			l = list(l, nod(OKEY, nodintconst(i++), nodintconst((uchar)*p++)));
 	} else {
-		// utf-8 []int
+		// utf-8 []rune
 		while(p < ep) {
 			p += chartorune(&r, p);
 			l = list(l, nod(OKEY, nodintconst(i++), nodintconst(r)));
diff --git a/src/cmd/gc/walk.c b/src/cmd/gc/walk.c
index 9d06f1b7f5..e94e043317 100644
--- a/src/cmd/gc/walk.c
+++ b/src/cmd/gc/walk.c
@@ -1151,8 +1151,8 @@ walkexpr(Node **np, NodeList **init)
 		goto ret;
 
 	case OARRAYRUNESTR:
-		// sliceinttostring([]int) string;
-		n = mkcall("sliceinttostring", n->type, init, n->left);
+		// slicerunetostring([]rune) string;
+		n = mkcall("slicerunetostring", n->type, init, n->left);
 		goto ret;
 
 	case OSTRARRAYBYTE:
@@ -1161,8 +1161,8 @@ walkexpr(Node **np, NodeList **init)
 		goto ret;
 
 	case OSTRARRAYRUNE:
-		// stringtosliceint(string) []int
-		n = mkcall("stringtosliceint", n->type, init, n->left);
+		// stringtoslicerune(string) []rune
+		n = mkcall("stringtoslicerune", n->type, init, n->left);
 		goto ret;
 
 	case OCMPIFACE:
diff --git a/src/cmd/gopack/ar.c b/src/cmd/gopack/ar.c
index 96f36605fd..c02903bc7a 100644
--- a/src/cmd/gopack/ar.c
+++ b/src/cmd/gopack/ar.c
@@ -612,6 +612,43 @@ qcmd(char *arname, int count, char **files)
 }
 
 /*
+ *	does the object header line p match the last one we saw?
+ *	update *lastp if it gets more specific.
+ */
+int
+matchhdr(char *p, char **lastp)
+{
+	int n;
+	char *last;
+	
+	// no information?
+	last = *lastp;
+	if(last == nil) {
+		*lastp = strdup(p);
+		return 1;
+	}
+
+	// identical match?
+	if(strcmp(last, p) == 0)
+		return 1;
+
+	// last has extra fields
+	n = strlen(p);
+	if(n < strlen(last) && last[n] == ' ')
+		return 1;
+
+	// p has extra fields - save in last
+	n = strlen(last);
+	if(n < strlen(p) && p[n] == ' ') {
+		free(last);
+		*lastp = strdup(p);
+		return 1;
+	}
+	
+	return 0;
+}	
+
+/*
  *	extract the symbol references from an object file
  */
 void
@@ -670,18 +707,23 @@ scanobj(Biobuf *b, Arfile *ap, long size)
 		return;
 	}
 	
-	if ((lastobj >= 0 && obj != lastobj) || (objhdr != nil && strcmp(p, objhdr) != 0)) {
-		fprint(2, "gopack: inconsistent object file %s\n", file);
+	if (!matchhdr(p, &objhdr)) {
+		fprint(2, "gopack: inconsistent object file %s: [%s] vs [%s]\n", file, p, objhdr);
 		errors++;
 		allobj = 0;
 		free(p);
 		return;
 	}
+	free(p);
+
+	// Old check.  Should be impossible since objhdrs match, but keep the check anyway.
+	if (lastobj >= 0 && obj != lastobj) {
+		fprint(2, "gopack: inconsistent object file %s\n", file);
+		errors++;
+		allobj = 0;
+		return;
+	}
 	lastobj = obj;
-	if(objhdr == nil)
-		objhdr = p;
-	else
-		free(p);
 		
 	if (!readar(b, obj, offset+size, 0)) {
 		fprint(2, "gopack: invalid symbol reference in file %s\n", file);
diff --git a/src/cmd/ld/lib.c b/src/cmd/ld/lib.c
index 456d41f5a1..8f95665b33 100644
--- a/src/cmd/ld/lib.c
+++ b/src/cmd/ld/lib.c
@@ -46,6 +46,7 @@ static int	cout = -1;
 char*	goroot;
 char*	goarch;
 char*	goos;
+char*	theline;
 
 void
 Lflag(char *arg)
@@ -478,12 +479,31 @@ ldobj(Biobuf *f, char *pkg, int64 len, char *pn, int whence)
 		diag("%s: not an object file", pn);
 		return;
 	}
-	t = smprint("%s %s %s", getgoos(), thestring, getgoversion());
-	if(strcmp(line+10, t) != 0 && !debug['f']) {
+	
+	// First, check that the basic goos, string, and version match.
+	t = smprint("%s %s %s ", getgoos(), thestring, getgoversion());
+	line[n] = ' ';
+	if(strncmp(line+10, t, strlen(t)) != 0 && !debug['f']) {
+		line[n] = '\0';
 		diag("%s: object is [%s] expected [%s]", pn, line+10, t);
 		free(t);
 		return;
 	}
+	
+	// Second, check that longer lines match each other exactly,
+	// so that the Go compiler and write additional information
+	// that must be the same from run to run.
+	line[n] = '\0';
+	if(n-10 > strlen(t)) {
+		if(theline == nil)
+			theline = strdup(line+10);
+		else if(strcmp(theline, line+10) != 0) {
+			line[n] = '\0';
+			diag("%s: object is [%s] expected [%s]", pn, line+10, theline);
+			free(t);
+			return;
+		}
+	}
 	free(t);
 	line[n] = '\n';
 
diff --git a/src/pkg/runtime/string.goc b/src/pkg/runtime/string.goc
index 53cd84c6e6..0e2dddda34 100644
--- a/src/pkg/runtime/string.goc
+++ b/src/pkg/runtime/string.goc
@@ -278,7 +278,7 @@ func stringtoslicebyte(s String) (b Slice) {
 	runtime·memmove(b.array, s.str, s.len);
 }
 
-func sliceinttostring(b Slice) (s String) {
+func slicerunetostring(b Slice) (s String) {
 	int32 siz1, siz2, i;
 	int32 *a;
 	byte dum[8];
@@ -301,13 +301,13 @@ func sliceinttostring(b Slice) (s String) {
 	s.str[s.len] = 0;
 }
 
-func stringtosliceint(s String) (b Slice) {
+func stringtoslicerune(s String) (b Slice) {
 	int32 n;
 	int32 dum, *r;
 	uint8 *p, *ep;
 
 	// two passes.
-	// unlike sliceinttostring, no race because strings are immutable.
+	// unlike slicerunetostring, no race because strings are immutable.
 	p = s.str;
 	ep = s.str+s.len;
 	n = 0;