build: remove tmp dir names from objects, support GOROOT_FINAL again

If we compile a generated file stored in a temporary
directory - let's say /tmp/12345/work/x.c - then by default
6c stores the full path and then the pcln table in the
final binary includes the full path. This makes repeated builds
(using different temporary directories) produce different
binaries, even if the inputs are the same.

In the old 'go tool pack', the P flag specified a prefix to remove
from all stored paths (if present), and cmd/go invoked
'go tool pack grcP $WORK' to remove references to the
temporary work directory.

We've changed the build to avoid pack as much as possible,
under the theory that instead of making pack convert from
.6 to .a, the tools should just write the .a directly and save a
round of I/O.

Instead of going back to invoking pack always, define a common
flag -trimpath in the assemblers, C compilers, and Go compilers,
implemented in liblink, and arrange for cmd/go to use the flag.
Then the object files being written out have the shortened paths
from the start.

While we are here, reimplement pcln support for GOROOT_FINAL.
A build in /tmp/go uses GOROOT=/tmp/go, but if GOROOT_FINAL=/usr/local/go
is set, then a source file named /tmp/go/x.go is recorded instead as
/usr/local/go/x.go. We use this so that we can prepare distributions
to be installed in /usr/local/go without actually working in that
directory. The conversion to liblink deleted all the old file name
handling code, including the GOROOT_FINAL translation.
Bring the GOROOT_FINAL translation back.

Before this CL, using GOROOT_FINAL=/goroot make.bash:

        g% strings $(which go) | grep -c $TMPDIR
        6
        g% strings $(which go) | grep -c $GOROOT
        793
        g%

After this CL:

        g% strings $(which go) | grep -c $TMPDIR
        0
        g% strings $(which go) | grep -c $GOROOT
        0
        g%

(The references to $TMPDIR tend to be cgo-generated source files.)

Adding the -trimpath flag to the assemblers required converting
them to the new Go-semantics flag parser. The text in go1.3.html
is copied and adjusted from go1.1.html, which is when we applied
that conversion to the compilers and linkers.

Fixes #6989.

LGTM=iant
R=r, iant
CC=golang-codereviews
https://golang.org/cl/88300045

doc/go1.3.html

@@ -145,6 +145,19 @@ Finally, the go command now supports packages that import Objective-C
 files (suffixed <code>.m</code>) through cgo.
 </p>
 
+<h3 id="gc_flag">Command-line flag parsing</h3>
+
+<p>
+In the gc tool chain, the assemblers now use the
+same command-line flag parsing rules as the Go flag package, a departure
+from the traditional Unix flag parsing. This may affect scripts that invoke
+the tool directly.
+For example,
+<code>go tool 6a -SDfoo</code> must now be written
+<code>go tool 6a -S -D foo</code>.
+(The same change was made to the compilers and linkers in <a href="/doc/go1.1#gc_flag">Go 1.1</a>.)
+</p>
+
 <h3 id="misc">Miscellany</h3>
 
 <p>

include/link.h

@@ -360,6 +360,9 @@ struct	Link
 	Biobuf*	bso;	// for -v flag
 	char*	pathname;
 	int32	windows;
+	char*	trimpath;
+	char*	goroot;
+	char*	goroot_final;
 
 	// hash table of all symbols
 	LSym*	hash[LINKHASH];

src/cmd/5a/a.h

@@ -97,7 +97,7 @@ enum
 	Always	= 14,
 };
 
-EXTERN	char	debug[256];
+EXTERN	int	debug[256];
 EXTERN	Sym*	hash[NHASH];
 EXTERN	char**	Dlist;
 EXTERN	int	nDlist;

src/cmd/5a/lex.c

@@ -57,11 +57,27 @@ Lconv(Fmt *fp)
 	return linklinefmt(ctxt, fp);
 }
 
+void
+dodef(char *p)
+{
+	if(nDlist%8 == 0)
+		Dlist = allocn(Dlist, nDlist*sizeof(char *),
+			8*sizeof(char *));
+	Dlist[nDlist++] = p;
+}
+
+void
+usage(void)
+{
+	print("usage: %ca [options] file.c...\n", thechar);
+	flagprint(1);
+	errorexit();
+}
+
 void
 main(int argc, char *argv[])
 {
 	char *p;
-	int c;
 
 	thechar = '5';
 	thestring = "arm";
@@ -84,49 +100,24 @@ main(int argc, char *argv[])
 	cinit();
 	outfile = 0;
 	setinclude(".");
-	ARGBEGIN {
-	default:
-		c = ARGC();
-		if(c >= 0 && c < sizeof(debug))
-			debug[c] = 1;
-		break;
+	
+	flagfn1("D", "name[=value]: add #define", dodef);
+	flagfn1("I", "dir: add dir to include path", setinclude);
+	flagcount("S", "print assembly and machine code", &debug['S']);
+	flagcount("m", "debug preprocessor macros", &debug['m']);
+	flagstr("o", "file: set output file", &outfile);
+	flagstr("trimpath", "prefix: remove prefix from recorded source file paths", &ctxt->trimpath);
 
-	case 'o':
-		outfile = ARGF();
-		break;
+	flagparse(&argc, &argv, usage);
+	ctxt->debugasm = debug['S'];
 
-	case 'D':
-		p = ARGF();
-		if(p) {
-			if (nDlist%8 == 0) 
-				Dlist = allocn(Dlist, nDlist*sizeof(char *), 
-					8*sizeof(char *));
-			Dlist[nDlist++] = p;
-		}
-		break;
-
-	case 'I':
-		p = ARGF();
-		setinclude(p);
-		break;
-
-	case 't':
-		thechar = 't';
-		thestring = "thumb";
-		break;
-
-	case 'S':
-		ctxt->debugasm++;
-		break;
-	} ARGEND
-	if(*argv == 0) {
-		print("usage: %ca [-options] file.s\n", thechar);
-		errorexit();
-	}
+	if(argc < 1)
+		usage();
 	if(argc > 1){
 		print("can't assemble multiple files\n");
 		errorexit();
 	}
+
 	if(assemble(argv[0]))
 		errorexit();
 	Bflush(&bstdout);

src/cmd/6a/a.h

@@ -109,7 +109,7 @@ enum
 	CPREPROC,
 };
 
-EXTERN	char	debug[256];
+EXTERN	int	debug[256];
 EXTERN	Sym*	hash[NHASH];
 EXTERN	char**	Dlist;
 EXTERN	int	nDlist;

src/cmd/6a/lex.c

@@ -63,13 +63,29 @@ Lconv(Fmt *fp)
 	return linklinefmt(ctxt, fp);
 }
 
+void
+dodef(char *p)
+{
+	if(nDlist%8 == 0)
+		Dlist = allocn(Dlist, nDlist*sizeof(char *),
+			8*sizeof(char *));
+	Dlist[nDlist++] = p;
+}
+
 LinkArch*       thelinkarch = &linkamd64;
 
+void
+usage(void)
+{
+	print("usage: %ca [options] file.c...\n", thechar);
+	flagprint(1);
+	errorexit();
+}
+
 void
 main(int argc, char *argv[])
 {
 	char *p;
-	int c;
 
 	thechar = '6';
 	thestring = "amd64";
@@ -94,45 +110,24 @@ main(int argc, char *argv[])
 	cinit();
 	outfile = 0;
 	setinclude(".");
-	ARGBEGIN {
-	default:
-		c = ARGC();
-		if(c >= 0 && c < sizeof(debug))
-			debug[c] = 1;
-		break;
+	
+	flagfn1("D", "name[=value]: add #define", dodef);
+	flagfn1("I", "dir: add dir to include path", setinclude);
+	flagcount("S", "print assembly and machine code", &debug['S']);
+	flagcount("m", "debug preprocessor macros", &debug['m']);
+	flagstr("o", "file: set output file", &outfile);
+	flagstr("trimpath", "prefix: remove prefix from recorded source file paths", &ctxt->trimpath);
 
-	case 'o':
-		outfile = ARGF();
-		break;
+	flagparse(&argc, &argv, usage);
+	ctxt->debugasm = debug['S'];
 
-	case 'D':
-		p = ARGF();
-		if(p) {
-			if (nDlist%8 == 0)
-				Dlist = allocn(Dlist, nDlist*sizeof(char *), 
-					8*sizeof(char *));
-			Dlist[nDlist++] = p;
-		}
-		break;
-
-	case 'I':
-		p = ARGF();
-		setinclude(p);
-		break;
-
-	case 'S':
-		ctxt->debugasm++;
-		break;
-	} ARGEND
-
-	if(*argv == 0) {
-		print("usage: %ca [-options] file.s\n", thechar);
-		errorexit();
-	}
+	if(argc < 1)
+		usage();
 	if(argc > 1){
 		print("can't assemble multiple files\n");
 		errorexit();
 	}
+
 	if(assemble(argv[0]))
 		errorexit();
 	Bflush(&bstdout);

src/cmd/8a/a.h

@@ -109,7 +109,7 @@ enum
 	CPREPROC,
 };
 
-EXTERN	char	debug[256];
+EXTERN	int	debug[256];
 EXTERN	Sym*	hash[NHASH];
 EXTERN	char**	Dlist;
 EXTERN	int	nDlist;

src/cmd/8a/lex.c

@@ -63,11 +63,26 @@ Lconv(Fmt *fp)
 	return linklinefmt(ctxt, fp);
 }
 
+void
+dodef(char *p)
+{
+	if(nDlist%8 == 0)
+		Dlist = allocn(Dlist, nDlist*sizeof(char *),
+			8*sizeof(char *));
+	Dlist[nDlist++] = p;
+}
+
+void
+usage(void)
+{
+	print("usage: %ca [options] file.c...\n", thechar);
+	flagprint(1);
+	errorexit();
+}
 void
 main(int argc, char *argv[])
 {
 	char *p;
-	int c;
 
 	thechar = '8';
 	thestring = "386";
@@ -90,44 +105,24 @@ main(int argc, char *argv[])
 	cinit();
 	outfile = 0;
 	setinclude(".");
-	ARGBEGIN {
-	default:
-		c = ARGC();
-		if(c >= 0 && c < sizeof(debug))
-			debug[c] = 1;
-		break;
+	
+	flagfn1("D", "name[=value]: add #define", dodef);
+	flagfn1("I", "dir: add dir to include path", setinclude);
+	flagcount("S", "print assembly and machine code", &debug['S']);
+	flagcount("m", "debug preprocessor macros", &debug['m']);
+	flagstr("o", "file: set output file", &outfile);
+	flagstr("trimpath", "prefix: remove prefix from recorded source file paths", &ctxt->trimpath);
 
-	case 'o':
-		outfile = ARGF();
-		break;
+	flagparse(&argc, &argv, usage);
+	ctxt->debugasm = debug['S'];
 
-	case 'D':
-		p = ARGF();
-		if(p) {
-			if (nDlist%8 == 0)
-				Dlist = allocn(Dlist, nDlist*sizeof(char *), 
-					8*sizeof(char *));
-			Dlist[nDlist++] = p;
-		}
-		break;
-
-	case 'I':
-		p = ARGF();
-		setinclude(p);
-		break;
-
-	case 'S':
-		ctxt->debugasm++;
-		break;
-	} ARGEND
-	if(*argv == 0) {
-		print("usage: %ca [-options] file.s\n", thechar);
-		errorexit();
-	}
+	if(argc < 1)
+		usage();
 	if(argc > 1){
 		print("can't assemble multiple files\n");
 		errorexit();
 	}
+
 	if(assemble(argv[0]))
 		errorexit();
 	Bflush(&bstdout);

src/cmd/cc/lex.c

@@ -195,6 +195,7 @@ main(int argc, char *argv[])
 	flagcount("q", "print Go definitions", &debug['q']);
 	flagcount("s", "print #define assembly offsets", &debug['s']);
 	flagcount("t", "debug code generation", &debug['t']);
+	flagstr("trimpath", "prefix: remove prefix from recorded source file paths", &ctxt->trimpath);
 	flagcount("w", "enable warnings", &debug['w']);
 	flagcount("v", "increase debug verbosity", &debug['v']);	
 	if(thechar == '6')

src/cmd/gc/lex.c

@@ -306,6 +306,7 @@ main(int argc, char *argv[])
 	flagcount("r", "debug generated wrappers", &debug['r']);
 	flagcount("race", "enable race detector", &flag_race);
 	flagcount("s", "warn about composite literals that can be simplified", &debug['s']);
+	flagstr("trimpath", "prefix: remove prefix from recorded source file paths", &ctxt->trimpath);
 	flagcount("u", "reject unsafe code", &safemode);
 	flagcount("v", "increase debug verbosity", &debug['v']);
 	flagcount("w", "debug type checking", &debug['w']);

src/cmd/go/build.go

@@ -1599,7 +1599,7 @@ func (gcToolchain) gc(b *builder, p *Package, archive, obj string, importArgs []
 		gcargs = append(gcargs, "-installsuffix", buildContext.InstallSuffix)
 	}
 
-	args := stringList(tool(archChar+"g"), "-o", ofile, buildGcflags, gcargs, "-D", p.localPrefix, importArgs)
+	args := stringList(tool(archChar+"g"), "-o", ofile, "-trimpath", b.work, buildGcflags, gcargs, "-D", p.localPrefix, importArgs)
 	if ofile == archive {
 		args = append(args, "-pack")
 	}
@@ -1613,7 +1613,7 @@ func (gcToolchain) gc(b *builder, p *Package, archive, obj string, importArgs []
 
 func (gcToolchain) asm(b *builder, p *Package, obj, ofile, sfile string) error {
 	sfile = mkAbs(p.Dir, sfile)
-	return b.run(p.Dir, p.ImportPath, nil, tool(archChar+"a"), "-I", obj, "-o", ofile, "-D", "GOOS_"+goos, "-D", "GOARCH_"+goarch, sfile)
+	return b.run(p.Dir, p.ImportPath, nil, tool(archChar+"a"), "-trimpath", b.work, "-I", obj, "-o", ofile, "-D", "GOOS_"+goos, "-D", "GOARCH_"+goarch, sfile)
 }
 
 func (gcToolchain) pkgpath(basedir string, p *Package) string {
@@ -1626,7 +1626,7 @@ func (gcToolchain) pack(b *builder, p *Package, objDir, afile string, ofiles []s
 	for _, f := range ofiles {
 		absOfiles = append(absOfiles, mkAbs(objDir, f))
 	}
-	cmd := "grcP"
+	cmd := "c"
 	absAfile := mkAbs(objDir, afile)
 	appending := false
 	if _, err := os.Stat(absAfile); err == nil {
@@ -1784,7 +1784,7 @@ func (gcToolchain) ld(b *builder, p *Package, out string, allactions []*action,
 func (gcToolchain) cc(b *builder, p *Package, objdir, ofile, cfile string) error {
 	inc := filepath.Join(goroot, "pkg", fmt.Sprintf("%s_%s", goos, goarch))
 	cfile = mkAbs(p.Dir, cfile)
-	args := stringList(tool(archChar+"c"), "-F", "-V", "-w", "-I", objdir, "-I", inc, "-o", ofile, buildCcflags, "-D", "GOOS_"+goos, "-D", "GOARCH_"+goarch, cfile)
+	args := stringList(tool(archChar+"c"), "-F", "-V", "-w", "-trimpath", b.work, "-I", objdir, "-I", inc, "-o", ofile, buildCcflags, "-D", "GOOS_"+goos, "-D", "GOARCH_"+goarch, cfile)
 	return b.run(p.Dir, p.ImportPath, nil, args)
 }
 

src/liblink/obj.c

@@ -87,6 +87,34 @@ linklinefmt(Link *ctxt, Fmt *fp)
 	return 0;
 }
 
+// Does s have t as a path prefix?
+// That is, does s == t or does s begin with t followed by a slash?
+// For portability, we allow ASCII case folding, so that haspathprefix("a/b/c", "A/B") is true.
+// Similarly, we allow slash folding, so that haspathprefix("a/b/c", "a\\b") is true.
+static int
+haspathprefix(char *s, char *t)
+{
+	int i, cs, ct;
+
+	if(t == nil)
+		return 0;
+	for(i=0; t[i]; i++) {
+		cs = s[i];
+		ct = t[i];
+		if('A' <= cs && cs <= 'Z')
+			cs += 'a' - 'A';
+		if('A' <= ct && ct <= 'Z')
+			ct += 'a' - 'A';
+		if(cs == '\\')
+			cs = '/';
+		if(ct == '\\')
+			ct = '/';
+		if(cs != ct)
+			return 0;
+	}
+	return s[i] == '\0' || s[i] == '/' || s[i] == '\\';
+}
+
 // This is a simplified copy of linklinefmt above.
 // It doesn't allow printing the full stack, and it returns the file name and line number separately.
 // TODO: Unify with linklinefmt somehow.
@@ -103,7 +131,7 @@ linkgetline(Link *ctxt, int32 line, LSym **f, int32 *l)
 	int32 lno, d, dlno;
 	int n;
 	Hist *h;
-	char buf[1024], *file;
+	char buf[1024], buf1[1024], *file;
 
 	lno = line;
 	n = 0;
@@ -159,6 +187,22 @@ linkgetline(Link *ctxt, int32 line, LSym **f, int32 *l)
 		snprint(buf, sizeof buf, "%s", file);
 	else
 		snprint(buf, sizeof buf, "%s/%s", ctxt->pathname, file);
+
+	// Remove leading ctxt->trimpath, or else rewrite $GOROOT to $GOROOT_FINAL.
+	if(haspathprefix(buf, ctxt->trimpath)) {
+		if(strlen(buf) == strlen(ctxt->trimpath))
+			strcpy(buf, "??");
+		else {
+			snprint(buf1, sizeof buf1, "%s", buf+strlen(ctxt->trimpath)+1);
+			if(buf1[0] == '\0')
+				strcpy(buf1, "??");
+			strcpy(buf, buf1);
+		}
+	} else if(ctxt->goroot_final != nil && haspathprefix(buf, ctxt->goroot)) {
+		snprint(buf1, sizeof buf1, "%s%s", ctxt->goroot_final, buf+strlen(ctxt->goroot));
+		strcpy(buf, buf1);
+	}
+
 	lno -= dlno;
 	*f = linklookup(ctxt, buf, HistVersion);
 	*l = lno;

src/liblink/sym.c

@@ -95,6 +95,10 @@ linknew(LinkArch *arch)
 	ctxt = emallocz(sizeof *ctxt);
 	ctxt->arch = arch;
 	ctxt->version = HistVersion;
+	ctxt->goroot = getgoroot();
+	ctxt->goroot_final = getenv("GOROOT_FINAL");
+	if(ctxt->goroot_final != nil && ctxt->goroot_final[0] == '\0')
+		ctxt->goroot_final = nil;
 
 	p = getgoarch();
 	if(strcmp(p, arch->name) != 0)