cmd/go: case-encode versions as well as module paths in files, URLs

While writing the GOPROXY docs it occurred to me that versions
can contain upper-case letters as well. The docs therefore say
that versions are case-encoded the same as paths in the proxy
protocol (and therefore in the cache as well). Make it so.

Change-Id: Ibc0c4af0192a4af251e5dd6f2d36cda7e529099a
Reviewed-on: https://go-review.googlesource.com/124795
Reviewed-by: Bryan C. Mills <bcmills@google.com>

src/cmd/go/internal/modfetch/cache.go

@@ -46,7 +46,11 @@ func CachePath(m module.Version, suffix string) (string, error) {
 	if module.CanonicalVersion(m.Version) != m.Version {
 		return "", fmt.Errorf("non-canonical module version %q", m.Version)
 	}
-	return filepath.Join(dir, m.Version+"."+suffix), nil
+	encVer, err := module.EncodeVersion(m.Version)
+	if err != nil {
+		return "", err
+	}
+	return filepath.Join(dir, encVer+"."+suffix), nil
 }
 
 func DownloadDir(m module.Version) (string, error) {
@@ -63,7 +67,11 @@ func DownloadDir(m module.Version) (string, error) {
 	if module.CanonicalVersion(m.Version) != m.Version {
 		return "", fmt.Errorf("non-canonical module version %q", m.Version)
 	}
-	return filepath.Join(SrcMod, enc+"@"+m.Version), nil
+	encVer, err := module.EncodeVersion(m.Version)
+	if err != nil {
+		return "", err
+	}
+	return filepath.Join(SrcMod, enc+"@"+encVer), nil
 }
 
 // A cachingRepo is a cache around an underlying Repo,

src/cmd/go/internal/modfetch/proxy.go

@@ -163,7 +163,11 @@ func (p *proxyRepo) latest() (*RevInfo, error) {
 
 func (p *proxyRepo) Stat(rev string) (*RevInfo, error) {
 	var data []byte
-	err := webGetBytes(p.url+"/@v/"+pathEscape(rev)+".info", &data)
+	encRev, err := module.EncodeVersion(rev)
+	if err != nil {
+		return nil, err
+	}
+	err = webGetBytes(p.url+"/@v/"+pathEscape(encRev)+".info", &data)
 	if err != nil {
 		return nil, err
 	}
@@ -191,7 +195,11 @@ func (p *proxyRepo) Latest() (*RevInfo, error) {
 
 func (p *proxyRepo) GoMod(version string) ([]byte, error) {
 	var data []byte
-	err := webGetBytes(p.url+"/@v/"+pathEscape(version)+".mod", &data)
+	encVer, err := module.EncodeVersion(version)
+	if err != nil {
+		return nil, err
+	}
+	err = webGetBytes(p.url+"/@v/"+pathEscape(encVer)+".mod", &data)
 	if err != nil {
 		return nil, err
 	}
@@ -200,7 +208,11 @@ func (p *proxyRepo) GoMod(version string) ([]byte, error) {
 
 func (p *proxyRepo) Zip(version string, tmpdir string) (tmpfile string, err error) {
 	var body io.ReadCloser
-	err = webGetBody(p.url+"/@v/"+pathEscape(version)+".zip", &body)
+	encVer, err := module.EncodeVersion(version)
+	if err != nil {
+		return "", err
+	}
+	err = webGetBody(p.url+"/@v/"+pathEscape(encVer)+".zip", &body)
 	if err != nil {
 		return "", err
 	}

src/cmd/go/internal/module/module.go

@@ -433,8 +433,22 @@ func EncodePath(path string) (encoding string, err error) {
 		return "", err
 	}
 
+	return encodeString(path)
+}
+
+// EncodeVersion returns the safe encoding of the given module version.
+// Versions are allowed to be in non-semver form but must be valid file names
+// and not contain exclamation marks.
+func EncodeVersion(v string) (encoding string, err error) {
+	if err := checkElem(v, true); err != nil || strings.Contains(v, "!") {
+		return "", fmt.Errorf("disallowed version string %q", v)
+	}
+	return encodeString(v)
+}
+
+func encodeString(s string) (encoding string, err error) {
 	haveUpper := false
-	for _, r := range path {
+	for _, r := range s {
 		if r == '!' || r >= utf8.RuneSelf {
 			// This should be disallowed by CheckPath, but diagnose anyway.
 			// The correctness of the encoding loop below depends on it.
@@ -446,11 +460,11 @@ func EncodePath(path string) (encoding string, err error) {
 	}
 
 	if !haveUpper {
-		return path, nil
+		return s, nil
 	}
 
 	var buf []byte
-	for _, r := range path {
+	for _, r := range s {
 		if 'A' <= r && r <= 'Z' {
 			buf = append(buf, '!', byte(r+'a'-'A'))
 		} else {
@@ -461,19 +475,45 @@ func EncodePath(path string) (encoding string, err error) {
 }
 
 // DecodePath returns the module path of the given safe encoding.
-// It fails if the encoding is invalid.
+// It fails if the encoding is invalid or encodes an invalid path.
 func DecodePath(encoding string) (path string, err error) {
+	path, ok := decodeString(encoding)
+	if !ok {
+		return "", fmt.Errorf("invalid module path encoding %q", encoding)
+	}
+	if err := CheckPath(path); err != nil {
+		return "", fmt.Errorf("invalid module path encoding %q: %v", encoding, err)
+	}
+	return path, nil
+}
+
+// DecodeVersion returns the version string for the given safe encoding.
+// It fails if the encoding is invalid or encodes an invalid version.
+// Versions are allowed to be in non-semver form but must be valid file names
+// and not contain exclamation marks.
+func DecodeVersion(encoding string) (v string, err error) {
+	v, ok := decodeString(encoding)
+	if !ok {
+		return "", fmt.Errorf("invalid version encoding %q", encoding)
+	}
+	if err := checkElem(v, true); err != nil {
+		return "", fmt.Errorf("disallowed version string %q", v)
+	}
+	return v, nil
+}
+
+func decodeString(encoding string) (string, bool) {
 	var buf []byte
 
 	bang := false
 	for _, r := range encoding {
 		if r >= utf8.RuneSelf {
-			goto BadEncoding
+			return "", false
 		}
 		if bang {
 			bang = false
 			if r < 'a' || 'z' < r {
-				goto BadEncoding
+				return "", false
 			}
 			buf = append(buf, byte(r+'A'-'a'))
 			continue
@@ -483,19 +523,12 @@ func DecodePath(encoding string) (path string, err error) {
 			continue
 		}
 		if 'A' <= r && r <= 'Z' {
-			goto BadEncoding
+			return "", false
 		}
 		buf = append(buf, byte(r))
 	}
 	if bang {
-		goto BadEncoding
+		return "", false
 	}
-	path = string(buf)
-	if err := CheckPath(path); err != nil {
-		return "", fmt.Errorf("invalid module path encoding %q: %v", encoding, err)
-	}
-	return path, nil
-
-BadEncoding:
-	return "", fmt.Errorf("invalid module path encoding %q", encoding)
+	return string(buf), true
 }

src/cmd/go/proxy_test.go

@@ -78,13 +78,18 @@ func readModList() {
 		if i < 0 {
 			continue
 		}
-		enc := strings.Replace(name[:i], "_", "/", -1)
-		path, err := module.DecodePath(enc)
+		encPath := strings.Replace(name[:i], "_", "/", -1)
+		path, err := module.DecodePath(encPath)
+		if err != nil {
+			fmt.Fprintf(os.Stderr, "go proxy_test: %v", err)
+			continue
+		}
+		encVers := name[i+1:]
+		vers, err := module.DecodeVersion(encVers)
 		if err != nil {
 			fmt.Fprintf(os.Stderr, "go proxy_test: %v", err)
 			continue
 		}
-		vers := name[i+1:]
 		modList = append(modList, module.Version{Path: path, Version: vers})
 	}
 }
@@ -132,7 +137,13 @@ func proxyHandler(w http.ResponseWriter, r *http.Request) {
 		http.NotFound(w, r)
 		return
 	}
-	vers, ext := file[:i], file[i+1:]
+	encVers, ext := file[:i], file[i+1:]
+	vers, err := module.DecodeVersion(encVers)
+	if err != nil {
+		fmt.Fprintf(os.Stderr, "go proxy_test: %v", err)
+		http.NotFound(w, r)
+		return
+	}
 
 	if codehost.AllHex(vers) {
 		var best string
@@ -239,9 +250,14 @@ func readArchive(path, vers string) *txtar.Archive {
 		fmt.Fprintf(os.Stderr, "go proxy: %v\n", err)
 		return nil
 	}
+	encVers, err := module.EncodeVersion(vers)
+	if err != nil {
+		fmt.Fprintf(os.Stderr, "go proxy: %v\n", err)
+		return nil
+	}
 
 	prefix := strings.Replace(enc, "/", "_", -1)
-	name := filepath.Join(cmdGoDir, "testdata/mod", prefix+"_"+vers+".txt")
+	name := filepath.Join(cmdGoDir, "testdata/mod", prefix+"_"+encVers+".txt")
 	a := archiveCache.Do(name, func() interface{} {
 		a, err := txtar.ParseFile(name)
 		if err != nil {

src/cmd/go/testdata/mod/rsc.io_!q!u!o!t!e_v1.5.2.txt

@@ -1,4 +1,4 @@
-rsc.io/quote@v2.0.0 && cp mod/rsc.io_quote_v0.0.0-20180709153244-fd906ed3b100.txt mod/rsc.io_quote_v2.0.0.txt
+rsc.io/QUOTE v1.5.2
 
 -- .mod --
 module rsc.io/QUOTE

src/cmd/go/testdata/mod/rsc.io_!q!u!o!t!e_v1.5.3-!p!r!e.txt

@@ -0,0 +1,88 @@
+rsc.io/QUOTE v1.5.3-PRE (sigh)
+
+-- .mod --
+module rsc.io/QUOTE
+
+require rsc.io/quote v1.5.2
+-- .info --
+{"Version":"v1.5.3-PRE","Name":"","Short":"","Time":"2018-07-15T16:25:34Z"}
+-- go.mod --
+module rsc.io/QUOTE
+
+require rsc.io/quote v1.5.2
+-- QUOTE/quote.go --
+// Copyright 2018 The Go 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 QUOTE COLLECTS LOUD SAYINGS.
+package QUOTE
+
+import (
+	"strings"
+
+	"rsc.io/quote"
+)
+
+// HELLO RETURNS A GREETING.
+func HELLO() string {
+	return strings.ToUpper(quote.Hello())
+}
+
+// GLASS RETURNS A USEFUL PHRASE FOR WORLD TRAVELERS.
+func GLASS() string {
+	return strings.ToUpper(quote.GLASS())
+}
+
+// GO RETURNS A GO PROVERB.
+func GO() string {
+	return strings.ToUpper(quote.GO())
+}
+
+// OPT RETURNS AN OPTIMIZATION TRUTH.
+func OPT() string {
+	return strings.ToUpper(quote.OPT())
+}
+-- QUOTE/quote_test.go --
+// Copyright 2018 The Go 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 QUOTE
+
+import (
+	"os"
+	"testing"
+)
+
+func init() {
+	os.Setenv("LC_ALL", "en")
+}
+
+func TestHELLO(t *testing.T) {
+	hello := "HELLO, WORLD"
+	if out := HELLO(); out != hello {
+		t.Errorf("HELLO() = %q, want %q", out, hello)
+	}
+}
+
+func TestGLASS(t *testing.T) {
+	glass := "I CAN EAT GLASS AND IT DOESN'T HURT ME."
+	if out := GLASS(); out != glass {
+		t.Errorf("GLASS() = %q, want %q", out, glass)
+	}
+}
+
+func TestGO(t *testing.T) {
+	go1 := "DON'T COMMUNICATE BY SHARING MEMORY, SHARE MEMORY BY COMMUNICATING."
+	if out := GO(); out != go1 {
+		t.Errorf("GO() = %q, want %q", out, go1)
+	}
+}
+
+func TestOPT(t *testing.T) {
+	opt := "IF A PROGRAM IS TOO SLOW, IT MUST HAVE A LOOP."
+	if out := OPT(); out != opt {
+		t.Errorf("OPT() = %q, want %q", out, opt)
+	}
+}

src/cmd/go/testdata/mod/rsc.io_quote_v2.0.0.txt

@@ -1,4 +1,4 @@
-rsc.io/quote@v2.0.0 && cp mod/rsc.io_quote_v0.0.0-20180709153244-fd906ed3b100.txt mod/rsc.io_quote_v2.0.0.txt
+rsc.io/quote@v2.0.0
 
 -- .mod --
 module "rsc.io/quote"

src/cmd/go/testdata/script/mod_case.txt

@@ -9,5 +9,12 @@ go list -f 'DIR {{.Dir}} DEPS {{.Deps}}' rsc.io/QUOTE/QUOTE
 stdout 'DEPS.*rsc.io/quote'
 stdout 'DIR.*!q!u!o!t!e'
 
+go get rsc.io/QUOTE@v1.5.3-PRE
+go list -m all
+stdout '^rsc.io/QUOTE v1.5.3-PRE'
+
+go list -f '{{.Dir}}' rsc.io/QUOTE/QUOTE
+stdout '!q!u!o!t!e@v1.5.3-!p!r!e'
+
 -- go.mod --
 module x