cgi: child support (e.g. Go CGI under Apache)

The http/cgi package now supports both being
a CGI host or being a CGI child process.

R=rsc, adg, bradfitzwork
CC=golang-dev
https://golang.org/cl/4245070

src/pkg/http/cgi/Makefile

@@ -6,6 +6,7 @@ include ../../../Make.inc
 
 TARG=http/cgi
 GOFILES=\
-	cgi.go\
+	child.go\
+	host.go\
 
 include ../../../Make.pkg

src/pkg/http/cgi/child.go

@@ -0,0 +1,202 @@
+// Copyright 2011 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.
+
+// This file implements CGI from the perspective of a child
+// process.
+
+package cgi
+
+import (
+	"bufio"
+	"fmt"
+	"http"
+	"io"
+	"os"
+	"strconv"
+	"strings"
+)
+
+// Request returns the HTTP request as represented in the current
+// environment. This assumes the current program is being run
+// by a web server in a CGI environment.
+func Request() (*http.Request, os.Error) {
+	return requestFromEnvironment(envMap(os.Environ()))
+}
+
+func envMap(env []string) map[string]string {
+	m := make(map[string]string)
+	for _, kv := range env {
+		if idx := strings.Index(kv, "="); idx != -1 {
+			m[kv[:idx]] = kv[idx+1:]
+		}
+	}
+	return m
+}
+
+// These environment variables are manually copied into Request
+var skipHeader = map[string]bool{
+	"HTTP_HOST":       true,
+	"HTTP_REFERER":    true,
+	"HTTP_USER_AGENT": true,
+}
+
+func requestFromEnvironment(env map[string]string) (*http.Request, os.Error) {
+	r := new(http.Request)
+	r.Method = env["REQUEST_METHOD"]
+	if r.Method == "" {
+		return nil, os.NewError("cgi: no REQUEST_METHOD in environment")
+	}
+	r.Close = true
+	r.Trailer = http.Header{}
+	r.Header = http.Header{}
+
+	r.Host = env["HTTP_HOST"]
+	r.Referer = env["HTTP_REFERER"]
+	r.UserAgent = env["HTTP_USER_AGENT"]
+
+	// CGI doesn't allow chunked requests, so these should all be accurate:
+	r.Proto = "HTTP/1.0"
+	r.ProtoMajor = 1
+	r.ProtoMinor = 0
+	r.TransferEncoding = nil
+
+	if lenstr := env["CONTENT_LENGTH"]; lenstr != "" {
+		clen, err := strconv.Atoi64(lenstr)
+		if err != nil {
+			return nil, os.NewError("cgi: bad CONTENT_LENGTH in environment: " + lenstr)
+		}
+		r.ContentLength = clen
+		r.Body = nopCloser{io.LimitReader(os.Stdin, clen)}
+	}
+
+	// Copy "HTTP_FOO_BAR" variables to "Foo-Bar" Headers
+	for k, v := range env {
+		if !strings.HasPrefix(k, "HTTP_") || skipHeader[k] {
+			continue
+		}
+		r.Header.Add(strings.Replace(k[5:], "_", "-", -1), v)
+	}
+
+	// TODO: cookies.  parsing them isn't exported, though.
+
+	if r.Host != "" {
+		// Hostname is provided, so we can reasonably construct a URL,
+		// even if we have to assume 'http' for the scheme.
+		r.RawURL = "http://" + r.Host + env["REQUEST_URI"]
+		url, err := http.ParseURL(r.RawURL)
+		if err != nil {
+			return nil, os.NewError("cgi: failed to parse host and REQUEST_URI into a URL: " + r.RawURL)
+		}
+		r.URL = url
+	}
+	// Fallback logic if we don't have a Host header or the URL
+	// failed to parse
+	if r.URL == nil {
+		r.RawURL = env["REQUEST_URI"]
+		url, err := http.ParseURL(r.RawURL)
+		if err != nil {
+			return nil, os.NewError("cgi: failed to parse REQUEST_URI into a URL: " + r.RawURL)
+		}
+		r.URL = url
+	}
+	return r, nil
+}
+
+// TODO: move this to ioutil or something.  It's copy/pasted way too often.
+type nopCloser struct {
+	io.Reader
+}
+
+func (nopCloser) Close() os.Error { return nil }
+
+// Serve executes the provided Handler on the currently active CGI
+// request, if any. If there's no current CGI environment
+// an error is returned. The provided handler may be nil to use
+// http.DefaultServeMux.
+func Serve(handler http.Handler) os.Error {
+	req, err := Request()
+	if err != nil {
+		return err
+	}
+	if handler == nil {
+		handler = http.DefaultServeMux
+	}
+	rw := &response{
+		req:    req,
+		header: make(http.Header),
+		bufw:   bufio.NewWriter(os.Stdout),
+	}
+	handler.ServeHTTP(rw, req)
+	if err = rw.bufw.Flush(); err != nil {
+		return err
+	}
+	return nil
+}
+
+type response struct {
+	req        *http.Request
+	header     http.Header
+	bufw       *bufio.Writer
+	headerSent bool
+}
+
+func (r *response) Flush() {
+	r.bufw.Flush()
+}
+
+func (r *response) RemoteAddr() string {
+	return os.Getenv("REMOTE_ADDR")
+}
+
+func (r *response) SetHeader(k, v string) {
+	if v == "" {
+		r.header.Del(k)
+	} else {
+		r.header.Set(k, v)
+	}
+}
+
+func (r *response) Write(p []byte) (n int, err os.Error) {
+	if !r.headerSent {
+		r.WriteHeader(http.StatusOK)
+	}
+	return r.bufw.Write(p)
+}
+
+func (r *response) WriteHeader(code int) {
+	if r.headerSent {
+		// Note: explicitly using Stderr, as Stdout is our HTTP output.
+		fmt.Fprintf(os.Stderr, "CGI attempted to write header twice on request for %s", r.req.URL)
+		return
+	}
+	r.headerSent = true
+	fmt.Fprintf(r.bufw, "Status: %d %s\r\n", code, http.StatusText(code))
+
+	// Set a default Content-Type
+	if _, hasType := r.header["Content-Type"]; !hasType {
+		r.header.Add("Content-Type", "text/html; charset=utf-8")
+	}
+
+	// TODO: add a method on http.Header to write itself to an io.Writer?
+	// This is duplicated code.
+	for k, vv := range r.header {
+		for _, v := range vv {
+			v = strings.Replace(v, "\n", "", -1)
+			v = strings.Replace(v, "\r", "", -1)
+			v = strings.TrimSpace(v)
+			fmt.Fprintf(r.bufw, "%s: %s\r\n", k, v)
+		}
+	}
+	r.bufw.Write([]byte("\r\n"))
+	r.bufw.Flush()
+}
+
+func (r *response) UsingTLS() bool {
+	// There's apparently a de-facto standard for this.
+	// http://docstore.mik.ua/orelly/linux/cgi/ch03_02.htm#ch03-35636
+	if s := os.Getenv("HTTPS"); s == "on" || s == "ON" || s == "1" {
+		return true
+	}
+	return false
+}

src/pkg/http/cgi/child_test.go

@@ -0,0 +1,83 @@
+// Copyright 2011 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.
+
+// Tests for CGI (the child process perspective)
+
+package cgi
+
+import (
+	"testing"
+)
+
+func TestRequest(t *testing.T) {
+	env := map[string]string{
+		"REQUEST_METHOD":  "GET",
+		"HTTP_HOST":       "example.com",
+		"HTTP_REFERER":    "elsewhere",
+		"HTTP_USER_AGENT": "goclient",
+		"HTTP_FOO_BAR":    "baz",
+		"REQUEST_URI":     "/path?a=b",
+		"CONTENT_LENGTH":  "123",
+	}
+	req, err := requestFromEnvironment(env)
+	if err != nil {
+		t.Fatalf("requestFromEnvironment: %v", err)
+	}
+	if g, e := req.UserAgent, "goclient"; e != g {
+		t.Errorf("expected UserAgent %q; got %q", e, g)
+	}
+	if g, e := req.Method, "GET"; e != g {
+		t.Errorf("expected Method %q; got %q", e, g)
+	}
+	if g, e := req.Header.Get("User-Agent"), ""; e != g {
+		// Tests that we don't put recognized headers in the map
+		t.Errorf("expected User-Agent %q; got %q", e, g)
+	}
+	if g, e := req.ContentLength, int64(123); e != g {
+		t.Errorf("expected ContentLength %d; got %d", e, g)
+	}
+	if g, e := req.Referer, "elsewhere"; e != g {
+		t.Errorf("expected Referer %q; got %q", e, g)
+	}
+	if req.Header == nil {
+		t.Fatalf("unexpected nil Header")
+	}
+	if g, e := req.Header.Get("Foo-Bar"), "baz"; e != g {
+		t.Errorf("expected Foo-Bar %q; got %q", e, g)
+	}
+	if g, e := req.RawURL, "http://example.com/path?a=b"; e != g {
+		t.Errorf("expected RawURL %q; got %q", e, g)
+	}
+	if g, e := req.URL.String(), "http://example.com/path?a=b"; e != g {
+		t.Errorf("expected URL %q; got %q", e, g)
+	}
+	if g, e := req.FormValue("a"), "b"; e != g {
+		t.Errorf("expected FormValue(a) %q; got %q", e, g)
+	}
+	if req.Trailer == nil {
+		t.Errorf("unexpected nil Trailer")
+	}
+}
+
+func TestRequestWithoutHost(t *testing.T) {
+	env := map[string]string{
+		"HTTP_HOST":      "",
+		"REQUEST_METHOD": "GET",
+		"REQUEST_URI":    "/path?a=b",
+		"CONTENT_LENGTH": "123",
+	}
+	req, err := requestFromEnvironment(env)
+	if err != nil {
+		t.Fatalf("requestFromEnvironment: %v", err)
+	}
+	if g, e := req.RawURL, "/path?a=b"; e != g {
+		t.Errorf("expected RawURL %q; got %q", e, g)
+	}
+	if req.URL == nil {
+		t.Fatalf("unexpected nil URL")
+	}
+	if g, e := req.URL.String(), "/path?a=b"; e != g {
+		t.Errorf("expected URL %q; got %q", e, g)
+	}
+}

src/pkg/http/cgi/host.go

@@ -2,6 +2,9 @@
 // Use of this source code is governed by a BSD-style
 // license that can be found in the LICENSE file.
 
+// This file implements the host side of CGI (being the webserver
+// parent process).
+
 // Package cgi implements CGI (Common Gateway Interface) as specified
 // in RFC 3875.
 //
@@ -12,6 +15,7 @@
 package cgi
 
 import (
+	"bytes"
 	"encoding/line"
 	"exec"
 	"fmt"
@@ -19,7 +23,7 @@ import (
 	"io"
 	"log"
 	"os"
-	"path"
+	"path/filepath"
 	"regexp"
 	"strconv"
 	"strings"
@@ -29,10 +33,12 @@ var trailingPort = regexp.MustCompile(`:([0-9]+)$`)
 
 // Handler runs an executable in a subprocess with a CGI environment.
 type Handler struct {
-	Path   string      // path to the CGI executable
-	Root   string      // root URI prefix of handler or empty for "/"
+	Path string // path to the CGI executable
+	Root string // root URI prefix of handler or empty for "/"
+
 	Env    []string    // extra environment variables to set, if any
 	Logger *log.Logger // optional log for errors or nil to use log.Print
+	Args   []string    // optional arguments to pass to child process
 }
 
 func (h *Handler) ServeHTTP(rw http.ResponseWriter, req *http.Request) {
@@ -73,9 +79,20 @@ func (h *Handler) ServeHTTP(rw http.ResponseWriter, req *http.Request) {
 		"SERVER_PORT=" + port,
 	}
 
-	for k, _ := range req.Header {
+	if len(req.Cookie) > 0 {
+		b := new(bytes.Buffer)
+		for idx, c := range req.Cookie {
+			if idx > 0 {
+				b.Write([]byte("; "))
+			}
+			fmt.Fprintf(b, "%s=%s", c.Name, c.Value)
+		}
+		env = append(env, "HTTP_COOKIE="+b.String())
+	}
+
+	for k, v := range req.Header {
 		k = strings.Map(upperCaseAndUnderscore, k)
-		env = append(env, "HTTP_"+k+"="+req.Header.Get(k))
+		env = append(env, "HTTP_"+k+"="+strings.Join(v, ", "))
 	}
 
 	if req.ContentLength > 0 {
@@ -89,15 +106,17 @@ func (h *Handler) ServeHTTP(rw http.ResponseWriter, req *http.Request) {
 		env = append(env, h.Env...)
 	}
 
-	// TODO: use filepath instead of path when available
-	cwd, pathBase := path.Split(h.Path)
+	cwd, pathBase := filepath.Split(h.Path)
 	if cwd == "" {
 		cwd = "."
 	}
 
+	args := []string{h.Path}
+	args = append(args, h.Args...)
+
 	cmd, err := exec.Run(
 		pathBase,
-		[]string{h.Path},
+		args,
 		env,
 		cwd,
 		exec.Pipe,        // stdin

src/pkg/http/cgi/host_test.go

@@ -176,6 +176,28 @@ func TestPathInfoDirRoot(t *testing.T) {
 	runCgiTest(t, h, "GET /myscript/bar?a=b HTTP/1.0\nHost: example.com\n\n", expectedMap)
 }
 
+func TestDupHeaders(t *testing.T) {
+	if skipTest(t) {
+		return
+	}
+	h := &Handler{
+		Path: "testdata/test.cgi",
+	}
+	expectedMap := map[string]string{
+		"env-REQUEST_URI":     "/myscript/bar?a=b",
+		"env-SCRIPT_FILENAME": "testdata/test.cgi",
+		"env-HTTP_COOKIE":     "nom=NOM; yum=YUM",
+		"env-HTTP_X_FOO":      "val1, val2",
+	}
+	runCgiTest(t, h, "GET /myscript/bar?a=b HTTP/1.0\n"+
+		"Cookie: nom=NOM\n"+
+		"Cookie: yum=YUM\n"+
+		"X-Foo: val1\n"+
+		"X-Foo: val2\n"+
+		"Host: example.com\n\n",
+		expectedMap)
+}
+
 func TestPathInfoNoRoot(t *testing.T) {
 	if skipTest(t) {
 		return

src/pkg/http/cgi/matryoshka_test.go

@@ -0,0 +1,74 @@
+// Copyright 2011 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.
+
+// Tests a Go CGI program running under a Go CGI host process.
+// Further, the two programs are the same binary, just checking
+// their environment to figure out what mode to run in.
+
+package cgi
+
+import (
+	"fmt"
+	"http"
+	"os"
+	"testing"
+)
+
+// This test is a CGI host (testing host.go) that runs its own binary
+// as a child process testing the other half of CGI (child.go).
+func TestHostingOurselves(t *testing.T) {
+	h := &Handler{
+		Path: os.Args[0],
+		Root: "/test.go",
+		Args: []string{"-test.run=TestBeChildCGIProcess"},
+	}
+	expectedMap := map[string]string{
+		"test":                  "Hello CGI-in-CGI",
+		"param-a":               "b",
+		"param-foo":             "bar",
+		"env-GATEWAY_INTERFACE": "CGI/1.1",
+		"env-HTTP_HOST":         "example.com",
+		"env-PATH_INFO":         "",
+		"env-QUERY_STRING":      "foo=bar&a=b",
+		"env-REMOTE_ADDR":       "1.2.3.4",
+		"env-REMOTE_HOST":       "1.2.3.4",
+		"env-REQUEST_METHOD":    "GET",
+		"env-REQUEST_URI":       "/test.go?foo=bar&a=b",
+		"env-SCRIPT_FILENAME":   os.Args[0],
+		"env-SCRIPT_NAME":       "/test.go",
+		"env-SERVER_NAME":       "example.com",
+		"env-SERVER_PORT":       "80",
+		"env-SERVER_SOFTWARE":   "go",
+	}
+	replay := runCgiTest(t, h, "GET /test.go?foo=bar&a=b HTTP/1.0\nHost: example.com\n\n", expectedMap)
+
+	if expected, got := "text/html; charset=utf-8", replay.Header.Get("Content-Type"); got != expected {
+		t.Errorf("got a Content-Type of %q; expected %q", got, expected)
+	}
+	if expected, got := "X-Test-Value", replay.Header.Get("X-Test-Header"); got != expected {
+		t.Errorf("got a X-Test-Header of %q; expected %q", got, expected)
+	}
+}
+
+// Note: not actually a test.
+func TestBeChildCGIProcess(t *testing.T) {
+	if os.Getenv("REQUEST_METHOD") == "" {
+		// Not in a CGI environment; skipping test.
+		return
+	}
+	Serve(http.HandlerFunc(func(rw http.ResponseWriter, req *http.Request) {
+		rw.SetHeader("X-Test-Header", "X-Test-Value")
+		fmt.Fprintf(rw, "test=Hello CGI-in-CGI\n")
+		req.ParseForm()
+		for k, vv := range req.Form {
+			for _, v := range vv {
+				fmt.Fprintf(rw, "param-%s=%s\n", k, v)
+			}
+		}
+		for _, kv := range os.Environ() {
+			fmt.Fprintf(rw, "env-%s\n", kv)
+		}
+	}))
+	os.Exit(0)
+}

src/pkg/http/cgi/testdata/test.cgi

@@ -12,7 +12,7 @@ my $q = CGI->new;
 my $params = $q->Vars;
 
 my $NL = "\r\n";
-$NL = "\n" if 1 || $params->{mode} eq "NL";
+$NL = "\n" if $params->{mode} eq "NL";
 
 my $p = sub {
   print "$_[0]$NL";

src/pkg/http/server.go

@@ -6,7 +6,6 @@
 
 // TODO(rsc):
 //	logging
-//	cgi support
 //	post support
 
 package http