From 42b7789a9277e7626e4ddaadcbaefb7689d9d8d4 Mon Sep 17 00:00:00 2001
From: Russ Cox <rsc@golang.org>
Date: Wed, 24 Sep 2008 15:26:55 -0700
Subject: [PATCH] early HTTP library and trivial server

R=r
OCL=15777
CL=15791
---
 src/lib/http/Makefile   |  59 ++++++++
 src/lib/http/conn.go    |  60 ++++++++
 src/lib/http/request.go | 300 ++++++++++++++++++++++++++++++++++++++++
 src/lib/http/server.go  |  65 +++++++++
 src/lib/http/triv.go    |  30 ++++
 src/lib/http/url.go     | 178 ++++++++++++++++++++++++
 6 files changed, 692 insertions(+)
 create mode 100644 src/lib/http/Makefile
 create mode 100644 src/lib/http/conn.go
 create mode 100644 src/lib/http/request.go
 create mode 100644 src/lib/http/server.go
 create mode 100644 src/lib/http/triv.go
 create mode 100644 src/lib/http/url.go

diff --git a/src/lib/http/Makefile b/src/lib/http/Makefile
new file mode 100644
index 00000000000..12153c40e37
--- /dev/null
+++ b/src/lib/http/Makefile
@@ -0,0 +1,59 @@
+# Copyright 2009 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.
+
+# DO NOT EDIT.  Automatically generated by gobuild.
+# gobuild -m http conn.go request.go server.go url.go
+O=6
+GC=$(O)g
+CC=$(O)c -w
+AS=$(O)a
+AR=$(O)ar
+
+PKG=$(GOROOT)/pkg/http.a
+
+install: $(PKG)
+
+nuke: clean
+	rm -f $(PKG)
+
+clean:
+	rm -f *.$O *.a
+
+%.$O: %.go
+	$(GC) $*.go
+
+%.$O: %.c
+	$(CC) $*.c
+
+%.$O: %.s
+	$(AS) $*.s
+
+
+O1=\
+	url.$O\
+
+O2=\
+	request.$O\
+
+O3=\
+	conn.$O\
+
+O4=\
+	server.$O\
+
+$(PKG): a1 a2 a3 a4
+a1:	$(O1)
+	$(AR) grc $(PKG) $(O1)
+a2:	$(O2)
+	$(AR) grc $(PKG) $(O2)
+a3:	$(O3)
+	$(AR) grc $(PKG) $(O3)
+a4:	$(O4)
+	$(AR) grc $(PKG) $(O4)
+
+$(O1): nuke
+$(O2): a1
+$(O3): a2
+$(O4): a3
+
diff --git a/src/lib/http/conn.go b/src/lib/http/conn.go
new file mode 100644
index 00000000000..ad1b7cc8660
--- /dev/null
+++ b/src/lib/http/conn.go
@@ -0,0 +1,60 @@
+// Copyright 2009 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 http
+
+import (
+	"io";
+	"bufio";
+	"http";
+	"os"
+)
+
+// Read/write/close interface.
+type RWC interface {
+	Read(p *[]byte) (n int, err *os.Error);
+	Write(p *[]byte) (n int, err *os.Error);
+	Close() *os.Error
+}
+
+// Active HTTP connection (server side).
+export type Conn struct {
+	rwc RWC;
+	br *bufio.BufRead;
+	bw *bufio.BufWrite;
+	close bool;
+	chunking bool;
+}
+
+// Create new connection from rwc.
+export func NewConn(rwc RWC) (c *Conn, err *os.Error) {
+	c = new(Conn);
+	c.rwc = rwc;
+	if c.br, err = bufio.NewBufRead(rwc); err != nil {
+		return nil, err
+	}
+	if c.bw, err = bufio.NewBufWrite(rwc); err != nil {
+		return nil, err
+	}
+	return c, nil
+}
+
+// Read next request from connection.
+func (c *Conn) ReadRequest() (req *Request, err *os.Error) {
+	if req, err = ReadRequest(c.br); err != nil {
+		return nil, err
+	}
+
+	// TODO: Proper handling of (lack of) Connection: close,
+	// and chunked transfer encoding on output.
+	c.close = true
+	return req, nil
+}
+
+// Close the connection.
+func (c *Conn) Close() {
+	c.bw.Flush();
+	c.rwc.Close();
+}
+
diff --git a/src/lib/http/request.go b/src/lib/http/request.go
new file mode 100644
index 00000000000..bed911eb4c1
--- /dev/null
+++ b/src/lib/http/request.go
@@ -0,0 +1,300 @@
+// Copyright 2009 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.
+
+// HTTP Request reading and parsing.
+
+package http
+
+import (
+	"bufio";
+	"http";
+	"os";
+	"strings"
+)
+
+const (
+	MaxLineLength = 1024;	// assumed < bufio.DefaultBufSize
+	MaxValueLength = 1024;
+	MaxHeaderLines = 1024;
+)
+
+export var (
+	LineTooLong = os.NewError("http header line too long");
+	ValueTooLong = os.NewError("http header value too long");
+	HeaderTooLong = os.NewError("http header too long");
+	BadHeader = os.NewError("malformed http header");
+	BadRequest = os.NewError("invalid http request");
+	BadHTTPVersion = os.NewError("unsupported http version");
+)
+
+// HTTP Request
+export type Request struct {
+	method string;		// GET, PUT,etc.
+	rawurl string;
+	url *URL;		// URI after GET, PUT etc.
+	proto string;	// "HTTP/1.0"
+	pmajor int;	// 1
+	pminor int;	// 0
+
+	header *map[string] string;
+
+	close bool;
+	host string;
+	referer string;
+	useragent string;
+}
+
+// Read a line of bytes (up to \n) from b.
+// Give up if the line exceeds MaxLineLength.
+// The returned bytes are a pointer into storage in
+// the bufio, so they are only valid until the next bufio read.
+func ReadLineBytes(b *bufio.BufRead) (p *[]byte, err *os.Error) {
+	if p, err = b.ReadLineSlice('\n'); err != nil {
+		return nil, err
+	}
+	if len(p) >= MaxLineLength {
+		return nil, LineTooLong
+	}
+
+	// Chop off trailing white space.
+	var i int
+	for i = len(p); i > 0; i-- {
+		if c := p[i-1]; c != ' ' && c != '\r' && c != '\t' && c != '\n' {
+			break
+		}
+	}
+	return p[0:i], nil
+}
+
+// ReadLineByte, but convert the bytes into a string.
+func ReadLine(b *bufio.BufRead) (s string, err *os.Error) {
+	p, e := ReadLineBytes(b)
+	if e != nil {
+		return "", e
+	}
+	return string(p), nil
+}
+
+// Read a key/value pair from b.
+// A key/value has the form Key: Value\r\n
+// and the Value can continue on multiple lines if each continuation line
+// starts with a space.
+func ReadKeyValue(b *bufio.BufRead) (key, value string, err *os.Error) {
+	line, e := ReadLineBytes(b)
+	if e != nil {
+		return "", "", e
+	}
+	if len(line) == 0 {
+		return "", "", nil
+	}
+
+	// Scan first line for colon.
+	for i := 0; i < len(line); i++ {
+		switch line[i] {
+		case ' ':
+			// Key field has space - no good.
+			return "", "", BadHeader
+		case ':':
+			key = string(line[0:i]);
+			// Skip initial space before value.
+			for i++; i < len(line); i++ {
+				if line[i] != ' ' {
+					break
+				}
+			}
+			value = string(line[i:len(line)])
+
+			// Look for extension lines, which must begin with space.
+			for {
+				var c byte;
+
+				if c, e = b.ReadByte(); e != nil {
+					return "", "", e
+				}
+				if c != ' ' {
+					// Not leading space; stop.
+					b.UnreadByte()
+					break
+				}
+
+				// Eat leading space.
+				for c == ' ' {
+					if c, e = b.ReadByte(); e != nil {
+						return "", "", e
+					}
+				}
+				b.UnreadByte()
+
+				// Read the rest of the line and add to value.
+				if line, e = ReadLineBytes(b); e != nil {
+					return "", "", e
+				}
+				value += " " + string(line)
+
+				if len(value) >= MaxValueLength {
+					return "", "", ValueTooLong
+				}
+			}
+			return key, value, nil
+		}
+	}
+
+	// Line ended before space or colon.
+	return "", "", BadHeader;
+}
+
+// Convert decimal at s[i:len(s)] to integer,
+// returning value, string position where the digits stopped,
+// and whether there was a valid number (digits, not too big).
+func atoi(s string, i int) (n, i1 int, ok bool) {
+	const Big = 1000000
+	if i >= len(s) || s[i] < '0' || s[i] > '9' {
+		return 0, 0, false
+	}
+	n = 0
+	for ; i < len(s) && '0' <= s[i] && s[i] <= '9'; i++ {
+		n = n*10 + int(s[i]-'0')
+		if n > Big {
+			return 0, 0, false
+		}
+	}
+	return n, i, true
+}
+
+// Parse HTTP version: "HTTP/1.2" -> (1, 2, true).
+func ParseHTTPVersion(vers string) (int, int, bool) {
+	if vers[0:5] != "HTTP/" {
+		return 0, 0, false
+	}
+	major, i, ok := atoi(vers, 5)
+	if !ok || i >= len(vers) || vers[i] != '.' {
+		return 0, 0, false
+	}
+	var minor int;
+	minor, i, ok = atoi(vers, i+1)
+	if !ok || i != len(vers) {
+		return 0, 0, false
+	}
+	return major, minor, true
+}
+
+// Read and parse a request from b.
+export func ReadRequest(b *bufio.BufRead) (req *Request, err *os.Error) {
+	req = new(Request);
+
+	// First line: GET /index.html HTTP/1.0
+	var s string
+	if s, err = ReadLine(b); err != nil {
+		return nil, err
+	}
+
+	var f *[]string
+	if f = strings.split(s, " "); len(f) != 3 {
+		return nil, BadRequest
+	}
+	req.method, req.rawurl, req.proto = f[0], f[1], f[2]
+	var ok bool;
+	if req.pmajor, req.pminor, ok = ParseHTTPVersion(req.proto); !ok {
+		return nil, BadHTTPVersion
+	}
+
+	if req.url, err = ParseURL(req.rawurl); err != nil {
+		return nil, err
+	}
+
+	// Subsequent lines: Key: value.
+	nheader := 0;
+	req.header = new(map[string] string)
+	for {
+		var key, value string
+		if key, value, err = ReadKeyValue(b); err != nil {
+			return nil, err
+		}
+		if key == "" {
+			break
+		}
+		if nheader++; nheader >= MaxHeaderLines {
+			return nil, HeaderTooLong
+		}
+
+		// RFC 2616 says that if you send the same header key
+		// multiple times, it has to be semantically equivalent
+		// to concatenating the values separated by commas.
+		oldvalue, present := req.header[key]
+		if present {
+			req.header[key] = oldvalue+","+value
+		} else {
+			req.header[key] = value
+		}
+	}
+
+	// RFC2616: Must treat
+	//	GET /index.html HTTP/1.1
+	//	Host: www.google.com
+	// and
+	//	GET http://www.google.com/index.html HTTP/1.1
+	//	Host: doesntmatter
+	// the same.  In the second case, any Host line is ignored.
+	if v, have := req.header["Host"]; have && req.url.host == "" {
+		req.host = v
+	}
+
+	// RFC2616: Should treat
+	//	Pragma: no-cache
+	// like
+	//	Cache-control: no-cache
+	if v, have := req.header["Pragma"]; have && v == "no-cache" {
+		if cc, havecc := req.header["Cache-control"]; !havecc {
+			req.header["Cache-control"] = "no-cache"
+		}
+	}
+
+	// Determine whether to hang up after sending the reply.
+	if req.pmajor < 1 || (req.pmajor == 1 && req.pminor < 1) {
+		req.close = true
+	} else if v, have := req.header["Connection"]; have {
+		// TODO: Should split on commas, toss surrounding white space,
+		// and check each field.
+		if v == "close" {
+			req.close = true
+		}
+	}
+
+	// Pull out useful fields as a convenience to clients.
+	if v, have := req.header["Referer"]; have {
+		req.referer = v
+	}
+	if v, have := req.header["User-Agent"]; have {
+		req.useragent = v
+	}
+
+
+	// TODO: Parse specific header values:
+	//	Accept
+	//	Accept-Encoding
+	//	Accept-Language
+	//	Authorization
+	//	Cache-Control
+	//	Connection
+	//	Date
+	//	Expect
+	//	From
+	//	If-Match
+	//	If-Modified-Since
+	//	If-None-Match
+	//	If-Range
+	//	If-Unmodified-Since
+	//	Max-Forwards
+	//	Proxy-Authorization
+	//	Referer [sic]
+	//	TE (transfer-codings)
+	//	Trailer
+	//	Transfer-Encoding
+	//	Upgrade
+	//	User-Agent
+	//	Via
+	//	Warning
+
+	return req, nil;
+}
diff --git a/src/lib/http/server.go b/src/lib/http/server.go
new file mode 100644
index 00000000000..0c9af6c0a4c
--- /dev/null
+++ b/src/lib/http/server.go
@@ -0,0 +1,65 @@
+// Copyright 2009 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.
+
+// Trivial HTTP server
+
+// TODO: Routines for writing responses.
+
+package http
+
+import (
+	"io";
+	"os";
+	"net";
+	"http";
+	"strings"
+)
+
+// Serve a new connection.
+func ServeConnection(fd net.Conn, raddr string, f *(*Conn, *Request)) {
+	c, err := NewConn(fd)
+	if err != nil {
+		return
+	}
+	for {
+		req, err := c.ReadRequest()
+		if err != nil {
+			break
+		}
+		f(c, req)
+		if c.close {
+			break
+		}
+	}
+	c.Close();
+}
+
+// Web server: already listening on l, call f for each request.
+export func Serve(l net.Listener, f *(*Conn, *Request)) *os.Error {
+	// TODO: Make this unnecessary
+	s, e := os.Getenv("GOMAXPROCS");
+	if n, ok := strings.atoi(s); n < 3 {
+		print("Warning: $GOMAXPROCS needs to be at least 3.\n");
+	}
+
+	for {
+		rw, raddr, e := l.Accept()
+		if e != nil {
+			return e
+		}
+		go ServeConnection(rw, raddr, f)
+	}
+	panic("not reached")
+}
+
+// Web server: listen on address, call f for each request.
+export func ListenAndServe(addr string, f *(*Conn, *Request)) *os.Error {
+	l, e := net.Listen("tcp", addr)
+	if e != nil {
+		return e
+	}
+	e = Serve(l, f);
+	l.Close()
+	return e
+}
diff --git a/src/lib/http/triv.go b/src/lib/http/triv.go
new file mode 100644
index 00000000000..19485f9359f
--- /dev/null
+++ b/src/lib/http/triv.go
@@ -0,0 +1,30 @@
+// Copyright 2009 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 main
+
+import (
+	"io";
+	"bufio";
+	"os";
+	"net";
+	"http"
+)
+
+func Echo(conn *http.Conn, req *http.Request) {
+	fd := conn.bw;
+	conn.close = true;
+	io.WriteString(fd, "HTTP/1.1 200 OK\r\n"
+		"Content-Type: text/plain\r\n"
+		"\r\n");
+	io.WriteString(fd, req.method+" "+req.rawurl+" "+req.proto+"\r\n")
+}
+
+func main() {
+	err := http.ListenAndServe("0.0.0.0:12345", &Echo)
+	if err != nil {
+		panic("ListenAndServe: ", err.String())
+	}
+}
+
diff --git a/src/lib/http/url.go b/src/lib/http/url.go
new file mode 100644
index 00000000000..07470e68cab
--- /dev/null
+++ b/src/lib/http/url.go
@@ -0,0 +1,178 @@
+// Copyright 2009 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.
+
+// Parse URLs (actually URIs, but that seems overly pedantic).
+
+package http
+
+import (
+	"os"
+)
+
+export var (
+	BadURL = os.NewError("bad url syntax")
+)
+
+func IsHex(c byte) bool {
+	switch {
+	case '0' <= c && c <= '9':
+		return true
+	case 'a' <= c && c <= 'f':
+		return true
+	case 'A' <= c && c <= 'F':
+		return true
+	}
+	return false
+}
+
+func UnHex(c byte) byte {
+	switch {
+	case '0' <= c && c <= '9':
+		return c - '0'
+	case 'a' <= c && c <= 'f':
+		return c - 'a' + 10
+	case 'A' <= c && c <= 'F':
+		return c - 'A' + 10
+	}
+	return 0
+}
+
+// Unescape %xx into hex.
+export func URLUnescape(s string) (string, *os.Error) {
+	// Count %, check that they're well-formed.
+	n := 0
+	for i := 0; i < len(s); {
+		if s[i] == '%' {
+			n++
+			if !IsHex(s[i+1]) || !IsHex(s[i+2]) {
+				return "", BadURL
+			}
+			i += 3
+		} else {
+			i++
+		}
+	}
+
+	if n == 0 {
+		return s, nil
+	}
+
+	t := new([]byte, len(s)-2*n);
+	j := 0
+	for i := 0; i < len(s); {
+		if s[i] == '%' {
+			t[j] = UnHex(s[i+1]) << 4 | UnHex(s[i+2]);
+			j++
+			i += 3
+		} else {
+			t[j] = s[i];
+			j++
+			i++
+		}
+	}
+	return string(t), nil
+}
+
+export type URL struct {
+	raw string;
+	scheme string;
+	rawpath string;
+	authority string;
+	userinfo string;
+	host string;
+	path string;
+	query string;
+	fragment string;
+}
+
+// Maybe rawurl is of the form scheme:path.
+// (Scheme must be [a-zA-Z][a-zA-Z0-9+-.]*)
+// If so, return scheme, path; else return "", rawurl.
+func GetScheme(rawurl string) (scheme, path string, err *os.Error) {
+	for i := 0; i < len(rawurl); i++ {
+		c := rawurl[i];
+		switch {
+		case 'a' <= c && c <= 'z' ||'A' <= c && c <= 'Z':
+			// do nothing
+		case '0' <= c && c <= '9' || c == '+' || c == '-' || c == '.':
+			if i == 0 {
+				return "", rawurl, nil
+			}
+		case c == ':':
+			if i == 0 {
+				return "", "", BadURL
+			}
+			return rawurl[0:i], rawurl[i+1:len(rawurl)], nil
+		}
+	}
+	return "", rawurl, nil
+}
+
+// Maybe s is of the form t c u.
+// If so, return t, c u (or t, u if cutc == true).
+// If not, return s, "".
+func Split(s string, c byte, cutc bool) (string, string) {
+	for i := 0; i < len(s); i++ {
+		if s[i] == c {
+			if cutc {
+				return s[0:i], s[i+1:len(s)]
+			}
+			return s[0:i], s[i:len(s)]
+		}
+	}
+	return s, ""
+}
+
+// Parse rawurl into a URL structure.
+export func ParseURL(rawurl string) (url *URL, err *os.Error) {
+	if rawurl == "" {
+		return nil, BadURL
+	}
+	url = new(URL);
+	url.raw = rawurl
+
+	// Split off possible leading "http:", "mailto:", etc.
+	var path string
+	if url.scheme, path, err = GetScheme(rawurl); err != nil {
+		return nil, err
+	}
+	url.rawpath = path
+
+	// RFC 2396: a relative URI (no scheme) has a ?query,
+	// but absolute URIs only have query if path begins with /
+	if url.scheme == "" || len(path) > 0 && path[0] == '/' {
+		path, url.query = Split(path, '?', true);
+		if url.query, err = URLUnescape(url.query); err != nil {
+			return nil, err
+		}
+	}
+
+	// Maybe path is //authority/path
+	if len(path) > 2 && path[0:2] == "//" {
+		url.authority, path = Split(path[2:len(path)], '/', false)
+	}
+	url.userinfo, url.host = Split(url.authority, '@', true);
+
+	// What's left is the path.
+	// TODO: Canonicalize (remove . and ..)?
+	if url.path, err = URLUnescape(url.path); err != nil {
+		return nil, err
+	}
+
+	return url, nil
+}
+
+// A URL reference is a URL with #frag potentially added.  Parse it.
+export func ParseURLReference(rawurlref string) (url *URL, err *os.Error) {
+	// Cut off #frag.
+	rawurl, frag := Split(rawurlref, '#', true);
+	if url, err = ParseURL(rawurl); err != nil {
+		return nil, err
+	}
+	if url.fragment, err = URLUnescape(frag); err != nil {
+		return nil, err
+	}
+	return url, nil
+}
+