From 0d32ab1356bd178925d15ba52df91daedfac2aad Mon Sep 17 00:00:00 2001
From: "Paul J. Davis" <paul.joseph.davis@gmail.com>
Date: Thu, 6 May 2010 18:14:45 -0400
Subject: [PATCH] Initial steps for importing simplehttp.

---
 gunicorn/http/__init__.py   |   3 +
 gunicorn/http/body.py       | 237 ++++++++++++++++++++++++++++++++++++
 gunicorn/http/errors.py     |  52 ++++++++
 gunicorn/http/message.py    | 187 ++++++++++++++++++++++++++++
 gunicorn/http/parser.py     |  38 ++++++
 gunicorn/http/unreader.py   |  73 +++++++++++
 gunicorn/{ => http}/wsgi.py |   0
 gunicorn/workers/sync.py    |  11 +-
 8 files changed, 595 insertions(+), 6 deletions(-)
 create mode 100644 gunicorn/http/__init__.py
 create mode 100644 gunicorn/http/body.py
 create mode 100644 gunicorn/http/errors.py
 create mode 100644 gunicorn/http/message.py
 create mode 100644 gunicorn/http/parser.py
 create mode 100644 gunicorn/http/unreader.py
 rename gunicorn/{ => http}/wsgi.py (100%)

diff --git a/gunicorn/http/__init__.py b/gunicorn/http/__init__.py
new file mode 100644
index 00000000..14e2939a
--- /dev/null
+++ b/gunicorn/http/__init__.py
@@ -0,0 +1,3 @@
+
+from message import Message, Request, Response
+from parser import Parser, RequestParser, ResponseParser
\ No newline at end of file
diff --git a/gunicorn/http/body.py b/gunicorn/http/body.py
new file mode 100644
index 00000000..06650aae
--- /dev/null
+++ b/gunicorn/http/body.py
@@ -0,0 +1,237 @@
+
+import re
+
+try:
+    from cStringIO import StringIO
+except ImportError:
+    from StringIO import StringIO
+
+from errors import *
+
+class ChunkedReader(object):
+    def __init__(self, req, unreader):
+        self.req = req
+        self.parser = self.parse_chunked(unreader)
+        self.buf = StringIO()
+    
+    def read(self, size=None):
+        if size == 0:
+            return ""
+        if size < 0:
+            size = None
+
+        if not self.parser:
+            return self.buf.getvalue()
+
+        while size is None or self.buf.tell() < size:
+            try:
+                self.buf.write(self.parser.next())
+            except StopIteration:
+                self.parser = None
+                break
+
+        if size is None or self.buf.tell() < size:
+            ret = self.buf.getvalue()
+            self.buf.truncate(0)
+            return ret
+        
+        data = self.buf.getvalue()
+        ret, rest = data[:size], data[size:]
+        self.buf.truncate(0)
+        self.buf.write(rest)
+        return ret
+    
+    def parse_trailers(self, unreader, data):
+        buf = StringIO()
+        buf.write(data)
+        
+        idx = buf.getvalue().find("\r\n\r\n")
+        done = buf.getvalue()[:2] == "\r\n"
+        while idx < 0 and not done:
+            self.get_data(unreader, buf)
+            idx = buf.getvalue().find("\r\n\r\n")
+            done = buf.getvalue()[:2] == "\r\n"
+        if done:
+            unreader.unread(buf.getvalue()[2:])
+            return ""
+        self.req.trailers = self.req.parse_headers(buf.getvalue()[:idx])
+        unreader.unread(buf.getvalue()[idx+4:])
+
+    def parse_chunked(self, unreader):
+        (size, rest) = self.parse_chunk_size(unreader)
+        while size > 0:
+            while size > len(rest):
+                size -= len(rest)
+                yield rest
+                rest = unreader.read()
+                if not rest:
+                    raise NoMoreData()
+            yield rest[:size]
+            # Remove \r\n after chunk
+            rest = rest[size:]
+            while len(rest) < 2:
+                rest += unreader.read()
+            if rest[:2] != '\r\n':
+                raise ChunkMissingTerminator(rest[:2])
+            (size, rest) = self.parse_chunk_size(unreader, data=rest[2:])          
+
+    def parse_chunk_size(self, unreader, data=None):
+        buf = StringIO()
+        if data is not None:
+            buf.write(data)
+
+        idx = buf.getvalue().find("\r\n")
+        while idx < 0:
+            self.get_data(unreader, buf)
+            idx = buf.getvalue().find("\r\n")
+
+        data = buf.getvalue()
+        line, rest_chunk = data[:idx], data[idx+2:]
+    
+        chunk_size = line.split(";", 1)[0].strip()
+        try:
+            chunk_size = int(chunk_size, 16)
+        except ValueError:
+            raise InvalidChunkSize(chunk_size)
+
+        if chunk_size == 0:
+            self.parse_trailers(unreader, rest_chunk)
+            return (0, None)
+        return (chunk_size, rest_chunk)
+
+    def get_data(self, unreader, buf):
+        data = unreader.read()
+        if not data:
+            raise NoMoreData()
+        buf.write(data)
+
+class LengthReader(object):
+    def __init__(self, unreader, length):
+        self.unreader = unreader
+        self.length = length
+    
+    def read(self, size=None):
+        if size is not None and not isinstance(size, (int, long)):
+            raise TypeError("size must be an integral type")
+
+        if size == 0 or self.length <= 0:
+            return ""
+        if size < 0 or size is None:
+            size = self.length
+        
+        buf = StringIO()
+        data = self.unreader.read()
+        while data:
+            buf.write(data)
+            if buf.tell() >= size:
+                break
+            data = self.unreader.read()
+        
+        buf = buf.getvalue()
+        ret, rest = buf[:size], buf[size:]
+        self.unreader.unread(rest)
+        self.length -= size
+        return ret
+
+class EOFReader(object):
+    def __init__(self, unreader):
+        self.unreader = unreader
+        self.buf = StringIO()
+        self.finished = False
+    
+    def read(self, size=None):
+        if size == 0 or self.finished:
+            return ""
+        if size < 0:
+            size = None
+        
+        data = self.unreader.read()
+        while data:
+            buf.write(data)
+            if size is not None and buf.tell() > size:
+                data = buf.getvalue()
+                ret, rest = data[:size], data[size:]
+                self.buf.truncate(0)
+                self.buf.write(rest)
+                return ret
+            data = self.unreader.read()
+
+        self.finished = True
+        ret = self.buf.getvalue()
+        self.buf.truncate(0)
+        return ret
+
+class Body(object):
+    def __init__(self, reader):
+        self.reader = reader
+        self.buf = StringIO()
+    
+    def __iter__(self):
+        return self
+    
+    def next(self):
+        ret = self.readline()
+        if not ret:
+            raise StopIteration()
+        return ret
+    
+    def read(self, size=None):
+        if size is not None and not isinstance(size, (int, long)):
+            raise TypeError("size must be an integral type")
+
+        if size is not None and size < self.buf.tell():
+            data = self.buf.getvalue()
+            ret, rest = data[:size], data[size:]
+            self.buf.truncate(0)
+            self.buf.write(rest)
+            return ret
+
+        if size > 0:
+            size -= self.buf.tell()
+        else:
+            size = None
+        
+        ret = self.buf.getvalue() + self.reader.read(size=size)
+        self.buf.truncate(0)
+        return ret
+    
+    def readline(self, size=None):
+        if size == 0:
+            return ""
+        if size < 0:
+            size = None
+        
+        idx = -1
+        while idx < 0:
+            data = self.reader.read(1024)
+            if not len(data):
+                break
+            self.buf.write(data)
+            if size is not None and self.buf.tell() > size:
+                break
+            idx = self.buf.getvalue().find("\r\n")
+
+        if idx < 0 and size is not None:
+            idx = size
+        elif idx < 0:
+            idx = self.buf.tell()
+        
+        data = self.buf.getvalue()
+        ret, rest = data[:idx], data[idx:]
+        self.buf.truncate(0)
+        self.buf.write(rest)
+        return ret
+    
+    def readlines(self, size=None):
+        ret = []
+        data = self.read()
+        while len(data):
+            pos = data.find("\n")
+            if pos < 0:
+                ret.append(data)
+                data = ""
+            else:
+                line, data = data[:pos+1], data[pos+1:]
+                ret.append(line)
+        return ret
+
diff --git a/gunicorn/http/errors.py b/gunicorn/http/errors.py
new file mode 100644
index 00000000..57cd9932
--- /dev/null
+++ b/gunicorn/http/errors.py
@@ -0,0 +1,52 @@
+
+class ParseException(Exception):
+    pass
+
+class NoMoreData(ParseException):
+    def __init__(self, buf):
+        self.buf = buf
+    def __str__(self):
+        return "No more data after: %r" % self.buf
+
+class InvalidRequestLine(ParseException):
+    def __init__(self, req):
+        self.req = req
+        self.code = 400
+
+    def __str__(self):
+        return "Invalid HTTP request line: %r" % self.req
+
+class InvalidRequestMethod(ParseException):
+    def __init__(self, method):
+        self.method = method
+
+    def __str__(self):
+        return "Invalid HTTP method: %r" % self.method
+
+class InvalidHeader(ParseException):
+    def __init__(self, hdr):
+        self.hdr = hdr
+    
+    def __str__(self):
+        return "Invalid HTTP Header: %r" % self.hdr
+
+class InvalidHeaderName(ParseException):
+    def __init__(self, hdr):
+        self.hdr = hdr
+
+    def __str__(self):
+        return "Invalid HTTP header name: %r" % self.hdr
+
+class InvalidChunkSize(ParseException):
+    def __init__(self, data):
+        self.data = data
+    
+    def __str__(self):
+        return "Invalid chunk size: %r" % self.data
+
+class ChunkMissingTerminator(ParseException):
+    def __init__(self, term):
+        self.term = term
+    
+    def __str__(self):
+        return "Invalid chunk terminator is not '\\r\\n': %r" % self.term
diff --git a/gunicorn/http/message.py b/gunicorn/http/message.py
new file mode 100644
index 00000000..cb703f83
--- /dev/null
+++ b/gunicorn/http/message.py
@@ -0,0 +1,187 @@
+
+import os
+import re
+import urlparse
+
+try:
+    from cStringIO import StringIO
+except ImportError:
+    from StringIO import StringIO
+
+from body import ChunkedReader, LengthReader, EOFReader, Body
+from errors import *
+
+class Message(object):
+    def __init__(self, unreader):
+        self.unreader = unreader
+        self.version = None
+        self.headers = []
+        self.trailers = []
+        self.body = None
+
+        self.hdrre = re.compile("[\x00-\x1F\x7F()<>@,;:\[\]={} \t\\\\\"]")
+
+        unused = self.parse(self.unreader)
+        self.unreader.unread(unused)
+        self.set_body_reader()
+    
+    def parse(self):
+        raise NotImplementedError()
+
+    def parse_headers(self, data):
+        headers = []
+
+        # Split lines on \r\n keeping the \r\n on each line
+        lines = []
+        while len(data):
+            pos = data.find("\r\n")
+            if pos < 0:
+                lines.append(data)
+                data = ""
+            else:
+                lines.append(data[:pos+2])
+                data = data[pos+2:]
+
+        # Parse headers into key/value pairs paying attention
+        # to continuation lines.
+        while len(lines):
+            # Parse initial header name : value pair.
+            curr = lines.pop(0)
+            if curr.find(":") < 0:
+                raise InvalidHeader(curr.strip())
+            name, value = curr.split(":", 1)
+            name = name.rstrip(" \t").upper()
+            if self.hdrre.search(name):
+                raise InvalidHeaderName(name)
+            name, value = name.strip(), [value.lstrip()]
+            
+            # Consume value continuation lines
+            while len(lines) and lines[0].startswith((" ", "\t")):
+                value.append(lines.pop(0))
+            value = ''.join(value).rstrip()
+            
+            headers.append((name, value))
+        return headers
+
+    def set_body_reader(self):
+        chunked = False
+        clength = None
+
+        for (name, value) in self.headers:
+            if name.upper() == "CONTENT-LENGTH":
+                try:
+                    clength = int(value)
+                except ValueError:
+                    clenth = None
+            elif name.upper() == "TRANSFER-ENCODING":
+                chunked = value.lower() == "chunked"
+        
+        if chunked:
+            self.body = Body(ChunkedReader(self, self.unreader))
+        elif clength is not None:
+            self.body = Body(LengthReader(self.unreader, clength))
+        else:
+            self.body = Body(EOFReader(self.unreader))
+
+    def should_close(self):
+        for (h, v) in self.headers:
+            if h.lower() == "connection":
+                if v.lower().strip() == "close":
+                    return True
+                elif v.lower().strip() == "keep-alive":
+                    return False
+        return self.version <= (1, 0)
+
+
+class Request(Message):
+    def __init__(self, unreader):
+        self.methre = re.compile("[A-Z0-9$-_.]{3,20}")
+        self.versre = re.compile("HTTP/(\d+).(\d+)")
+    
+        self.method = None
+        self.uri = None
+        self.scheme = None
+        self.host = None
+        self.port = 80
+        self.path = None
+        self.query = None
+        self.fragment = None
+
+        super(Request, self).__init__(unreader)
+
+    
+    def get_data(self, unreader, buf, stop=False):
+        data = unreader.read()
+        if not data:
+            if stop:
+                raise StopIteration()
+            raise NoMoreData(buf.getvalue())
+        buf.write(data)
+    
+    def parse(self, unreader):
+        buf = StringIO()
+
+        self.get_data(unreader, buf, stop=True)
+        
+        # Request line
+        idx = buf.getvalue().find("\r\n")
+        while idx < 0:
+            self.get_data(unreader, buf)
+            idx = buf.getvalue().find("\r\n")
+        self.parse_request_line(buf.getvalue()[:idx])
+        rest = buf.getvalue()[idx+2:] # Skip \r\n
+        buf.truncate(0)
+        buf.write(rest)
+        
+        # Headers
+        idx = buf.getvalue().find("\r\n\r\n")
+        done = buf.getvalue()[:2] == "\r\n"
+        while idx < 0 and not done:
+            self.get_data(unreader, buf)
+            idx = buf.getvalue().find("\r\n\r\n")
+            done = buf.getvalue()[:2] == "\r\n"
+        if done:
+            self.unreader.unread(buf.getvalue()[2:])
+            return ""
+        self.headers = self.parse_headers(buf.getvalue()[:idx])
+
+        ret = buf.getvalue()[idx+4:]
+        buf.truncate(0)
+        return ret
+    
+    def parse_request_line(self, line):
+        bits = line.split(None, 2)
+        if len(bits) != 3:
+            raise InvalidRequestLine(line)
+
+        # Method
+        if not self.methre.match(bits[0]):
+            raise InvalidRequestMethod(bits[0])
+        self.method = bits[0].upper()
+
+        # URI
+        self.uri = bits[1]
+        parts = urlparse.urlparse(bits[1])
+        self.scheme = parts.scheme or None
+        self.host = parts.netloc or None
+        if parts.port is None:
+            self.port = 80
+        else:
+            self.host = self.host.rsplit(":", 1)[0]
+            self.port = parts.port
+        self.path = parts.path or None
+        self.query = parts.query or None
+        self.fragment = parts.fragment or None
+
+        # Version
+        match = self.versre.match(bits[2])
+        if match is None:
+            raise InvalidHTTPVersion(bits[2])
+        self.version = (int(match.group(1)), int(match.group(2)))
+
+    def set_body_reader(self):
+        super(Request, self).set_body_reader()
+        if isinstance(self.body.reader, EOFReader):
+            self.body = Body(LengthReader(self.unreader, 0))
+
+
diff --git a/gunicorn/http/parser.py b/gunicorn/http/parser.py
new file mode 100644
index 00000000..a5e31bc3
--- /dev/null
+++ b/gunicorn/http/parser.py
@@ -0,0 +1,38 @@
+
+import socket
+
+from message import Request
+from unreader import SocketUnreader, IterUnreader
+
+class Parser(object):
+    def __init__(self, mesg_class, source):
+        self.mesg_class = mesg_class
+        if isinstance(source, socket.socket):
+            self.unreader = SocketUnreader(source)
+        else:
+            self.unreader = IterUnreader(source)
+        self.mesg = None
+
+    def __iter__(self):
+        return self
+    
+    def next(self):
+        if self.mesg.should_close():
+            raise StopIteration()
+        self.discard()
+        self.mesg = self.mesg_class(self.unreader)
+        if not self.mesg:
+            raise StopIteration()
+        return self.mesg
+
+    def discard(self):
+        if self.mesg is not None:
+            data = self.mesg.read(8192)
+            while data:
+                self.mesg.read(8192)
+        self.mesg = None
+
+class RequestParser(Parser):
+    def __init__(self, *args, **kwargs):
+        super(RequestParser, self).__init__(Request, *args, **kwargs)
+
diff --git a/gunicorn/http/unreader.py b/gunicorn/http/unreader.py
new file mode 100644
index 00000000..838452e9
--- /dev/null
+++ b/gunicorn/http/unreader.py
@@ -0,0 +1,73 @@
+
+import os
+
+try:
+    from cStringIO import StringIO
+except ImportError:
+    from StringIO import StringIO
+
+# Classes that can undo reading data from
+# a given type of data source.
+
+class Unreader(object):
+    def __init__(self):
+        self.buf = StringIO()
+    
+    def chunk(self):
+        raise NotImplementedError()
+    
+    def read(self, size=None):
+        if size is not None and not isinstance(size, (int, long)):
+            raise TypeError("size parameter must be an int or long.")
+        if size == 0:
+            return ""
+        if size < 0:
+            size = None
+
+        self.buf.seek(0, os.SEEK_END)
+
+        if size is None and self.buf.tell():
+            ret = self.buf.getvalue()
+            self.buf.truncate(0)
+            return ret
+        if size is None:
+            return self.chunk()
+
+        while self.buf.tell() < size:
+            chunk = self.chunk()
+            if not len(chunk):
+                ret = self.buf.getvalue()
+                self.buf.truncate(0)
+                return ret
+            self.buf.write(chunk)
+        data = self.buf.getvalue()
+        self.buf.truncate(0)
+        self.buf.write(data[size:])
+        return data[:size]
+    
+    def unread(self, data):
+        self.buf.seek(0, os.SEEK_END)
+        self.buf.write(data)
+
+class SocketUnreader(Unreader):
+    def __init__(self, sock, max_chunk=8192):
+        super(SocketUnreader, self).__init__()
+        self.sock = sock
+        self.mxchunk = max_chunk
+    
+    def chunk(self):
+        return self.sock.recv(self.mxchunk)
+
+class IterUnreader(Unreader):
+    def __init__(self, iterable):
+        super(IterUnreader, self).__init__()
+        self.iter = iter(iterable)
+
+    def chunk(self):
+        if not self.iter:
+            return ""
+        try:
+            return self.iter.next()
+        except StopIteration:
+            self.iter = None
+            return ""
diff --git a/gunicorn/wsgi.py b/gunicorn/http/wsgi.py
similarity index 100%
rename from gunicorn/wsgi.py
rename to gunicorn/http/wsgi.py
diff --git a/gunicorn/workers/sync.py b/gunicorn/workers/sync.py
index 32d25daf..7380ae44 100644
--- a/gunicorn/workers/sync.py
+++ b/gunicorn/workers/sync.py
@@ -10,13 +10,12 @@ import select
 import socket
 import traceback
 
-from simplehttp import RequestParser
-
+import gunicorn.http.parser as parser
+import gunicorn.http.wsgi as wsgi
 import gunicorn.util as util
-import gunicorn.wsgi as wsgi
-from gunicorn.workers.base import Worker
+import gunicorn.workers.base as base
 
-class SyncWorker(Worker):
+class SyncWorker(base.Worker):
     
     def run(self):
         self.nr = 0
@@ -71,7 +70,7 @@ class SyncWorker(Worker):
         
     def handle(self, client, addr):
         try:
-            parser = RequestParser(client)
+            parser = parser.RequestParser(client)
             req = parser.next()
             self.handle_request(req, client, addr)
         except socket.error, e: