Bump version.

We get confusing output on terminals if we leave \r unescaped.

README

@@ -1,11 +1,7 @@
 
-Netlib is a collection of network utility classes, used by pathod and mitmproxy
-projects. It differs from other projects in some fundamental respects, because
-both pathod and mitmproxy often need to violate standards. This means that
-protocols are implemented as small, well-contained and flexible functions, and
-servers are implemented to allow misbehaviour when needed.
-
-At this point, I have no plans to make netlib useful beyond mitmproxy and
-pathod. Please get in touch if you think parts of netlib might have broader
-utility.
+Netlib is a collection of network utility classes, used by the pathod and
+mitmproxy projects. It differs from other projects in some fundamental
+respects, because both pathod and mitmproxy often need to violate standards.
+This means that protocols are implemented as small, well-contained and flexible
+functions, and are designed to allow misbehaviour when needed.
 

netlib/http.py

@@ -36,8 +36,8 @@ def parse_url(url):
 
 def read_headers(fp):
     """
-        Read a set of headers from a file pointer. Stop once a blank line
-        is reached. Return a ODictCaseless object.
+        Read a set of headers from a file pointer. Stop once a blank line is
+        reached. Return a ODictCaseless object, or None if headers are invalid.
     """
     ret = []
     name = ''
@@ -46,6 +46,8 @@ def read_headers(fp):
         if not line or line == '\r\n' or line == '\n':
             break
         if line[0] in ' \t':
+            if not ret:
+                return None
             # continued header
             ret[-1][1] = ret[-1][1] + '\r\n ' + line.strip()
         else:
@@ -55,6 +57,8 @@ def read_headers(fp):
                 name = line[:i]
                 value = line[i+1:].strip()
                 ret.append([name, value])
+            else:
+                return None
     return odict.ODictCaseless(ret)
 
 
@@ -282,6 +286,8 @@ def read_response(rfile, method, body_size_limit):
     except ValueError:
         raise HttpError(502, "Invalid server response: %s"%repr(line))
     headers = read_headers(rfile)
+    if headers is None:
+        raise HttpError(502, "Invalid headers.")
     if code >= 100 and code <= 199:
         return read_response(rfile, method, body_size_limit)
     if method == "HEAD" or code == 204 or code == 304:

netlib/http_uastrings.py

@@ -0,0 +1,77 @@
+"""
+    A small collection of useful user-agent header strings. These should be
+    kept reasonably current to reflect common usage.
+"""
+
+# A collection of (name, shortcut, string) tuples.
+
+UASTRINGS = [
+    (
+        "android",
+        "a",
+        "Mozilla/5.0 (Linux; U; Android 4.1.1; en-gb; Nexus 7 Build/JRO03D) AFL/01.04.02"
+    ),
+
+    (
+        "blackberry",
+        "l",
+        "Mozilla/5.0 (BlackBerry; U; BlackBerry 9900; en) AppleWebKit/534.11+ (KHTML, like Gecko) Version/7.1.0.346 Mobile Safari/534.11+"
+    ),
+
+    (
+        "bingbot",
+        "b",
+        "Mozilla/5.0 (compatible; bingbot/2.0; +http://www.bing.com/bingbot.htm)"
+    ),
+
+    (
+        "chrome",
+        "c",
+        "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.1 (KHTML, like Gecko) Chrome/22.0.1207.1 Safari/537.1"
+    ),
+
+    (
+        "firefox",
+        "f",
+        "Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:14.0) Gecko/20120405 Firefox/14.0a1"
+    ),
+
+    (
+        "googlebot",
+        "g",
+        "Googlebot/2.1 (+http://www.googlebot.com/bot.html)"
+    ),
+
+    (
+        "ie9",
+        "i",
+        "Mozilla/5.0 (Windows; U; MSIE 9.0; WIndows NT 9.0; en-US))"
+    ),
+
+    (
+        "ipad",
+        "p",
+        "Mozilla/5.0 (iPad; CPU OS 5_1 like Mac OS X) AppleWebKit/534.46 (KHTML, like Gecko ) Version/5.1 Mobile/9B176 Safari/7534.48.3"
+    ),
+
+    (
+        "iphone",
+        "h",
+        "Mozilla/5.0 (iPhone; CPU iPhone OS 4_2_1 like Mac OS X) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8C148a Safari/6533.18.5",
+    ),
+
+    (
+        "safari",
+        "s",
+        "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_3) AppleWebKit/534.55.3 (KHTML, like Gecko) Version/5.1.3 Safari/534.53.10"
+    )
+]
+
+
+def get_by_shortcut(s):
+    """
+        Retrieve a user agent entry by shortcut.
+    """
+    for i in UASTRINGS:
+        if s == i[1]:
+            return i

netlib/odict.py

@@ -80,6 +80,12 @@ class ODict:
         else:
             return d
 
+    def get_first(self, k, d=None):
+        if k in self:
+            return self[k][0]
+        else:
+            return d
+
     def items(self):
         return self.lst[:]
 

netlib/tcp.py

@@ -39,27 +39,84 @@ class NetLibDisconnect(Exception): pass
 class NetLibTimeout(Exception): pass
 
 
-class FileLike:
+class _FileLike:
     BLOCKSIZE = 1024 * 32
     def __init__(self, o):
         self.o = o
+        self._log = None
+
+    def set_descriptor(self, o):
+        self.o = o
 
     def __getattr__(self, attr):
         return getattr(self.o, attr)
 
-    def flush(self):
-        if hasattr(self.o, "flush"):
-            self.o.flush()
+    def start_log(self):
+        """
+            Starts or resets the log.
 
+            This will store all bytes read or written.
+        """
+        self._log = []
+
+    def stop_log(self):
+        """
+            Stops the log.
+        """
+        self._log = None
+
+    def is_logging(self):
+        return self._log is not None
+
+    def get_log(self):
+        """
+            Returns the log as a string.
+        """
+        if not self.is_logging():
+            raise ValueError("Not logging!")
+        return "".join(self._log)
+
+    def add_log(self, v):
+        if self.is_logging():
+            self._log.append(v)
+
+
+class Writer(_FileLike):
+    def flush(self):
+        try:
+            if hasattr(self.o, "flush"):
+                self.o.flush()
+        except socket.error, v:
+            raise NetLibDisconnect(str(v))
+
+    def write(self, v):
+        if v:
+            try:
+                if hasattr(self.o, "sendall"):
+                    self.add_log(v)
+                    return self.o.sendall(v)
+                else:
+                    r = self.o.write(v)
+                    self.add_log(v[:r])
+                    return r
+            except (SSL.Error, socket.error), v:
+                raise NetLibDisconnect(str(v))
+
+
+class Reader(_FileLike):
     def read(self, length):
         """
-            If length is None, we read until connection closes.
+            If length is -1, we read until connection closes.
         """
         result = ''
         start = time.time()
         while length == -1 or length > 0:
+            if length == -1 or length > self.BLOCKSIZE:
+                rlen = self.BLOCKSIZE
+            else:
+                rlen = length
             try:
-                data = self.o.read(self.BLOCKSIZE if length == -1 else length)
+                data = self.o.read(rlen)
             except SSL.ZeroReturnError:
                 break
             except SSL.WantReadError:
@@ -79,19 +136,9 @@ class FileLike:
             result += data
             if length != -1:
                 length -= len(data)
+        self.add_log(result)
         return result
 
-    def write(self, v):
-        if v:
-            try:
-                if hasattr(self.o, "sendall"):
-                    return self.o.sendall(v)
-                else:
-                    r = self.o.write(v)
-                    return r
-            except (SSL.Error, socket.error):
-                raise NetLibDisconnect()
-
     def readline(self, size = None):
         result = ''
         bytes_read = 0
@@ -137,16 +184,16 @@ class TCPClient:
         except SSL.Error, v:
             raise NetLibError("SSL handshake error: %s"%str(v))
         self.cert = certutils.SSLCert(self.connection.get_peer_certificate())
-        self.rfile = FileLike(self.connection)
-        self.wfile = FileLike(self.connection)
+        self.rfile.set_descriptor(self.connection)
+        self.wfile.set_descriptor(self.connection)
 
     def connect(self):
         try:
             addr = socket.gethostbyname(self.host)
             connection = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
             connection.connect((addr, self.port))
-            self.rfile = FileLike(connection.makefile('rb', self.rbufsize))
-            self.wfile = FileLike(connection.makefile('wb', self.wbufsize))
+            self.rfile = Reader(connection.makefile('rb', self.rbufsize))
+            self.wfile = Writer(connection.makefile('wb', self.wbufsize))
         except socket.error, err:
             raise NetLibError('Error connecting to "%s": %s' % (self.host, err))
         self.connection = connection
@@ -180,8 +227,8 @@ class BaseHandler:
     wbufsize = -1
     def __init__(self, connection, client_address, server):
         self.connection = connection
-        self.rfile = FileLike(self.connection.makefile('rb', self.rbufsize))
-        self.wfile = FileLike(self.connection.makefile('wb', self.wbufsize))
+        self.rfile = Reader(self.connection.makefile('rb', self.rbufsize))
+        self.wfile = Writer(self.connection.makefile('wb', self.wbufsize))
 
         self.client_address = client_address
         self.server = server
@@ -206,8 +253,8 @@ class BaseHandler:
             self.connection.do_handshake()
         except SSL.Error, v:
             raise NetLibError("SSL handshake error: %s"%str(v))
-        self.rfile = FileLike(self.connection)
-        self.wfile = FileLike(self.connection)
+        self.rfile.set_descriptor(self.connection)
+        self.wfile.set_descriptor(self.connection)
 
     def finish(self):
         self.finished = True
@@ -241,6 +288,9 @@ class BaseHandler:
     def handle(self): # pragma: no cover
         raise NotImplementedError
 
+    def settimeout(self, n):
+        self.connection.settimeout(n)
+
     def close(self):
         """
             Does a hard close of the socket, i.e. a shutdown, followed by a close.

netlib/utils.py

@@ -0,0 +1,36 @@
+
+def cleanBin(s, fixspacing=False):
+    """
+        Cleans binary data to make it safe to display. If fixspacing is True,
+        tabs, newlines and so forth will be maintained, if not, they will be
+        replaced with a placeholder.
+    """
+    parts = []
+    for i in s:
+        o = ord(i)
+        if (o > 31 and o < 127):
+            parts.append(i)
+        elif i in "\n\t" and not fixspacing:
+            parts.append(i)
+        else:
+            parts.append(".")
+    return "".join(parts)
+
+
+def hexdump(s):
+    """
+        Returns a set of tuples:
+            (offset, hex, str)
+    """
+    parts = []
+    for i in range(0, len(s), 16):
+        o = "%.10x"%i
+        part = s[i:i+16]
+        x = " ".join("%.2x"%ord(i) for i in part)
+        if len(part) < 16:
+            x += " "
+            x += " ".join("  " for i in range(16 - len(part)))
+        parts.append(
+            (o, x, cleanBin(part, True))
+        )
+    return parts

netlib/version.py

@@ -1,4 +1,4 @@
-IVERSION = (0, 2)
+IVERSION = (0, 2, 2)
 VERSION = ".".join(str(i) for i in IVERSION)
 NAME = "netlib"
 NAMEVERSION = NAME + " " + VERSION

test/test_http.py

@@ -169,16 +169,20 @@ def test_parse_init_http():
 
 
 class TestReadHeaders:
+    def _read(self, data, verbatim=False):
+        if not verbatim:
+            data = textwrap.dedent(data)
+            data = data.strip()
+        s = cStringIO.StringIO(data)
+        return http.read_headers(s)
+
     def test_read_simple(self):
         data = """
             Header: one
             Header2: two
             \r\n
         """
-        data = textwrap.dedent(data)
-        data = data.strip()
-        s = cStringIO.StringIO(data)
-        h = http.read_headers(s)
+        h = self._read(data)
         assert h.lst == [["Header", "one"], ["Header2", "two"]]
 
     def test_read_multi(self):
@@ -187,10 +191,7 @@ class TestReadHeaders:
             Header: two
             \r\n
         """
-        data = textwrap.dedent(data)
-        data = data.strip()
-        s = cStringIO.StringIO(data)
-        h = http.read_headers(s)
+        h = self._read(data)
         assert h.lst == [["Header", "one"], ["Header", "two"]]
 
     def test_read_continued(self):
@@ -200,12 +201,19 @@ class TestReadHeaders:
             Header2: three
             \r\n
         """
-        data = textwrap.dedent(data)
-        data = data.strip()
-        s = cStringIO.StringIO(data)
-        h = http.read_headers(s)
+        h = self._read(data)
         assert h.lst == [["Header", "one\r\n two"], ["Header2", "three"]]
 
+    def test_read_continued_err(self):
+        data = "\tfoo: bar\r\n"
+        assert self._read(data, True) is None
+
+    def test_read_err(self):
+        data = """
+            foo
+        """
+        assert self._read(data) is None
+
 
 def test_read_response():
     def tst(data, method, limit):
@@ -248,6 +256,14 @@ def test_read_response():
     assert tst(data, "GET", None)[4] == 'foo'
     assert tst(data, "HEAD", None)[4] == ''
 
+    data = """
+        HTTP/1.1 200 OK
+        \tContent-Length: 3
+
+        foo
+    """
+    tutils.raises("invalid headers", tst, data, "GET", None)
+
 
 def test_parse_url():
     assert not http.parse_url("")

test/test_http_uastrings.py

@@ -0,0 +1,7 @@
+from netlib import http_uastrings
+
+
+def test_get_shortcut():
+    assert http_uastrings.get_by_shortcut("c")[0] == "chrome"
+    assert not http_uastrings.get_by_shortcut("_")
+

test/test_odict.py

@@ -85,6 +85,12 @@ class TestODict:
         assert self.od.get("one") == ["two"]
         assert self.od.get("two") == None
 
+    def test_get_first(self):
+        self.od.add("one", "two")
+        self.od.add("one", "three")
+        assert self.od.get_first("one") == "two"
+        assert self.od.get_first("two") == None
+
 
 class TestODictCaseless:
     def setUp(self):

test/test_tcp.py

@@ -28,6 +28,11 @@ class ServerTestBase:
         cls.server.shutdown()
 
 
+    @property
+    def last_handler(self):
+        return self.server.server.last_handler
+
+
 class SNIHandler(tcp.BaseHandler):
     sni = None
     def handle_sni(self, connection):
@@ -63,15 +68,27 @@ class HangHandler(tcp.BaseHandler):
             time.sleep(1)
 
 
+class TimeoutHandler(tcp.BaseHandler):
+    def handle(self):
+        self.timeout = False
+        self.settimeout(0.01)
+        try:
+            self.rfile.read(10)
+        except tcp.NetLibTimeout:
+            self.timeout = True
+
+
 class TServer(tcp.TCPServer):
-    def __init__(self, addr, ssl, q, handler, v3_only=False):
+    def __init__(self, addr, ssl, q, handler_klass, v3_only=False):
         tcp.TCPServer.__init__(self, addr)
         self.ssl, self.q = ssl, q
         self.v3_only = v3_only
-        self.handler = handler
+        self.handler_klass = handler_klass
+        self.last_handler = None
 
     def handle_connection(self, request, client_address):
-        h = self.handler(request, client_address, self)
+        h = self.handler_klass(request, client_address, self)
+        self.last_handler = h
         if self.ssl:
             if self.v3_only:
                 method = tcp.SSLv3_METHOD
@@ -194,12 +211,24 @@ class TestDisconnect(ServerTestBase):
         c.close()
 
 
+class TestServerTimeOut(ServerTestBase):
+    @classmethod
+    def makeserver(cls):
+        return TServer(("127.0.0.1", 0), False, cls.q, TimeoutHandler)
+
+    def test_timeout(self):
+        c = tcp.TCPClient("127.0.0.1", self.port)
+        c.connect()
+        time.sleep(0.3)
+        assert self.last_handler.timeout
+
+
 class TestTimeOut(ServerTestBase):
     @classmethod
     def makeserver(cls):
         return TServer(("127.0.0.1", 0), False, cls.q, HangHandler)
 
-    def test_timeout_client(self):
+    def test_timeout(self):
         c = tcp.TCPClient("127.0.0.1", self.port)
         c.connect()
         c.settimeout(0.1)
@@ -226,10 +255,21 @@ class TestTCPClient:
 
 
 class TestFileLike:
+    def test_blocksize(self):
+        s = cStringIO.StringIO("1234567890abcdefghijklmnopqrstuvwxyz")
+        s = tcp.Reader(s)
+        s.BLOCKSIZE = 2
+        assert s.read(1) == "1"
+        assert s.read(2) == "23"
+        assert s.read(3) == "456"
+        assert s.read(4) == "7890"
+        d = s.read(-1)
+        assert d.startswith("abc") and d.endswith("xyz")
+
     def test_wrap(self):
         s = cStringIO.StringIO("foobar\nfoobar")
-        s = tcp.FileLike(s)
         s.flush()
+        s = tcp.Reader(s)
         assert s.readline() == "foobar\n"
         assert s.readline() == "foobar"
         # Test __getattr__
@@ -237,11 +277,39 @@ class TestFileLike:
 
     def test_limit(self):
         s = cStringIO.StringIO("foobar\nfoobar")
-        s = tcp.FileLike(s)
+        s = tcp.Reader(s)
         assert s.readline(3) == "foo"
 
     def test_limitless(self):
         s = cStringIO.StringIO("f"*(50*1024))
-        s = tcp.FileLike(s)
+        s = tcp.Reader(s)
         ret = s.read(-1)
         assert len(ret) == 50 * 1024
+
+    def test_readlog(self):
+        s = cStringIO.StringIO("foobar\nfoobar")
+        s = tcp.Reader(s)
+        assert not s.is_logging()
+        s.start_log()
+        assert s.is_logging()
+        s.readline()
+        assert s.get_log() == "foobar\n"
+        s.read(1)
+        assert s.get_log() == "foobar\nf"
+        s.start_log()
+        assert s.get_log() == ""
+        s.read(1)
+        assert s.get_log() == "o"
+        s.stop_log()
+        tutils.raises(ValueError, s.get_log)
+
+    def test_writelog(self):
+        s = cStringIO.StringIO()
+        s = tcp.Writer(s)
+        s.start_log()
+        assert s.is_logging()
+        s.write("x")
+        assert s.get_log() == "x"
+        s.write("x")
+        assert s.get_log() == "xx"
+

test/test_utils.py

@@ -0,0 +1,13 @@
+from netlib import utils
+
+
+def test_hexdump():
+    assert utils.hexdump("one\0"*10)
+
+
+def test_cleanBin():
+    assert utils.cleanBin("one") == "one"
+    assert utils.cleanBin("\00ne") == ".ne"
+    assert utils.cleanBin("\nne") == "\nne"
+    assert utils.cleanBin("\nne", True) == ".ne"
+