From jython-checkins at python.org Sun Feb 9 21:22:03 2014 From: jython-checkins at python.org (jeff.allen) Date: Sun, 9 Feb 2014 21:22:03 +0100 (CET) Subject: [Jython-checkins] =?utf-8?q?jython=3A_Add_PYTHONIOENCODING_env_va?= =?utf-8?q?r_addressing_issue_=231876=2C_and_-E_option_to_suppress=2E?= Message-ID: <3fMhVg27P2z7Ljd@mail.python.org> http://hg.python.org/jython/rev/6e438088c0e3 changeset: 7181:6e438088c0e3 user: Jeff Allen date: Sun Feb 09 19:26:34 2014 +0000 summary: Add PYTHONIOENCODING env var addressing issue #1876, and -E option to suppress. Also introduces registry items python.io.encoding and python.io.errors, with appropriate sequence of priority for site, user, environment variable and command-line values. Additions to test.test_sys (from CPython 2.7) and test.test_sys_jy for registry items. files: Lib/test/test_sys.py | 20 ++ Lib/test/test_sys_jy.py | 63 ++++++- NEWS | 4 + src/org/python/core/Console.java | 8 + src/org/python/core/Options.java | 1 - src/org/python/core/PlainConsole.java | 9 +- src/org/python/core/PyFile.java | 21 +- src/org/python/core/PySystemState.java | 54 ++++- src/org/python/core/StdoutWrapper.java | 5 +- src/org/python/util/jython.java | 125 ++++++++++-- 10 files changed, 256 insertions(+), 54 deletions(-) diff --git a/Lib/test/test_sys.py b/Lib/test/test_sys.py --- a/Lib/test/test_sys.py +++ b/Lib/test/test_sys.py @@ -251,6 +251,26 @@ self.assert_(vi[3] in ("alpha", "beta", "candidate", "final")) self.assert_(isinstance(vi[4], int)) + def test_ioencoding(self): # from v2.7 test + import subprocess,os + env = dict(os.environ) + + # Test character: cent sign, encoded as 0x4A (ASCII J) in CP424, + # not representable in ASCII. + + env["PYTHONIOENCODING"] = "cp424" + p = subprocess.Popen([sys.executable, "-c", 'print unichr(0xa2)'], + stdout = subprocess.PIPE, env=env) + out = p.stdout.read().strip() + self.assertEqual(out, unichr(0xa2).encode("cp424")) + + env["PYTHONIOENCODING"] = "ascii:replace" + p = subprocess.Popen([sys.executable, "-c", 'print unichr(0xa2)'], + stdout = subprocess.PIPE, env=env) + out = p.stdout.read().strip() + self.assertEqual(out, '?') + + def test_main(): if test.test_support.is_jython: del SysModuleTest.test_lost_displayhook diff --git a/Lib/test/test_sys_jy.py b/Lib/test/test_sys_jy.py --- a/Lib/test/test_sys_jy.py +++ b/Lib/test/test_sys_jy.py @@ -1,3 +1,4 @@ +# -*- coding: iso-8859-1 -*- from __future__ import with_statement import os import re @@ -185,13 +186,69 @@ finally: os.rmdir(moduleDir) self.assertFalse(os.path.exists(moduleDir)) - + +class SysEncodingTest(unittest.TestCase): + + # Adapted from CPython 2.7 test_sys to exercise setting Jython registry + # values related to encoding and error policy. + + def test_ioencoding(self): # adapted from CPython v2.7 test_sys + import subprocess, os + env = dict(os.environ) + + def check(code, encoding=None, errors=None): + # Execute with encoding and errors optionally set via Java properties + command = [sys.executable] + if (encoding): + command.append('-Dpython.io.encoding={}'.format(encoding)) + if (errors): + command.append('-Dpython.io.errors={}'.format(errors)) + command.append('-c') + command.append('print unichr({:#x})'.format(code)) + #print "\n ", " ".join(command), " ... ", + p = subprocess.Popen(command, stdout = subprocess.PIPE, env=env) + return p.stdout.read().strip() + + env.pop("PYTHONIOENCODING", None) + self.assertEqual(check(ord(u'A')), b"A") + + # Test character: U+00a2 cent sign (?) is: + # not representable in ASCII. + # xml: ¢ + # cp1252: a2 + # cp850: bd + # cp424: 4a + # utf-8: c2 a2 + + self.assertEqual(check(0xa2, "iso-8859-1"), "?") # same as this file + + # self.assertEqual(check(0xa2, "ascii"), "") # and an error message + self.assertEqual(check(0xa2, "ascii", "ignore"),"") + self.assertEqual(check(0xa2, "ascii", "replace"), "?") + self.assertEqual(check(0xa2, "ascii", "backslashreplace"), r"\xa2") + self.assertEqual(check(0xa2, "ascii", "xmlcharrefreplace"), "¢") + + self.assertEqual(check(0xa2, "Cp1252"), "\xa2") + self.assertEqual(check(0xa2, "Cp424"), "\x4a") + self.assertEqual(check(0xa2, "utf-8"), "\xc2\xa2") + + self.assertEqual(check(0xa2, "iso8859-5", "backslashreplace"), r"\xa2") + + # Now check that PYTHONIOENCODING can be superseded piecemeal + env["PYTHONIOENCODING"] = "ascii:xmlcharrefreplace" + self.assertEqual(check(0xa2, "iso8859-5"), "¢") + self.assertEqual(check(0xa2, None, "backslashreplace"), r"\xa2") + self.assertEqual(check(0xa2, "cp850"), "\xbd") + def test_main(): - test_support.run_unittest(SysTest, + test_support.run_unittest( + SysTest, ShadowingTest, SyspathResourceTest, - SyspathUnicodeTest) + SyspathUnicodeTest, + SysEncodingTest, + ) if __name__ == "__main__": test_main() diff --git a/NEWS b/NEWS --- a/NEWS +++ b/NEWS @@ -5,6 +5,7 @@ - [ 1753 ] zlib doesn't call end() on compress and decompress - [ 1860 ] test failures in test_array.py - [ 1862 ] cStringIO does not support arrays as arguments + - [ 1876 ] PYTHONIOENCODING unsupported, used (among others) by PyDev - [ 1926 ] Adjust MutableSet.pop test so we do not need to skip it - [ 1964 ] time.strptime() does not support %f in format - [ 2005 ] threading.Event object's wait([timeout]) function returns null instead of True/False. @@ -18,6 +19,9 @@ - [ 2075 ] Incorrect padding for hex format strings - [ 2082 ] Unexpected (Pdb) prompt during regression tests - [ 2083 ] os.unlink() can delete directories + New Features + - Command line option -E (ignore environment variables) + - Environment variable PYTHONIOENCODING, and corresponding registry items Jython 2.7b1 Bugs Fixed diff --git a/src/org/python/core/Console.java b/src/org/python/core/Console.java --- a/src/org/python/core/Console.java +++ b/src/org/python/core/Console.java @@ -2,6 +2,7 @@ package org.python.core; import java.io.IOException; +import java.nio.charset.Charset; /** * A class named in configuration as the value of python.console must implement this @@ -29,4 +30,11 @@ */ public void uninstall() throws UnsupportedOperationException; + /** + * Accessor for encoding to use for line input as a Charset. + * + * @return Charset of the encoding in use. + */ + public Charset getEncodingCharset(); + } diff --git a/src/org/python/core/Options.java b/src/org/python/core/Options.java --- a/src/org/python/core/Options.java +++ b/src/org/python/core/Options.java @@ -83,7 +83,6 @@ public static boolean dont_write_bytecode = false; /** Whether -E (ignore environment) was enabled via the command line. */ - //XXX: place holder, not implemented yet. public static boolean ignore_environment = false; //XXX: place holder, not implemented yet. diff --git a/src/org/python/core/PlainConsole.java b/src/org/python/core/PlainConsole.java --- a/src/org/python/core/PlainConsole.java +++ b/src/org/python/core/PlainConsole.java @@ -37,7 +37,7 @@ * must be one supported by the JVM. The PlainConsole does not replace System.in or * System.out, and does not add any line-editing capability to what is standard for * your OS console. - * + * * @param encoding name of a supported encoding or null for * Charset.defaultCharset() */ @@ -59,7 +59,7 @@ * A PlainConsole may be uninstalled. This method assumes any sub-class may not be * uninstalled. Sub-classes that permit themselves to be uninstalled must override (and * not call) this method. - * + * * @throws UnsupportedOperationException unless this class is exactly PlainConsole */ @Override @@ -71,4 +71,9 @@ } } + @Override + public Charset getEncodingCharset() { + return encodingCharset; + } + } diff --git a/src/org/python/core/PyFile.java b/src/org/python/core/PyFile.java --- a/src/org/python/core/PyFile.java +++ b/src/org/python/core/PyFile.java @@ -48,6 +48,9 @@ @ExposedGet(doc = BuiltinDocs.file_encoding_doc) public String encoding; + @ExposedGet(doc = BuiltinDocs.file_errors_doc) + public String errors; + /** Indicator dictating whether a space should be written to this * file on the next print statement (not currently implemented in * print ) */ @@ -170,6 +173,18 @@ } /** + * Set the strings defining the encoding and error handling policy. Setting these strings + * affects behaviour of the {@link #writelines(PyObject)} when passed a {@link PyUnicode} value. + * + * @param encoding the encoding property of file. + * @param errors the errors property of file (or null). + */ + void setEncoding(String encoding, String errors) { + this.encoding = encoding; + this.errors = errors; + } + + /** * Wrap the given RawIOBase with a BufferedIOBase according to the * mode and given bufsize. * @@ -446,13 +461,13 @@ * * @param obj to write * @param message for TypeError if raised (or null for default message) - * @return bytes representing tha value (as a String in the Jython convention) + * @return bytes representing the value (as a String in the Jython convention) */ private String asWritable(PyObject obj, String message) { if (obj instanceof PyUnicode) { - // By convention, use platform default encoding to bytes - return ((PyUnicode)obj).encode(); + // Unicode must be encoded into bytes (null arguments here invoke the default values) + return ((PyUnicode)obj).encode(encoding, errors); } else if (obj instanceof PyString) { // Take a short cut diff --git a/src/org/python/core/PySystemState.java b/src/org/python/core/PySystemState.java --- a/src/org/python/core/PySystemState.java +++ b/src/org/python/core/PySystemState.java @@ -53,6 +53,8 @@ public static final String PYTHON_CACHEDIR = "python.cachedir"; public static final String PYTHON_CACHEDIR_SKIP = "python.cachedir.skip"; public static final String PYTHON_CONSOLE_ENCODING = "python.console.encoding"; + public static final String PYTHON_IO_ENCODING = "python.io.encoding"; + public static final String PYTHON_IO_ERRORS = "python.io.errors"; protected static final String CACHEDIR_DEFAULT_NAME = "cachedir"; public static final String JYTHON_JAR = "jython.jar"; @@ -256,18 +258,25 @@ } } + /** + * Initialise the encoding of sys.stdin, sys.stdout, and + * sys.stderr, and their error handling policy, from registry variables. + * Under the console app util.jython, values reflect PYTHONIOENCODING if not overridden. + * Note that the encoding must name a Python codec, as in codecs.encode(). + */ private void initEncoding() { - String encoding = registry.getProperty(PYTHON_CONSOLE_ENCODING); - if (encoding == null) { - return; + // Two registry variables, counterparts to PYTHONIOENCODING = [encoding][:errors] + String encoding = registry.getProperty(PYTHON_IO_ENCODING); + String errors = registry.getProperty(PYTHON_IO_ERRORS); + + if (encoding==null) { + // We still don't have an explicit selection for this: match the console. + encoding = Py.getConsole().getEncodingCharset().name(); } - for (PyFile stdStream : new PyFile[] {(PyFile)this.stdin, (PyFile)this.stdout, - (PyFile)this.stderr}) { - if (stdStream.isatty()) { - stdStream.encoding = encoding; - } - } + ((PyFile)stdin).setEncoding(encoding, errors); + ((PyFile)stdout).setEncoding(encoding, errors); + ((PyFile)stderr).setEncoding(encoding, "backslashreplace"); } // might be nice to have something general here, but for now these @@ -683,6 +692,8 @@ } catch (SecurityException e) { // Continue } + + // Now the post properties (possibly set by custom JythonInitializer). registry.putAll(postProperties); if (standalone) { // set default standalone property (if not yet set) @@ -690,24 +701,34 @@ registry.put(PYTHON_CACHEDIR_SKIP, "true"); } } + + /* + * The console encoding is the one used by line-editing consoles to decode on the OS side and + * encode on the Python side. It must be a Java codec name, so any relationship to + * python.io.encoding is dubious. + */ if (!registry.containsKey(PYTHON_CONSOLE_ENCODING)) { String encoding = getPlatformEncoding(); if (encoding != null) { registry.put(PYTHON_CONSOLE_ENCODING, encoding); } } + // Set up options from registry Options.setFromRegistry(); } /** - * @return the encoding of the underlying platform; can be null + * Return the encoding of the underlying platform, if we can work it out by any means at all. + * + * @return the encoding of the underlying platform */ private static String getPlatformEncoding() { // first try to grab the Console encoding String encoding = getConsoleEncoding(); if (encoding == null) { try { + // Not quite the console encoding (differs on Windows) encoding = System.getProperty("file.encoding"); } catch (SecurityException se) { // ignore, can't do anything about it @@ -722,7 +743,7 @@ private static String getConsoleEncoding() { String encoding = null; try { - Method encodingMethod = Console.class.getDeclaredMethod("encoding"); + Method encodingMethod = java.io.Console.class.getDeclaredMethod("encoding"); encodingMethod.setAccessible(true); // private static method encoding = (String)encodingMethod.invoke(Console.class); } catch (Exception e) { @@ -731,6 +752,12 @@ return encoding; } + /** + * Merge the contents of a property file into the registry without overriding any values already + * set there. + * + * @param file + */ private static void addRegistryFile(File file) { if (file.exists()) { if (!file.isDirectory()) { @@ -922,9 +949,6 @@ } Py.initClassExceptions(getDefaultBuiltins()); - // defaultSystemState can't init its own encoding, see its constructor - Py.defaultSystemState.initEncoding(); - // Make sure that Exception classes have been loaded new PySyntaxError("", 1, 1, "", ""); @@ -1077,7 +1101,7 @@ Class consoleClass = Class.forName(consoleName); // Ensure it can be cast to the interface type of all consoles - if (! consoleType.isAssignableFrom(consoleClass)) { + if (!consoleType.isAssignableFrom(consoleClass)) { throw new ClassCastException(); } diff --git a/src/org/python/core/StdoutWrapper.java b/src/org/python/core/StdoutWrapper.java --- a/src/org/python/core/StdoutWrapper.java +++ b/src/org/python/core/StdoutWrapper.java @@ -103,8 +103,9 @@ private String printToFile(PyFile file, PyObject o) { String s; - if (o instanceof PyUnicode && file.encoding != null) { - s = ((PyUnicode)o).encode(file.encoding, "strict"); + if (o instanceof PyUnicode) { + // Use the encoding and policy defined for the stream. (Each may be null.) + s = ((PyUnicode)o).encode(file.encoding, file.errors); } else { s = o.__str__().toString(); } diff --git a/src/org/python/util/jython.java b/src/org/python/util/jython.java --- a/src/org/python/util/jython.java +++ b/src/org/python/util/jython.java @@ -55,8 +55,7 @@ + "-c cmd : program passed in as string (terminates option list)\n" // + "-d : debug output from parser (also PYTHONDEBUG=x)\n" + "-Dprop=v : Set the property `prop' to value `v'\n" - // + "-E : ignore environment variables (such as PYTHONPATH)\n" - + "-C codec : Use a different codec when reading from the console.\n" + + "-E : ignore environment variables (such as JYTHONPATH)\n" + "-h : print this help message and exit (also --help)\n" + "-i : inspect interactively after running script\n" // + ", (also PYTHONINSPECT=x)\n" @@ -83,9 +82,11 @@ + "file : program read from script file\n" + "- : program read from stdin (default; interactive mode if a tty)\n" + "arg ... : arguments passed to program in sys.argv[1:]\n" + "\n" - + "Other environment variables:\n" + "JYTHONPATH: '" + File.pathSeparator + + "Other environment variables:\n" // + + "JYTHONPATH: '" + File.pathSeparator + "'-separated list of directories prefixed to the default module\n" - + " search path. The result is sys.path."; + + " search path. The result is sys.path.\n" + + "PYTHONIOENCODING: Encoding[:errors] used for stdin/stdout/stderr."; public static boolean shouldRestart; @@ -94,7 +95,7 @@ * root of the JAR archive. Note that the __name__ is set to the base name of the JAR file and * not to "__main__" (for historic reasons). This method do NOT handle exceptions. the caller * SHOULD handle any (Py)Exceptions thrown by the code. - * + * * @param filename The path to the filename to run. */ public static void runJar(String filename) { @@ -211,16 +212,22 @@ // Get system properties (or empty set if we're prevented from accessing them) Properties preProperties = PySystemState.getBaseProperties(); + // Read environment variable PYTHONIOENCODING into properties (registry) + String pythonIoEncoding = getenv("PYTHONIOENCODING"); + if (pythonIoEncoding != null) { + String[] spec = splitString(pythonIoEncoding, ':', 2); + // Note that if encoding or errors is blank (=null), the registry value wins. + addDefault(preProperties, PySystemState.PYTHON_IO_ENCODING, spec[0]); + addDefault(preProperties, PySystemState.PYTHON_IO_ERRORS, spec[1]); + } + // Decide if System.in is interactive if (!opts.fixInteractive || opts.interactive) { // The options suggest System.in is interactive: but only if isatty() agrees opts.interactive = Py.isInteractive(); if (opts.interactive) { // Set the default console type if nothing else has - String consoleClassName = preProperties.getProperty("python.console"); - if (consoleClassName==null) { - preProperties.setProperty("python.console", PYTHON_CONSOLE_CLASS); - } + addDefault(preProperties, "python.console", PYTHON_CONSOLE_CLASS); } } @@ -230,7 +237,9 @@ PyList warnoptions = new PyList(); addWarnings(opts.warnoptions, warnoptions); - addWarnings(warnOptionsFromEnv(), warnoptions); + if (!Options.ignore_environment) { + addWarnings(warnOptionsFromEnv(), warnoptions); + } systemState.setWarnoptions(warnoptions); // Make sure warnings module is loaded if there are warning options @@ -378,24 +387,18 @@ } if (opts.fixInteractive || (opts.filename == null && opts.command == null)) { - if (opts.encoding == null) { - opts.encoding = PySystemState.registry.getProperty("python.console.encoding"); - } - if (opts.encoding != null) { - if (!Charset.isSupported(opts.encoding)) { - System.err.println(opts.encoding - + " is not a supported encoding on this JVM, so it can't " - + "be used in python.console.encoding."); - System.exit(1); - } - interp.cflags.encoding = opts.encoding; - } + // Go interactive with the console: the parser needs to know the encoding. + String encoding = Py.getConsole().getEncodingCharset().name(); + + // Run the interpreter interactively try { + interp.cflags.encoding = encoding; interp.interact(null, null); } catch (Throwable t) { Py.printException(t); } } + interp.cleanup(); } @@ -414,9 +417,79 @@ // continue } } + + /** + * Return an array of trimmed strings by splitting the argument at each occurrence of a + * separator character. (Helper for configuration variable processing.) Segments of zero length + * after trimming emerge as null. If there are more than the specified number of + * segments the last element of the array contains all of the source string after the + * (n-1)th occurrence of sep. + * + * @param spec to split + * @param sep character on which to split + * @param n number of parts to split into + * @return n-element array of strings (or nulls) + */ + private static String[] splitString(String spec, char sep, int n) { + String[] list = new String[n]; + int p = 0, i = 0, L = spec.length(); + while (p < L) { + int c = spec.indexOf(sep, p); + if (c < 0 || i >= n - 1) { + // No more seps, or no more space: i.th piece is the rest of spec. + c = L; + } + String s = spec.substring(p, c).trim(); + list[i++] = (s.length() > 0) ? s : null; + p = c + 1; + } + return list; + } + + /** + * If the key is not currently present and the passed value is not null, sets the + * key to the value in the given Properties object. Thus, + * it provides a default value for a subsequent getProperty(). + * + * @param registry to be (possibly) updated + * @param key at which to set value + * @param value to set (or null for no setting) + * @return true iff a value was set + */ + private static boolean addDefault(Properties registry, String key, String value) { + // Set value at key if nothing else has set it + if (value == null || registry.containsKey(key)) { + return false; + } else { + registry.setProperty(key, value); + return true; + } + } + + /** + * Get the value of an environment variable, if we are allowed to and it exists; otherwise + * return null. We are allowed to access the environment variable if the -E flag + * was not given and the application has permission to read environment variables. The -E flag + * is reflected in {@link Options#ignore_environment}, and will be set automatically if it turns + * out we do not have permission. + * + * @param varname name to access in the environment + * @return the value or null. + */ + private static String getenv(String varname) { + if (!Options.ignore_environment) { + try { + return System.getenv(varname); + } catch (SecurityException e) { + // We're not allowed to access them after all + Options.ignore_environment = true; + } + } + return null; + } + } - class CommandLineOptions { public String filename; @@ -515,12 +588,8 @@ } else { return argumentExpected(arg); } - } else if (arg.equals("-C")) { - encoding = args[++index]; - setProperty("python.console.encoding", encoding); } else if (arg.equals("-E")) { - // XXX: accept -E (ignore environment variables) to be compatible with - // CPython. do nothing for now (we could ignore the registry) + // -E (ignore environment variables) Options.ignore_environment = true; } else if (arg.startsWith("-D")) { String key = null; -- Repository URL: http://hg.python.org/jython From jython-checkins at python.org Wed Feb 19 23:48:03 2014 From: jython-checkins at python.org (jeff.allen) Date: Wed, 19 Feb 2014 23:48:03 +0100 (CET) Subject: [Jython-checkins] =?utf-8?q?jython=3A_Tweak_test=5Ffile2k_test=5F?= =?utf-8?q?unicode_for_Jython_on_Windows_=28=5Cn_-=3E_=3B=29=2E?= Message-ID: <3fTvGW0854z7LjS@mail.python.org> http://hg.python.org/jython/rev/409c52e174f4 changeset: 7182:409c52e174f4 user: Jeff Allen date: Fri Feb 14 22:52:20 2014 +0000 summary: Tweak test_file2k test_unicode for Jython on Windows (\n -> ;). The test relates to PYTHONIOENCODING but incidentally requires a -c command be passed containing newlines. Use of jython.bat wrapper seems to preclude this. files: Lib/test/test_file2k.py | 3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/Lib/test/test_file2k.py b/Lib/test/test_file2k.py --- a/Lib/test/test_file2k.py +++ b/Lib/test/test_file2k.py @@ -711,12 +711,11 @@ finally: sys.stdout = save_stdout - @unittest.skipIf(test_support.is_jython, "FIXME: Not working on Jython") def test_unicode(self): import subprocess def get_message(encoding, *code): - code = '\n'.join(code) + code = ';'.join(code) # jython.bat cannot cope with '\n' in arguments env = os.environ.copy() env['PYTHONIOENCODING'] = encoding process = subprocess.Popen([sys.executable, "-c", code], -- Repository URL: http://hg.python.org/jython From jython-checkins at python.org Wed Feb 19 23:48:04 2014 From: jython-checkins at python.org (jeff.allen) Date: Wed, 19 Feb 2014 23:48:04 +0100 (CET) Subject: [Jython-checkins] =?utf-8?q?jython=3A_Fixes_failure_to_close_file?= =?utf-8?q?_in_test=5Fcookielib=2EFileCookieJarTests=2Etest=5Fbad=5Fmagic?= =?utf-8?q?=2E?= Message-ID: <3fTvGX6Ykmz7Ljj@mail.python.org> http://hg.python.org/jython/rev/df8eb4bcfa70 changeset: 7183:df8eb4bcfa70 user: Jeff Allen date: Fri Feb 14 23:53:38 2014 +0000 summary: Fixes failure to close file in test_cookielib.FileCookieJarTests.test_bad_magic. Failure to close causes failure to delete, and erratic behaviour in regression tests. files: Lib/test/test_cookielib.py | 1762 ++++++++++++++++++++++++ 1 files changed, 1762 insertions(+), 0 deletions(-) diff --git a/Lib/test/test_cookielib.py b/Lib/test/test_cookielib.py new file mode 100644 --- /dev/null +++ b/Lib/test/test_cookielib.py @@ -0,0 +1,1762 @@ +# -*- coding: latin-1 -*- +"""Tests for cookielib.py.""" + +import cookielib +import os +import re +import time + +from unittest import TestCase + +from test import test_support + + +class DateTimeTests(TestCase): + + def test_time2isoz(self): + from cookielib import time2isoz + + base = 1019227000 + day = 24*3600 + self.assertEqual(time2isoz(base), "2002-04-19 14:36:40Z") + self.assertEqual(time2isoz(base+day), "2002-04-20 14:36:40Z") + self.assertEqual(time2isoz(base+2*day), "2002-04-21 14:36:40Z") + self.assertEqual(time2isoz(base+3*day), "2002-04-22 14:36:40Z") + + az = time2isoz() + bz = time2isoz(500000) + for text in (az, bz): + self.assertTrue(re.search(r"^\d{4}-\d\d-\d\d \d\d:\d\d:\d\dZ$", text), + "bad time2isoz format: %s %s" % (az, bz)) + + def test_http2time(self): + from cookielib import http2time + + def parse_date(text): + return time.gmtime(http2time(text))[:6] + + self.assertEqual(parse_date("01 Jan 2001"), (2001, 1, 1, 0, 0, 0.0)) + + # this test will break around year 2070 + self.assertEqual(parse_date("03-Feb-20"), (2020, 2, 3, 0, 0, 0.0)) + + # this test will break around year 2048 + self.assertEqual(parse_date("03-Feb-98"), (1998, 2, 3, 0, 0, 0.0)) + + def test_http2time_formats(self): + from cookielib import http2time, time2isoz + + # test http2time for supported dates. Test cases with 2 digit year + # will probably break in year 2044. + tests = [ + 'Thu, 03 Feb 1994 00:00:00 GMT', # proposed new HTTP format + 'Thursday, 03-Feb-94 00:00:00 GMT', # old rfc850 HTTP format + 'Thursday, 03-Feb-1994 00:00:00 GMT', # broken rfc850 HTTP format + + '03 Feb 1994 00:00:00 GMT', # HTTP format (no weekday) + '03-Feb-94 00:00:00 GMT', # old rfc850 (no weekday) + '03-Feb-1994 00:00:00 GMT', # broken rfc850 (no weekday) + '03-Feb-1994 00:00 GMT', # broken rfc850 (no weekday, no seconds) + '03-Feb-1994 00:00', # broken rfc850 (no weekday, no seconds, no tz) + + '03-Feb-94', # old rfc850 HTTP format (no weekday, no time) + '03-Feb-1994', # broken rfc850 HTTP format (no weekday, no time) + '03 Feb 1994', # proposed new HTTP format (no weekday, no time) + + # A few tests with extra space at various places + ' 03 Feb 1994 0:00 ', + ' 03-Feb-1994 ', + ] + + test_t = 760233600 # assume broken POSIX counting of seconds + result = time2isoz(test_t) + expected = "1994-02-03 00:00:00Z" + self.assertEqual(result, expected, + "%s => '%s' (%s)" % (test_t, result, expected)) + + for s in tests: + t = http2time(s) + t2 = http2time(s.lower()) + t3 = http2time(s.upper()) + + self.assertTrue(t == t2 == t3 == test_t, + "'%s' => %s, %s, %s (%s)" % (s, t, t2, t3, test_t)) + + def test_http2time_garbage(self): + from cookielib import http2time + + for test in [ + '', + 'Garbage', + 'Mandag 16. September 1996', + '01-00-1980', + '01-13-1980', + '00-01-1980', + '32-01-1980', + '01-01-1980 25:00:00', + '01-01-1980 00:61:00', + '01-01-1980 00:00:62', + ]: + self.assertTrue(http2time(test) is None, + "http2time(%s) is not None\n" + "http2time(test) %s" % (test, http2time(test)) + ) + + +class HeaderTests(TestCase): + + def test_parse_ns_headers_expires(self): + from cookielib import parse_ns_headers + + # quotes should be stripped + expected = [[('foo', 'bar'), ('expires', 2209069412L), ('version', '0')]] + for hdr in [ + 'foo=bar; expires=01 Jan 2040 22:23:32 GMT', + 'foo=bar; expires="01 Jan 2040 22:23:32 GMT"', + ]: + self.assertEqual(parse_ns_headers([hdr]), expected) + + def test_parse_ns_headers_version(self): + from cookielib import parse_ns_headers + + # quotes should be stripped + expected = [[('foo', 'bar'), ('version', '1')]] + for hdr in [ + 'foo=bar; version="1"', + 'foo=bar; Version="1"', + ]: + self.assertEqual(parse_ns_headers([hdr]), expected) + + def test_parse_ns_headers_special_names(self): + # names such as 'expires' are not special in first name=value pair + # of Set-Cookie: header + from cookielib import parse_ns_headers + + # Cookie with name 'expires' + hdr = 'expires=01 Jan 2040 22:23:32 GMT' + expected = [[("expires", "01 Jan 2040 22:23:32 GMT"), ("version", "0")]] + self.assertEqual(parse_ns_headers([hdr]), expected) + + def test_join_header_words(self): + from cookielib import join_header_words + + joined = join_header_words([[("foo", None), ("bar", "baz")]]) + self.assertEqual(joined, "foo; bar=baz") + + self.assertEqual(join_header_words([[]]), "") + + def test_split_header_words(self): + from cookielib import split_header_words + + tests = [ + ("foo", [[("foo", None)]]), + ("foo=bar", [[("foo", "bar")]]), + (" foo ", [[("foo", None)]]), + (" foo= ", [[("foo", "")]]), + (" foo=", [[("foo", "")]]), + (" foo= ; ", [[("foo", "")]]), + (" foo= ; bar= baz ", [[("foo", ""), ("bar", "baz")]]), + ("foo=bar bar=baz", [[("foo", "bar"), ("bar", "baz")]]), + # doesn't really matter if this next fails, but it works ATM + ("foo= bar=baz", [[("foo", "bar=baz")]]), + ("foo=bar;bar=baz", [[("foo", "bar"), ("bar", "baz")]]), + ('foo bar baz', [[("foo", None), ("bar", None), ("baz", None)]]), + ("a, b, c", [[("a", None)], [("b", None)], [("c", None)]]), + (r'foo; bar=baz, spam=, foo="\,\;\"", bar= ', + [[("foo", None), ("bar", "baz")], + [("spam", "")], [("foo", ',;"')], [("bar", "")]]), + ] + + for arg, expect in tests: + try: + result = split_header_words([arg]) + except: + import traceback, StringIO + f = StringIO.StringIO() + traceback.print_exc(None, f) + result = "(error -- traceback follows)\n\n%s" % f.getvalue() + self.assertEqual(result, expect, """ +When parsing: '%s' +Expected: '%s' +Got: '%s' +""" % (arg, expect, result)) + + def test_roundtrip(self): + from cookielib import split_header_words, join_header_words + + tests = [ + ("foo", "foo"), + ("foo=bar", "foo=bar"), + (" foo ", "foo"), + ("foo=", 'foo=""'), + ("foo=bar bar=baz", "foo=bar; bar=baz"), + ("foo=bar;bar=baz", "foo=bar; bar=baz"), + ('foo bar baz', "foo; bar; baz"), + (r'foo="\"" bar="\\"', r'foo="\""; bar="\\"'), + ('foo,,,bar', 'foo, bar'), + ('foo=bar,bar=baz', 'foo=bar, bar=baz'), + + ('text/html; charset=iso-8859-1', + 'text/html; charset="iso-8859-1"'), + + ('foo="bar"; port="80,81"; discard, bar=baz', + 'foo=bar; port="80,81"; discard, bar=baz'), + + (r'Basic realm="\"foo\\\\bar\""', + r'Basic; realm="\"foo\\\\bar\""') + ] + + for arg, expect in tests: + input = split_header_words([arg]) + res = join_header_words(input) + self.assertEqual(res, expect, """ +When parsing: '%s' +Expected: '%s' +Got: '%s' +Input was: '%s' +""" % (arg, expect, res, input)) + + +class FakeResponse: + def __init__(self, headers=[], url=None): + """ + headers: list of RFC822-style 'Key: value' strings + """ + import mimetools, StringIO + f = StringIO.StringIO("\n".join(headers)) + self._headers = mimetools.Message(f) + self._url = url + def info(self): return self._headers + +def interact_2965(cookiejar, url, *set_cookie_hdrs): + return _interact(cookiejar, url, set_cookie_hdrs, "Set-Cookie2") + +def interact_netscape(cookiejar, url, *set_cookie_hdrs): + return _interact(cookiejar, url, set_cookie_hdrs, "Set-Cookie") + +def _interact(cookiejar, url, set_cookie_hdrs, hdr_name): + """Perform a single request / response cycle, returning Cookie: header.""" + from urllib2 import Request + req = Request(url) + cookiejar.add_cookie_header(req) + cookie_hdr = req.get_header("Cookie", "") + headers = [] + for hdr in set_cookie_hdrs: + headers.append("%s: %s" % (hdr_name, hdr)) + res = FakeResponse(headers, url) + cookiejar.extract_cookies(res, req) + return cookie_hdr + + +class FileCookieJarTests(TestCase): + def test_lwp_valueless_cookie(self): + # cookies with no value should be saved and loaded consistently + from cookielib import LWPCookieJar + filename = test_support.TESTFN + c = LWPCookieJar() + interact_netscape(c, "http://www.acme.com/", 'boo') + self.assertEqual(c._cookies["www.acme.com"]["/"]["boo"].value, None) + try: + c.save(filename, ignore_discard=True) + c = LWPCookieJar() + c.load(filename, ignore_discard=True) + finally: + try: os.unlink(filename) + except OSError: pass + self.assertEqual(c._cookies["www.acme.com"]["/"]["boo"].value, None) + + def test_bad_magic(self): + from cookielib import LWPCookieJar, MozillaCookieJar, LoadError + # IOErrors (eg. file doesn't exist) are allowed to propagate + filename = test_support.TESTFN + for cookiejar_class in LWPCookieJar, MozillaCookieJar: + c = cookiejar_class() + try: + c.load(filename="for this test to work, a file with this " + "filename should not exist") + except IOError, exc: + # exactly IOError, not LoadError + self.assertEqual(exc.__class__, IOError) + else: + self.fail("expected IOError for invalid filename") + # Invalid contents of cookies file (eg. bad magic string) + # causes a LoadError. + try: + with open(filename, "w") as f: + f.write("oops\n") + for cookiejar_class in LWPCookieJar, MozillaCookieJar: + c = cookiejar_class() + self.assertRaises(LoadError, c.load, filename) + finally: + try: os.unlink(filename) + except OSError: pass + +class CookieTests(TestCase): + # XXX + # Get rid of string comparisons where not actually testing str / repr. + # .clear() etc. + # IP addresses like 50 (single number, no dot) and domain-matching + # functions (and is_HDN)? See draft RFC 2965 errata. + # Strictness switches + # is_third_party() + # unverifiability / third-party blocking + # Netscape cookies work the same as RFC 2965 with regard to port. + # Set-Cookie with negative max age. + # If turn RFC 2965 handling off, Set-Cookie2 cookies should not clobber + # Set-Cookie cookies. + # Cookie2 should be sent if *any* cookies are not V1 (ie. V0 OR V2 etc.). + # Cookies (V1 and V0) with no expiry date should be set to be discarded. + # RFC 2965 Quoting: + # Should accept unquoted cookie-attribute values? check errata draft. + # Which are required on the way in and out? + # Should always return quoted cookie-attribute values? + # Proper testing of when RFC 2965 clobbers Netscape (waiting for errata). + # Path-match on return (same for V0 and V1). + # RFC 2965 acceptance and returning rules + # Set-Cookie2 without version attribute is rejected. + + # Netscape peculiarities list from Ronald Tschalar. + # The first two still need tests, the rest are covered. +## - Quoting: only quotes around the expires value are recognized as such +## (and yes, some folks quote the expires value); quotes around any other +## value are treated as part of the value. +## - White space: white space around names and values is ignored +## - Default path: if no path parameter is given, the path defaults to the +## path in the request-uri up to, but not including, the last '/'. Note +## that this is entirely different from what the spec says. +## - Commas and other delimiters: Netscape just parses until the next ';'. +## This means it will allow commas etc inside values (and yes, both +## commas and equals are commonly appear in the cookie value). This also +## means that if you fold multiple Set-Cookie header fields into one, +## comma-separated list, it'll be a headache to parse (at least my head +## starts hurting every time I think of that code). +## - Expires: You'll get all sorts of date formats in the expires, +## including emtpy expires attributes ("expires="). Be as flexible as you +## can, and certainly don't expect the weekday to be there; if you can't +## parse it, just ignore it and pretend it's a session cookie. +## - Domain-matching: Netscape uses the 2-dot rule for _all_ domains, not +## just the 7 special TLD's listed in their spec. And folks rely on +## that... + + def test_domain_return_ok(self): + # test optimization: .domain_return_ok() should filter out most + # domains in the CookieJar before we try to access them (because that + # may require disk access -- in particular, with MSIECookieJar) + # This is only a rough check for performance reasons, so it's not too + # critical as long as it's sufficiently liberal. + import cookielib, urllib2 + pol = cookielib.DefaultCookiePolicy() + for url, domain, ok in [ + ("http://foo.bar.com/", "blah.com", False), + ("http://foo.bar.com/", "rhubarb.blah.com", False), + ("http://foo.bar.com/", "rhubarb.foo.bar.com", False), + ("http://foo.bar.com/", ".foo.bar.com", True), + ("http://foo.bar.com/", "foo.bar.com", True), + ("http://foo.bar.com/", ".bar.com", True), + ("http://foo.bar.com/", "com", True), + ("http://foo.com/", "rhubarb.foo.com", False), + ("http://foo.com/", ".foo.com", True), + ("http://foo.com/", "foo.com", True), + ("http://foo.com/", "com", True), + ("http://foo/", "rhubarb.foo", False), + ("http://foo/", ".foo", True), + ("http://foo/", "foo", True), + ("http://foo/", "foo.local", True), + ("http://foo/", ".local", True), + ]: + request = urllib2.Request(url) + r = pol.domain_return_ok(domain, request) + if ok: self.assertTrue(r) + else: self.assertTrue(not r) + + def test_missing_value(self): + from cookielib import MozillaCookieJar, lwp_cookie_str + + # missing = sign in Cookie: header is regarded by Mozilla as a missing + # name, and by cookielib as a missing value + filename = test_support.TESTFN + c = MozillaCookieJar(filename) + interact_netscape(c, "http://www.acme.com/", 'eggs') + interact_netscape(c, "http://www.acme.com/", '"spam"; path=/foo/') + cookie = c._cookies["www.acme.com"]["/"]["eggs"] + self.assertTrue(cookie.value is None) + self.assertEqual(cookie.name, "eggs") + cookie = c._cookies["www.acme.com"]['/foo/']['"spam"'] + self.assertTrue(cookie.value is None) + self.assertEqual(cookie.name, '"spam"') + self.assertEqual(lwp_cookie_str(cookie), ( + r'"spam"; path="/foo/"; domain="www.acme.com"; ' + 'path_spec; discard; version=0')) + old_str = repr(c) + c.save(ignore_expires=True, ignore_discard=True) + try: + c = MozillaCookieJar(filename) + c.revert(ignore_expires=True, ignore_discard=True) + finally: + os.unlink(c.filename) + # cookies unchanged apart from lost info re. whether path was specified + self.assertEqual( + repr(c), + re.sub("path_specified=%s" % True, "path_specified=%s" % False, + old_str) + ) + self.assertEqual(interact_netscape(c, "http://www.acme.com/foo/"), + '"spam"; eggs') + + def test_rfc2109_handling(self): + # RFC 2109 cookies are handled as RFC 2965 or Netscape cookies, + # dependent on policy settings + from cookielib import CookieJar, DefaultCookiePolicy + + for rfc2109_as_netscape, rfc2965, version in [ + # default according to rfc2965 if not explicitly specified + (None, False, 0), + (None, True, 1), + # explicit rfc2109_as_netscape + (False, False, None), # version None here means no cookie stored + (False, True, 1), + (True, False, 0), + (True, True, 0), + ]: + policy = DefaultCookiePolicy( + rfc2109_as_netscape=rfc2109_as_netscape, + rfc2965=rfc2965) + c = CookieJar(policy) + interact_netscape(c, "http://www.example.com/", "ni=ni; Version=1") + try: + cookie = c._cookies["www.example.com"]["/"]["ni"] + except KeyError: + self.assertTrue(version is None) # didn't expect a stored cookie + else: + self.assertEqual(cookie.version, version) + # 2965 cookies are unaffected + interact_2965(c, "http://www.example.com/", + "foo=bar; Version=1") + if rfc2965: + cookie2965 = c._cookies["www.example.com"]["/"]["foo"] + self.assertEqual(cookie2965.version, 1) + + def test_ns_parser(self): + from cookielib import CookieJar, DEFAULT_HTTP_PORT + + c = CookieJar() + interact_netscape(c, "http://www.acme.com/", + 'spam=eggs; DoMain=.acme.com; port; blArgh="feep"') + interact_netscape(c, "http://www.acme.com/", 'ni=ni; port=80,8080') + interact_netscape(c, "http://www.acme.com:80/", 'nini=ni') + interact_netscape(c, "http://www.acme.com:80/", 'foo=bar; expires=') + interact_netscape(c, "http://www.acme.com:80/", 'spam=eggs; ' + 'expires="Foo Bar 25 33:22:11 3022"') + + cookie = c._cookies[".acme.com"]["/"]["spam"] + self.assertEqual(cookie.domain, ".acme.com") + self.assertTrue(cookie.domain_specified) + self.assertEqual(cookie.port, DEFAULT_HTTP_PORT) + self.assertTrue(not cookie.port_specified) + # case is preserved + self.assertTrue(cookie.has_nonstandard_attr("blArgh") and + not cookie.has_nonstandard_attr("blargh")) + + cookie = c._cookies["www.acme.com"]["/"]["ni"] + self.assertEqual(cookie.domain, "www.acme.com") + self.assertTrue(not cookie.domain_specified) + self.assertEqual(cookie.port, "80,8080") + self.assertTrue(cookie.port_specified) + + cookie = c._cookies["www.acme.com"]["/"]["nini"] + self.assertTrue(cookie.port is None) + self.assertTrue(not cookie.port_specified) + + # invalid expires should not cause cookie to be dropped + foo = c._cookies["www.acme.com"]["/"]["foo"] + spam = c._cookies["www.acme.com"]["/"]["foo"] + self.assertTrue(foo.expires is None) + self.assertTrue(spam.expires is None) + + def test_ns_parser_special_names(self): + # names such as 'expires' are not special in first name=value pair + # of Set-Cookie: header + from cookielib import CookieJar + + c = CookieJar() + interact_netscape(c, "http://www.acme.com/", 'expires=eggs') + interact_netscape(c, "http://www.acme.com/", 'version=eggs; spam=eggs') + + cookies = c._cookies["www.acme.com"]["/"] + self.assertTrue('expires' in cookies) + self.assertTrue('version' in cookies) + + def test_expires(self): + from cookielib import time2netscape, CookieJar + + # if expires is in future, keep cookie... + c = CookieJar() + future = time2netscape(time.time()+3600) + interact_netscape(c, "http://www.acme.com/", 'spam="bar"; expires=%s' % + future) + self.assertEqual(len(c), 1) + now = time2netscape(time.time()-1) + # ... and if in past or present, discard it + interact_netscape(c, "http://www.acme.com/", 'foo="eggs"; expires=%s' % + now) + h = interact_netscape(c, "http://www.acme.com/") + self.assertEqual(len(c), 1) + self.assertTrue('spam="bar"' in h and "foo" not in h) + + # max-age takes precedence over expires, and zero max-age is request to + # delete both new cookie and any old matching cookie + interact_netscape(c, "http://www.acme.com/", 'eggs="bar"; expires=%s' % + future) + interact_netscape(c, "http://www.acme.com/", 'bar="bar"; expires=%s' % + future) + self.assertEqual(len(c), 3) + interact_netscape(c, "http://www.acme.com/", 'eggs="bar"; ' + 'expires=%s; max-age=0' % future) + interact_netscape(c, "http://www.acme.com/", 'bar="bar"; ' + 'max-age=0; expires=%s' % future) + h = interact_netscape(c, "http://www.acme.com/") + self.assertEqual(len(c), 1) + + # test expiry at end of session for cookies with no expires attribute + interact_netscape(c, "http://www.rhubarb.net/", 'whum="fizz"') + self.assertEqual(len(c), 2) + c.clear_session_cookies() + self.assertEqual(len(c), 1) + self.assertIn('spam="bar"', h) + + # XXX RFC 2965 expiry rules (some apply to V0 too) + + def test_default_path(self): + from cookielib import CookieJar, DefaultCookiePolicy + + # RFC 2965 + pol = DefaultCookiePolicy(rfc2965=True) + + c = CookieJar(pol) + interact_2965(c, "http://www.acme.com/", 'spam="bar"; Version="1"') + self.assertIn("/", c._cookies["www.acme.com"]) + + c = CookieJar(pol) + interact_2965(c, "http://www.acme.com/blah", 'eggs="bar"; Version="1"') + self.assertIn("/", c._cookies["www.acme.com"]) + + c = CookieJar(pol) + interact_2965(c, "http://www.acme.com/blah/rhubarb", + 'eggs="bar"; Version="1"') + self.assertIn("/blah/", c._cookies["www.acme.com"]) + + c = CookieJar(pol) + interact_2965(c, "http://www.acme.com/blah/rhubarb/", + 'eggs="bar"; Version="1"') + self.assertIn("/blah/rhubarb/", c._cookies["www.acme.com"]) + + # Netscape + + c = CookieJar() + interact_netscape(c, "http://www.acme.com/", 'spam="bar"') + self.assertIn("/", c._cookies["www.acme.com"]) + + c = CookieJar() + interact_netscape(c, "http://www.acme.com/blah", 'eggs="bar"') + self.assertIn("/", c._cookies["www.acme.com"]) + + c = CookieJar() + interact_netscape(c, "http://www.acme.com/blah/rhubarb", 'eggs="bar"') + self.assertIn("/blah", c._cookies["www.acme.com"]) + + c = CookieJar() + interact_netscape(c, "http://www.acme.com/blah/rhubarb/", 'eggs="bar"') + self.assertIn("/blah/rhubarb", c._cookies["www.acme.com"]) + + def test_default_path_with_query(self): + cj = cookielib.CookieJar() + uri = "http://example.com/?spam/eggs" + value = 'eggs="bar"' + interact_netscape(cj, uri, value) + # default path does not include query, so is "/", not "/?spam" + self.assertIn("/", cj._cookies["example.com"]) + # cookie is sent back to the same URI + self.assertEqual(interact_netscape(cj, uri), value) + + def test_escape_path(self): + from cookielib import escape_path + cases = [ + # quoted safe + ("/foo%2f/bar", "/foo%2F/bar"), + ("/foo%2F/bar", "/foo%2F/bar"), + # quoted % + ("/foo%%/bar", "/foo%%/bar"), + # quoted unsafe + ("/fo%19o/bar", "/fo%19o/bar"), + ("/fo%7do/bar", "/fo%7Do/bar"), + # unquoted safe + ("/foo/bar&", "/foo/bar&"), + ("/foo//bar", "/foo//bar"), + ("\176/foo/bar", "\176/foo/bar"), + # unquoted unsafe + ("/foo\031/bar", "/foo%19/bar"), + ("/\175foo/bar", "/%7Dfoo/bar"), + # unicode + (u"/foo/bar\uabcd", "/foo/bar%EA%AF%8D"), # UTF-8 encoded + ] + for arg, result in cases: + self.assertEqual(escape_path(arg), result) + + def test_request_path(self): + from urllib2 import Request + from cookielib import request_path + # with parameters + req = Request("http://www.example.com/rheum/rhaponticum;" + "foo=bar;sing=song?apples=pears&spam=eggs#ni") + self.assertEqual(request_path(req), + "/rheum/rhaponticum;foo=bar;sing=song") + # without parameters + req = Request("http://www.example.com/rheum/rhaponticum?" + "apples=pears&spam=eggs#ni") + self.assertEqual(request_path(req), "/rheum/rhaponticum") + # missing final slash + req = Request("http://www.example.com") + self.assertEqual(request_path(req), "/") + + def test_request_port(self): + from urllib2 import Request + from cookielib import request_port, DEFAULT_HTTP_PORT + req = Request("http://www.acme.com:1234/", + headers={"Host": "www.acme.com:4321"}) + self.assertEqual(request_port(req), "1234") + req = Request("http://www.acme.com/", + headers={"Host": "www.acme.com:4321"}) + self.assertEqual(request_port(req), DEFAULT_HTTP_PORT) + + def test_request_host(self): + from urllib2 import Request + from cookielib import request_host + # this request is illegal (RFC2616, 14.2.3) + req = Request("http://1.1.1.1/", + headers={"Host": "www.acme.com:80"}) + # libwww-perl wants this response, but that seems wrong (RFC 2616, + # section 5.2, point 1., and RFC 2965 section 1, paragraph 3) + #self.assertEqual(request_host(req), "www.acme.com") + self.assertEqual(request_host(req), "1.1.1.1") + req = Request("http://www.acme.com/", + headers={"Host": "irrelevant.com"}) + self.assertEqual(request_host(req), "www.acme.com") + # not actually sure this one is valid Request object, so maybe should + # remove test for no host in url in request_host function? + req = Request("/resource.html", + headers={"Host": "www.acme.com"}) + self.assertEqual(request_host(req), "www.acme.com") + # port shouldn't be in request-host + req = Request("http://www.acme.com:2345/resource.html", + headers={"Host": "www.acme.com:5432"}) + self.assertEqual(request_host(req), "www.acme.com") + + def test_is_HDN(self): + from cookielib import is_HDN + self.assertTrue(is_HDN("foo.bar.com")) + self.assertTrue(is_HDN("1foo2.3bar4.5com")) + self.assertTrue(not is_HDN("192.168.1.1")) + self.assertTrue(not is_HDN("")) + self.assertTrue(not is_HDN(".")) + self.assertTrue(not is_HDN(".foo.bar.com")) + self.assertTrue(not is_HDN("..foo")) + self.assertTrue(not is_HDN("foo.")) + + def test_reach(self): + from cookielib import reach + self.assertEqual(reach("www.acme.com"), ".acme.com") + self.assertEqual(reach("acme.com"), "acme.com") + self.assertEqual(reach("acme.local"), ".local") + self.assertEqual(reach(".local"), ".local") + self.assertEqual(reach(".com"), ".com") + self.assertEqual(reach("."), ".") + self.assertEqual(reach(""), "") + self.assertEqual(reach("192.168.0.1"), "192.168.0.1") + + def test_domain_match(self): + from cookielib import domain_match, user_domain_match + self.assertTrue(domain_match("192.168.1.1", "192.168.1.1")) + self.assertTrue(not domain_match("192.168.1.1", ".168.1.1")) + self.assertTrue(domain_match("x.y.com", "x.Y.com")) + self.assertTrue(domain_match("x.y.com", ".Y.com")) + self.assertTrue(not domain_match("x.y.com", "Y.com")) + self.assertTrue(domain_match("a.b.c.com", ".c.com")) + self.assertTrue(not domain_match(".c.com", "a.b.c.com")) + self.assertTrue(domain_match("example.local", ".local")) + self.assertTrue(not domain_match("blah.blah", "")) + self.assertTrue(not domain_match("", ".rhubarb.rhubarb")) + self.assertTrue(domain_match("", "")) + + self.assertTrue(user_domain_match("acme.com", "acme.com")) + self.assertTrue(not user_domain_match("acme.com", ".acme.com")) + self.assertTrue(user_domain_match("rhubarb.acme.com", ".acme.com")) + self.assertTrue(user_domain_match("www.rhubarb.acme.com", ".acme.com")) + self.assertTrue(user_domain_match("x.y.com", "x.Y.com")) + self.assertTrue(user_domain_match("x.y.com", ".Y.com")) + self.assertTrue(not user_domain_match("x.y.com", "Y.com")) + self.assertTrue(user_domain_match("y.com", "Y.com")) + self.assertTrue(not user_domain_match(".y.com", "Y.com")) + self.assertTrue(user_domain_match(".y.com", ".Y.com")) + self.assertTrue(user_domain_match("x.y.com", ".com")) + self.assertTrue(not user_domain_match("x.y.com", "com")) + self.assertTrue(not user_domain_match("x.y.com", "m")) + self.assertTrue(not user_domain_match("x.y.com", ".m")) + self.assertTrue(not user_domain_match("x.y.com", "")) + self.assertTrue(not user_domain_match("x.y.com", ".")) + self.assertTrue(user_domain_match("192.168.1.1", "192.168.1.1")) + # not both HDNs, so must string-compare equal to match + self.assertTrue(not user_domain_match("192.168.1.1", ".168.1.1")) + self.assertTrue(not user_domain_match("192.168.1.1", ".")) + # empty string is a special case + self.assertTrue(not user_domain_match("192.168.1.1", "")) + + def test_wrong_domain(self): + # Cookies whose effective request-host name does not domain-match the + # domain are rejected. + + # XXX far from complete + from cookielib import CookieJar + c = CookieJar() + interact_2965(c, "http://www.nasty.com/", + 'foo=bar; domain=friendly.org; Version="1"') + self.assertEqual(len(c), 0) + + def test_strict_domain(self): + # Cookies whose domain is a country-code tld like .co.uk should + # not be set if CookiePolicy.strict_domain is true. + from cookielib import CookieJar, DefaultCookiePolicy + + cp = DefaultCookiePolicy(strict_domain=True) + cj = CookieJar(policy=cp) + interact_netscape(cj, "http://example.co.uk/", 'no=problemo') + interact_netscape(cj, "http://example.co.uk/", + 'okey=dokey; Domain=.example.co.uk') + self.assertEqual(len(cj), 2) + for pseudo_tld in [".co.uk", ".org.za", ".tx.us", ".name.us"]: + interact_netscape(cj, "http://example.%s/" % pseudo_tld, + 'spam=eggs; Domain=.co.uk') + self.assertEqual(len(cj), 2) + + def test_two_component_domain_ns(self): + # Netscape: .www.bar.com, www.bar.com, .bar.com, bar.com, no domain + # should all get accepted, as should .acme.com, acme.com and no domain + # for 2-component domains like acme.com. + from cookielib import CookieJar, DefaultCookiePolicy + + c = CookieJar() + + # two-component V0 domain is OK + interact_netscape(c, "http://foo.net/", 'ns=bar') + self.assertEqual(len(c), 1) + self.assertEqual(c._cookies["foo.net"]["/"]["ns"].value, "bar") + self.assertEqual(interact_netscape(c, "http://foo.net/"), "ns=bar") + # *will* be returned to any other domain (unlike RFC 2965)... + self.assertEqual(interact_netscape(c, "http://www.foo.net/"), + "ns=bar") + # ...unless requested otherwise + pol = DefaultCookiePolicy( + strict_ns_domain=DefaultCookiePolicy.DomainStrictNonDomain) + c.set_policy(pol) + self.assertEqual(interact_netscape(c, "http://www.foo.net/"), "") + + # unlike RFC 2965, even explicit two-component domain is OK, + # because .foo.net matches foo.net + interact_netscape(c, "http://foo.net/foo/", + 'spam1=eggs; domain=foo.net') + # even if starts with a dot -- in NS rules, .foo.net matches foo.net! + interact_netscape(c, "http://foo.net/foo/bar/", + 'spam2=eggs; domain=.foo.net') + self.assertEqual(len(c), 3) + self.assertEqual(c._cookies[".foo.net"]["/foo"]["spam1"].value, + "eggs") + self.assertEqual(c._cookies[".foo.net"]["/foo/bar"]["spam2"].value, + "eggs") + self.assertEqual(interact_netscape(c, "http://foo.net/foo/bar/"), + "spam2=eggs; spam1=eggs; ns=bar") + + # top-level domain is too general + interact_netscape(c, "http://foo.net/", 'nini="ni"; domain=.net') + self.assertEqual(len(c), 3) + +## # Netscape protocol doesn't allow non-special top level domains (such +## # as co.uk) in the domain attribute unless there are at least three +## # dots in it. + # Oh yes it does! Real implementations don't check this, and real + # cookies (of course) rely on that behaviour. + interact_netscape(c, "http://foo.co.uk", 'nasty=trick; domain=.co.uk') +## self.assertEqual(len(c), 2) + self.assertEqual(len(c), 4) + + def test_two_component_domain_rfc2965(self): + from cookielib import CookieJar, DefaultCookiePolicy + + pol = DefaultCookiePolicy(rfc2965=True) + c = CookieJar(pol) + + # two-component V1 domain is OK + interact_2965(c, "http://foo.net/", 'foo=bar; Version="1"') + self.assertEqual(len(c), 1) + self.assertEqual(c._cookies["foo.net"]["/"]["foo"].value, "bar") + self.assertEqual(interact_2965(c, "http://foo.net/"), + "$Version=1; foo=bar") + # won't be returned to any other domain (because domain was implied) + self.assertEqual(interact_2965(c, "http://www.foo.net/"), "") + + # unless domain is given explicitly, because then it must be + # rewritten to start with a dot: foo.net --> .foo.net, which does + # not domain-match foo.net + interact_2965(c, "http://foo.net/foo", + 'spam=eggs; domain=foo.net; path=/foo; Version="1"') + self.assertEqual(len(c), 1) + self.assertEqual(interact_2965(c, "http://foo.net/foo"), + "$Version=1; foo=bar") + + # explicit foo.net from three-component domain www.foo.net *does* get + # set, because .foo.net domain-matches .foo.net + interact_2965(c, "http://www.foo.net/foo/", + 'spam=eggs; domain=foo.net; Version="1"') + self.assertEqual(c._cookies[".foo.net"]["/foo/"]["spam"].value, + "eggs") + self.assertEqual(len(c), 2) + self.assertEqual(interact_2965(c, "http://foo.net/foo/"), + "$Version=1; foo=bar") + self.assertEqual(interact_2965(c, "http://www.foo.net/foo/"), + '$Version=1; spam=eggs; $Domain="foo.net"') + + # top-level domain is too general + interact_2965(c, "http://foo.net/", + 'ni="ni"; domain=".net"; Version="1"') + self.assertEqual(len(c), 2) + + # RFC 2965 doesn't require blocking this + interact_2965(c, "http://foo.co.uk/", + 'nasty=trick; domain=.co.uk; Version="1"') + self.assertEqual(len(c), 3) + + def test_domain_allow(self): + from cookielib import CookieJar, DefaultCookiePolicy + from urllib2 import Request + + c = CookieJar(policy=DefaultCookiePolicy( + blocked_domains=["acme.com"], + allowed_domains=["www.acme.com"])) + + req = Request("http://acme.com/") + headers = ["Set-Cookie: CUSTOMER=WILE_E_COYOTE; path=/"] + res = FakeResponse(headers, "http://acme.com/") + c.extract_cookies(res, req) + self.assertEqual(len(c), 0) + + req = Request("http://www.acme.com/") + res = FakeResponse(headers, "http://www.acme.com/") + c.extract_cookies(res, req) + self.assertEqual(len(c), 1) + + req = Request("http://www.coyote.com/") + res = FakeResponse(headers, "http://www.coyote.com/") + c.extract_cookies(res, req) + self.assertEqual(len(c), 1) + + # set a cookie with non-allowed domain... + req = Request("http://www.coyote.com/") + res = FakeResponse(headers, "http://www.coyote.com/") + cookies = c.make_cookies(res, req) + c.set_cookie(cookies[0]) + self.assertEqual(len(c), 2) + # ... and check is doesn't get returned + c.add_cookie_header(req) + self.assertTrue(not req.has_header("Cookie")) + + def test_domain_block(self): + from cookielib import CookieJar, DefaultCookiePolicy + from urllib2 import Request + + pol = DefaultCookiePolicy( + rfc2965=True, blocked_domains=[".acme.com"]) + c = CookieJar(policy=pol) + headers = ["Set-Cookie: CUSTOMER=WILE_E_COYOTE; path=/"] + + req = Request("http://www.acme.com/") + res = FakeResponse(headers, "http://www.acme.com/") + c.extract_cookies(res, req) + self.assertEqual(len(c), 0) + + p = pol.set_blocked_domains(["acme.com"]) + c.extract_cookies(res, req) + self.assertEqual(len(c), 1) + + c.clear() + req = Request("http://www.roadrunner.net/") + res = FakeResponse(headers, "http://www.roadrunner.net/") + c.extract_cookies(res, req) + self.assertEqual(len(c), 1) + req = Request("http://www.roadrunner.net/") + c.add_cookie_header(req) + self.assertTrue((req.has_header("Cookie") and + req.has_header("Cookie2"))) + + c.clear() + pol.set_blocked_domains([".acme.com"]) + c.extract_cookies(res, req) + self.assertEqual(len(c), 1) + + # set a cookie with blocked domain... + req = Request("http://www.acme.com/") + res = FakeResponse(headers, "http://www.acme.com/") + cookies = c.make_cookies(res, req) + c.set_cookie(cookies[0]) + self.assertEqual(len(c), 2) + # ... and check is doesn't get returned + c.add_cookie_header(req) + self.assertTrue(not req.has_header("Cookie")) + + def test_secure(self): + from cookielib import CookieJar, DefaultCookiePolicy + + for ns in True, False: + for whitespace in " ", "": + c = CookieJar() + if ns: + pol = DefaultCookiePolicy(rfc2965=False) + int = interact_netscape + vs = "" + else: + pol = DefaultCookiePolicy(rfc2965=True) + int = interact_2965 + vs = "; Version=1" + c.set_policy(pol) + url = "http://www.acme.com/" + int(c, url, "foo1=bar%s%s" % (vs, whitespace)) + int(c, url, "foo2=bar%s; secure%s" % (vs, whitespace)) + self.assertTrue( + not c._cookies["www.acme.com"]["/"]["foo1"].secure, + "non-secure cookie registered secure") + self.assertTrue( + c._cookies["www.acme.com"]["/"]["foo2"].secure, + "secure cookie registered non-secure") + + def test_quote_cookie_value(self): + from cookielib import CookieJar, DefaultCookiePolicy + c = CookieJar(policy=DefaultCookiePolicy(rfc2965=True)) + interact_2965(c, "http://www.acme.com/", r'foo=\b"a"r; Version=1') + h = interact_2965(c, "http://www.acme.com/") + self.assertEqual(h, r'$Version=1; foo=\\b\"a\"r') + + def test_missing_final_slash(self): + # Missing slash from request URL's abs_path should be assumed present. + from cookielib import CookieJar, DefaultCookiePolicy + from urllib2 import Request + url = "http://www.acme.com" + c = CookieJar(DefaultCookiePolicy(rfc2965=True)) + interact_2965(c, url, "foo=bar; Version=1") + req = Request(url) + self.assertEqual(len(c), 1) + c.add_cookie_header(req) + self.assertTrue(req.has_header("Cookie")) + + def test_domain_mirror(self): + from cookielib import CookieJar, DefaultCookiePolicy + + pol = DefaultCookiePolicy(rfc2965=True) + + c = CookieJar(pol) + url = "http://foo.bar.com/" + interact_2965(c, url, "spam=eggs; Version=1") + h = interact_2965(c, url) + self.assertNotIn("Domain", h, + "absent domain returned with domain present") + + c = CookieJar(pol) + url = "http://foo.bar.com/" + interact_2965(c, url, 'spam=eggs; Version=1; Domain=.bar.com') + h = interact_2965(c, url) + self.assertIn('$Domain=".bar.com"', h, "domain not returned") + + c = CookieJar(pol) + url = "http://foo.bar.com/" + # note missing initial dot in Domain + interact_2965(c, url, 'spam=eggs; Version=1; Domain=bar.com') + h = interact_2965(c, url) + self.assertIn('$Domain="bar.com"', h, "domain not returned") + + def test_path_mirror(self): + from cookielib import CookieJar, DefaultCookiePolicy + + pol = DefaultCookiePolicy(rfc2965=True) + + c = CookieJar(pol) + url = "http://foo.bar.com/" + interact_2965(c, url, "spam=eggs; Version=1") + h = interact_2965(c, url) + self.assertNotIn("Path", h, "absent path returned with path present") + + c = CookieJar(pol) + url = "http://foo.bar.com/" + interact_2965(c, url, 'spam=eggs; Version=1; Path=/') + h = interact_2965(c, url) + self.assertIn('$Path="/"', h, "path not returned") + + def test_port_mirror(self): + from cookielib import CookieJar, DefaultCookiePolicy + + pol = DefaultCookiePolicy(rfc2965=True) + + c = CookieJar(pol) + url = "http://foo.bar.com/" + interact_2965(c, url, "spam=eggs; Version=1") + h = interact_2965(c, url) + self.assertNotIn("Port", h, "absent port returned with port present") + + c = CookieJar(pol) + url = "http://foo.bar.com/" + interact_2965(c, url, "spam=eggs; Version=1; Port") + h = interact_2965(c, url) + self.assertTrue(re.search("\$Port([^=]|$)", h), + "port with no value not returned with no value") + + c = CookieJar(pol) + url = "http://foo.bar.com/" + interact_2965(c, url, 'spam=eggs; Version=1; Port="80"') + h = interact_2965(c, url) + self.assertIn('$Port="80"', h, + "port with single value not returned with single value") + + c = CookieJar(pol) + url = "http://foo.bar.com/" + interact_2965(c, url, 'spam=eggs; Version=1; Port="80,8080"') + h = interact_2965(c, url) + self.assertIn('$Port="80,8080"', h, + "port with multiple values not returned with multiple " + "values") + + def test_no_return_comment(self): + from cookielib import CookieJar, DefaultCookiePolicy + + c = CookieJar(DefaultCookiePolicy(rfc2965=True)) + url = "http://foo.bar.com/" + interact_2965(c, url, 'spam=eggs; Version=1; ' + 'Comment="does anybody read these?"; ' + 'CommentURL="http://foo.bar.net/comment.html"') + h = interact_2965(c, url) + self.assertTrue( + "Comment" not in h, + "Comment or CommentURL cookie-attributes returned to server") + + def test_Cookie_iterator(self): + from cookielib import CookieJar, Cookie, DefaultCookiePolicy + + cs = CookieJar(DefaultCookiePolicy(rfc2965=True)) + # add some random cookies + interact_2965(cs, "http://blah.spam.org/", 'foo=eggs; Version=1; ' + 'Comment="does anybody read these?"; ' + 'CommentURL="http://foo.bar.net/comment.html"') + interact_netscape(cs, "http://www.acme.com/blah/", "spam=bar; secure") + interact_2965(cs, "http://www.acme.com/blah/", + "foo=bar; secure; Version=1") + interact_2965(cs, "http://www.acme.com/blah/", + "foo=bar; path=/; Version=1") + interact_2965(cs, "http://www.sol.no", + r'bang=wallop; version=1; domain=".sol.no"; ' + r'port="90,100, 80,8080"; ' + r'max-age=100; Comment = "Just kidding! (\"|\\\\) "') + + versions = [1, 1, 1, 0, 1] + names = ["bang", "foo", "foo", "spam", "foo"] + domains = [".sol.no", "blah.spam.org", "www.acme.com", + "www.acme.com", "www.acme.com"] + paths = ["/", "/", "/", "/blah", "/blah/"] + + for i in range(4): + i = 0 + for c in cs: + self.assertIsInstance(c, Cookie) + self.assertEqual(c.version, versions[i]) + self.assertEqual(c.name, names[i]) + self.assertEqual(c.domain, domains[i]) + self.assertEqual(c.path, paths[i]) + i = i + 1 + + def test_parse_ns_headers(self): + from cookielib import parse_ns_headers + + # missing domain value (invalid cookie) + self.assertEqual( + parse_ns_headers(["foo=bar; path=/; domain"]), + [[("foo", "bar"), + ("path", "/"), ("domain", None), ("version", "0")]] + ) + # invalid expires value + self.assertEqual( + parse_ns_headers(["foo=bar; expires=Foo Bar 12 33:22:11 2000"]), + [[("foo", "bar"), ("expires", None), ("version", "0")]] + ) + # missing cookie value (valid cookie) + self.assertEqual( + parse_ns_headers(["foo"]), + [[("foo", None), ("version", "0")]] + ) + # shouldn't add version if header is empty + self.assertEqual(parse_ns_headers([""]), []) + + def test_bad_cookie_header(self): + + def cookiejar_from_cookie_headers(headers): + from cookielib import CookieJar + from urllib2 import Request + c = CookieJar() + req = Request("http://www.example.com/") + r = FakeResponse(headers, "http://www.example.com/") + c.extract_cookies(r, req) + return c + + # none of these bad headers should cause an exception to be raised + for headers in [ + ["Set-Cookie: "], # actually, nothing wrong with this + ["Set-Cookie2: "], # ditto + # missing domain value + ["Set-Cookie2: a=foo; path=/; Version=1; domain"], + # bad max-age + ["Set-Cookie: b=foo; max-age=oops"], + # bad version + ["Set-Cookie: b=foo; version=spam"], + ]: + c = cookiejar_from_cookie_headers(headers) + # these bad cookies shouldn't be set + self.assertEqual(len(c), 0) + + # cookie with invalid expires is treated as session cookie + headers = ["Set-Cookie: c=foo; expires=Foo Bar 12 33:22:11 2000"] + c = cookiejar_from_cookie_headers(headers) + cookie = c._cookies["www.example.com"]["/"]["c"] + self.assertTrue(cookie.expires is None) + + +class LWPCookieTests(TestCase): + # Tests taken from libwww-perl, with a few modifications and additions. + + def test_netscape_example_1(self): + from cookielib import CookieJar, DefaultCookiePolicy + from urllib2 import Request + + #------------------------------------------------------------------- + # First we check that it works for the original example at + # http://www.netscape.com/newsref/std/cookie_spec.html + + # Client requests a document, and receives in the response: + # + # Set-Cookie: CUSTOMER=WILE_E_COYOTE; path=/; expires=Wednesday, 09-Nov-99 23:12:40 GMT + # + # When client requests a URL in path "/" on this server, it sends: + # + # Cookie: CUSTOMER=WILE_E_COYOTE + # + # Client requests a document, and receives in the response: + # + # Set-Cookie: PART_NUMBER=ROCKET_LAUNCHER_0001; path=/ + # + # When client requests a URL in path "/" on this server, it sends: + # + # Cookie: CUSTOMER=WILE_E_COYOTE; PART_NUMBER=ROCKET_LAUNCHER_0001 + # + # Client receives: + # + # Set-Cookie: SHIPPING=FEDEX; path=/fo + # + # When client requests a URL in path "/" on this server, it sends: + # + # Cookie: CUSTOMER=WILE_E_COYOTE; PART_NUMBER=ROCKET_LAUNCHER_0001 + # + # When client requests a URL in path "/foo" on this server, it sends: + # + # Cookie: CUSTOMER=WILE_E_COYOTE; PART_NUMBER=ROCKET_LAUNCHER_0001; SHIPPING=FEDEX + # + # The last Cookie is buggy, because both specifications say that the + # most specific cookie must be sent first. SHIPPING=FEDEX is the + # most specific and should thus be first. + + year_plus_one = time.localtime()[0] + 1 + + headers = [] + + c = CookieJar(DefaultCookiePolicy(rfc2965 = True)) + + #req = Request("http://1.1.1.1/", + # headers={"Host": "www.acme.com:80"}) + req = Request("http://www.acme.com:80/", + headers={"Host": "www.acme.com:80"}) + + headers.append( + "Set-Cookie: CUSTOMER=WILE_E_COYOTE; path=/ ; " + "expires=Wednesday, 09-Nov-%d 23:12:40 GMT" % year_plus_one) + res = FakeResponse(headers, "http://www.acme.com/") + c.extract_cookies(res, req) + + req = Request("http://www.acme.com/") + c.add_cookie_header(req) + + self.assertEqual(req.get_header("Cookie"), "CUSTOMER=WILE_E_COYOTE") + self.assertEqual(req.get_header("Cookie2"), '$Version="1"') + + headers.append("Set-Cookie: PART_NUMBER=ROCKET_LAUNCHER_0001; path=/") + res = FakeResponse(headers, "http://www.acme.com/") + c.extract_cookies(res, req) + + req = Request("http://www.acme.com/foo/bar") + c.add_cookie_header(req) + + h = req.get_header("Cookie") + self.assertIn("PART_NUMBER=ROCKET_LAUNCHER_0001", h) + self.assertIn("CUSTOMER=WILE_E_COYOTE", h) + + headers.append('Set-Cookie: SHIPPING=FEDEX; path=/foo') + res = FakeResponse(headers, "http://www.acme.com") + c.extract_cookies(res, req) + + req = Request("http://www.acme.com/") + c.add_cookie_header(req) + + h = req.get_header("Cookie") + self.assertIn("PART_NUMBER=ROCKET_LAUNCHER_0001", h) + self.assertIn("CUSTOMER=WILE_E_COYOTE", h) + self.assertNotIn("SHIPPING=FEDEX", h) + + req = Request("http://www.acme.com/foo/") + c.add_cookie_header(req) + + h = req.get_header("Cookie") + self.assertIn("PART_NUMBER=ROCKET_LAUNCHER_0001", h) + self.assertIn("CUSTOMER=WILE_E_COYOTE", h) + self.assertTrue(h.startswith("SHIPPING=FEDEX;")) + + def test_netscape_example_2(self): + from cookielib import CookieJar + from urllib2 import Request + + # Second Example transaction sequence: + # + # Assume all mappings from above have been cleared. + # + # Client receives: + # + # Set-Cookie: PART_NUMBER=ROCKET_LAUNCHER_0001; path=/ + # + # When client requests a URL in path "/" on this server, it sends: + # + # Cookie: PART_NUMBER=ROCKET_LAUNCHER_0001 + # + # Client receives: + # + # Set-Cookie: PART_NUMBER=RIDING_ROCKET_0023; path=/ammo + # + # When client requests a URL in path "/ammo" on this server, it sends: + # + # Cookie: PART_NUMBER=RIDING_ROCKET_0023; PART_NUMBER=ROCKET_LAUNCHER_0001 + # + # NOTE: There are two name/value pairs named "PART_NUMBER" due to + # the inheritance of the "/" mapping in addition to the "/ammo" mapping. + + c = CookieJar() + headers = [] + + req = Request("http://www.acme.com/") + headers.append("Set-Cookie: PART_NUMBER=ROCKET_LAUNCHER_0001; path=/") + res = FakeResponse(headers, "http://www.acme.com/") + + c.extract_cookies(res, req) + + req = Request("http://www.acme.com/") + c.add_cookie_header(req) + + self.assertEqual(req.get_header("Cookie"), + "PART_NUMBER=ROCKET_LAUNCHER_0001") + + headers.append( + "Set-Cookie: PART_NUMBER=RIDING_ROCKET_0023; path=/ammo") + res = FakeResponse(headers, "http://www.acme.com/") + c.extract_cookies(res, req) + + req = Request("http://www.acme.com/ammo") + c.add_cookie_header(req) + + self.assertTrue(re.search(r"PART_NUMBER=RIDING_ROCKET_0023;\s*" + "PART_NUMBER=ROCKET_LAUNCHER_0001", + req.get_header("Cookie"))) + + def test_ietf_example_1(self): + from cookielib import CookieJar, DefaultCookiePolicy + #------------------------------------------------------------------- + # Then we test with the examples from draft-ietf-http-state-man-mec-03.txt + # + # 5. EXAMPLES + + c = CookieJar(DefaultCookiePolicy(rfc2965=True)) + + # + # 5.1 Example 1 + # + # Most detail of request and response headers has been omitted. Assume + # the user agent has no stored cookies. + # + # 1. User Agent -> Server + # + # POST /acme/login HTTP/1.1 + # [form data] + # + # User identifies self via a form. + # + # 2. Server -> User Agent + # + # HTTP/1.1 200 OK + # Set-Cookie2: Customer="WILE_E_COYOTE"; Version="1"; Path="/acme" + # + # Cookie reflects user's identity. + + cookie = interact_2965( + c, 'http://www.acme.com/acme/login', + 'Customer="WILE_E_COYOTE"; Version="1"; Path="/acme"') + self.assertTrue(not cookie) + + # + # 3. User Agent -> Server + # + # POST /acme/pickitem HTTP/1.1 + # Cookie: $Version="1"; Customer="WILE_E_COYOTE"; $Path="/acme" + # [form data] + # + # User selects an item for ``shopping basket.'' + # + # 4. Server -> User Agent + # + # HTTP/1.1 200 OK + # Set-Cookie2: Part_Number="Rocket_Launcher_0001"; Version="1"; + # Path="/acme" + # + # Shopping basket contains an item. + + cookie = interact_2965(c, 'http://www.acme.com/acme/pickitem', + 'Part_Number="Rocket_Launcher_0001"; ' + 'Version="1"; Path="/acme"'); + self.assertTrue(re.search( + r'^\$Version="?1"?; Customer="?WILE_E_COYOTE"?; \$Path="/acme"$', + cookie)) + + # + # 5. User Agent -> Server + # + # POST /acme/shipping HTTP/1.1 + # Cookie: $Version="1"; + # Customer="WILE_E_COYOTE"; $Path="/acme"; + # Part_Number="Rocket_Launcher_0001"; $Path="/acme" + # [form data] + # + # User selects shipping method from form. + # + # 6. Server -> User Agent + # + # HTTP/1.1 200 OK + # Set-Cookie2: Shipping="FedEx"; Version="1"; Path="/acme" + # + # New cookie reflects shipping method. + + cookie = interact_2965(c, "http://www.acme.com/acme/shipping", + 'Shipping="FedEx"; Version="1"; Path="/acme"') + + self.assertTrue(re.search(r'^\$Version="?1"?;', cookie)) + self.assertTrue(re.search(r'Part_Number="?Rocket_Launcher_0001"?;' + '\s*\$Path="\/acme"', cookie)) + self.assertTrue(re.search(r'Customer="?WILE_E_COYOTE"?;\s*\$Path="\/acme"', + cookie)) + + # + # 7. User Agent -> Server + # + # POST /acme/process HTTP/1.1 + # Cookie: $Version="1"; + # Customer="WILE_E_COYOTE"; $Path="/acme"; + # Part_Number="Rocket_Launcher_0001"; $Path="/acme"; + # Shipping="FedEx"; $Path="/acme" + # [form data] + # + # User chooses to process order. + # + # 8. Server -> User Agent + # + # HTTP/1.1 200 OK + # + # Transaction is complete. + + cookie = interact_2965(c, "http://www.acme.com/acme/process") + self.assertTrue( + re.search(r'Shipping="?FedEx"?;\s*\$Path="\/acme"', cookie) and + "WILE_E_COYOTE" in cookie) + + # + # The user agent makes a series of requests on the origin server, after + # each of which it receives a new cookie. All the cookies have the same + # Path attribute and (default) domain. Because the request URLs all have + # /acme as a prefix, and that matches the Path attribute, each request + # contains all the cookies received so far. + + def test_ietf_example_2(self): + from cookielib import CookieJar, DefaultCookiePolicy + + # 5.2 Example 2 + # + # This example illustrates the effect of the Path attribute. All detail + # of request and response headers has been omitted. Assume the user agent + # has no stored cookies. + + c = CookieJar(DefaultCookiePolicy(rfc2965=True)) + + # Imagine the user agent has received, in response to earlier requests, + # the response headers + # + # Set-Cookie2: Part_Number="Rocket_Launcher_0001"; Version="1"; + # Path="/acme" + # + # and + # + # Set-Cookie2: Part_Number="Riding_Rocket_0023"; Version="1"; + # Path="/acme/ammo" + + interact_2965( + c, "http://www.acme.com/acme/ammo/specific", + 'Part_Number="Rocket_Launcher_0001"; Version="1"; Path="/acme"', + 'Part_Number="Riding_Rocket_0023"; Version="1"; Path="/acme/ammo"') + + # A subsequent request by the user agent to the (same) server for URLs of + # the form /acme/ammo/... would include the following request header: + # + # Cookie: $Version="1"; + # Part_Number="Riding_Rocket_0023"; $Path="/acme/ammo"; + # Part_Number="Rocket_Launcher_0001"; $Path="/acme" + # + # Note that the NAME=VALUE pair for the cookie with the more specific Path + # attribute, /acme/ammo, comes before the one with the less specific Path + # attribute, /acme. Further note that the same cookie name appears more + # than once. + + cookie = interact_2965(c, "http://www.acme.com/acme/ammo/...") + self.assertTrue( + re.search(r"Riding_Rocket_0023.*Rocket_Launcher_0001", cookie)) + + # A subsequent request by the user agent to the (same) server for a URL of + # the form /acme/parts/ would include the following request header: + # + # Cookie: $Version="1"; Part_Number="Rocket_Launcher_0001"; $Path="/acme" + # + # Here, the second cookie's Path attribute /acme/ammo is not a prefix of + # the request URL, /acme/parts/, so the cookie does not get forwarded to + # the server. + + cookie = interact_2965(c, "http://www.acme.com/acme/parts/") + self.assertIn("Rocket_Launcher_0001", cookie) + self.assertNotIn("Riding_Rocket_0023", cookie) + + def test_rejection(self): + # Test rejection of Set-Cookie2 responses based on domain, path, port. + from cookielib import DefaultCookiePolicy, LWPCookieJar + + pol = DefaultCookiePolicy(rfc2965=True) + + c = LWPCookieJar(policy=pol) + + max_age = "max-age=3600" + + # illegal domain (no embedded dots) + cookie = interact_2965(c, "http://www.acme.com", + 'foo=bar; domain=".com"; version=1') + self.assertTrue(not c) + + # legal domain + cookie = interact_2965(c, "http://www.acme.com", + 'ping=pong; domain="acme.com"; version=1') + self.assertEqual(len(c), 1) + + # illegal domain (host prefix "www.a" contains a dot) + cookie = interact_2965(c, "http://www.a.acme.com", + 'whiz=bang; domain="acme.com"; version=1') + self.assertEqual(len(c), 1) + + # legal domain + cookie = interact_2965(c, "http://www.a.acme.com", + 'wow=flutter; domain=".a.acme.com"; version=1') + self.assertEqual(len(c), 2) + + # can't partially match an IP-address + cookie = interact_2965(c, "http://125.125.125.125", + 'zzzz=ping; domain="125.125.125"; version=1') + self.assertEqual(len(c), 2) + + # illegal path (must be prefix of request path) + cookie = interact_2965(c, "http://www.sol.no", + 'blah=rhubarb; domain=".sol.no"; path="/foo"; ' + 'version=1') + self.assertEqual(len(c), 2) + + # legal path + cookie = interact_2965(c, "http://www.sol.no/foo/bar", + 'bing=bong; domain=".sol.no"; path="/foo"; ' + 'version=1') + self.assertEqual(len(c), 3) + + # illegal port (request-port not in list) + cookie = interact_2965(c, "http://www.sol.no", + 'whiz=ffft; domain=".sol.no"; port="90,100"; ' + 'version=1') + self.assertEqual(len(c), 3) + + # legal port + cookie = interact_2965( + c, "http://www.sol.no", + r'bang=wallop; version=1; domain=".sol.no"; ' + r'port="90,100, 80,8080"; ' + r'max-age=100; Comment = "Just kidding! (\"|\\\\) "') + self.assertEqual(len(c), 4) + + # port attribute without any value (current port) + cookie = interact_2965(c, "http://www.sol.no", + 'foo9=bar; version=1; domain=".sol.no"; port; ' + 'max-age=100;') + self.assertEqual(len(c), 5) + + # encoded path + # LWP has this test, but unescaping allowed path characters seems + # like a bad idea, so I think this should fail: +## cookie = interact_2965(c, "http://www.sol.no/foo/", +## r'foo8=bar; version=1; path="/%66oo"') + # but this is OK, because '<' is not an allowed HTTP URL path + # character: + cookie = interact_2965(c, "http://www.sol.no/ http://hg.python.org/jython/rev/ba650dd24933 changeset: 7184:ba650dd24933 user: Jeff Allen date: Sat Feb 15 17:25:32 2014 +0000 summary: Fixes failure to close file in test_old_mailbox.MboxTestCase.test_from_regex. Failure to close causes failure to delete, and erratic behaviour in regression tests. files: Lib/test/test_old_mailbox.py | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/Lib/test/test_old_mailbox.py b/Lib/test/test_old_mailbox.py --- a/Lib/test/test_old_mailbox.py +++ b/Lib/test/test_old_mailbox.py @@ -144,12 +144,12 @@ body4 """) f.close() - box = mailbox.UnixMailbox(open(self._path, 'r')) + box = mailbox.UnixMailbox(open(self._path, 'rb')) messages = list(iter(box)) self.assert_(len(messages) == 4) for message in messages: message.fp.close() - + box.fp.close() # Jython addition: explicit close needed # XXX We still need more tests! -- Repository URL: http://hg.python.org/jython From jython-checkins at python.org Wed Feb 19 23:48:07 2014 From: jython-checkins at python.org (jeff.allen) Date: Wed, 19 Feb 2014 23:48:07 +0100 (CET) Subject: [Jython-checkins] =?utf-8?q?jython=3A_Skip_testing_os=2Elink_wher?= =?utf-8?q?e_it_doesn=27t_exist=2E?= Message-ID: <3fTvGb2ltCz7Ljr@mail.python.org> http://hg.python.org/jython/rev/b3d9ae5b6942 changeset: 7185:b3d9ae5b6942 user: Jeff Allen date: Sat Feb 15 18:30:40 2014 +0000 summary: Skip testing os.link where it doesn't exist. files: Lib/test/test_os_jy.py | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/Lib/test/test_os_jy.py b/Lib/test/test_os_jy.py --- a/Lib/test/test_os_jy.py +++ b/Lib/test/test_os_jy.py @@ -22,6 +22,7 @@ os.remove(test_support.TESTFN) self.assertRaises(OSError, os.utime, test_support.TESTFN, None) + @unittest.skipUnless(hasattr(os, 'link'), "os.link not available") def test_issue1824(self): os.remove(test_support.TESTFN) self.assertRaises(OSError, os.link, -- Repository URL: http://hg.python.org/jython From jython-checkins at python.org Wed Feb 19 23:48:08 2014 From: jython-checkins at python.org (jeff.allen) Date: Wed, 19 Feb 2014 23:48:08 +0100 (CET) Subject: [Jython-checkins] =?utf-8?q?jython=3A_Add_custom_test=5Fpopen2_so?= =?utf-8?q?_that_we_test_with_=27more=27_not_=27cat=27_on_Windows=2E?= Message-ID: <3fTvGc55pqz7Ljd@mail.python.org> http://hg.python.org/jython/rev/cd84a196c282 changeset: 7186:cd84a196c282 user: Jeff Allen date: Sun Feb 16 18:02:55 2014 +0000 summary: Add custom test_popen2 so that we test with 'more' not 'cat' on Windows. files: Lib/test/test_popen2.py | 130 ++++++++++++++++++++++++++++ 1 files changed, 130 insertions(+), 0 deletions(-) diff --git a/Lib/test/test_popen2.py b/Lib/test/test_popen2.py new file mode 100644 --- /dev/null +++ b/Lib/test/test_popen2.py @@ -0,0 +1,130 @@ +#! /usr/bin/env python +"""Test script for popen2.py""" + +import warnings +warnings.filterwarnings("ignore", ".*popen2 module is deprecated.*", + DeprecationWarning) +warnings.filterwarnings("ignore", "os\.popen. is deprecated.*", + DeprecationWarning) + +import os +import sys +import unittest +import popen2 + +from test.test_support import run_unittest, reap_children + +if sys.platform[:4] == 'beos' or sys.platform[:6] == 'atheos': + # Locks get messed up or something. Generally we're supposed + # to avoid mixing "posix" fork & exec with native threads, and + # they may be right about that after all. + raise unittest.SkipTest("popen2() doesn't work on " + sys.platform) + +# if we don't have os.popen, check that +# we have os.fork. if not, skip the test +# (by raising an ImportError) +try: + from os import popen + del popen +except ImportError: + from os import fork + del fork + +class Popen2Test(unittest.TestCase): + cmd = "cat" + if os.name == "nt" or (os.name == "java" and os._name == "nt"): + cmd = "more" + teststr = "ab cd\n" + # "more" doesn't act the same way across Windows flavors, + # sometimes adding an extra newline at the start or the + # end. So we strip whitespace off both ends for comparison. + expected = teststr.strip() + + def setUp(self): + popen2._cleanup() + # When the test runs, there shouldn't be any open pipes + self.assertFalse(popen2._active, "Active pipes when test starts" + + repr([c.cmd for c in popen2._active])) + + def tearDown(self): + for inst in popen2._active: + inst.wait() + popen2._cleanup() + self.assertFalse(popen2._active, "popen2._active not empty") + # The os.popen*() API delegates to the subprocess module (on Unix) + import subprocess + for inst in subprocess._active: + inst.wait() + subprocess._cleanup() + self.assertFalse(subprocess._active, "subprocess._active not empty") + reap_children() + + def validate_output(self, teststr, expected_out, r, w, e=None): + w.write(teststr) + w.close() + got = r.read() + self.assertEqual(expected_out, got.strip(), "wrote %r read %r" % + (teststr, got)) + + if e is not None: + got = e.read() + self.assertFalse(got, "unexpected %r on stderr" % got) + + def test_popen2(self): + r, w = popen2.popen2(self.cmd) + self.validate_output(self.teststr, self.expected, r, w) + + def test_popen3(self): + if os.name == 'posix': + r, w, e = popen2.popen3([self.cmd]) + self.validate_output(self.teststr, self.expected, r, w, e) + + r, w, e = popen2.popen3(self.cmd) + self.validate_output(self.teststr, self.expected, r, w, e) + + def test_os_popen2(self): + # same test as test_popen2(), but using the os.popen*() API + if os.name == 'posix': + w, r = os.popen2([self.cmd]) + self.validate_output(self.teststr, self.expected, r, w) + + w, r = os.popen2(["echo", self.teststr]) + got = r.read() + self.assertEqual(got, self.teststr + "\n") + + w, r = os.popen2(self.cmd) + self.validate_output(self.teststr, self.expected, r, w) + + def test_os_popen3(self): + # same test as test_popen3(), but using the os.popen*() API + if os.name == 'posix': + w, r, e = os.popen3([self.cmd]) + self.validate_output(self.teststr, self.expected, r, w, e) + + w, r, e = os.popen3(["echo", self.teststr]) + got = r.read() + self.assertEqual(got, self.teststr + "\n") + got = e.read() + self.assertFalse(got, "unexpected %r on stderr" % got) + + w, r, e = os.popen3(self.cmd) + self.validate_output(self.teststr, self.expected, r, w, e) + + def test_os_popen4(self): + if os.name == 'posix': + w, r = os.popen4([self.cmd]) + self.validate_output(self.teststr, self.expected, r, w) + + w, r = os.popen4(["echo", self.teststr]) + got = r.read() + self.assertEqual(got, self.teststr + "\n") + + w, r = os.popen4(self.cmd) + self.validate_output(self.teststr, self.expected, r, w) + + +def test_main(): + run_unittest(Popen2Test) + +if __name__ == "__main__": + test_main() -- Repository URL: http://hg.python.org/jython From jython-checkins at python.org Wed Feb 19 23:48:10 2014 From: jython-checkins at python.org (jeff.allen) Date: Wed, 19 Feb 2014 23:48:10 +0100 (CET) Subject: [Jython-checkins] =?utf-8?q?jython=3A_Use_the_lib-python_version_?= =?utf-8?q?of_test=5Funivnewlines=2E?= Message-ID: <3fTvGf0Sgxz7Ljt@mail.python.org> http://hg.python.org/jython/rev/cd23b01ffa52 changeset: 7187:cd23b01ffa52 user: Jeff Allen date: Mon Feb 17 20:04:04 2014 +0000 summary: Use the lib-python version of test_univnewlines. This fixes non-deletion of the temporary file, through use of 'with'. All test pass ok. files: Lib/test/test_univnewlines.py | 132 ---------------------- 1 files changed, 0 insertions(+), 132 deletions(-) diff --git a/Lib/test/test_univnewlines.py b/Lib/test/test_univnewlines.py deleted file mode 100644 --- a/Lib/test/test_univnewlines.py +++ /dev/null @@ -1,132 +0,0 @@ -# Tests universal newline support for both reading and parsing files. -from __future__ import with_statement -import unittest -import os -import sys -from test import test_support - -if not hasattr(sys.stdin, 'newlines'): - raise test_support.TestSkipped, \ - "This Python does not have universal newline support" - -FATX = 'x' * (2**14) - -DATA_TEMPLATE = [ - "line1=1", - "line2='this is a very long line designed to go past the magic " + - "hundred character limit that is inside fileobject.c and which " + - "is meant to speed up the common case, but we also want to test " + - "the uncommon case, naturally.'", - "def line3():pass", - "line4 = '%s'" % FATX, - ] - -DATA_LF = "\n".join(DATA_TEMPLATE) + "\n" -DATA_CR = "\r".join(DATA_TEMPLATE) + "\r" -DATA_CRLF = "\r\n".join(DATA_TEMPLATE) + "\r\n" - -# Note that DATA_MIXED also tests the ability to recognize a lone \r -# before end-of-file. -DATA_MIXED = "\n".join(DATA_TEMPLATE) + "\r" -DATA_SPLIT = [x + "\n" for x in DATA_TEMPLATE] -del x - -class TestGenericUnivNewlines(unittest.TestCase): - # use a class variable DATA to define the data to write to the file - # and a class variable NEWLINE to set the expected newlines value - READMODE = 'U' - WRITEMODE = 'wb' - - def setUp(self): - fp = open(test_support.TESTFN, self.WRITEMODE) - fp.write(self.DATA) - fp.close() - - def tearDown(self): - try: - os.unlink(test_support.TESTFN) - except: - pass - - def test_read(self): - fp = open(test_support.TESTFN, self.READMODE) - data = fp.read() - self.assertEqual(data, DATA_LF) - self.assertEqual(repr(fp.newlines), repr(self.NEWLINE)) - - def test_readlines(self): - fp = open(test_support.TESTFN, self.READMODE) - data = fp.readlines() - self.assertEqual(data, DATA_SPLIT) - self.assertEqual(repr(fp.newlines), repr(self.NEWLINE)) - - def test_readline(self): - fp = open(test_support.TESTFN, self.READMODE) - data = [] - d = fp.readline() - while d: - data.append(d) - d = fp.readline() - self.assertEqual(data, DATA_SPLIT) - self.assertEqual(repr(fp.newlines), repr(self.NEWLINE)) - - def test_seek(self): - fp = open(test_support.TESTFN, self.READMODE) - fp.readline() - pos = fp.tell() - data = fp.readlines() - self.assertEqual(data, DATA_SPLIT[1:]) - fp.seek(pos) - data = fp.readlines() - self.assertEqual(data, DATA_SPLIT[1:]) - - def test_execfile(self): - namespace = {} - with test_support.check_py3k_warnings(): - execfile(test_support.TESTFN, namespace) - func = namespace['line3'] - self.assertEqual(func.func_code.co_firstlineno, 3) - self.assertEqual(namespace['line4'], FATX) - - -class TestNativeNewlines(TestGenericUnivNewlines): - NEWLINE = None - DATA = DATA_LF - READMODE = 'r' - WRITEMODE = 'w' - -class TestCRNewlines(TestGenericUnivNewlines): - NEWLINE = '\r' - DATA = DATA_CR - -class TestLFNewlines(TestGenericUnivNewlines): - NEWLINE = '\n' - DATA = DATA_LF - -class TestCRLFNewlines(TestGenericUnivNewlines): - NEWLINE = '\r\n' - DATA = DATA_CRLF - - def test_tell(self): - fp = open(test_support.TESTFN, self.READMODE) - self.assertEqual(repr(fp.newlines), repr(None)) - data = fp.readline() - pos = fp.tell() - self.assertEqual(repr(fp.newlines), repr(self.NEWLINE)) - -class TestMixedNewlines(TestGenericUnivNewlines): - NEWLINE = ('\r', '\n') - DATA = DATA_MIXED - - -def test_main(): - test_support.run_unittest( - TestNativeNewlines, - TestCRNewlines, - TestLFNewlines, - TestCRLFNewlines, - TestMixedNewlines - ) - -if __name__ == '__main__': - test_main() -- Repository URL: http://hg.python.org/jython From jython-checkins at python.org Sun Feb 23 23:13:51 2014 From: jython-checkins at python.org (jeff.allen) Date: Sun, 23 Feb 2014 23:13:51 +0100 (CET) Subject: [Jython-checkins] =?utf-8?q?jython=3A_Work-around_in_test=5Fzipim?= =?utf-8?q?port=5Fsupport_for_attempt_to_delete_still-open_file?= Message-ID: <3fXLKC6T4hz7Lk5@mail.python.org> http://hg.python.org/jython/rev/93701c9c5f45 changeset: 7188:93701c9c5f45 user: Jeff Allen date: Sun Feb 23 15:42:00 2014 +0000 summary: Work-around in test_zipimport_support for attempt to delete still-open file (Windows). files: Lib/test/test_zipimport_support.py | 19 ++++++++++++++++++ 1 files changed, 19 insertions(+), 0 deletions(-) diff --git a/Lib/test/test_zipimport_support.py b/Lib/test/test_zipimport_support.py --- a/Lib/test/test_zipimport_support.py +++ b/Lib/test/test_zipimport_support.py @@ -34,6 +34,24 @@ from test import test_doctest, sample_doctest from test.test_importhooks import ImportHooksBaseTestCase +if is_jython and os._name=="nt": + # Jython holds open zip/jar files placed on its sys.path. (Assume there + # is a good reason for this.) Windows will not then allow the script + # directory to be cleaned up on context exit, resulting in test failures + # unrelated to the purpose of the test. + + # Replace test.script_helper.temp_dir with this copy. + import contextlib, tempfile, shutil, warnings + @contextlib.contextmanager + def temp_dir(): + dirname = os.path.realpath(tempfile.mkdtemp()) + try: + yield dirname + finally: + try: + shutil.rmtree(dirname) + except OSError: + warnings.warn("Failed to remove "+dirname) def _run_object_doctest(obj, module): # Direct doctest output (normally just errors) to real stdout; doctest @@ -88,6 +106,7 @@ self.assertEqual(inspect.getsource(zip_pkg.foo), test_src) @unittest.skipIf(is_jython, "FIXME: not working on Jython") + # Failure possibly due to sys.path not passing to sub-process in test_doctest. def test_doctest_issue4197(self): # To avoid having to keep two copies of the doctest module's # unit tests in sync, this test works by taking the source of -- Repository URL: http://hg.python.org/jython