<br><br>---------- Forwarded message ----------<br><span class="gmail_quote">From: <b class="gmail_sendername">Richard Tew</b> <<a href="mailto:richard.m.tew@gmail.com">richard.m.tew@gmail.com</a>><br>Date: Apr 15, 2007 1:55 PM
<br>Subject: Re: [Stackless] Stackless 2.5 branch updated for Python 2.5.1c1<br>To: Santiago Gala <<a href="mailto:sgala@apache.org">sgala@apache.org</a>><br>Cc: stackless <<a href="mailto:stackless@stackless.com">
stackless@stackless.com</a>><br><br></span>On 4/15/07, Santiago Gala <<a href="mailto:sgala@apache.org">sgala@apache.org</a>> wrote:<br>> linux amd64 (Core 2 Duo) results:<br>> $ ../../python runAll.py<br>> ..........E......F.E..............
<br>> ...................EF.............<br><br>My suspicion was that the amd64 hard switching support was not working<br>correctly for Hieu's setup. This reinforces that belief. Given that<br>it obviously has worked for someone in the past, based on past
<br>experience, I would assume the problem lies with gcc.<br><br>So for either Hieu or yourself: What version of gcc did you compile<br>with? Can you switch to an older gcc version and recompile and then<br>run the Stackless tests again? If so, do they pass or fail still?
<br><br>Oops, I forgot to paste:<br><br>$ ./python<br>Python 2.5.1c1 Stackless 3.1b3 060516 (release25-maint, Apr 15 2007, 13:23:04) <br>[GCC 4.1.2 (Gentoo 4.1.2)] on linux2<br>Type "help", "copyright", "credits" or "license" for more information.
<br>>>> <br><br>I have around just an old gcc-3.4.6, but I would not recommend using gcc-3 to anybody currently, the quality of the generated code pales in the comparison.<br><br>The python2.5 regressions are the same. Stackless unittests work like:
<br><br>$ ../../python runAll.py <br>.................F................<br>======================================================================<br>FAIL: testXrange (test_pickle.TestConcretePickledTasklets)<br>----------------------------------------------------------------------
<br>Traceback (most recent call last):<br> File "./test_pickle.py", line 197, in testXrange<br> self.run_pickled(xrangetest, 20, 13)<br> File "./test_pickle.py", line 164, in run_pickled<br> self.assertEquals
(new_rval, old_rval)<br>AssertionError: 9159019261958553619 != 19<br><br>----------------------------------------------------------------------<br>Ran 34 tests in 1.400s<br><br>FAILED (failures=1)<br>(and then the process stays forever at 100% CPU until I kill it:
<br><br>FAILED (failures=1)<br>Traceback (most recent call last):<br> File "runAll.py", line 55, in <module><br> unittest.TextTestRunner().run(testSuite)<br> File "/home/sgala/newcode/stackless/Lib/unittest.py", line 705, in run
<br> test(result)<br> File "/home/sgala/newcode/stackless/Lib/unittest.py", line 437, in __call__<br> return self.run(*args, **kwds)<br>)<br><br>Now:<br><br>$ ../../python <br>Python 2.5.1c1 Stackless 3.1b3
060516 (release25-maint, Apr 15 2007, 14:09:48) <br>[GCC 3.4.6 (Gentoo 3.4.6-r2, ssp-3.4.6-1.0, pie-8.7.10)] on linux2<br>Type "help", "copyright", "credits" or "license" for more information.
<br><br><br>One of the most useful pieces of information I could have when people<br>paste test output (and I should have asked for this) is the header<br>when you run the interpreter.<br><br>e.g. Python 2.5 Stackless 3.1b3
060516 (python-2.5:53557:53567, Jan 25<br>2007, 22:01:46) [MSC v.1310 32 bit (Intel)] on win32<br><br>Thanks,<br>Richard.<br>