Mailman 3 October 2009 - Python-checkins

r75993 - python/branches/release26-maint
by mark.dickinson 31 Oct '09

31 Oct '09

Author: mark.dickinson Date: Sat Oct 31 13:48:51 2009 New Revision: 75993 Log: Blocked revisions 75991 via svnmerge ........ r75991 | mark.dickinson | 2009-10-31 12:47:47 +0000 (Sat, 31 Oct 2009) | 1 line Set retval on PyOS_string_to_double failure. ........ Modified: python/branches/release26-maint/ (props changed)

1 0

r75992 - python/branches/py3k
by mark.dickinson 31 Oct '09

31 Oct '09

Author: mark.dickinson Date: Sat Oct 31 13:48:28 2009 New Revision: 75992 Log: Blocked revisions 75991 via svnmerge ........ r75991 | mark.dickinson | 2009-10-31 12:47:47 +0000 (Sat, 31 Oct 2009) | 1 line Set retval on PyOS_string_to_double failure. ........ Modified: python/branches/py3k/ (props changed)

1 0

r75991 - python/trunk/Python/marshal.c
by mark.dickinson 31 Oct '09

31 Oct '09

Author: mark.dickinson Date: Sat Oct 31 13:47:47 2009 New Revision: 75991 Log: Set retval on PyOS_string_to_double failure. Modified: python/trunk/Python/marshal.c Modified: python/trunk/Python/marshal.c ============================================================================== --- python/trunk/Python/marshal.c (original) +++ python/trunk/Python/marshal.c Sat Oct 31 13:47:47 2009 @@ -699,8 +699,10 @@ } buf[n] = '\0'; dx = PyOS_string_to_double(buf, NULL, NULL); - if (dx == -1.0 && PyErr_Occurred()) + if (dx == -1.0 && PyErr_Occurred()) { + retval = NULL; break; + } retval = PyFloat_FromDouble(dx); break; } @@ -738,8 +740,10 @@ } buf[n] = '\0'; c.real = PyOS_string_to_double(buf, NULL, NULL); - if (c.real == -1.0 && PyErr_Occurred()) + if (c.real == -1.0 && PyErr_Occurred()) { + retval = NULL; break; + } n = r_byte(p); if (n == EOF || r_string(buf, (int)n, p) != n) { PyErr_SetString(PyExc_EOFError, @@ -749,8 +753,10 @@ } buf[n] = '\0'; c.imag = PyOS_string_to_double(buf, NULL, NULL); - if (c.imag == -1.0 && PyErr_Occurred()) + if (c.imag == -1.0 && PyErr_Occurred()) { + retval = NULL; break; + } retval = PyComplex_FromCComplex(c); break; }

1 0

r75990 - tracker/roundup-src/roundup/cgi/actions.py
by martin.v.loewis 31 Oct '09

31 Oct '09

Author: martin.v.loewis Date: Sat Oct 31 13:18:40 2009 New Revision: 75990 Log: reject invalid numbers, so that postgres doesn't have to Modified: tracker/roundup-src/roundup/cgi/actions.py Modified: tracker/roundup-src/roundup/cgi/actions.py ============================================================================== --- tracker/roundup-src/roundup/cgi/actions.py (original) +++ tracker/roundup-src/roundup/cgi/actions.py Sat Oct 31 13:18:40 2009 @@ -245,6 +245,11 @@ # replace the single value with the split list for v in l: self.form.value.append(cgi.MiniFieldStorage(key, v)) + elif isinstance(prop, hyperdb.Number): + try: + float(self.form[key].value) + except ValueError: + raise exceptions.FormError, "Invalid number: "+self.form[key].value self.form.value.append(cgi.MiniFieldStorage('@filter', key))

1 0

Python Regression Test Failures refleak (1)
by Neal Norwitz 31 Oct '09

31 Oct '09

More important issues: ---------------------- test_warnings leaked [0, 0, 120] references, sum=120 Less important issues: ---------------------- test_popen2 leaked [-50, 0, -29] references, sum=-79 test_threading leaked [48, 48, 48] references, sum=144

1 0

r75989 - in python/branches/release31-maint: Lib/test/test_signal.py
by mark.dickinson 31 Oct '09

31 Oct '09

Author: mark.dickinson Date: Sat Oct 31 11:39:21 2009 New Revision: 75989 Log: Merged revisions 75988 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/py3k ................ r75988 | mark.dickinson | 2009-10-31 10:38:43 +0000 (Sat, 31 Oct 2009) | 9 lines Merged revisions 75986 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r75986 | mark.dickinson | 2009-10-31 10:36:06 +0000 (Sat, 31 Oct 2009) | 2 lines Issue #7042: Use a better mechanism for testing timers in test_signal. ........ ................ Modified: python/branches/release31-maint/ (props changed) python/branches/release31-maint/Lib/test/test_signal.py Modified: python/branches/release31-maint/Lib/test/test_signal.py ============================================================================== --- python/branches/release31-maint/Lib/test/test_signal.py (original) +++ python/branches/release31-maint/Lib/test/test_signal.py Sat Oct 31 11:39:21 2009 @@ -360,11 +360,14 @@ signal.signal(signal.SIGVTALRM, self.sig_vtalrm) signal.setitimer(self.itimer, 0.3, 0.2) - for i in range(100000000): + start_time = time.time() + while time.time() - start_time < 5.0: # use up some virtual time by doing real work _ = pow(12345, 67890, 10000019) if signal.getitimer(self.itimer) == (0.0, 0.0): break # sig_vtalrm handler stopped this itimer + else: + self.fail('timeout waiting for sig_vtalrm signal') # virtual itimer should be (0.0, 0.0) now self.assertEquals(signal.getitimer(self.itimer), (0.0, 0.0)) @@ -376,9 +379,14 @@ signal.signal(signal.SIGPROF, self.sig_prof) signal.setitimer(self.itimer, 0.2, 0.2) - for i in range(100000000): + start_time = time.time() + while time.time() - start_time < 5.0: + # do some work + _ = pow(12345, 67890, 10000019) if signal.getitimer(self.itimer) == (0.0, 0.0): break # sig_prof handler stopped this itimer + else: + self.fail('timeout waiting for sig_prof signal') # profiling itimer should be (0.0, 0.0) now self.assertEquals(signal.getitimer(self.itimer), (0.0, 0.0))

1 0

r75988 - in python/branches/py3k: Lib/test/test_signal.py
by mark.dickinson 31 Oct '09

31 Oct '09

Author: mark.dickinson Date: Sat Oct 31 11:38:43 2009 New Revision: 75988 Log: Merged revisions 75986 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r75986 | mark.dickinson | 2009-10-31 10:36:06 +0000 (Sat, 31 Oct 2009) | 2 lines Issue #7042: Use a better mechanism for testing timers in test_signal. ........ Modified: python/branches/py3k/ (props changed) python/branches/py3k/Lib/test/test_signal.py Modified: python/branches/py3k/Lib/test/test_signal.py ============================================================================== --- python/branches/py3k/Lib/test/test_signal.py (original) +++ python/branches/py3k/Lib/test/test_signal.py Sat Oct 31 11:38:43 2009 @@ -360,11 +360,14 @@ signal.signal(signal.SIGVTALRM, self.sig_vtalrm) signal.setitimer(self.itimer, 0.3, 0.2) - for i in range(100000000): + start_time = time.time() + while time.time() - start_time < 5.0: # use up some virtual time by doing real work _ = pow(12345, 67890, 10000019) if signal.getitimer(self.itimer) == (0.0, 0.0): break # sig_vtalrm handler stopped this itimer + else: + self.fail('timeout waiting for sig_vtalrm signal') # virtual itimer should be (0.0, 0.0) now self.assertEquals(signal.getitimer(self.itimer), (0.0, 0.0)) @@ -376,9 +379,14 @@ signal.signal(signal.SIGPROF, self.sig_prof) signal.setitimer(self.itimer, 0.2, 0.2) - for i in range(100000000): + start_time = time.time() + while time.time() - start_time < 5.0: + # do some work + _ = pow(12345, 67890, 10000019) if signal.getitimer(self.itimer) == (0.0, 0.0): break # sig_prof handler stopped this itimer + else: + self.fail('timeout waiting for sig_prof signal') # profiling itimer should be (0.0, 0.0) now self.assertEquals(signal.getitimer(self.itimer), (0.0, 0.0))

1 0

r75987 - in python/branches/release26-maint: Lib/test/test_signal.py
by mark.dickinson 31 Oct '09

31 Oct '09

Author: mark.dickinson Date: Sat Oct 31 11:37:15 2009 New Revision: 75987 Log: Merged revisions 75986 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r75986 | mark.dickinson | 2009-10-31 10:36:06 +0000 (Sat, 31 Oct 2009) | 2 lines Issue #7042: Use a better mechanism for testing timers in test_signal. ........ Modified: python/branches/release26-maint/ (props changed) python/branches/release26-maint/Lib/test/test_signal.py Modified: python/branches/release26-maint/Lib/test/test_signal.py ============================================================================== --- python/branches/release26-maint/Lib/test/test_signal.py (original) +++ python/branches/release26-maint/Lib/test/test_signal.py Sat Oct 31 11:37:15 2009 @@ -360,11 +360,14 @@ signal.signal(signal.SIGVTALRM, self.sig_vtalrm) signal.setitimer(self.itimer, 0.3, 0.2) - for i in xrange(100000000): + start_time = time.time() + while time.time() - start_time < 5.0: # use up some virtual time by doing real work _ = pow(12345, 67890, 10000019) if signal.getitimer(self.itimer) == (0.0, 0.0): break # sig_vtalrm handler stopped this itimer + else: + self.fail('timeout waiting for sig_vtalrm signal') # virtual itimer should be (0.0, 0.0) now self.assertEquals(signal.getitimer(self.itimer), (0.0, 0.0)) @@ -376,9 +379,14 @@ signal.signal(signal.SIGPROF, self.sig_prof) signal.setitimer(self.itimer, 0.2, 0.2) - for i in xrange(100000000): + start_time = time.time() + while time.time() - start_time < 5.0: + # do some work + _ = pow(12345, 67890, 10000019) if signal.getitimer(self.itimer) == (0.0, 0.0): break # sig_prof handler stopped this itimer + else: + self.fail('timeout waiting for sig_prof signal') # profiling itimer should be (0.0, 0.0) now self.assertEquals(signal.getitimer(self.itimer), (0.0, 0.0))

1 0

r75986 - python/trunk/Lib/test/test_signal.py
by mark.dickinson 31 Oct '09

31 Oct '09

Author: mark.dickinson Date: Sat Oct 31 11:36:06 2009 New Revision: 75986 Log: Issue #7042: Use a better mechanism for testing timers in test_signal. Modified: python/trunk/Lib/test/test_signal.py Modified: python/trunk/Lib/test/test_signal.py ============================================================================== --- python/trunk/Lib/test/test_signal.py (original) +++ python/trunk/Lib/test/test_signal.py Sat Oct 31 11:36:06 2009 @@ -360,11 +360,14 @@ signal.signal(signal.SIGVTALRM, self.sig_vtalrm) signal.setitimer(self.itimer, 0.3, 0.2) - for i in xrange(100000000): + start_time = time.time() + while time.time() - start_time < 5.0: # use up some virtual time by doing real work _ = pow(12345, 67890, 10000019) if signal.getitimer(self.itimer) == (0.0, 0.0): break # sig_vtalrm handler stopped this itimer + else: + self.fail('timeout waiting for sig_vtalrm signal') # virtual itimer should be (0.0, 0.0) now self.assertEquals(signal.getitimer(self.itimer), (0.0, 0.0)) @@ -376,9 +379,14 @@ signal.signal(signal.SIGPROF, self.sig_prof) signal.setitimer(self.itimer, 0.2, 0.2) - for i in xrange(100000000): + start_time = time.time() + while time.time() - start_time < 5.0: + # do some work + _ = pow(12345, 67890, 10000019) if signal.getitimer(self.itimer) == (0.0, 0.0): break # sig_prof handler stopped this itimer + else: + self.fail('timeout waiting for sig_prof signal') # profiling itimer should be (0.0, 0.0) now self.assertEquals(signal.getitimer(self.itimer), (0.0, 0.0))

1 0

r75985 - in python/branches/release31-maint: Misc/NEWS Python/ceval.c
by mark.dickinson 31 Oct '09

31 Oct '09

Author: mark.dickinson Date: Sat Oct 31 11:20:38 2009 New Revision: 75985 Log: Merged revisions 75984 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/py3k ................ r75984 | mark.dickinson | 2009-10-31 10:18:44 +0000 (Sat, 31 Oct 2009) | 12 lines Merged revisions 75982 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r75982 | mark.dickinson | 2009-10-31 10:11:28 +0000 (Sat, 31 Oct 2009) | 5 lines Issue #6603: Fix --with-tsc build failures on x86-64 that resulted from a gcc inline assembler peculiarity. (gcc's "A" constraint apparently means 'rax or rdx' in 64-bit mode, not edx:eax or rdx:rax as one might expect.) ........ ................ Modified: python/branches/release31-maint/ (props changed) python/branches/release31-maint/Misc/NEWS python/branches/release31-maint/Python/ceval.c Modified: python/branches/release31-maint/Misc/NEWS ============================================================================== --- python/branches/release31-maint/Misc/NEWS (original) +++ python/branches/release31-maint/Misc/NEWS Sat Oct 31 11:20:38 2009 @@ -130,6 +130,10 @@ Build ----- +- Issue #6603: Change READ_TIMESTAMP macro in ceval.c so that it + compiles correctly under gcc on x86-64. This fixes a reported + problem with the --with-tsc build on x86-64. + - Issue #6802: Fix build issues on MacOSX 10.6 - Issue #6801 : symmetric_difference_update also accepts |. Modified: python/branches/release31-maint/Python/ceval.c ============================================================================== --- python/branches/release31-maint/Python/ceval.c (original) +++ python/branches/release31-maint/Python/ceval.c Sat Oct 31 11:20:38 2009 @@ -51,11 +51,29 @@ ((long*)(v))[1] = tb; } -#else /* this is for linux/x86 (and probably any other GCC/x86 combo) */ +#elif defined(__i386__) + +/* this is for linux/x86 (and probably any other GCC/x86 combo) */ #define READ_TIMESTAMP(val) \ __asm__ __volatile__("rdtsc" : "=A" (val)) +#elif defined(__x86_64__) + +/* for gcc/x86_64, the "A" constraint in DI mode means *either* rax *or* rdx; + not edx:eax as it does for i386. Since rdtsc puts its result in edx:eax + even in 64-bit mode, we need to use "a" and "d" for the lower and upper + 32-bit pieces of the result. */ + +#define READ_TIMESTAMP(val) \ + __asm__ __volatile__("rdtsc" : \ + "=a" (((int*)&(val))[0]), "=d" (((int*)&(val))[1])); + + +#else + +#error "Don't know how to implement timestamp counter for this architecture" + #endif void dump_tsc(int opcode, int ticked, uint64 inst0, uint64 inst1,

1 0