msg23661 - (view) |
Author: Axel Kaiser (axel_kaiser) |
Date: 2004-12-12 09:25 |
The new Python 2.4 crashes when starting some of my programs. This happens under win xp as well as under win nt. The crashes are reproduceble, but I didn't succeed in finding the exact place where it happens. In the attachment you find a Dr. Watson logfile. I hope it helps. Otherwise I think I will be able to help with more information. Thanks in advance Axel Kaiser Germany |
|
|
msg23662 - (view) |
Author: Raymond Hettinger (rhettinger) *  |
Date: 2004-12-12 09:56 |
Logged In: YES user_id=80475 The attachment didn't make it. Be sure to check the upload box or it won't make it. |
|
|
msg23663 - (view) |
Author: salyee (salyee) |
Date: 2004-12-15 06:17 |
Logged In: YES user_id=1178573 I have the same or similar problem. While python24.dll(Python 2.4 for Windows) uses msvcr71.dll for CRT, host C++ program embeding Python 2.4 as a script engine, in my project, is compiled by VC++ 6.0 (msvcrt.dll). So, the following code snip makes a crash. FILE* fp = fopen("somecode.py", "rt"); // uses msvcrt.dll PyRun_SimpleFile(fp, "somecode.py); // uses msvcr71.dll inside. Cause of a crash. In my case, I must use VC++ 6.0 and I want clients of my program to install Python 2.4 from www.python.org. How do I solve this problem ? |
|
|
msg23664 - (view) |
Author: Wolfgang Langner (tds333) * |
Date: 2004-12-15 15:16 |
Logged In: YES user_id=600792 The FILE pointer is specific to your c-lib. Different versions of the ms c-lib have incompatible implementations of this structure. To work arround such problems use python to get the right FILE* Example (not tested): /* use pythons std c library to produce a FILE object to avoid mess with different c libs in debug or release mode and different versions used to compile python */ PyObject* pyFile = PyFile_FromString(fileName, "r"); FILE* const file = PyFile_AsFile(pyFile); assert(0 != file); const int succ = PyRun_SimpleFile(file, fileName); // now remove py file object Py_XDECREF(pyFile); |
|
|
msg23665 - (view) |
Author: Raymond Hettinger (rhettinger) *  |
Date: 2004-12-15 18:02 |
Logged In: YES user_id=80475 Any objections to marking this as Invalid and closing the bug? |
|
|
msg23666 - (view) |
Author: Axel Kaiser (axel_kaiser) |
Date: 2004-12-15 20:58 |
Logged In: YES user_id=1176563 Sorry, the attachment was not appended. By the way, the problem occurs without any external modules/libraries. It just juses python-source. The problem might be connected with the use of module array. |
|
|
msg23667 - (view) |
Author: Martin v. Löwis (loewis) *  |
Date: 2004-12-15 22:39 |
Logged In: YES user_id=21627 Can you create a small test case that demonstrates the problem? |
|
|
msg23668 - (view) |
Author: Axel Kaiser (axel_kaiser) |
Date: 2004-12-16 13:56 |
Logged In: YES user_id=1176563 Okay, I finally was able to produce a small example. When I start python from the commandline with file "Lebenserwartung.py" python crashes. It doesn't crash when I remove the coding-line (1st line) of file "Sterbetafeln.py". (The example can't run properly due to missing files, but these don't influence the crashes) Best regards, Axel |
|
|
msg23669 - (view) |
Author: Walter Dörwald (doerwalter) *  |
Date: 2004-12-16 14:34 |
Logged In: YES user_id=89016 This sounds like it might be related to bug http://www.python.org/sf/1076985. Can you retry your scripts with the patch from http://www.python.org/sf/1076985? What happens if you replace the "cp1252" in the encoding header with "iso-8859- 1"? A quick workaround might be to wrap the long source lines in Sterbetafeln.py. |
|
|
msg23670 - (view) |
Author: Axel Kaiser (axel_kaiser) |
Date: 2004-12-17 12:12 |
Logged In: YES user_id=1176563 That's it!!! I applied the patch to codecs.py and now it works without crash! Thanks a lot! Let's close this thread. |
|
|