Issue 2550503
 
 
 
              
              
Created on 2009-02-15 20:49 by tobias-herp, last changed 2009-03-12 02:53 by richard. 
 |
 
  | File name | Uploaded | Description | Edit | Remove |  
  | anypy-test.asc | tobias-herp,
   2009-02-22 22:45 | Test output (run_tests.py anypy) |  |  |  
  | anypy-2009-03-10.patch | tobias-herp,
   2009-03-10 14:21 | updated patch (2009-03-10) |  |  |  
 |
 
   | msg3544 | Author: [hidden] (tobias-herp) | Date: 2009-02-15 20:49 |  |  
   | Roundup supports Python versions 2.3 ff; this yields some compatibility
issues:
- the md5 and sha modules are deprecated as of Python 2.6,
  but the hashlib modules is not available before Python 2.5
- as of Python 2.6, the sets module is deprecated;
  the built-in set type should be used instead of sets.Set
This patch adds a compatibility layer to handle the Python version
issues; the modules in the anypy package provide only the functionality
which is in fact used by roundup. |  
   | msg3545 | Author: [hidden] (tobias-herp) | Date: 2009-02-16 17:45 |  |  
   | test/test_anypy_hashlib.py now integrates nicely in run_tests.py; no
other changes (revert the other version first) |  
   | msg3574 | Author: [hidden] (tobias-herp) | Date: 2009-02-22 23:43 |  |  
   | Maybe someone could apply this patch (updated to match rev 4159, minimal
changes); currently I can't do it myself...
BTW,
 from roundup.anypy.sets_ import set
 ...
 set(...)
should perform better than
 import sets
 sets.Set(...)
for at least two reasons:
- the builtin set type performs better than sets.Set
- dots are evaluated at run time
<http://wiki.python.org/moin/PythonSpeed/PerformanceTips#Avoidingdots...> |  
   | msg3622 | Author: [hidden] (tobias-herp) | Date: 2009-03-10 14:21 |  |  
   | Updated patch against revision 4178; obsoletes all older patches.
Can *please* someone apply the patch (created with svn diff in the
source root directory) and commit it?  My sourceforge account is not
fixed yet, I can't do it myself.  IMO, we *should* get going with the
compatibility-with-current-Python issue.
The round.anypy package is added in setup.py.  The hashlib compatibility
is tested thoroughly, and if anyone considers it necessary, I can
provide a test for anypy.sets_, too.  Since "set" is a builtin type, I
renamed the "set" boolean variable in hyperdb.py to "do_set". |  
   | msg3634 | Author: [hidden] (richard) | Date: 2009-03-12 02:53 |  |  
   | Applied in r4181, thanks.
One major outstanding issue is the failure of
test.test_sqlite.sqliteDBTest.testStringUnicode |  |
 
| Date | User | Action | Args |  | 2009-03-12 02:53:24 | richard | set | status: new -> closed assignee: richard
 resolution: accepted
 messages:
  + msg3634
 |  | 2009-03-10 14:21:58 | tobias-herp | set | files:
  - anypy-2009-02-22b.patch |  | 2009-03-10 14:21:45 | tobias-herp | set | files:
  - anypy.patch |  | 2009-03-10 14:21:35 | tobias-herp | set | files:
  + anypy-2009-03-10.patch messages:
  + msg3622
 |  | 2009-02-23 02:51:11 | tobias-herp | set | files:
  - test_anypy_hashlib.patch |  | 2009-02-22 23:43:09 | tobias-herp | set | files:
  + anypy-2009-02-22b.patch messages:
  + msg3574
 |  | 2009-02-22 22:45:38 | tobias-herp | set | files:
  + anypy-test.asc nosy:
  + richard, stefan
 |  | 2009-02-16 17:45:57 | tobias-herp | set | files:
  + test_anypy_hashlib.patch messages:
  + msg3545
 |  | 2009-02-15 20:49:48 | tobias-herp | create |  | 
 |