Opened 7 years ago
Closed 7 years ago
#4312 closed defect (fixed)
[with patch; with positive review] major @paralllel (hence pyprocessing) new bug in 3.1.3, still in 3.1.4
Reported by: | was | Owned by: | cwitty |
---|---|---|---|
Priority: | blocker | Milestone: | sage-3.2 |
Component: | misc | Keywords: | |
Cc: | Merged in: | ||
Authors: | Reviewers: | ||
Report Upstream: | Work issues: | ||
Branch: | Commit: | ||
Dependencies: | Stopgaps: |
Description
In 3.1.2 this worked fine. It's totally broken in 3.1.4.
sage@modular:~/build/sage-3.1.4$ more a.sage @parallel(8) def f(p): print p t = cputime() M = ModularSymbols(p^2,sign=1) w = M.atkin_lehner_operator(p) K = (w-1).kernel() N = K.new_subspace() D = N.decomposition() print cputime(t) M.save(str(p)) save(D, '%s-decomp'%p) sage@modular:~/build/sage-3.1.4$ ./sage ---------------------------------------------------------------------- | SAGE Version 3.1.4, Release Date: 2008-10-16 | | Type notebook() for the GUI, and license() for information. | ---------------------------------------------------------------------- sage: load a.sage sage: list(f([11,17])) 17 11 --------------------------------------------------------------------------- NameError Traceback (most recent call last) /home2/sage/.sage/temp/modular/25347/_home2_sage_build_sage_3_1_4_a_sage_0.py in <module>() ----> 1 2 3 4 5 /home2/sage/build/sage-3.1.4/local/lib/python2.5/site-packages/sage/parallel/multiprocessing.pyc in parallel_iter(processes, f, inputs) 64 65 result = p.imapUnordered(call_pickled_function, [ (fp, t) for t in inputs ]) ---> 66 for res in result: 67 yield res 68 /home2/sage/build/sage-3.1.4/local/lib/python2.5/site-packages/processing/pool.pyc in next(self, timeout) 468 if success: 469 return value --> 470 raise value 471 472 def _set(self, i, obj): NameError: global name '_sage_const_2' is not defined sage:
This is a pyprocessing problem since:
sage: load a.sage sage: f(11) 11 0.168011 sage: f(13) 13 0.244015 sage:
}}}
Attachments (2)
Change History (9)
comment:1 Changed 7 years ago by jason
comment:2 Changed 7 years ago by was
- Summary changed from major @interact (hence pyprocessing) new bug in 3.1.3, still in 3.1.4 to major @paralllel (hence pyprocessing) new bug in 3.1.3, still in 3.1.4
comment:3 Changed 7 years ago by mabshoff
- Milestone changed from sage-3.2.1 to sage-3.2
Bug Day 15 material? This certainly should get fixed in 3.2.
Cheers,
Michael
comment:4 Changed 7 years ago by mabshoff
- Milestone changed from sage-3.2.1 to sage-3.2
comment:5 Changed 7 years ago by was
- Summary changed from major @paralllel (hence pyprocessing) new bug in 3.1.3, still in 3.1.4 to [with patch; needs review] major @paralllel (hence pyprocessing) new bug in 3.1.3, still in 3.1.4
I've posted a hack-ish patch that at least fixes this problem. Something better might be to modify @parallel to more properly transfer all _sage_const*'s to subprocesses or something.
Changed 7 years ago by was
Changed 7 years ago by craigcitro
comment:6 Changed 7 years ago by craigcitro
- Summary changed from [with patch; needs review] major @paralllel (hence pyprocessing) new bug in 3.1.3, still in 3.1.4 to [with patch; with positive review] major @paralllel (hence pyprocessing) new bug in 3.1.3, still in 3.1.4
Looks good.
I made a new ticket to have someone find an actual fix, which is #4545. I also edited the patch to have this info in it, and posted that above.
comment:7 Changed 7 years ago by mabshoff
- Resolution set to fixed
- Status changed from new to closed
Merged sage-4312-edited.patch in Sage 3.2.rc2
Note: See
TracTickets for help on using
tickets.
So does this issue have anything to do with @interact? Or is it @parallel that has the problem?