Ticket #3942 (closed defect: fixed)

Opened 3 months ago

Last modified 3 months ago

[with patch, positive review] Sage interfaces vs. pyprocessing

Reported by: rlm Assigned to: was
Priority: major Milestone: sage-3.1.2
Component: interfaces Keywords:
Cc: cwitty

Description

When using pyprocessing, if you use an interface to some external function, such as GAP, sometimes the separate subprocesses will collide, since they all share the same temporary file.

A temporary fix (due to cwitty) was to add

sage.interfaces.expect.tmp_expect_interface_local = '/tmp/interface' + str(os.getpid())

to the beginning of the function succeeding the @parallel decorator. This way, the temp files are separate. This should probably be part of the parallel function itself.

Attachments

trac_3942_pid-interfaces.patch (1.3 kB) - added by rlm on 08/24/2008 07:40:24 AM.

Change History

08/24/2008 04:40:41 AM changed by malb

How about replacing tmp_expect_interface_local by a function that returns,

'/tmp/interface' + str(os.getpid())

this way, it should always just work?

08/24/2008 07:40:03 AM changed by rlm 

rank4:sage-main rlmill$ ../../sage -tp 2 sage/interfaces/
Global iterations: 1
File iterations: 1
TeX files: 0
 
----------------------------------------------------------------------
sage -t  devel/sage-main/sage/interfaces/tachyon.py
	 [2.0 s]
sage -t  devel/sage-main/sage/interfaces/tests.py
	 [3.5 s]
sage -t  devel/sage-main/sage/interfaces/singular.py
	 [3.5 s]
sage -t  devel/sage-main/sage/interfaces/sage0.py
	 [9.4 s]
sage -t  devel/sage-main/sage/interfaces/r.py
	 [7.0 s]
sage -t  devel/sage-main/sage/interfaces/rubik.py
	 [15.8 s]
sage -t  devel/sage-main/sage/interfaces/qsieve.py
	 [3.4 s]
sage -t  devel/sage-main/sage/interfaces/povray.py
	 [2.2 s]
sage -t  devel/sage-main/sage/interfaces/phc.py
	 [3.2 s]
sage -t  devel/sage-main/sage/interfaces/octave.py
	 [2.9 s]
sage -t  devel/sage-main/sage/interfaces/psage.py
	 [11.8 s]
sage -t  devel/sage-main/sage/interfaces/mwrank.py
	 [2.0 s]
sage -t  devel/sage-main/sage/interfaces/mupad.py
	 [2.1 s]
sage -t  devel/sage-main/sage/interfaces/monitor.py
	 [2.1 s]
sage -t  devel/sage-main/sage/interfaces/matlab.py
	 [2.0 s]
sage -t  devel/sage-main/sage/interfaces/mathematica.py
	 [2.0 s]
sage -t  devel/sage-main/sage/interfaces/maple.py
	 [2.0 s]
sage -t  devel/sage-main/sage/interfaces/magma_sim.py
	 [1.9 s]
sage -t  devel/sage-main/sage/interfaces/magma_free.py
	 [2.0 s]
sage -t  devel/sage-main/sage/interfaces/magma.py
	 [1.9 s]
sage -t  devel/sage-main/sage/interfaces/macaulay2.py
	 [1.9 s]
sage -t  devel/sage-main/sage/interfaces/lie.py
	 [2.1 s]
sage -t  devel/sage-main/sage/interfaces/maxima.py
	 [17.0 s]
sage -t  devel/sage-main/sage/interfaces/kash.py
	 [2.0 s]
sage -t  devel/sage-main/sage/interfaces/gnuplot.py
	 [2.1 s]
sage -t  devel/sage-main/sage/interfaces/gp.py
	 [3.4 s]
sage -t  devel/sage-main/sage/interfaces/gfan.py
	 [2.0 s]
sage -t  devel/sage-main/sage/interfaces/get_sigs.py
	 [2.0 s]
sage -t  devel/sage-main/sage/interfaces/genus2reduction.py
	 [2.7 s]
sage -t  devel/sage-main/sage/interfaces/gap.py
	 [3.4 s]
sage -t  devel/sage-main/sage/interfaces/frobby.py
	 [2.0 s]
sage -t  devel/sage-main/sage/interfaces/expect.py
	 [6.9 s]
sage -t  devel/sage-main/sage/interfaces/axiom.py
	 [1.9 s]
sage -t  devel/sage-main/sage/interfaces/ecm.py
	 [14.5 s]
All tests passed!
Total time for all tests: 76.7 seconds
rank4:sage-main rlmill$

08/24/2008 07:40:24 AM changed by rlm

  • attachment trac_3942_pid-interfaces.patch added.

08/24/2008 07:41:51 AM changed by rlm

  • summary changed from Sage interfaces vs. pyprocessing to [with patch, needs review] Sage interfaces vs. pyprocessing.

My one concern with this patch is that the temp files should be cleaned up properly. I don't know the mechanisms for this...

08/24/2008 10:15:59 AM changed by was

  • summary changed from [with patch, needs review] Sage interfaces vs. pyprocessing to [with patch, positive review] Sage interfaces vs. pyprocessing.

My one concern with this patch is that the temp files should be cleaned up properly. I don't know the mechanisms for this...

That's all taken care of because the files are in SAGE_TMP. No worries.

I think this is a good patch.

08/24/2008 08:52:14 PM changed by mabshoff

  • status changed from new to closed.
  • resolution set to fixed.

Merged in Sage 3.1.2.alpha1