[Python-bugs-list] [Bug #116546] test_longexp.py causes apparently endless page thrashing

noreply@sourceforge.net noreply@sourceforge.net
Wed, 11 Oct 2000 12:40:13 -0700


Bug #116546, was updated on 2000-Oct-10 14:59
Here is a current snapshot of the bug.

Project: Python
Category: None
Status: Closed
Resolution: Wont Fix
Bug Group: Platform-specific
Priority: 5
Summary: test_longexp.py causes apparently endless page thrashing

Details: 2.0 release candidate 1, built from source on a SPARCstation LX w/64 MB,
fails to run the test_longexp.py test.  The disk became very active (page
thrashing?), the window system became non-responsive to mouse button
clicks, and the cursor followed mouse movement only sluggishly.

The system also became non-responsive to network requests, attempts to
ping and to telnet both failing.

At present values of up to 10000 for REPS have been validated as OK.

I will leave an overnight test running with the original value to see if it
completes (having terminated the original test after 30 min).

Follow-Ups:

Date: 2000-Oct-11 06:18
By: holdenweb

Comment:
The overnight output tends to suggest the test is causing severe problems for
the operating system.  I believe the termination was due to over-consumption
of resources, but if it can't complete in 2:41 then I have a suspicion it never will.

[sholden@holden Python-2.0c1]$ time ./python ./Lib/test/test_longexp.py
Command terminated by signal 9
7.37user 492.86system 2:41:00elapsed 5%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (54713major+3735minor)pagefaults 6965swaps

-------------------------------------------------------

Date: 2000-Oct-11 06:48
By: gvanrossum

Comment:
It seems the test clearly uses more memory than the machine has.

So is this a Python bug or should the user just give up?

Is it worth that the test uses that much memory?

Could the test be changed to use less memory when invoked as part of the regression test?
-------------------------------------------------------

Date: 2000-Oct-11 12:40
By: fdrake

Comment:
This is simply due to memory starvation -- not a Python problem.
-------------------------------------------------------

For detailed info, follow this link:
http://sourceforge.net/bugs/?func=detailbug&bug_id=116546&group_id=5470