Opened 11 years ago

Last modified 11 years ago

#10098 closed defect

Flaky doctest in sage/interfaces/expect.py — at Version 3

Reported by: mpatel Owned by: mvngu
Priority: blocker Milestone: sage-4.6
Component: doctest coverage Keywords:
Cc: boothby, drkirkby, malb, mariah, SimonKing, vbraun, was Merged in:
Authors: Reviewers:
Report Upstream: N/A Work issues:
Branch: Commit:
Dependencies: Stopgaps:

Status badges

Description (last modified by mpatel)

With #10004 merged into 4.6.alpha3, I get the doctest failures

#!
sage -t -long  -force_lib devel/sage/sage/interfaces/expect.py
**********************************************************************
File "/home/buildbot/build/sage/cleo-1/cleo_full/build/sage-4.6.alpha3/devel/sage-main/sage/interfaces/expect.py", line 596:
    sage: L = [t[1] for t in f(range(5))]
Expected nothing
Got:
    [Errno 16] Device or resource busy: '/home/buildbot/build/sage/cleo-1/cleo_full/build/sage-4.6.alpha3/.sage/temp/cleo/32313/dir_0/.nfs00000000007d541a0000abe9'
**********************************************************************
File "/home/buildbot/build/sage/cleo-1/cleo_full/build/sage-4.6.alpha3/devel/sage-main/sage/interfaces/expect.py", line 603:
    sage: L = [t[1] for t in f(range(5))]
Expected nothing
Got:
    [Errno 16] Device or resource busy: '/home/buildbot/build/sage/cleo-1/cleo_full/build/sage-4.6.alpha3/.sage/temp/cleo/32313/dir_1/.nfs00000000007d541a0000abf1'

on the Skynet machines cleo, eno, fulvia, sextus, and taurus.

Change History (3)

comment:1 follow-up: Changed 11 years ago by mpatel

  • Summary changed from Flaky doctest in to Flaky doctest in sage/interfaces/expect.py

This is reminiscent of #9501 / #9616. Since the fix at #10004 works (right?), maybe we should tag these tests as "not tested" instead of reverting the patch? Or is there another way?

comment:2 in reply to: ↑ 1 Changed 11 years ago by SimonKing

Replying to mpatel:

This is reminiscent of #9501 / #9616. Since the fix at #10004 works (right?),

I hope so...

maybe we should tag these tests as "not tested" instead of reverting the patch? Or is there another way?

What doees "device busy" exactly mean? Can one trac this down? Actually the purpose of #10004 was to avoid conflicts (race conditions) on the hard disk.

comment:3 Changed 11 years ago by mpatel

  • Cc drkirkby mariah vbraun added
  • Description modified (diff)

I think this is a filesystem quirk (a problem?), but I'm not sure. David, Mariah, and Volker, any thoughts?

Note: See TracTickets for help on using tickets.