Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo

Close

#2340 Failing Regression Test on OpenBSD/i386

None
closed
nobody
None
5
2012-12-11
2012-01-19
Edd Barrett
No

Hi,

A single regression test is failing on OpenBSD/i386 but *not* on OpenBSD/amd64:

> Running tests in rtest8:
> ********************** Problem 104 ***************
> Input:
> ev(e5, au = 0, omega = 2)
>
>
> Result:
> - u
> quad_qawf(%e , u, 0, 2, sin, epsabs = 1.e-10, limit = 32, maxp1 = 100,
> limlst =
> 10
> )
>
> This differed from the expected result:
> [.4000000000000001, 2.216570948815925e-11, 175, 0]

Discussion on ports@openbsd.org is here: http://comments.gmane.org/gmane.os.openbsd.ports/53209

If a fix is found, I would appreciate a patch.

Cheers.

PS.

Thanks for the comprehensive regression tests in maxima. When all of the tests are passing, this gives me great confidence in maxima :)

Discussion

  • Raymond Toy
    Raymond Toy
    2012-01-19

    Please specify what lisp you were using. Or provide the output of build_info();

     
  • Edd Barrett
    Edd Barrett
    2012-01-19

    Maxima version: 5.25.1
    Maxima build date: 8:25 1/13/2012
    Host type: i386-unknown-openbsd5.1
    Lisp implementation type: ECL
    Lisp implementation version: 11.1.1

     
  • Edd Barrett
    Edd Barrett
    2012-01-23

    Any luck with this? Do you need any other info?

     
  • Raymond Toy
    Raymond Toy
    2012-01-24

    No progress thus far. I just need to get some time to get a version of ecl 11.1.1 to run on my machine that will build maxima. (Oh, I won't be testing this on OpenBSD, but Linux and/or Mac.)

     
  • Edd Barrett
    Edd Barrett
    2012-05-01

    Any joy?

     
  • Raymond Toy
    Raymond Toy
    2012-05-02

    No progress because on my Mac machine with ecl 11.1.1, this problem passes. But I suspect the epsabs is too small on ecl for some reason. Try changing it to 1e-9 and see if the test passes:

    quad_qawf(exp(-u), u, 0, 2, sin, epsabs = 1.e-9);

     
  • Robert Dodier
    Robert Dodier
    2012-12-11

    I'm informed via email from Edd Barrett that ECL 11.1.1 + Maxima 5.29.1 passes the test suite with no unexpected failures. Closing this report accordingly.

     
  • Robert Dodier
    Robert Dodier
    2012-12-11

    • status: open --> closed
    • milestone: --> None