Re: [Bug-gsl] GSL on Solaris problems

[Henry Tillotson]

Dear Brian Gough,

On Mon, 23 Apr 2007, Brian Gough wrote:

[snip]
>
> > 2)  make test problem.
> >   After the successful make, we tried 'make check'. This failed in the
> >   'sort' test:
> > make  check-TESTS
> > FAIL: indexing gsl_vector_char, n = 128, stride = 1, ordered [19999]
> > FAIL: sorting, gsl_vector_char, n = 128, stride = 1, ordered [20000]
> >   I'm not really sure where to go from here in investigating this: if you
> >   have any suggestions for further things I could try I'd be most grateful.
>
> I haven't seen this problem before.  To investigate further you'd need
> to dump the contents of the vectors that are being tested to see what
> is going wrong.  The patch below will do that.
>
> --
> Brian Gough
> (GSL Maintainer)

thanks for the diagnostic patch - I have now applied it and run the

  make check

again. I have attached the output (rather than include it in this message),
since it is over 8000 lines long.

The thing I immediately notice is that all the 2nd and 3rd values on the
diagnostic lines are in the range -128 to 127, as if they are signed one-byte
integers: I don't know whether this is correct or not.

If you want my try any further diagnostics, please let me know.

regards

Henry Tillotson
=====================================================================
Henry Tillotson                         e-mail: address@hidden
Applications and Academic Support       phone: 020 7679 7827
Information Systems                     fax: 020 7388 5406
University College London, Gower Street, London WC1 E 6BT
=====================================================================