Re: bug in a(a<3)?

[John W. Eaton]

On  6-Sep-1999, address@hidden <address@hidden> wrote:

| If you look at (a<3) it comes out all ones, while (a<0.4) with a high
| probability comes out with at least one zero.  Octave can in the latter
| case deduce that you want to select items, but in the first case it
| is ambiguous and the result depends on how you've set a certain built-in
| variable (another reply to your query gives the name of the variable).
| Your default setting obviously treats [1 1 1] as three copies of '1'
| and not as three cases of "yes, include me."

FWIW, this is fixed in 2.1.x by introducing a new boolean type.

Results of comparisons are boolean, so even if the result is all
`true' values (sill represented by an integer value of 1, but now with
the additional information that it is a boolean value) there is no
ambiguity.

jwe



---------------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.  To ensure
that development continues, see www.che.wisc.edu/octave/giftform.html
Instructions for unsubscribing: www.che.wisc.edu/octave/archive.html
---------------------------------------------------------------------