Re: use of "Octave:invalid-input-arg" as warning id

[Juan Pablo Carbajal]

On Tue, Dec 15, 2015 at 4:21 PM, Juan Pablo Carbajal
<address@hidden> wrote:
> On Tue, Dec 15, 2015 at 3:53 PM, Carlo De Falco <address@hidden> wrote:
>> Hi,
>>
>> The ode solvers newly introduced in core raise a warning
>> with id "Octave:invalid-input-arg" when an otion is not
>> set in the input to tell the user that the default value
>> is being used.
>>
>> This behaviour seems to have been introduced by this changeset
>>
>> $ hg log -r 20631
>> changeset:   20631:00caf63edcdf
>> parent:      20629:ab705b42cfd8
>> user:        Rik <address@hidden>
>> date:        Fri Oct 16 10:31:02 2015 -0700
>> summary:     maint: Remove obsolete ODE options from odeset/odeget/ode45.
>>
>> In order to replace warning id "OdePkg:InvalidArgument" which is
>> not correct in a core function.
>>
>> I'm not sure the current behaviour is correct either, though, as
>>
>> 1) "Octave:invalid-input-arg" is listed in the docs as an error_id but not 
>> as a warning_id
>> 2) the documentation says
>> 'Octave:invalid-input-arg'
>>      Indicates that a function was called with invalid input arguments.
>> but actually this warning is raised when no value is beeing supplied.
>>
>> Any idea how to deal with this cleanly?
>>
>> c.
>>
>>
>>
>
> It was always quite annoying to get a warning when a default value is
> used. Using a default value is not an error, and the user has to be
> made aware of the defaults value via the documentation.
> I think the cleanest thing is to not generate any warnings, unless of
> course, we are attaching to mathworks idiosyncrasy.

To add, if you are going to warn users when they use default values,
why you put the defaults in the first place?