Re: waterfall implementation

[Juan Pablo Carbajal]

On Thu, Apr 25, 2013 at 12:00 AM, Mike Miller <address@hidden> wrote:
> On Wed, Apr 24, 2013 at 4:52 AM, Juan Pablo Carbajal wrote:
>> On Wed, Apr 24, 2013 at 4:11 AM, Mike Miller <address@hidden> wrote:
>>> I have a changeset ready to push that implements the function
>>> waterfall (inspired by a question on the help list today). It's all
>>> good, very simple function, except it doesn't work with gnuplot since
>>> it relies on the "meshstyle" property which apparently doesn't work
>>> with gnuplot. On gnuplot it will just look identical to meshz. Worth
>>> pushing this anyway? Should I mention in the docstring that it won't
>>> work with gnuplot? Thanks.
>>>
>>> --
>>> mike
>>
>> I do not know if is it worth pushing or not (maybe fits in a Forge
>> package?), but you can give a warning if gnuplot is the current
>> graphics toolkit using
>>
>> if strcmpi ("gnuplot", graphics_toolkit ())
>> warining ("Octave:matlab-incompatible", "This function does not
>> produce the expected output with gnuplot as graphic toolkit.\n");
>> endif
>
> Good idea, I added something similar to that. I took out the
> matlab-incompatible id because it seems to deal mostly with the
> behavior of the parser. Someone can add an id if it makes sense to.
>
> http://hg.savannah.gnu.org/hgweb/octave/rev/34a9a4e48f9b
>
> --
> mike


Please add an ID to the warning, we should not add warning with id
anymore. If you have not a better id for the wanring then follow my
suggestion cause according to warning_ids that is the best I could
find.