Re: Bugs in new key placement code

[<br...@ph...>]

Petr Mikulik wrote:
>> 1) add more 'coordinate' classes for 3D plots, if only for debugging. 
>> We need commands like
>>
>>     set arrow from view 0,0,0 to view 1,1,1  # 'camera' coordinates
>>     set arrow from canvas 0,0 to canvas 1,1  # the 'plot_boundary'
> 
> Aha, there is no coordinate system for the bounding rectangle of the 3D?
> Do we need it?

I think so.  It would certainly help us in modifying the layout code, 
and it could be useful for users, too.  A 3D plot has more separate 
coordinate systems than a 2D one, but we only offer coordinate systems 
for those that also exist in 2D.

> This looks like a bug:
>     set arrow 1 from graph 0,0,0 to graph 1,1,1
>     splot x*y
>     pause -1
>     set view map
>     replot
>     pause -1
>     plot x
> 
> I would expect a line in another diagonal.

Well, I did point out that the 'set view map' implementation is an 
unholy mess, didn't I?

>> 2) Extend boundary3d(): add handling for the other 3 margins, 'set 
>> offsets', and completely re-think space reservations for ticks, 
>> ticklabels, and axis labels.
> 
> You mean 'set margins'? I think this is the most easy and helpful fix. 
> This would allow to scale down the 3D plot area manually in order to 
> allow enough space for the color box when needed.

Scaling it further down isn't exactly the most pressing need, IMHO.  The 
real problem are horrors like 'set view ,,1.3' many people use in an 
attempt to scale *up* the 3D plot area.  The notion that 3D plots were 
not utilizing enough of the available space was a good part of the 
motivation of 'set view map', and look where that got us.


> 
> ---
> PM
>