Re: Scale problem using plot()

[Ben Abbott]

On Oct 25, 2010, at 4:12 PM, Ben Abbott wrote:

> On Oct 25, 2010, at 4:03 PM, Michael D Godfrey wrote:
> 
>> On 10/25/2010 12:55 AM, Ben Abbott wrote:
>>> but what do you imply is wrong with the gnuplot version?
>>> 
>> The ymin should appear as 56, not  blank.  But, as you point out matlab does
>> it wrong, too.  In other cases gnuplot is like fltk in that it draws below
>> the xaxis line.  I hope that the fix to fltk will also fix gnuplot.  If not,
>> I will post a more complex case where both fltk and gnuplot draw
>> junk below the xaxis.
>> 
>> Michael
> 
> The scaling of the axes is determined by the "xlim" and "ylim" property 
> values. The values of the tickmarks aren't intended to serve that function. 
> Meaning they are not treated as line objects when the {x,y}limmode is "auto". 
>  So, I think gnuplot and matlab are doing the what is intended (i.e. no bug) 
> ... at least they seem to do the correct thing for this example.
> 
> If you can give me an example where gnuplot doesn't clip the tick marks, grid 
> lines its children, I can take a look at fixing that.
> 
> I've cc'd Shai. Perhaps, it is a simple task to crop the tick marks and grid 
> lines for the fltk backend.
> 
> Ben

I took a look at the sources and put a patch together for gl-render.cc

I'm skeptical I got everything correct, but it does compile and resolves the 
problem with the grid, tickmarks, and ticklabels being drawn beyond the axis 
limits.

Can someone take a look and let me know I haven't done anything t0o messy ... 
or worse, broke something.

Ben

tick_fix.patch
Description: Binary data