Re: 4.6.3 bug: tic labels disappear on minor tics

[Juhász P. <pet...@gm...>]

On Fri, 2013-06-28 at 11:12 -0700, Ethan Merritt wrote:
> On Friday, 28 June 2013, Juhász Péter wrote:
[...]
> 
> Another option is to relax the restriction that tic levels can only be 
> "major" (level 0) and "minor" (level 1).  The code already tracks tic level
> as an integer but ignores requests for tic level 2 or 3 etc.  It should be
> easy to modify the code that so that you can add labels using
>       set xtics add ("Level 2" 2 2)
> I'm not sure what length tic (if any) to assign at higher levels.

I think the goal is to have minor tics that nevertheless could have
labels.

> 
> > Allow the third number (the tic level) in the "set tics" specification
> > to be a real number instead of an integer. There would still be two tic
> > length settings, one for major, the other for minor tics, and the actual
> > length of a tic would be calculated by interpolating between the two
> > pre-set lengths (tic_len = tic_level * (minor - major) + major).
> 
> Sounds too complicated.
> 
> > Any tic level that is not 1 could have labels.
> 
> Yes - this.   Want to send a patch?

I'll try, but of course it's more complicated than it first appeared.
The xtick2d_... etc. callback functions test the text pointer being null
to distinguish between minor and major tics. To break this coupling,
they need to get the tic level as a separate parameter. I'll do it
tomorrow.

Peter

> 
> 	Ethan