Re: pslatex code contaminating the gnuplot core

[Ethan M. <merritt@u.washington.edu>]

On Sunday 13 March 2005 01:16 pm, Hans-Bernhard Broeker wrote:
>  A proposition:  let's say we do add a
> term->changelayer(int layer) function, to be called by the core at
> appropriate places (and in strictly ascending order of 'layer').  The
> numbers passed to this function would have a meaning only to the core,
> but be documented somewhere.  What currently has options 'front' /
> 'back' would become layers 1 and 30000, e.g., and be documented as such.

I am not very enthusiastic about this.  
I can see a multitude of questions that arise immediately.

Is the core code always required to generate a series of these benchmark
calls, on the off-chance that a terminal driver is looking for a particular
transition?  
For example, does the terminal test routing test_command() have to call
term->changelayer(N) many times? Once? Never?
Is the value of N guaranteed to be strictly increasing?

Any driver implementing (term->changelayer)() must now keep extra state
information, possibly a lot of extra state information.
Suppose there is an int_error() or some other event that terminates
the sequence of layering calls prematurely.  Does the terminal driver
have to recognize this and reset its state when it receives the next
plot command?

Can this scheme co-exist with multiplot mode?
I would expect that the component graphs in a multiplot would each
step through some identical sequence of layering numbers, so that
the overall sequence would be something like
a,b,c,...,j        Operations setting up the page as a whole
k,l,...,n          Operations producing subplot 1
k,l,...,n          Operations producing subplot 2
...
k,l,...,n          Operations producing subplot <N>
o,p,...,z          Operations closing out the page as a whole
But this would mean that the layering calls are *not* strictly monotonic.
So now what does a driver do if it notices that the layer has jumped back
from n to k?  Ignore it?  Assume multiplotting is in progress?  
If somewhere in the range k...n there is a transition point, should the
driver allow it to trigger multiple times?

Would these term->changelayer calls have to be passed through to
outboard drivers like gnuplot_x11?  

Would the calls be regenerated during zoom commands
or other operations triggered by hot-keys?

-- 
Ethan A Merritt
Biomolecular Structure Center
University of Washington 98195-7742