Re: v0.84 patchlevel 2 dial and xyoverlay problems

Dr. T.C. Zhao (zhao@bragg.phys.uwm.edu)
Tue, 24 Dec 1996 00:04:59 -0600

To subscribers of the xforms list from "Dr. T.C. Zhao" <zhao@bragg.phys.uwm.edu> :

From: Dimitris Pantzartzis <panas@rain.org>
>1. I cannot run fdesign:
> fdesign: can't resolve symbol 'snprintf'

Remember seeing this a couple of days ago and the solution
was to run the a.out version of fdesign. Xforms does not
call snprintf, must be a libc problem.

>2. the dial object now turns CW. Is there a way to have the dial object
> behave as in 0.81 (CCW) with some routine call? I am trying to have
> a dial turning from 9 oclock to 3 oclock CCW (bottom hemisphere of the
> circle) and after trying various combinations I awasn't able to reproduce
> the 0.81 behavior.
I will see what I can do. I got many requests for clock-wise rotations
and citing the fact that virtually all controls (volume control
on TV, radio etc.) works this way. Is uper hemisphere really
unacceptable ?

>3. There is a problem with xyplot overlays. The plot overlay is not redrawn
> after resetting the x/y bounds of the xyplot.
Thanks for the sample program. Not really a bug. You supplied two
points for you overlay plot and one of them is way outside of
the plotting area and got clipped before being mapped to the
screen (this has to be done otherwise wrap-around might occur
as the X screen coordinates only have a resolution of a short).
I will special-case the two point plot in next release.

>4. Setting the x/ybounds to 0.0 doesn't make the xyplot end at the
> last x data point but rather at multiples of tic label boundaries.

This behavior most of the times makes the plot look better. Maybe
adding a threshold (so that this is done only if the last point
is closer to a tic than the threshold) is the right thing to do.