#1948 error in edge3d_intersect

[dima]

Hi. I consistently see a bug in 3d vector plots. This is with the bleeding gnuplot from the repos AND with older builds I have lying around. To reproduce, run this script:

set view equal xyz
set view map

splot '-' using 1:2:3:4:5:6 notitle with vectors ,'-' using 1:2:3:4:5:6 notitle with vectors ,'-' using 1:2:3:4:5:6 notitle with vectors ,'-' using 1:2:3:4:5:6 notitle with vectors ,'-' using 1:2:3:4:5:6 notitle with vectors
0.095 -0.581 -2.94 0.332110639476 0.943079108763 0.017452737884
0.095 -0.581 -2.94 -0.00579643957202 -0.016462057844 0.999847829541
0.095 -0.581 -2.94 0.943222775771 -0.332161219236 -7.23491987209e-07
e
0.119105016471 0.420544462544 -2.93079632762 0.332012099611 0.943180562645 0.0133668342611
0.119105016471 0.420544462544 -2.93079632762 -0.00759522547312 -0.0114971202323 0.999905200044
0.119105016471 0.420544462544 -2.93079632762 0.943244697075 -0.332082102458 0.00334648553835
e
0.095 -0.567 -2.952 -0.3090944652 0.95095203906 0.012216688544
0.095 -0.567 -2.952 -0.00452367945605 -0.0143157022401 0.9998865003
0.095 -0.567 -2.952 0.95101974974 0.309004363344 0.00872671344008
e
0.0792652681939 0.42916557448 -2.93639569994 -0.315191119053 0.948993084282 0.00806851559619
0.0792652681939 0.42916557448 -2.93639569994 -0.00372523495315 -0.00973898724515 0.999944844099
0.0792652681939 0.42916557448 -2.93639569994 0.949020072276 0.315143926804 0.00660486972555
e
0.0 0.0 -2.0 1.0 0.0 0.0
0.0 0.0 -2.0 0.0 1.0 0.0
0.0 0.0 -2.0 0.0 0.0 1.0
e

set xrange [ -0.0118404 : 0.240412 ]
set yrange [ -0.714327 : -0.409803 ]
replot

Here we plot some 3d vectors in a 'map' view (this works). Then we change the x,y extents to zoom in and replot. This produces the "error in edge3d_intersect" message, and the plot window behaves strangely from that point on.

[Ethan Merritt]

There seem to be multiple things going wrong here.

The initial plot is already incorrect. The "set view equal" and "set view map" interact badly.

Removing "set view equal" fixes the initial plot, but subsequent replot or refresh fails and anything after that is just weird.

I don't have any immediate diagnosis or fix for these problems, although using a datablock rather than inline '-' would probably fix the replot/refresh.

[Ethan Merritt]

Update:
I don't know what the proper fix is. Earlier fixes appear to have prevented the error from causing a persistent problem ("behaves strangely from that point on"), but the error message itself still occurs. Maybe it would be OK to make this warning rather than an error and thus let command processing continue even though the current plot shows a clipping failure.

[Ethan Merritt]

Found it.
Fixed now.