Help needed with printing bug

[John W. Eaton]

I need help fixing this bug:

  https://savannah.gnu.org/bugs/?44916

I am unable to duplicate the problem, but have narrowed it down 
somewhat.  The issue seems to be that when gl2ps gathers the info it 
needs to draw the plot from OpenGL, some of the data points are not 
correct.  Apparently printing worked in 3.8.2 but fails now.  I tried 
undoing all the changes to the gl2ps rendering code that have been made 
since 3.8.2 but am told that did not solve the problem.

I thought perhaps we had changed the options for gl2ps or OpenGL in some 
way that triggered the bug.  But now I'm just clueless about what the 
problem could be, other than a driver issue, especially since the 
problem seems to appear on some systems but not others.

Maybe someone else has a better plan, but the only thing I can think of 
now to debug this is to bisect and try to determine which changeset 
introduced the problem.  But that will be a somewhat painful process as 
I'm unable to test whether a given changeset is good or bad myself.

So, first I need someone who can duplicate the error and who is willing 
to run a simple test for me.  We would do something like

  1. I build a binary
  2. You download and test it and report the result
  3. I bisect and go to 1 until we find the bad changeset

Given that there have been thousands of changes committed between 3.8.x 
and now, I guess this could take 12 iterations or so.

jwe