Re: [Gnuplot-py-users] Fwd: Test.py suite errors

[Michael H. <mh...@al...>]

Trevor Law wrote:
> Regarding error described below in reply history.  I followed the svn
> instructions (and did essentially what I believe I did before).  The
> same error occurred.  So something does appear to fail in the tests (as
> described below).  If someone with experience working with the source
> could let me know whether or not this test error is serious, I would
> appreciate it very much.
> 
> On 6/22/07, * bm...@ca... <mailto:bm...@ca...>* <
> bm...@ca... <mailto:bm...@ca...>> wrote:
>     Quoting Trevor Law < law...@gm... <mailto:law...@gm...>>:
> 
>     > I recently obtained Gnuplot.py and it appears to be
>     importable.  However,
>     > when I ran the test.py script, I got
>     > the following error message (I believe from gnuplot):
>     >
>     > gnuplot> set terminal postscript default enhanced
>     > gnuplot> set output "gp_test.ps"
>     > gnuplot> plot cos(0.5*x*x) title "cos(0.5*x^2)" with linespoints 2 2
>     > gnuplot> set terminal x11
>     > gnuplot> set output
>     > Testing hardcopy options: enhanced=1
>     > Press return to show results...
>     >         line 714: undefined variable: default
>     > "
>     >
>     > Otherwise the tests seemed to work fine (although some did not
>     produce the
>     > visual changes I expected,
>     > I do not know enough about gnuplot to know if there are any
>     issues).  The
>     > command in test.py that results
>     > in the error appears to be line 223.
>     >
>     > I would just like to make sure this is not a major issue that would
>     indicate
>     > that gnuplot.py is not functioning correctly
>     > on my machine.

It appears that the error message is emitted by gnuplot when Gnuplot.py
sends it the command "set terminal postscript default enhanced".
Probably the "default" option has been removed.

According to my version of gnuplot's "help set terminal postscript",
"default" is a valid option.  Obviously it must have been valid when the
test suite was written.  But presumably this option has been removed
from gnuplot without the documentation having been updated.

I wouldn't worry about it.  If the rest of the test suite runs OK, then
you shouldn't have any problems (as long as you don't use the "default"
option for hard copies!)

Patches to the test suite and/or Gnuplot.hardcopy() method (which
actually requires changes in termdefs.py) would be welcome.

Michael