in reply to Chart::Gnuplot dataset under Windows XP empty GIF

is there a way to find out if gnuplot is complaining about something through Chart::Gnuplot ?
You could make a copy of the source code of Chart::Gnuplot and edit the code to check the return value of the call to system in sub _execute:
if (system("$gnuplot $self->{_script}")) { die "Error: $gnuplot $self->{_script}: $?" }

Replies are listed 'Best First'.
Re^2: Chart::Gnuplot dataset under Windows XP empty GIF
by anthonyt (Novice) on Feb 15, 2009 at 14:35 UTC
    Thanks, that brings me a step closer. I am now getting:
    Error: e:\apps\gnuplot\bin\wgnuplot.exe E:\tmp\xhxbHAg9oN\plot: 256 at + C:/strawberry/perl/site/lib/Chart/Gnuplot.pm line 841.
    I'm not sure what the 256 is yet, but there are many temporary directories and files created and destroyed along the way. I can break the program at various points to see what is happening.

    Note: I used "set TMP=e:\tmp" to override placing the temporary files deep in my docs and settings folder.

    Note: Strawberry perl stores copies of its CPAN modules in three places, using Windows "Find" revealed C:\strawberry\perl\site\lib\Chart\Gnuplot.pm to be the working version.
      There seems a bug in the block "Data in points" in the subroutine _thaw. Please try to replace the lines
      my $dirTmp = tempdir(CLEANUP => 1, DIR => '/tmp'); my $fileTmp = "$dirTmp/data";
      and by the following line
      my $fileTmp = $self->{_data};
        Getting closer I think.

        Now the data temp directory is the same as the script directory was, but they both do not seem to exist at the same time. I'll try to find out why the data file is gone when the script file is ready to be used.

        Also, another bug: $baseTmp construction varies with $ENV(TMP) having priority over $ENV(TEMP) and vice versa elsewhere.