Re: Need help/advice with NIST dataplot
On 20101219_145122, Riccardo Stagni wrote:
> On Sun, Dec 19, 2010 at 02:54:54AM -0700, Paul E Condon wrote:
> > It worked. It took a while to get libgd2-xpm-dev installed on my system
> > because I had previously installed gnuplot. Gnuplot install had automatically
> > installed libgd2-noxpm and the two are incompatible. Of course, gnuplot
> > counld have installed libgd2-xpm, but that is not how the package install
> > script is written.
> > I don't know much at all about writing a successful Debian package, but I
> > imagine this behavior of the gnuplot package will provide an interesting
> > extra challenge.
>
> gnuplot depends on "libgd2-noxpm (>= 2.0.36~rc1~dfsg) | libgd2-xpm (>=
> 2.0.36~rc1~dfsg)", so it should not be a problem. Installing
> libgd2-xpm-dev with aptitude, it just asks me if it should remove libgd2-noxpm
> (without saying anything about gnuplot because of its dependencies).
>
> Anyway, if dataplot doesn't need to write xpm files, you could simply have
> installed libgd2-noxpm-dev.
>
> ciao
> Riccardo
With libgd2-noxpm-dev installed, I got errors during compile complaining that compiler
could not include gd.h . This, I think, is a defect in the dataplot compile script, but
it is a fact. My work-around was to figure out how to get libgd2-xpm-dev installed,
and then re-install gnuplot, which seems to have no problem using -xpm. Not having ever
seen Dataplot functioning, I cannot know whether it needs to write xpm files. Because
of the compile error, I never got to confronting that issue. I supposed I could have
told the script that libgd was not available, but then I would not get to see the full
capability of the software. I am installing it in order to see how useful it is. It
seems unfair to conduct such a test with a deliberately stripped down version.
Cheers,
--
Paul E Condon
pecondon@mesanetworks.net
Reply to: