www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - More PLPlot stuff

reply dsimcha <dsimcha yahoo.com> writes:
So I've started the PLPlot project by building the original version and
playing around with it and...

That's where the trouble begins.  PLPlot has an amazingly complicated Rube
Goldberg build system.  After fighting with it for a good hour or two, I
finally got it to build.  It output a bunch of DLL files.  (Side rant:  Why do
people use DLLs?  Static linking is so much less prone to Murphy's Law and
memory and storage are cheap nowadays.)  I used implib to generate .lib stubs
for all these files and passed them all to the linker.  I still can't even get
a program that uses the original PLPlot to link.  I get weird symbol undefined
messages, yet I found the name of the function for which I'm getting this
error in one of the .lib files output by implib.

Has anyone else gotten this contraption to work with D?  If so, how?  I'm
starting to think this project might not be the best idea, because if everyone
else has half the trouble I've had so far with this Rube Goldberg build system
and DLL/implib crap, the barrier to entry will just be too high to have anyone
actually use a D-ified PLPlot.  I'm halfway thinking it would be easier to
translate the whole #( ! lib to D than make a sane build system for the mixed
language version.
May 11 2009
parent Mike W <cf725b-0 yahoo.com> writes:
Ha!  I spent nearly 3 hours this past Saturday trying to get
PLPlot to build (Windows, MinGW) before finally giving up.  I have
to agree that without some serious changes to their install
process (or maybe at least some clearer instructions), few people
are going to use it.  Seems a shame, it looks like it's pretty
nice once it's up and running.
May 11 2009