From: Alan W. I. <ir...@be...> - 2011-05-03 05:59:14
|
On 2011-05-02 16:22-0600 Doug Hunt wrote: > Hi Alan: I've looked at your updated instructions for running the perl/PDL > tests. I don't see any easy way to improve upon them. > > I typically install plplot first (with no mention of perl/PDL) and then PDL > (with no mention of plplot, so the older internal version of > PDL::Graphics::PLplot is not installed) and finally install > PDL::Graphics::PLplot. > > At this point, if I want to run the perl/PDL examples from plplot, I would > just go to the plplot/examples/perl directory in the plplot build area and > run the perl scripts there manually. > > Perhaps it would make sense to package these test files with > PDL::Graphics::PLplot in addition to plplot/examples/perl? There are several > down-sides to this (mostly dual maintenance) but it would allow one to > incorporate all the plplot tests with the 'make test' part of the > PDL::Graphics::PLplot install. One could even put md5sums of expected output > files (maybe in xfig format or some other device that is most always > available) in PDL::Graphics::PLplot so the test could be rated as > definitively passing or failing. Hi Doug: We pretty much rely upon you for Perl/PDL expertise so I assume you will be doing most of the Perl/PDL maintenance of the examples in any case. After a round of such maintenance I would encourage you to go ahead and copy those updated examples from PLplot to the PDL::Graphics::PLplot package since those examples are useful documentation of how to use PDL::Graphics::PLplot. If you try your md5sum idea, then probably PostScript results from the psc device (with datestamp excluded which can be straightforwardly arranged) or SVG results from the svg device are the ones you should use. Both those file devices have no external dependencies, and they are much better maintained than the xfig device. Hopefully, you will find the md5sum will be invariant from one platform to the next for all our test example results. That depends on whether platform floating-point differences propagate to the rounded results in PostScript (or SVG) results or not, but you will have to find that out by experiment. Alan __________________________ Alan W. Irwin Astronomical research affiliation with Department of Physics and Astronomy, University of Victoria (astrowww.phys.uvic.ca). Programming affiliations with the FreeEOS equation-of-state implementation for stellar interiors (freeeos.sf.net); PLplot scientific plotting software package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net); and the Linux Brochure Project (lbproject.sf.net). __________________________ Linux-powered Science __________________________ |