Hi Alan: Happy new year (almost) and Merry Christmas! I fixed plgfnam and added plcalc_world to the just-released PDL-Graphics-PLplot-0.47. I've got rid of your FIXMEs in x31.pl and checked it in. Both psc output and stdout now match the C version.
I'm still mulling over what to do about the scrambled argument order in many calls, fixing that (and cleaning up the rest of the interface) will be a big task! Best Regards, Doug dh...@ucar.edu Software Engineer IV UCAR - COSMIC, Tel. (303) 497-2611 On Thu, 18 Dec 2008, Alan W. Irwin wrote: > On 2008-12-18 13:19-0800 Alan W. Irwin wrote: > >> Hi Doug: >> >> The subject line pretty much says it all, but I wanted to make sure you >> got that information so we didn't duplicate effort. >> >> Once I have completed x31.pl, perl/pdl will have a perfect ctest compare >> result again. > > Well, I completed an initial version of x31.pl (revision 9199), but it turns > out that PDL::Graphics::PLplot needs to define plcalc_world and also needs > to revise plgfnam so that it works. (See comments near the various FIXME's > in x31.pl.) Could you please make the appropriate changes to > PDL::Graphics::PLplot? I would also appreciate you looking over x31.pl and > letting me know (a) what to do about the SYNC comment, and (b) letting me > know if you prefer a different style. > > There is one style issue I should discuss further with you. I found out by > happy accident that I didn't need to use the PDL->new(0) construct to set up > variables that will be filled by passing back arguments of plg functions. > > So instead of > > my $compression2 = PDL->new(0); > plgcompression ($compression2); > > I used the simpler method > > my $compression2 = plgcompression (); > > At least the simpler method appears to work for all the many plg functions > that are exercised by example 31, and I assume it takes care of all memory > management issues transparently. It's not clear that memory management is > okay if you pass back values through arguments specified by the user as in > the more complex method above. For example, what happens if some of the > passed back arguments are floats and others integers on platforms where > those are different sizes? > > I also notice x23.pl uses the more complex method while x01.pl uses the > simpler method (for plgver) and x14.pl uses the simpler method (for plgdev). > I just checked that "perl/x14.pl -dev xwin" worked fine. > Do you think it is a good idea to change x23.pl to use the simpler method? > > I also noticed for plgver that it will not accept any arguments. I assume > that is a good thing since passing back strings through arguments is > notorious for buffer overflow problems. So probably something has to be done > to plgfnam in PDL::Graphics::PLplot to make it conform to the plgver (and > plgdev?) style of passing back the string. > > Once you have updated PDL::Graphics::PLplot, you may want to also update > x31.pl or alternatively I can do that. In any case I will be happy to do > the required build-system changes for the new version of > PDL::Graphics::PLplot. > > I am looking forward to having an absolutely clean ctest compare result > for perl/PDL again. > > 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 > __________________________ > ------------------------------------------------------------------------------ _______________________________________________ Plplot-devel mailing list Plplot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/plplot-devel
