On Fri, Jan 16, 2009 at 10:40:24AM +0000, Andrew Ross wrote: > On Thu, Jan 15, 2009 at 11:31:24AM -0800, Alan Irwin wrote: > > Due to Andrew and Arjen's extremely recent efforts, java, python, and tcl > > join c++, f77, f95, perl, and ocaml to give perfect PostScript and stdout > > comparison results with the corresponding C examples. Here are the only > > remaining issues: > > > > octave > > Missing examples : 19 > > Differing postscript output : > > Missing stdout : > > Differing stdout : > > > > ada > > Missing examples : > > Differing postscript output : 17 > > Missing stdout : > > Differing stdout : > > adathick > > Missing examples : > > Differing postscript output : 17 > > Missing stdout : > > Differing stdout : > > > > This summary looks substantially better than our last release, and when you > > consider how much more extensive the testing now is behind the scenes, this > > result actually represents a tremendous improvement over our last release. > > > > Thanks to everyone who helped get us this far. > > On a 32-bit Linux system results are not (quite) so good. > > I note the problems above, plus > > python > Missing examples : > Differing postscript output : > Missing stdout : > Differing stdout : 23 > tcl > Missing examples : > Differing postscript output : 03 14a 17 21 > Missing stdout : > Differing stdout : 21 > > For python > > $ diff x23c_psc.txt x23p_psc.txt > 1c1 > < For example 23 prior to page 12 the FCI is 0x80000000 > --- > > For example 23 prior to page 12 the FCI is 0x80000000L > > This is with python 2.5.2. I imagine this is because the FCI constant > is a long rather than standard integer on a 32-bit system since python > does not have unsigned int. Or it may be a python version thing? > > For tcl the differences with 03 and 14a are just a single line where an > integer differs by one. I put this down to rounding error. This is with > tcl8.4. Example 21 requires tcl 8.5. Example 17 I see large differences, > but again often only rounding errors. This is hard to compare because a > small error can lead to the axes rescaling at a different time which > produces large differences in the postscript output. Visually the final > results looks good.
To follow up my own post - switching to tcl8.5 gives perfect results, so this is not a 32-bit issue. Perhaps we should add a note that tcl 8.5 is recommended, although previous versions will work? Andrew ------------------------------------------------------------------------------ This SF.net email is sponsored by: SourcForge Community SourceForge wants to tell your story. http://p.sf.net/sfu/sf-spreadtheword _______________________________________________ Plplot-devel mailing list Plplot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/plplot-devel
