Dear PDL folks,

PDL 2.037 has just been uploaded to CPAN. Key features since the last 
announcement:

  *   New type qualifiers for pp_def Pars: “real” and “complex”, motivated by 
PDL::FFTW3 so it can return the real or complex version of the input 
floating-point data type (so “real” = cfloat->float, “complex” = float->cfloat, 
etc) as necessary
  *   Bad value support is now always on, and no longer configurable to off, 
which simplified a fair bit of code
  *   To help with maximum reproducibility, support for ~/.perldl.conf has been 
removed (and the whole perldl.conf will go away in due course)
  *   pthread detection now works better, and works on MacOS and Windows
  *   GenericTypes in pp_def functions defaults if not specified to only the 
real datatypes, for backwards-compatibility; functions have to “opt in” by 
specifying complex datatypes
  *   The Proj4 modules work again, by using the CPAN Alien::Proj4 module – 
needed because general Linux distributions now package versions of proj that 
are not compatible with v4 anymore
  *   Makefile.PL only prints when things don’t work/aren’t available, which 
makes developing nicer

As usual, please give it a try and let me/us know of any problems.

Next steps:

  *   I intend to permanently enable the “badval per pdl” feature, and also 
incorporate the ability to use NaN as the badvalue, so those two configuration 
options will also disappear
  *   I intend to stop calling PDL objects “piddles”, because that’s terribly 
unclear what it actually means, and update all the docs to call them “ndarrays”
  *   Similarly, I intend to rename “threading” to “broadcasting” to eliminate 
unnecessary confusion
  *   oodler577 on the #pdl channel is helping towards PDL supporting OpenMP
  *   Better integration of native complex numbers

Thoughts welcome on any and all of the above!

Best regards,
Ed

_______________________________________________
pdl-devel mailing list
pdl-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/pdl-devel

Reply via email to