Well pointed-out! This is an artifact of PDL’s design of exporting lots of stuff, and especially having random libraries load functions into the “PDL” namespace. PDL::Complex should be considered deprecated, though I haven’t put that in the docs yet.
When I was looking at the lu_backsub stuff, I also replaced various “xchg(0,1)” and “mv(0,1)” with “transpose”, because that’s just clearer (and it automatically inserts dummy dimensions already as needed). I was tempted to implement a correct PDL::Complex::transpose, but quailed at the thought of the breakage that would cause for all the people who have code that expects the current behaviour. My direction of travel is that PDL::Complex will continue to behave as it does now, but users are recommended to write new code using native-complex. Best regards, Ed From: Luis Mochan<mailto:moc...@icf.unam.mx> Sent: 01 June 2021 01:49 To: pdl-devel@lists.sourceforge.net<mailto:pdl-devel@lists.sourceforge.net>; perldl<mailto:pdl-gene...@lists.sourceforge.net> Subject: Re: [Pdl-devel] PDL 2.049 released One detail to be aware of: pdl> p i()->info PDL: CDouble D [] pdl> use PDL::Complex pdl> p i()->info PDL::Complex: Double D [2] Thus, the function i() is modified when loading PDL::Complex! Regards, Luis On Mon, May 31, 2021 at 11:20:47PM +0000, Ed . wrote: > Dear PDL folks, > > PDL 2.049 has just been released. Notable changes: > > Matrix stuff: > - MatrixOps::identity now preserves higher dims on multi-dim input > - MatrixOps::lu_decomp now preserves inplace flag on input > - MatrixOps::lu_backsub now gets output dims right including inplace - thanks > @wlmb for report > > Complex numbers: > - PDL::Complex->from_native / as_native > - parse simple complex-number strings e.g. ‘[1+i 2-3i]’ > - replace PDL::Complex stringification with native-complex version > - PDL::r2C able to take Perl scalars > - add PDL::i2C > - replace creal, cimag with re, im, for easier PDL::Complex -> native > transition > - change default r2C and i2C input types to D from F > - add czip($re, $im) -> native complex > - i() as function (not constant) like nan() > - ci() removed > - add abs2 > - pdl([native-complex]) now creates a native-complex ndarray > - sumover/prodover now also process native complex > - comparing PDL::Complex and native complex other than for equality now a > fatal error, can == with real > - PDL::Complex slices that affect 0-th dim get re-blessed as PDL > - PDL::Complex can assign real values and threading won't also assign to imag > - PDL::Constants only loads PDL::Complex on demand > > General stuff: > - copy now copies ndarray and leaves original ndarray's inplace alone > - SvPDLV gives more comprehensible error if given a hash without "PDL" key > - Inline now not a dep but if installed gets updated to suitable version > - new pdlperl.h for Perl-specific functions/macros > - $PDL::undefval can be ndarray > - overloading now done in PDL::Ops/Primitive where functions defined > - overloading honours subclass methods > > As usual, please give it a try and report problems. > > Best regards, > Ed > _______________________________________________ > pdl-devel mailing list > pdl-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/pdl-devel -- o W. Luis Mochán, | tel:(52)(777)329-1734 /<(*) Instituto de Ciencias Físicas, UNAM | fax:(52)(777)317-5388 `>/ /\ Av. Universidad s/n CP 62210 | (*)/\/ \ Cuernavaca, Morelos, México | moc...@fis.unam.mx /\_/\__/ GPG: 791EB9EB, C949 3F81 6D9B 1191 9A16 C2DF 5F0A C52B 791E B9EB _______________________________________________ pdl-devel mailing list pdl-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/pdl-devel
_______________________________________________ pdl-devel mailing list pdl-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/pdl-devel
