Converting between PDL::Complex and native is still on the to-do list. What your expression created there is a PDL::Complex (which is a thin layer over a PDL with an extra dim of 2) whose underlying numbers were native complex numbers, hence your output. That would be straightforward enough to detect and allow for. Want to open an issue and/or a PR with just a test case (in t/complex.t please)? š
More broadly, interoperability between the two is conceptually a bit tricky, because of the dimensions thing, which was kind of the motivation for implementing this in the first place. As you know, I plan to allow the virtual slice of āreā and āimā with dataflow, but because thatās quite hard I havenāt started it yet (Iām learning more about PDL and complex numbers by converting Photonic to using the native stuff). I am open to ideas! By the way, folks whoāve kindly PR-ed stuff in the past from their own GitHub forks: could I ask you to delete branches that arenāt yours (such as indexops)? Itās very easy to do in the web UI (the trashcan icon on the āBranchesā list) and makes things tidier for me when Iām pulling your branches to my local repo (if I want to bring your work up to date before merging it, for instance). From: Luis Mochan<mailto:moc...@icf.unam.mx> Sent: 03 April 2021 20:07 To: pdl-devel@lists.sourceforge.net<mailto:pdl-devel@lists.sourceforge.net> Subject: [Pdl-devel] complex numbers Trying to find out how safe is it to mix different types of complex numbers (for migration purposes), I tried: pdl> use PDL::Complex pdl> print i*ci() 0+0i +0+1ii So, I guess not surprisingly, that they are not. I don't know if these mixed operations should be forbidden (I guess it would be the easiest course) or if PDL::Complex'es should be promoted to native complexes before the operation or the other way around.
_______________________________________________ pdl-devel mailing list pdl-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/pdl-devel
