Hi David, I write this although it probably is unrelated, but who knows... I've implemented native complex numbers (as in access to the C types) and am running into issues when compiling for Windows with bad value support and, I think, use nan as bad (or the other one). I've never gotten to the bottom of this. Saying this, I suspect the implicit assumption (spread across the core/pp code base) of relying on type orders (integers come first ... ) is causing issues.
The other thing that I can think of would be word size interpretation issues between perl, PDL and the OS, something similar to what had to be done for the index data type due to 32/64 bits. Sorry, but I am just guessing. On 19.06.20 15:29, David Mertens wrote:
Hello everyone, In my research I needed the uncertainties in my linear fits. I've always used PDL::Fit::OO for linear fitting (because I wrote it, it has no Fortran dependencies, and know it well enough to work even on my students' machines), so I finally got around to implementing a covariance matrix calculation. I encountered an error (segfault) when my student tried to run the covariance calculation on his Windows 10 machine. It seems that the combination of GenericTypes and a temporary piddle cause things to go bad. To fix the problem, we just commented-out the GenericTypes line before compiling on his machine. As such, I don't have access to the exact error anymore. The student has a fairly new version of Perl (5.20s) and installed PDL within the last 12 months, so everything is fairly fresh. Still, it seems like a GenericTypes constraint together with a temporary piddle must be *somewhere* in the PDL codebase, so this kind of issue should have cropped up. Does this ring any bells? Is there something I'm doing wrong in my code? Thanks! David -- "Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." -- Brian Kernighan _______________________________________________ pdl-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/pdl-general
_______________________________________________ pdl-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/pdl-general
