Hi, On 15/09/15 22:32, Anton Gladky wrote: > tags 786351 +patch > thanks
> in attachment there is a proposed patch, which removes > eigen2_support for this package. The patch is relatively > large, but it should work and it will be very good to ask > upstream to merge it there. > There is one problem place in this patch (see comments > starting from "This term was disabled...."). There is a > Vector2d constructed as Vector2d(double), what is > strange and leads to undefined behavior, because it > requires 2 values. Such construction fails to build with > a 3.3~alpha1 of eigen3 (in experimental). > Is it OK for you if I apply this patch and upload it to > experimental to test this version? I've adapted your patch missing piece to something that makes sense gut wise, that is, the first Vector2d is the variance related to the charges and the second one to the position. In any case, the previous version was segfaulting in this calculation. I've uploaded this patched version to experimental, and it needs to be tested. At the same time we need to send this patch upstream, could you do this? It's always better to be able to give some input on the patch if upstream has some question regarding it, and eigen3 wise, you are the expert here. Happy hacking, -- "Can you imagine what I would do if I could do all I can?" -- Sun Tzu Saludos /\/\ /\ >< `/
signature.asc
Description: OpenPGP digital signature
