Goran Flegar

> I would prefer templating it on T, just like shared_ptr or unique_ptr, rather than on the pointer type. It's a different kind of pointer. Behavior should be the same,...

> Alas, there is no standard C++ library way of doing that at compile time. If you think about it, it's impossible to do at compile time anyway. You only...

> How do you handle non-null objects when you do transfer ownership? The solution doesn't handle the case of non-null uniqe_ptr or shared_ptr, just a plain pointer. I still don't...

> nullptr is a constexpr, and new (by default) must throw if allocation fails. Thus, in theory, anything that is not nullptr and that comes from (default) new, can be...

@tcojean @hartwiganzt just want to remind you of this one. It's quite an important feature, which should not be that difficult to implement (there's a prototype implementation above), but constitutes...

Somewhat related is also #179.

Here is my rough idea for now, given the information I have. Since an eigensolver produces a (partial) eigendecomposition, we could treat it the same way as we plan to...

Btw, everyone should feel free to edit the description of the issue however they see fit (think of it as a wiki page). The idea of it is to provide...

Do they have support for different architectures, e.g. GPUs? EDIT: Ah, you're talking about an implementation we could use as a reference, correct? I though you were answering __Q6__.

> Last update: May 27, 1998. Did they forget to update the update date? :confused: