inv(Matrix4x4) not working about oldimmutablearrays.jl

[twadleigh]

Update: I tried to see if switching the immutable array types to being subtypes of DenseArray instead being immediate subtypes of AbstractArray would help, but that resulted in a bazillion warnings concerning ambiguous definitions for the .+, etc. operators, as well as an error for not having an Array => Matrix[n]x[n] conversion.

I should have a definition for the conversion, anyway. I'll add it and see how far that gets us.

[twadleigh]

Turns out I did attempt to define a convert from AbstractArray, but it was done incorrectly such that it was never found. That didn't fix the problem, though.

I have a question pending on the mailing list about just how much linear algebra stuff are my types supposed to define to get all the linear algebra in Base to "just work".

[SimonDanisch]

Cool!
Thanks a lot for the effort!
I'd love to help, but I'm not familiar enough with the issue to quickly get
results.

2014-06-21 14:57 GMT+02:00 Tracy Wadleigh [email protected]:

Turns out I did attempt to define a convert from AbstractArray, but it was
done incorrectly such that it was never found. That didn't fix the problem,
though.

I have a question pending on the mailing list about just how much linear
algebra stuff are my types supposed to define to get all the linear algebra
in Base to "just work".

—
Reply to this email directly or view it on GitHub
#30 (comment)
.

[twadleigh]

I've added conversions from AbstractArray and to Array, so that, in the short term at least, you can round-trip the types to do fancier linear algebra.

For the longer term, I'll open an issue for adding an unrolled (and probably unpivoted) Gaussian elimination on which to base an inversion appropriate to immutable types.