Norms

[yoongkang]

I'm trying to pick up where @MpoMp left off with #204. Basically minor modifications to the specs because I don't feel there is much of a point of comparing exact values of floating point numbers.

[yoongkang]

Well that didn't go well. I'll look at them when I get home.

[yoongkang]

Got them passing.

[translunar]

Hey, could you please rebase into a single commit?

[yoongkang]

Sure. I didn't want to squash @MpoMp's commits, but I suppose you could add his name to the changelog.

[wlevine]

I wonder if this should be called matrix_norm instead of norm, so people won't get it confused with the vector norm (since NMatrix also supports vectors).

[wlevine]

What is the problem with nrm2? Can you file a bug on this? Maybe you should avoid using nrm2 since it doesn't work for complex types either (#389).

[yoongkang]

I honestly do not know. From the commit history at the other pull request (see #204) it seems that @MpoMp used nrm2 at some stage and changed it later but the reasoning wasn't documented.

I'd be happy to look into it though. But it won't be tonight.

[translunar]

Bumping this almost-completed pull request to the top for GSOC students who'd like to finish it. :)

[npriyadarshi]

@MohawkJohn I would like to work on this one if nobody is already doing it.

[translunar]

Go for it.

[translunar]

Oops, think this was closed by #586. Correct me if I'm wrong.