Skip to content

Conversation

@ararslan
Copy link
Contributor

@ararslan ararslan commented Nov 7, 2018

The vecdot function was deprecated in favor of dot in JuliaLang/julia#27401; dot on matrices now treats them like vectors.

The change here updates our extension of dot to match this behavior and deprecates vecnorm.

@ararslan ararslan force-pushed the aa/vecdot branch 2 times, most recently from e43ba22 to 22caba8 Compare November 8, 2018 03:21
@omus
Copy link
Contributor

omus commented Nov 30, 2018

@ararslan what's left to do here?

@ararslan
Copy link
Contributor Author

I need to change this back to something like it was when I first pushed the branch to ensure that we don't end up using broadcast behaviors like extending singleton dimensions, as Eric noted.

The `vecdot` function was deprecated in favor of `dot` in Julia PR
27401; `dot` on matrices now treats them like vectors.

The change here updates our extension of `dot` to match this behavior
and deprecates `vecnorm`.
@iamed2
Copy link
Contributor

iamed2 commented Dec 4, 2018

What if the Value passed in is a 3+-dimensional array?

@ararslan
Copy link
Contributor Author

ararslan commented Dec 4, 2018

Those aren't supported by Convex, which is documented and assumed but not strictly enforced.

@iamed2
Copy link
Contributor

iamed2 commented Dec 4, 2018

Uh. Okay then I guess.

@ararslan ararslan merged commit db1b509 into master Dec 4, 2018
@ararslan ararslan deleted the aa/vecdot branch December 4, 2018 17:54
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Development

Successfully merging this pull request may close these issues.

5 participants