Skip to content

Conversation

KristofferC
Copy link

Unfortunately, quite a lot of packages seem to be stuck at ArrayLayouts < 1.0 which errors on 1.11 due to the number of parameters change in the triangular array.

This backports the fixes so that this package passes test on 0.18 so that the packages stuff on older version keep working. Normally I wouldn't bother with this but there were enough packages that I felt it made sense.

A maintainer here would have to:

  • create a release-0.18 branch from this commit 449a3e5
  • merge this PR into that branch
  • make a new release from that merge commit

This should release 0.18.19 which should work on 1.11.

dkarrasch and others added 3 commits April 22, 2024 16:32
* remove usage of 2-arg `LA.AbstractTriangular`

* fix

* another fix

* fix v1.9

* include review comments

* Update Project.toml

* Update src/ArrayLayouts.jl

Co-authored-by: Alex Arslan <[email protected]>

---------

Co-authored-by: Alex Arslan <[email protected]>
(cherry picked from commit 32e5ab8)
@dlfivefifty
Copy link
Member

Maybe I’ve had a few times I many pints tonight but this PR boggles my mind… the current version is v1.9.x but you want to down grade it on the master branch to c0.18.x???

@dlfivefifty
Copy link
Member

Ah ok I now saw the comment about the “maintainer” doing a bunch of stuff.

I believe you have admin access through the org. so please feel free to do what needs to be done. Perhaps @jishnub is happy to do this otherwise. (If it’s not clear already: I’m not going to do it.)

it would be good to have a list of the packages stuck on earlier versions to try to get them up dated.

@jishnub jishnub changed the base branch from master to release-0.18 April 23, 2024 06:04
@jishnub jishnub merged commit 3aa78c6 into release-0.18 Apr 23, 2024
@jishnub jishnub deleted the kc/triag_backport branch April 23, 2024 06:34
@KristofferC
Copy link
Author

Maybe I’ve had a few times I many pints tonight

Cheers ;)

@jishnub
Copy link
Member

jishnub commented Apr 23, 2024

Unfortunately, this seems to lead to sporadic test failures (https://github.com/JuliaLinearAlgebra/ArrayLayouts.jl/actions/runs/8795905802/job/24137784781). The RHS seems completely incorrect (and possibly uninitialized). I haven't been able to replicate this locally.

@dlfivefifty
Copy link
Member

Does calling Random.seed!(32324) make it consistent?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants