search

Circuitscape / Circuitscape.jl

Algorithms from circuit theory to predict connectivity in heterogeneous landscapes
https://circuitscape.org
MIT License
128 stars 35 forks source link

Get CI to pass #245

Closed ViralBShah closed 3 years ago

ViralBShah commented 4 years ago

We should get CI to pass before tagging a new version.

I think given the tiff stuff, all the dependency updates, and the pmap fix (among other things), we should go to 6.0.

vlandau commented 4 years ago

I think the issue with CI not passing on Windows has to do with SharedArrays right? I got the sense that it won't be solved until multithreading is implemented from @ranjanan.

vlandau commented 4 years ago

Looks like AlgebraicMultigrid v0.3 requires Julia 1.4 or greater https://travis-ci.org/github/Circuitscape/Circuitscape.jl/jobs/697038000

vlandau commented 4 years ago

So, we need to either allow AlgebraicMultigrid v0.2 or bump the Julia compat to 1.4.

ViralBShah commented 4 years ago

We should bump the Julia compat. Especially with all the threading omniscape uses, it is best to keep moving up, since lots of issues keep getting fixed.

ViralBShah commented 4 years ago

Fixed in 21e5f61a3a63fa950828fca37df69d1dd1958eb3. Bumped CI versions as well.

ViralBShah commented 4 years ago

https://github.com/JuliaLinearAlgebra/AlgebraicMultigrid.jl/issues/65

ViralBShah commented 4 years ago

To make CI more reliable, we should revisit how many tests need to be run during CI.

ranjanan commented 4 years ago

To make CI pass we need to move to multi-threading. When that works, let's release 6.0. But till then we can make minor releases.

ViralBShah commented 4 years ago

Why do we need multi-threading to pass CI? Can we not make the CI test cases slightly less comprehensive for now?

ViralBShah commented 3 years ago

CI is still failing. We should get it to pass...

vlandau commented 3 years ago

Things seem to be passing now.