Update for julia 1.x #5
Conversation
| - name: CompatHelper.main() | ||
| env: | ||
| GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} | ||
| COMPATHELPER_PRIV: ${{ secrets.DOCUMENTER_KEY }} |
There was a problem hiding this comment.
I’m not sure this package has a DOCUMENTER_KEY, since it doesn’t have any documentation. I guess we’ll see what this does in that case. In my opinion, we don’t really need CompatHelper.
| runs-on: ubuntu-latest | ||
| steps: | ||
| - uses: JuliaRegistries/TagBot@v1 | ||
| with: | ||
| token: ${{ secrets.GITHUB_TOKEN }} | ||
| ssh: ${{ secrets.DOCUMENTER_KEY }} |
There was a problem hiding this comment.
See above, I guess we’ll see whether this works or not. TagBot would be really good to have.
.vscode/settings.json
Outdated
| @@ -0,0 +1 @@ | |||
| {} | |||
There was a problem hiding this comment.
I'm sorry, I forgot to add the folder to gitignore, I will remove it
| @@ -0,0 +1,19 @@ | |||
| name = "UnitfulIntegration" | |||
| uuid = "d000b014-1cca-4823-a42a-b615c320b663" | |||
| authors = ["contributors"] | |||
There was a problem hiding this comment.
| authors = ["contributors"] |
IMO, just writing "contributors" doesn’t add much value. I would just remove the line (the authors field is not required, Unitful doesn’t have one either).
|
|
||
| PRs for other integration packages are welcome. |
There was a problem hiding this comment.
Why remove this? Do you think that they shouldn’t be welcome any more?
There was a problem hiding this comment.
That was my mistake, I intended to remove the part concearning the old julia version and ended up removing the PR part too
| # Test integration over an axis with units | ||
| @test QuadGK.quadgk(ustrip, 0.0m, 1.0m, atol=0.0m)[1] ≈ 0.5m | ||
|
|
||
| # Test integration where the unitful domain is infinite or semi-infinite | ||
| @test QuadGK.quadgk(x->exp(-x/(1.0m)), 0.0m, Inf*m, atol=0.0m)[1] ≈ 1.0m | ||
| @test QuadGK.quadgk(x->exp(x/(1.0m)), -Inf*m, 0.0m, atol=0.0m)[1] ≈ 1.0m | ||
| @test QuadGK.quadgk(x->exp(-abs(x/(1.0m))), | ||
| -Inf*m, Inf*m, atol=0.0m)[1] ≈ 2.0m | ||
|
|
||
| # Test mixed case (physical quantity-valued f and unitful domain) | ||
| @test QuadGK.quadgk(t->ustrip(t)*m/s, 0.0s, 2.0s, atol=0.0m)[1] ≈ 2.0m | ||
|
|
||
| # Test that errors are thrown when dimensionally unsound | ||
| @test_throws DimensionError QuadGK.quadgk(ustrip, 0.0m, 1.0s)[1] | ||
| @test_throws DimensionError QuadGK.quadgk(ustrip, 0.0, 1.0m)[1] | ||
|
|
||
| # Test that we throw an error when atol is not specified (at present | ||
| # I believe it is only possible to check when the domain is unitful) | ||
| @test_throws ErrorException QuadGK.quadgk(ustrip, 0.0m, 1.0m) |
There was a problem hiding this comment.
Concerning the test, are they all still relevant? I don't have any particular reason against more tests, I'm just not sure if all the old test are necessary. I guess it's better to have more tests than lesser though, I'll amend my commit
There was a problem hiding this comment.
Edit: it seems they were, as there are issues with the infinites handling (more broken stuff), I'm working on it
|
|
||
| # Test physical quantity-valued functions | ||
| @test QuadGK.quadgk(x->x*m, 0.0, 1.0, atol=0.0m)[1] ≈ 0.5m |
|
@ajkeller34 @giordano Can one of you approve running the workflows? I don’t have access to this repo, so I cannot do it. |
aurelio-amerio
force-pushed
the
master
branch
from
3d70547 to
ee1eec8
Compare
Co-authored-by: Sebastian Stock <[email protected]>
Co-authored-by: Sebastian Stock <[email protected]>
Co-authored-by: Sebastian Stock <[email protected]>
Co-authored-by: Sebastian Stock <[email protected]>
Co-authored-by: Sebastian Stock <[email protected]>
Co-authored-by: Sebastian Stock <[email protected]>
|
QuadGK 2.7.0 supports Unitful quantities out of the box. My suggestion would be to add a deprecation note to the README (#6) and close this PR. |
Following PainterQubits/Unitful.jl#597 , I have updated this library to let QuadGK work with Unitful.
Have a great day!