repository transfer

[ranocha]

I updated the URL accordingly. I also bumped the version number so that we can check the registration process.

I also relaxed the work-precision time tests since they failed sometimes, e.g., in https://github.com/NumericalMathematics/PositiveIntegrators.jl/actions/runs/14188778930/job/39748822294#step:6:848. I have seen this a few times before and it is annoying to have to restart the CI jobs just because of this.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

📢 Thoughts on this report? Let us know!

[SKopecz]

There is one work_precision test failing on MacOS.

work-precision fixed: Test Failed at /Users/runner/work/PositiveIntegrators.jl/PositiveIntegrators.jl/test/runtests.jl:2594
  Expression: maximum((v .- m1) ./ m1) < 2.5
   Evaluated: 8.10910718891217 < 2.5

If I recall correctly, we had issues with work_precision tests on MacOS before, and only on MacOS. The whole point of these tests is to ensure that two work_precision runs create more or less the same diagram. I don't think we should increase the relative tolerance any further in general.

I see two possibilities:

1. We increase the relative tolerance to about 10 for MacOS systems and decrease the tolerance again for all other systems. Then we know that everything works in general and consider this to be a MacOS issue.

2. We could play with the settings inside work_precision. For instance, we could see if increasing numruns helps. But I tried this in the past and it didn't have the desired effect.

So I would vote for Option 1.

[SKopecz]

Thanks! Please see #156 (comment).

[ranocha]

Good point. I made a change as suggested in the last commit.

[SKopecz]

Thanks a lot!

[JoshuaLampert]

Thanks!