Small change/fix in Mixed Riemanian Trust region solver

The former implementation tests if the norm of the correction of the trust region sub problem is smaller than innerTolerance_ which is usually a small number. If the correction dX has a large magnitude, e.g. ||dX|| = 2. The for loop with innerIterations_ runs till the end because ||dX|| will always stay larger than innerTolerance_. And the full set of innerIterations_ is performed even if corr_global is already converged within machine precision. Therefore, to terminate the loop at a better time comparing innerTolerance_ to diff which measures the change of corr_global within the current iteration.

I discussed this with @jaap and @lnebel . Maybe you two can check if i missed something?

What do you think about this? @osander

@amueller: Can you restart the pipeline? I created a branch on my fork of gfe and there it's passing: https://gitlab.mn.tu-dresden.de/lnebel/dune-gfe/-/pipelines/5760.

@osander: Can you merge this as soon as the pipeline here passes or let us know what needs to be changed?

I restarted the pipeline.But somehow dune:git parmg dune-vtk clang and dune:git parmg dune-vtk gcc always fails. This is also true for !70 (merged) .

The pipes fails with

duneci-install-module https://gitlab-ci-token:${CI_JOB_TOKEN}@gitlab.mn.tu-dresden.de/paraphase/dune-parmg.git Cloning into 'dune-parmg'... remote: The project you were looking for could not be found. fatal: repository 'https://gitlab.mn.tu-dresden.de/paraphase/dune-parmg.git/' not found

How can this be fixed?

Can you restart it once more? I added you to the repo https://gitlab.mn.tu-dresden.de/paraphase/dune-parmg.git/ as a Developer.

This fixed it! Thanks!

Indeed, this does look bogus. Thanks for reporting it!

merged

mentioned in commit 9e5d2428