As the title states, absolute values of quantities need to be compared while picking the maximum value from a pair of values, to use as the normalizing factor for relative norms in the adaptivity computation. This is important when two negative values, or a negative value and a zero value are being compared in the adaptivity computation. @mathiskelm pointed this out.
Checklist:
[x] I made sure that the CI passed before I ask for a review.
[x] I added a summary of the changes (compared to the last release) in the CHANGELOG.md.
[x] If necessary, I made changes to the documentation and/or added new content.
[x] I will remember to squash-and-merge, providing a useful summary of the changes of this PR.
As the title states, absolute values of quantities need to be compared while picking the maximum value from a pair of values, to use as the normalizing factor for relative norms in the adaptivity computation. This is important when two negative values, or a negative value and a zero value are being compared in the adaptivity computation. @mathiskelm pointed this out.
Checklist:
CHANGELOG.md.