Skip to content

Conversation

@wence-
Copy link
Contributor

@wence- wence- commented Dec 9, 2016

More careful attempt to fix #920. See 00d9066 for explanation.

This exercises a corner case where we should kill both components but
only end up killing one.
If a boundary node appears in one BC object that sets all the
components, but another that only sets one component, we will set some
high bits.  Later, in PyOP2, we use the presence of high bits to decide
which components to kill.  Therefore, if any of the BC objects are going
to set a component, we should "unroll" the full BCs to set all the high
bits so that we kill all the components.  Fixes #920, hopefully for good.
@wence- wence- merged commit 00d9066 into master Dec 12, 2016
wence- added a commit that referenced this pull request Dec 12, 2016
* more-indexed-vfs:
  More careful fix of component bcs
  Add test of component bcs where full bc overlaps
@wence- wence- deleted the more-indexed-vfs branch December 12, 2016 17:37
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Applying DirichletBC on component of a VectorFunction that is in a mixed function space does not work

2 participants