Firedrake meeting 2020 05 05

Date and time 2020-05-05 15:00UTC (16:00BST)

Action Items

1. ALL: (ongoing) triage the open issues and confirm if they are indeed still open (and perhaps provide labels)
2. KS, (DH, LM): Document describing what we think the mixed domain interface should look like (and hence what is needed in UFL, and whether it matches the existing fenics efforts).
3. DRS: Open PR for stream plot See PR #1670. Review required
4. DH: Tackle the remaining complex sprint items
5. FW: Add the hack to have shape derivatives tests working in complex
6. SV: Expunge references to COFFEE before tsslac merge
7. LM: Go through the tsslac PRs
8. JB: Add mechanism for python3.8

Minutes

Present:

Apologies:

DRS - Stream Plot PR

Correct metric for various cell geometric quantities discussed. "cell size" vs "cell diameter" where the latter is needed to work on quads. DECISION: "cell size" will return "cell diameter" on quads.

DH Complex Sprint

Now passing in almost all cases - 1 bug (where a test function and trial function get mixed up) needs to be fixed. Action DH to find time to fix to get final test passing. Noted that real mode could be improved with better implementation of PETSc real matrices. Merging plan discussed. Action Someone to look at TSFC merge. Action DH to find time to sort out Jenkins to run in real and complex mode. PSA: Merging complex will double testing load. Jenkins will stop testing all commits. Only merge commits will be tested and only branches with active pull requests will be tested.

SV tsslac merge

Work ongoing.

JB: Status of issue #1451

In light of dropping COFFEE in the (near?) future, what is the current status of: https://github.com/firedrakeproject/firedrake/issues/1451

RWH: further discussion needed on #1667

Proposed solution from last meeting introduces new problems without obvious solutions. See https://github.com/firedrakeproject/firedrake/pull/1667

Date of next meeting