add!(::ConstraintHandler, ::Dirichlet) for nonconcrete celltypes #427
Conversation
Codecov ReportBase: 92.92% // Head: 92.87% // Decreases project coverage by
Additional details and impacted files@@ Coverage Diff @@
## master #427 +/- ##
==========================================
- Coverage 92.92% 92.87% -0.05%
==========================================
Files 27 27
Lines 4111 4127 +16
==========================================
+ Hits 3820 3833 +13
- Misses 291 294 +3
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report at Codecov. |
Add new warning if there is no overlap, but the warning for only partial overlap is silenced
|
Could you review @kimauth? |
|
The current version doesn't support the strategy from #459 (comment). |
| @@ -43,6 +43,7 @@ SHA = "ea8e919c-243c-51af-8825-aaa63cd721ce" | |||
| StableRNGs = "860ef19b-820b-49d6-a774-d7a799459cd3" | |||
| Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40" | |||
| TimerOutputs = "a759f4b9-e2f1-59dc-863e-4aeb61b1ea8f" | |||
| Logging = "56ddb016-857b-54e1-b83d-db4d58db5568" | |||
There was a problem hiding this comment.
@test_logs comes from Test so I don't think this is needed?
There was a problem hiding this comment.
Hm, must have missed that when testing (probably tried without using Test) :)
Will remove
There was a problem hiding this comment.
Edit: It seems actually to be required for min_level=Logging.Warn
src/Dofs/ConstraintHandler.jl
Outdated
| if overlaps(fh, dbc) | ||
| # If dbc have dofs not in fh, then these will be removed from dbc, hence deepcopy | ||
| # In this case, add! will warn, unless warn_check_cellset=false | ||
| add!(ch, fh, deepcopy(dbc), warn_check_cellset=false) |
There was a problem hiding this comment.
I believe I asked @kimauth this some time -- but why is there a need to copy at all? See also
Ferrite.jl/src/Dofs/ConstraintHandler.jl
Line 43 in 70f69dd
There was a problem hiding this comment.
The deepcopy is required due to
Ferrite.jl/src/Dofs/ConstraintHandler.jl
Line 318 in 70f69dd
There was a problem hiding this comment.
But perhaps it would be better to rather recreate the entire Dirichlet like in your link, but perhaps here:
Ferrite.jl/src/Dofs/ConstraintHandler.jl
Line 313 in 70f69dd
There was a problem hiding this comment.
Yea I don't really understand why this doesn't just skip.
|
I recently ran to the issue which this PR resolves (I wanted to add a constraint on a surface/faceset which had cells in different fieldhandlers). Looks good to me :) |
This patch cleans up the code under ``` add!(::ConstraintHandler, ::Dirichlet) ``` a bit after the DofHandler merge. In particular, it deprecates the method ``` add!(::ConstraintHandler, ::FieldHandler, ::Dirichlet) ``` in favor of just ``` add!(::ConstraintHandler, ::Dirichlet) ``` which was added in #427. There is no need to be able to specify the constrained set by both the set passed to `Dirichlet` and by the set given to the `FieldHandler`. Looking at the tests modified in this patch, it seems this was never the intention anyway. This patch also removes the copy of the set in the `Dirichlet` constructor in favor of explicit partitioning of the set over the `FieldHandler`s.
This patch cleans up the code for adding `Dirichlet` constraints to the `ConstraintHandler` a bit after the DofHandler merge. In particular, it deprecates the method ``` add!(::ConstraintHandler, ::FieldHandler, ::Dirichlet) ``` in favor of just ``` add!(::ConstraintHandler, ::Dirichlet) ``` which was added in #427. There is no need to be able to specify the constrained set by both the set passed to `Dirichlet` and by the set given to the `FieldHandler`. Looking at the tests modified in this patch, it seems this was never the intention anyway. This patch also removes the copy of the set in the `Dirichlet` constructor in favor of explicit partitioning of the set over the `FieldHandler`s.
This patch cleans up the code for adding `Dirichlet` constraints to the `ConstraintHandler` a bit after the DofHandler merge. In particular, it deprecates the method ``` add!(::ConstraintHandler, ::FieldHandler, ::Dirichlet) ``` in favor of just ``` add!(::ConstraintHandler, ::Dirichlet) ``` which was added in #427. There is no need to be able to specify the constrained set by both the set passed to `Dirichlet` and by the set given to the `FieldHandler`. Looking at the tests modified in this patch, it seems this was never the intention anyway. This patch also removes the copy of the set in the `Dirichlet` constructor in favor of explicit partitioning of the set over the `FieldHandler`s.
This patch cleans up the code for adding `Dirichlet` constraints to the `ConstraintHandler` a bit after the DofHandler merge. In particular, it deprecates the method ``` add!(::ConstraintHandler, ::FieldHandler, ::Dirichlet) ``` in favor of just ``` add!(::ConstraintHandler, ::Dirichlet) ``` which was added in #427. There is no need to be able to specify the constrained set by both the set passed to `Dirichlet` and by the set given to the `FieldHandler`. Looking at the tests modified in this patch, it seems this was never the intention anyway. This patch also removes the copy of the set in the `Dirichlet` constructor in favor of explicit partitioning of the set over the `FieldHandler`s.
Allow the following syntax also for a Dirichlet condition covering nonconcrete celltypes
add!(ch, dbc)Before, this had to be done for each fieldhandler in
ch::ConstraintHandlerwith a uniquedbc::Dirichletfor each fieldhandler.