[9.5] Refactor grid fixup functions

[drwells]

Followup to #13765 and #13766.

We typically (but not always) call the same three grid fixup functions after loading in data. This PR adds a utility function for calling all of them in the right order under the correct circumstances. As of #13766 we can detect twisted pyramids and wedges, so we can call GridTools::invert_cells_with_negative_measure() for all types of supported meshes.

[drwells]

On second thought, this might break things in a weird way (it at least substantially changes the behavior of loading mixed meshes) so lets hold off until 9.5. If necessary users can always fix meshes themselves.

[marcfehling]

One test fails:

1315/4553 Test #1319: bits/distorted_cells_02.debug .........................................................***Failed    9.84 sec
make[3]: *** [distorted_cells_02.debug/diff] Error 1
make[2]: *** [CMakeFiles/bits.distorted_cells_02.debug.test.dir/all] Error 2
make[1]: *** [CMakeFiles/bits.distorted_cells_02.debug.test.dir/rule] Error 2
make: *** [bits.distorted_cells_02.debug.test] Error 2
Test bits/distorted_cells_02.debug: DIFF
===============================   OUTPUT BEGIN  ===============================
[  0%] Generating distorted_cells_02.debug/interrupt_guard.cc
Scanning dependencies of target bits.distorted_cells_02.debug
[  0%] Building CXX object CMakeFiles/bits.distorted_cells_02.debug.dir/distorted_cells_02.cc.o
Building CXX object CMakeFiles/bits.distorted_cells_02.debug.dir/distorted_cells_02.debug/interrupt_guard.cc.o
Linking CXX executable distorted_cells_02.debug/distorted_cells_02.debug
Built target bits.distorted_cells_02.debug
Scanning dependencies of target bits.distorted_cells_02.debug.test
Generating distorted_cells_02.debug/output
Generating distorted_cells_02.debug/diff
bits/distorted_cells_02.debug: BUILD successful.
bits/distorted_cells_02.debug: RUN successful.
bits/distorted_cells_02.debug: DIFF failed. ------ Source: /jenkins/workspace/dealii_PR-13820/tests/bits/distorted_cells_02.debug.output
bits/distorted_cells_02.debug: DIFF failed. ------ Result: /home/dealii/build/tests/bits/distorted_cells_02.debug/output
Check /home/dealii/build/tests/bits/distorted_cells_02.debug/output /jenkins/workspace/dealii_PR-13820/tests/bits/distorted_cells_02.debug.output
bits/distorted_cells_02.debug: DIFF failed. ------ Diff:   /home/dealii/build/tests/bits/distorted_cells_02.debug/failing_diff
bits/distorted_cells_02.debug: DIFF failed. ------ First 20 lines of numdiff output:
----------------
##2       #:2   <== ExcGridHasInvalidCell
##2       #:2   ==> ExcInternalError
@                                                     @@
##2       #>3   <== cell_no)
##2             ==>
@ Line 2 in file "output" is shorter than expected!
----------------
##3       #:2   <== ExcGridHasInvalidCell
##3       #:2   ==> ExcInternalError
@                                                     @@
##3       #>3   <== cell_no)
##3             ==>
@ Line 3 in file "output" is shorter than expected!

+++  File "/jenkins/workspace/dealii_PR-13820/tests/bits/distorted_cells_02.debug.output" differs from file "output"
CMakeFiles/bits.distorted_cells_02.debug.test.dir/build.make:64: recipe for target 'distorted_cells_02.debug/diff' failed
CMakeFiles/Makefile2:19075: recipe for target 'CMakeFiles/bits.distorted_cells_02.debug.test.dir/all' failed
CMakeFiles/Makefile2:19082: recipe for target 'CMakeFiles/bits.distorted_cells_02.debug.test.dir/rule' failed
Makefile:7851: recipe for target 'bits.distorted_cells_02.debug.test' failed


bits/distorted_cells_02.debug: ******    DIFF failed    *******

===============================    OUTPUT END   ===============================
Expected stage PASSED - aborting
CMake Error at /home/dealii/build/share/deal.II/scripts/run_test.cmake:144 (MESSAGE):
  *** abort

[drwells]

Hm. We might want to keep throwing that exception - it is a lot more informative than an internal error. I'll see if I can get that to work.

[bangerth]

@drwells Any news?

[drwells]

Not yet. I've been traveling for most of the last two weeks and I haven't had time to look at this.

[drwells]

That was pretty easy to address - all we need to do is switch to the correct exception type here.

[drwells]

/rebuild

[drwells]

Can someone take another look?

[kronbichler]

Nice!