Make is_prepared() more fine-grained #3759

roystgnr · 2024-01-04T00:21:39Z

I'm still trying to figure out how to make this easy to use, but I'd like to at least test that what I've got so far still passes CI.

lindsayad · 2024-01-04T18:00:30Z

Does "do not merge" also mean "do not review"?

moosebuild · 2024-01-04T18:16:11Z

Job Coverage on 7c8c1d1 wanted to post the following:

Coverage

	ffa3a3	#3759 7c8c1d
	Total	Total	+/-	New
Rate	62.35%	62.36%	+0.02%	98.67%
Hits	68275	68328	+53	74
Misses	41235	41236	+1	1

Diff coverage report

Full coverage report

This comment will be updated on new commits.

roystgnr · 2024-01-04T20:06:16Z

Does "do not merge" also mean "do not review"?

It's already bad enough to need fixing, huh? ;-)

Nah, please jump in as you see fit. I've got more I'm wanting to add here, but if there's flaws in the foundation then it's probably better to fix them before I put more layers on top.

lindsayad

Just looking at the header this seems like a great foundation to me

roystgnr · 2025-12-05T21:29:02Z

This might actually be in shape to merge soon. My remaining concerns:

The additional internal testing I've added might trigger dbg and/or devel failures downstream, in tests of code that doesn't properly prepare a mesh before solving (so it'll fail the new assertions) but either doesn't rely on the unprepared data or just compares to a gold solution that mis-relied on it (so it's not already failing tests). I'll add every recipe I think might be affected to the CI here.
The additional internal testing I've added might be too expensive in dbg mode, where it tests preparation validity before every assembly. I think we can afford that (prepare_for_use() is much cheaper if we don't repartition, which we don't when checking validity). If we see CI timeouts then I'll try moving my testing from assembly methods to solve() methods. If we still see timeouts after that then I'll remove the intensive testing entirely for now.

moosebuild · 2025-12-11T10:21:52Z

Job Coverage, step Generate coverage on b0367ec wanted to post the following:

Coverage

Inconsistent report tags were found between the head and base reports.
This can happen when reports are missing from either the head or the base.

Inconsistent tags:
32bit-np2_threads2
Full coverage report

This comment will be updated on new commits.

We need things like neighbor pointers, and in general we'd like to assert that we're never assembling or solving on unprepared meshes.

We can rebuild point locators or element caches, but we don't want to use an already-built invalid cache by accident.

If we're skipping partitioning, we shouldn't mark ourselves as partitioned if we're not, and we shouldn't expect that we're marked as partitioned.

roystgnr · 2026-02-11T20:35:16Z

I'm moving the "incredibly careful assertions" part of this PR to a new branch, https://github.com/roystgnr/libmesh/tree/mesh_preparation_assertions, which I'll put into a new PR for after this one goes in. That (plus the other PRs I've just factored out, plus one more bugfix PR that'll wait on #4394 because it requires a commit there) should get this PR down to just the functionality promised in the title, and should leave it in a ready-to-review state.

jwpeterson

This looks good to me, but @roystgnr and I discussed changing the names of the new APIs, e.g. set_hasnt_neighbor_ptrs() -> unset_has_neighbor_ptrs(), which I preferred. This would also include eventually deprecating set_isnt_prepared() in favor of unset_is_prepared().

jwpeterson · 2026-02-12T13:54:26Z

include/mesh/mesh_base.h

+  /**
+   * Flags indicating in what ways a mesh has been prepared for use.
+   */
+  struct Preparation {


We usually do all opening curly brace on next line.

roystgnr · 2026-02-12T18:02:44Z

I feel like 90% of the work on this PR has been fixes, not for bugs in the PR, but for bugs exposed by the PR.

Along with the API name changes I figured I'd throw in a couple more unit tests, despite the fact that they would surely be redundant with all the test coverage I got in the branch I'm following this one with ... and, long story short, there'll be a couple more non-trivial commits coming along with the name changes and the new tests: a bug fix in the MeshBase copy constructor, and some changes to allow copying and cloning of some types of badly-unprepared meshes.

Not sure how we missed this for so long.

My unit tests (in an upcoming commit) uncovered that it was impossible to copy a mesh with orphaned nodes that have boundary conditions, because copy_nodes_and_elements would helpfully delete them despite BoundaryInfo expecting to find them. But we do want to be able to copy even unprepared meshes without throwing an error. Instead of whining that it would be hard to break backwards compatibility here, I should have just added an option to drop backwards compatibility here.

If we have a parallel-inconsistent partitioning then we may have a real problem, but we might temporarily have nodes that need to be repartitioned, and we'd like to be able to copy them anyway.

This hits the easiest-to-set-up partially-unprepared cases, and even that was enough to expose a subtle bug or two.

roystgnr added the do not merge label Jan 4, 2024

roystgnr self-assigned this Jan 4, 2024

lindsayad mentioned this pull request Jan 4, 2024

Update PETSc to 3.20.3, update libmesh idaholab/moose#26462

Merged

lindsayad reviewed Jan 5, 2024

View reviewed changes

roystgnr force-pushed the mesh_preparation_finer branch from 7c8c1d1 to fc6ee94 Compare June 21, 2024 15:16

roystgnr force-pushed the mesh_preparation_finer branch from fc6ee94 to e39fc75 Compare December 5, 2025 21:17

roystgnr removed the do not merge label Dec 5, 2025

roystgnr mentioned this pull request Dec 5, 2025

[WIP] Debugging check that variables of discontinuous family live on just one elem->dim() #4345

Closed

lindsayad mentioned this pull request Dec 10, 2025

Refactor BoundaryInfo sync and add elems #4349

Draft

roystgnr mentioned this pull request Dec 16, 2025

Mesh preparedness fixes idaholab/moose#32111

Merged

roystgnr closed this in idaholab/moose@a3def84 Dec 17, 2025

roystgnr reopened this Dec 17, 2025

roystgnr force-pushed the mesh_preparation_finer branch 2 times, most recently from cc7fd1a to 9e6f0db Compare January 15, 2026 19:28

This was referenced Jan 29, 2026

Simple mesh fixes #4380

Merged

Distributed mesh fixes #4381

Merged

roystgnr force-pushed the mesh_preparation_finer branch from 7143591 to 6febe53 Compare February 11, 2026 00:00

This was referenced Feb 11, 2026

Distributed infinite element mesh building fixes #4394

Open

Allow for remote_elem interior_parent() values #4395

Open

GhostingFunctor clone() should be pure virtual #4396

Merged

roystgnr added 4 commits February 11, 2026 13:26

Make is_prepared more fine-grained

0016a5a

Make PerfLog name accurate

8e5fd88

More detailed preparation comments

809be07

Update comments

296bcd1

roystgnr added 4 commits February 11, 2026 13:30

Prepare mesh_copy before smoother solves on it

9386791

We need things like neighbor pointers, and in general we'd like to assert that we're never assembling or solving on unprepared meshes.

MeshModification shouldn't leave invalid caches

b0a29d3

We can rebuild point locators or element caches, but we don't want to use an already-built invalid cache by accident.

Clarify prepare_for_use() doxygen

14d6378

Handle skip_partitioning() in assertions

b0367ec

If we're skipping partitioning, we shouldn't mark ourselves as partitioned if we're not, and we shouldn't expect that we're marked as partitioned.

roystgnr force-pushed the mesh_preparation_finer branch from 6febe53 to b0367ec Compare February 11, 2026 20:35

jwpeterson reviewed Feb 12, 2026

View reviewed changes

roystgnr added 7 commits February 12, 2026 11:03

Copy subdomains in copy constructor+assignment

d43cd7e

Not sure how we missed this for so long.

set_hasnt_ -> unset_has_ (/is) name changes

30adbe8

Whitespace fix

730d2c0

Use skip_preparation option in copy constructors

ba88db9

Weaken copy_nodes_and_elements assertion

a7258c0

If we have a parallel-inconsistent partitioning then we may have a real problem, but we might temporarily have nodes that need to be repartitioned, and we'd like to be able to copy them anyway.

Add some unit tests for partial mesh preparation

a3571d7

This hits the easiest-to-set-up partially-unprepared cases, and even that was enough to expose a subtle bug or two.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Make is_prepared() more fine-grained #3759

Make is_prepared() more fine-grained #3759

Uh oh!

roystgnr commented Jan 4, 2024

Uh oh!

lindsayad commented Jan 4, 2024

Uh oh!

moosebuild commented Jan 4, 2024

Uh oh!

roystgnr commented Jan 4, 2024

Uh oh!

lindsayad left a comment

Uh oh!

roystgnr commented Dec 5, 2025

Uh oh!

moosebuild commented Dec 11, 2025 •

edited

Loading

Uh oh!

roystgnr commented Feb 11, 2026

Uh oh!

jwpeterson left a comment •

edited

Loading

Uh oh!

jwpeterson Feb 12, 2026

Uh oh!

roystgnr commented Feb 12, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Make is_prepared() more fine-grained #3759

Are you sure you want to change the base?

Make is_prepared() more fine-grained #3759

Uh oh!

Conversation

roystgnr commented Jan 4, 2024

Uh oh!

lindsayad commented Jan 4, 2024

Uh oh!

moosebuild commented Jan 4, 2024

Coverage

Uh oh!

roystgnr commented Jan 4, 2024

Uh oh!

lindsayad left a comment

Choose a reason for hiding this comment

Uh oh!

roystgnr commented Dec 5, 2025

Uh oh!

moosebuild commented Dec 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Coverage

Uh oh!

roystgnr commented Feb 11, 2026

Uh oh!

jwpeterson left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jwpeterson Feb 12, 2026

Choose a reason for hiding this comment

Uh oh!

roystgnr commented Feb 12, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

moosebuild commented Dec 11, 2025 •

edited

Loading

jwpeterson left a comment •

edited

Loading