Repository navigation
Remove various Core.Boxes being captured by closures in tests - #5965
Merged
Merged
Conversation
The stratified Couette flow validation script sets initial conditions
via closures that reached into `model.grid` for the domain extents, and
a nested `print_progress` function reassigned `model`, which turned the
captured variable into a `Core.Box`. On KernelAbstractions 0.10 the CPU
backend compiles kernels like GPU backends do, so these closures were
rejected as non-bitstype kernel arguments:
KernelError: passing non-bitstype argument
.ε is of type var"#ε#20" which is not isbits.
.model is of type Core.Box which is not isbits.
Capture `Lx` and `Lz` as plain numbers instead, and stop reassigning
`model` inside `print_progress`, so the closures only hold numbers and
compile on every backend.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VHciC8x39gm97sABrSvBkt
Collaborator
|
I think all the validation scripts need to be refreshed, most of them are outdated |
Member
Author
|
This is sufficient to make the |
simone-silvestri
approved these changes
Sep 10, 2026
`z_immersed_boundary` had two methods, the second calling the first by
name. A local function that refers to its own name captures itself
through a `Core.Box`, so the closure is not isbits and cannot be a
kernel argument. `set!(field, func)` evaluates the function on the CPU,
which with KernelAbstractions 0.10 is a compiled backend with the same
isbits requirement as GPUs:
KernelError: passing non-bitstype argument
.func is of type var"#z_immersed_boundary#..." which is not isbits.
.z_immersed_boundary is of type Core.Box which is not isbits.
Only the 3-argument method is called, so define that one alone.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VHciC8x39gm97sABrSvBkt
`bᵢ` captured the whole `grid` to read `grid.Lx`. On the GPU the grid
holds `CuArray`s, so the closure is not isbits and cannot be a kernel
argument. `set!(field, func)` evaluates the function on the CPU, which
with KernelAbstractions 0.10 is a compiled backend with the same isbits
requirement as GPUs:
KernelError: passing non-bitstype argument
.func is of type var"#bᵢ#bᵢ##0"{ImmersedBoundaryGrid{...}} which is not isbits.
.grid is of type ImmersedBoundaryGrid{...} which is not isbits.
Capture the domain length as a number instead.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VHciC8x39gm97sABrSvBkt
giordano
changed the title
Core.Boxes being captured by closures in tests
Member
Author
|
I cherry picked a few other changes from #5799 getting rid of |
simone-silvestri
approved these changes
Sep 11, 2026
…ction
The immersed boundary function is called with `(x, z)` on grids with a
`Flat` y direction and with `(x, y, z)` otherwise:
MethodError: no method matching (::var"#z_immersed_boundary#...")(::Float64, ::Float64)
Define both methods with independent bodies; unlike a method calling the
other by name, this keeps the closure isbits.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VHciC8x39gm97sABrSvBkt
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #5965 +/- ##
==========================================
+ Coverage 67.90% 74.11% +6.20%
==========================================
Files 428 428
Lines 25517 26573 +1056
==========================================
+ Hits 17327 19694 +2367
+ Misses 8190 6879 -1311
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
I mentioned multiple times that reusing same label is problematic, now it's biting us 🙂
Taken out of #5799.