Fix scalar indexing issue in the callbacks

[huiyuxie]

Includes fix for #156.

[Copilot]

Pull Request Overview

Fixes scalar indexing in the callback stage by introducing a temporary CPU‐side Jacobian helper and updating the 1D integration API.

• Add volume_jacobian_temp as a stopgap before a GPU kernel implementation
• Swap usage of volume_jacobian for 1D callbacks and provide a new integrate wrapper
• Refactor comments in stable.jl and introduce get_node_vars_view for GPU array slicing

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

File	Description
src/solvers/dg.jl	Added volume_jacobian_temp; corrected GPU-dimension comment
src/callbacks_step/analysis_dg_1d.jl	Replaced volume_jacobian usage; added integrate for 1D
src/auxiliary/stable.jl	Reformatted comments; cleaned helpers; added get_node_vars_view

Comments suppressed due to low confidence (3)

src/callbacks_step/analysis_dg_1d.jl:74

• The positional arguments to integrate_via_indices are misaligned: the first parameter should be the function func, followed by u. Use integrate_via_indices(func, u, mesh, ...) to match its signature.

integrate_via_indices(u, mesh, equations, dg, cache;                          normalize = normalize) do u, i, element, equations, dg

src/callbacks_step/analysis_dg_1d.jl:71

• [nitpick] The name integrate may collide with existing Base or common library functions; consider a more specific name or module namespace to prevent ambiguity.

function integrate(func::Func, u,

src/auxiliary/stable.jl:28

• The new helper get_node_vars_view lacks unit tests. Adding coverage will catch slicing errors early.

# Call to get slice of GPU arrays outside of GPU kernels.