refactor: extract s_write_field helper in post_process s_save_data

[SVS87]

Description

Collapse the 38 repeated fill->name->write->clear stanzas in s_save_data into calls to a new private helper s_write_field. The helper optionally slices a scalar_field into out%q_sf, writes varname to the database, and clears varname on return.

Also fixes the ib_markers block which was missing the varname clear, and moves the elasticity/hyperelasticity varname clear inside the prim_vars_wrt guard where it belongs.

Part of #1523

Type of change (delete unused ones)

• Bug fix
• Refactor

Testing

Ran ./mfc.sh test -j 4 locally on Ubuntu/WSL with GNU compiler and MPI. 578 tests passed, 0 failed. Output-neutral by construction — the helper performs the exact same slice, write, and clear in the same order, so database files are bit-identical.

Checklist

Check these like this [x] to indicate which of the below applies.

• I added or updated tests for new behavior
• I updated documentation if user-facing behavior changed

See the developer guide for full coding standards.

AI code reviews

Reviews are not retriggered automatically. To request a review, comment on the PR:

• @claude full review — Claude full review (also triggers on PR open/reopen/ready)
• Or add label claude-full-review — Claude full review via label

[codecov]

Codecov Report

❌ Patch coverage is 88.88889% with 5 lines in your changes missing coverage. Please review.
✅ Project coverage is 60.45%. Comparing base (b4be438) to head (0926d3d).
⚠️ Report is 5 commits behind head on master.

Files with missing lines	Patch %	Lines
src/post_process/m_start_up.fpp	88.88%	5 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1600      +/-   ##
==========================================
- Coverage   60.51%   60.45%   -0.07%     
==========================================
  Files          83       83              
  Lines       19905    19852      -53     
  Branches     2950     2951       +1     
==========================================
- Hits        12046    12001      -45     
+ Misses       5866     5858       -8     
  Partials     1993     1993              

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[sbryngelson]

Reviewed as a behavior-equivalence audit of the refactor. Traced all 38 converted stanzas against their originals:

• Every src-passing call uses the same field, index expression, and x_beg:x_end,y_beg:y_end,z_beg:z_end slice as the original inline fill (checked the index-bearing ones: mom, vel, Y, alpha_rho_e, tau, xi, alpha, color_function, all qbmm rs/vs/ps/ms(i)).
• Every no-src call still has out%q_sf populated beforehand by a retained inline slice, an s_derive_* call, or a manual loop — none orphaned.
• s_write_field is a sibling module procedure; out resolves via use-association from m_data_output (public module var), and name_len/scalar_field/wp from their respective modules. ✓
• The varname-clear relocation is inert: every DB write is preceded by a full internal write(varname,...), which blank-pads the whole fixed-length scalar, so the clears were always redundant.

Output-neutral, as claimed. One latent (non-blocking) note inline.

[sbryngelson]

Latent interface weakness (non-blocking): the slice is guarded on present(src) only, not on the six bound args. The interface therefore permits a src-without-bounds call — e.g. call s_write_field(varname, t_step, q_cons_vf(k)) — which would dereference non-present optional integers x_beg..z_end (undefined behavior), with no compiler diagnostic. None of the current call sites do this (all pass src and the six bounds together, or omit both), so there's no live bug. Worth either tightening the guard (also assert present(x_beg)) or documenting that src and the bounds are all-or-nothing, to protect future callers.

[sbryngelson]

remove empty line