[Fix]: NewtonRaphsonSolver convergence check and error handling

[Ady0333]

Summary

Fixes #1284

Two related fixes for NewtonRaphsonSolver:

1. Initial convergence check - Actually test if starting residual is below tolerance
2. Error handling - Replace @unreachable with informative error message

---

Changes

In src/Algebra/NLSolvers.jl:

Line 81 - Initial check:

# Before: return (false, m0)
# After:  return (m0 < nls.tol, m0)

Line 67 - Non-convergence handling:

# Before: @unreachable
# After:  error("NewtonRaphsonSolver failed to converge in $(nls.max_nliters) iterations. Final residual: $m")

---

Behavior

Before:

• Starting from converged state → crash with "@unreachable"
• Relative tolerance m < tol*m0 unsatisfiable when m0 ≈ 0

After:

• Already-converged initial state → exits immediately
• Non-convergence → informative error with diagnostics

---

Testing

Tested with Newton starting at exact solution (zero residual):

• Previously: crashed with "@unreachable"
• Now: exits cleanly on first convergence check

All existing Gridap tests pass.

---

Related

Related to #1282 - part of broader Newton solver convergence testing

[Ady0333]

Hello @JordiManyer , @Antoinemarteau and @oriolcg !!! Please review this pr and let me know if any changes are needed...

~Thanks

[codecov]

Codecov Report

❌ Patch coverage is 50.00000% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 88.82%. Comparing base (670f802) to head (a6cb5dd).
⚠️ Report is 5 commits behind head on master.

Files with missing lines	Patch %	Lines
src/Algebra/NLSolvers.jl	50.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1285   +/-   ##
=======================================
  Coverage   88.82%   88.82%           
=======================================
  Files         227      227           
  Lines       29753    29753           
=======================================
  Hits        26429    26429           
  Misses       3324     3324           

Flag	Coverage Δ
drivers	39.35% <0.00%> (ø)
extensions	5.13% <0.00%> (ø)
unit-adaptivity	40.11% <0.00%> (ø)
unit-basics	14.67% <50.00%> (ø)
unit-celldata	21.15% <0.00%> (ø)
unit-fespaces-1	32.34% <0.00%> (ø)
unit-fespaces-2	38.83% <0.00%> (ø)
unit-fields	17.59% <0.00%> (ø)
unit-geometry	28.80% <0.00%> (ø)
unit-multifield	30.86% <0.00%> (-0.12%)	⬇️
unit-odes	28.75% <0.00%> (ø)
unit-referencefes	34.28% <0.00%> (ø)
unit-visualization	11.92% <0.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 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.

[JordiManyer]

Thanks @Ady0333 for the contribution. Changes look good, please update NEWS.md so that I can merge. This PR could have been bundled with the other one, it's addressing the same kind of things.

Also, kindly stop tagging everyone when you open a PR. People have their own lives and our inboxes are already quite full. When you leave a PR, someone will eventually look into it. Unless your PR has been sitting without response for over a week, it is really not necessary.

[Ady0333]

Thanks @Ady0333 for the contribution. Changes look good, please update NEWS.md so that I can merge. This PR could have been bundled with the other one, it's addressing the same kind of things.

Also, kindly stop tagging everyone when you open a PR. People have their own lives and our inboxes are already quite full. When you leave a PR, someone will eventually look into it. Unless your PR has been sitting without response for over a week, it is really not necessary.

Thanks for the review. I have updated the NEWS.md and will make sure not to tag anyone...apologies for that!!!