Always re-solve eq in ProximalProjection before first computation of value and jacobian

[dpanici]

Resolves #1831

[github-actions]

Memory benchmark result

|               Test Name                |      %Δ      |    Master (MB)     |      PR (MB)       |    Δ (MB)    |    Time PR (s)     |  Time Master (s)   |
| -------------------------------------- | ------------ | ------------------ | ------------------ | ------------ | ------------------ | ------------------ |
  test_objective_jac_w7x                 |    6.25 %    |     3.843e+03      |     4.084e+03      |    240.34    |       33.94        |       31.23        |
- test_proximal_jac_w7x_with_eq_update   |   19.83 %    |     5.713e+03      |     6.846e+03      |   1133.09    |       160.30       |       53.58        |
  test_proximal_freeb_jac                |   -0.23 %    |     1.320e+04      |     1.317e+04      |    -30.03    |       77.52        |       77.72        |
  test_proximal_freeb_jac_blocked        |   -0.66 %    |     7.666e+03      |     7.615e+03      |    -50.71    |       68.59        |       68.60        |
  test_proximal_freeb_jac_batched        |    0.62 %    |     7.595e+03      |     7.642e+03      |    47.35     |       68.07        |       68.99        |
  test_proximal_jac_ripple               |    2.79 %    |     7.475e+03      |     7.683e+03      |    208.23    |       69.41        |       71.33        |
  test_proximal_jac_ripple_spline        |    0.09 %    |     3.471e+03      |     3.474e+03      |     3.23     |       71.30        |       71.22        |
  test_eq_solve                          |   -3.31 %    |     2.073e+03      |     2.004e+03      |    -68.59    |       125.16       |       125.07       |

For the memory plots, go to the summary of Memory Benchmarks workflow and download the artifact.

[github-actions]

|             benchmark_name             |         dt(%)          |         dt(s)          |        t_new(s)        |        t_old(s)        | 
| -------------------------------------- | ---------------------- | ---------------------- | ---------------------- | ---------------------- |
 test_build_transform_fft_midres         |     +0.22 +/- 2.46     | +1.47e-03 +/- 1.63e-02 |  6.65e-01 +/- 8.3e-03  |  6.63e-01 +/- 1.4e-02  |
 test_build_transform_fft_highres        |     -0.15 +/- 3.87     | -1.36e-03 +/- 3.50e-02 |  9.02e-01 +/- 2.6e-02  |  9.03e-01 +/- 2.4e-02  |
 test_equilibrium_init_lowres            |     -2.88 +/- 1.79     | -1.29e-01 +/- 8.00e-02 |  4.34e+00 +/- 4.9e-02  |  4.46e+00 +/- 6.3e-02  |
 test_objective_compile_atf              |     -3.50 +/- 3.62     | -2.18e-01 +/- 2.25e-01 |  6.00e+00 +/- 1.6e-01  |  6.22e+00 +/- 1.6e-01  |
 test_objective_compute_atf              |     -7.20 +/- 17.63    | -1.59e-04 +/- 3.89e-04 |  2.05e-03 +/- 2.5e-04  |  2.21e-03 +/- 3.0e-04  |
 test_objective_jac_atf                  |     -1.62 +/- 2.26     | -2.80e-02 +/- 3.90e-02 |  1.70e+00 +/- 2.4e-02  |  1.73e+00 +/- 3.0e-02  |
 test_perturb_1                          |     -4.57 +/- 3.00     | -6.48e-01 +/- 4.26e-01 |  1.35e+01 +/- 3.7e-01  |  1.42e+01 +/- 2.1e-01  |
 test_proximal_jac_atf                   |     -0.52 +/- 1.47     | -2.89e-02 +/- 8.23e-02 |  5.55e+00 +/- 6.6e-02  |  5.58e+00 +/- 4.9e-02  |
 test_proximal_freeb_compute             |     -1.24 +/- 2.95     | -2.02e-03 +/- 4.79e-03 |  1.60e-01 +/- 3.2e-03  |  1.62e-01 +/- 3.6e-03  |
 test_solve_fixed_iter                   |     -3.89 +/- 2.33     | -1.15e+00 +/- 6.89e-01 |  2.84e+01 +/- 2.9e-01  |  2.95e+01 +/- 6.2e-01  |
 test_objective_compute_ripple           |     -0.34 +/- 1.04     | -8.93e-03 +/- 2.72e-02 |  2.61e+00 +/- 2.0e-02  |  2.61e+00 +/- 1.9e-02  |
 test_objective_grad_ripple              |     -0.37 +/- 2.48     | -1.76e-02 +/- 1.17e-01 |  4.68e+00 +/- 4.6e-02  |  4.69e+00 +/- 1.1e-01  |

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.81%. Comparing base (7e334c3) to head (7fa9267).
⚠️ Report is 61 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1842      +/-   ##
==========================================
- Coverage   95.81%   95.81%   -0.01%     
==========================================
  Files         100      100              
  Lines       27536    27539       +3     
==========================================
+ Hits        26384    26386       +2     
- Misses       1152     1153       +1     

Files with missing lines	Coverage Δ
desc/optimize/_constraint_wrappers.py	97.31% <100.00%> (+0.02%)	⬆️

... and 3 files with indirect coverage changes

🚀 New features to boost your workflow:

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

[YigitElma]

Hmm, I don't see how we can prevent the failing benchmarks since the kwarg is recently added. If you can merge it by overriding somehow, I am fine with it.

Should update the changelog though