Alternative fix to extra allocations in get_gradient! from Jacobian

[mateuszbaran]

Alternative fix to #300 , as discussed on Slack.

[codecov]

Codecov Report

Merging #301 (b3e9c88) into master (48f5ab0) will decrease coverage by 0.15%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #301      +/-   ##
==========================================
- Coverage   99.88%   99.74%   -0.15%     
==========================================
  Files          77       77              
  Lines        7083     7087       +4     
==========================================
- Hits         7075     7069       -6     
- Misses          8       18      +10     

Files	Coverage Δ
src/plans/nonlinear_least_squares_plan.jl	100.00% <100.00%> (ø)
src/solvers/LevenbergMarquardt.jl	100.00% <100.00%> (ø)

... and 2 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[kellertuer]

I fixed the docs, but I am not yet sure what is missing in code coverage by now.

[mateuszbaran]

Thanks for adding docstring to docs! I've added a couple of new tests that may help with coverage loss but some of it looks unrelated to this PR.

[kellertuer]

Then you have to fix them as well ;)

No the ±10 lines is still something that bothers me since sometimes one of the Wolfe cases is hit, sometimes not and I am not able to consistently provide a test that always hits that. Wolfe is not easy. So 18 lines is still fine due to this.

[mateuszbaran]

OK, I will ignore those 10 lines.

[kellertuer]

This looks fine to me and we can surely merge that :)

[mateuszbaran]

OK, I will bump version then.

[Affie]

Thanks for fixing, this takes a big step in making Manopt usable to us.

[mateuszbaran]

You're welcome 🙂 . Let me know if you encounter any other issues with Manifolds.jl or Manopt.jl.