Some lessons learned:

• Parenthesis matter. They can prevent precision loss and overflow if done correctly.
• lda is the "leading dimension" and used to index matrices, not n/m (actual dimensions)
• Documentation must be taken care of properly
• Add bounds/length checks not present in reference

As for the test, I need to generate not only the general matrix but also ipiv and jpiv. I'll take a look at other tests which do this i.e. dlaqp2.go and dlaqps.go

As for the test, I need to generate not only the general matrix but also ipiv and jpiv.

It will be much easier (and more correct) if you first implement Dgetc2 and use it in the test to generate input for Dgesc2.

It will be much easier (and more correct) if you first implement Dgetc2 and use it in the test to generate input for Dgesc2.

should I do it in this PR? should I start another?

Dgetc2 should be a separate PR.

Codecov Report

Merging #1652 (dc7652f) into master (0164be0) will increase coverage by 0.01%.
The diff coverage is 39.39%.

@@            Coverage Diff             @@
##           master    #1652      +/-   ##
==========================================
+ Coverage   73.24%   73.25%   +0.01%     
==========================================
  Files         506      507       +1     
  Lines       59955    60006      +51     
==========================================
+ Hits        43913    43958      +45     
- Misses      13556    13562       +6     
  Partials     2486     2486              

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0164be0...dc7652f. Read the comment docs.

Dgesc2 is passing albeit without the test you mentioned. I tried programming it and ended up leaving the t.Error declaration commented because it was failing although X was being calculated correctly.

pinging. Needed for #1658 and #1669.

worth mentioning #1665 is ready for merge too.

@vladimir-ch Thanks for taking a look!