backport of QR to 1.x

[palday]

from #507

[codecov[bot]]

Codecov Report

Patch coverage: 86.79% and project coverage change: +0.72 :tada:

Comparison is base (6406a70) 87.39% compared to head (cd187fb) 88.11%.

Additional details and impacted files

@@            Coverage Diff             @@
##               v1     #537      +/-   ##
==========================================
+ Coverage   87.39%   88.11%   +0.72%     
==========================================
  Files           7        7              
  Lines         952     1077     +125     
==========================================
+ Hits          832      949     +117     
- Misses        120      128       +8     

Impacted Files	Coverage Δ
src/GLM.jl	60.00% <50.00%> (+10.00%)	:arrow_up:
src/negbinfit.jl	81.81% <66.66%> (-0.85%)	:arrow_down:
src/linpred.jl	88.55% <83.63%> (+2.94%)	:arrow_up:
src/lm.jl	94.07% <97.14%> (+0.74%)	:arrow_up:
src/glmfit.jl	81.33% <100.00%> (+0.31%)	:arrow_up:

... and 1 file with indirect coverage changes

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.

[nalimilan]

Is there any particular place we should look at (i.e. where you had to adapt the original PR)?

[palday]

@nalimilan if you look at the diff excluding the first commit, then you can see things that were non obvious in the cherry-pick -- one exception (not including a local fit(::Type, ::FormulaTerm,...)) method, I accepted all changes from the original PR. The follow-up modifications were related to the additional redirection from TableRegressionModel.