CI improvements

[jangorecki]

This PR does not

• include heavy jobs to be run on schedule: add rchk and/or torture CI jobs #3618
• make separate pipelines for artifacts/heavy jobs

Both are blocked on GLCI ability to run multiple pipelines for same branch.

This PR

• adds new Windows job "oldrel" that runs R 3.6, produce and publish binaries
• prepare Check table to support proving information from scheduled jobs (different version, commit sha)
• improves consistency to CRAN about quality/quantity of information provided (thanks @sindribaldur)
• updates CI yaml for newer, more robust syntax

---

Also removes plm from our extended suggests other.Rraw as it require new OS dependency which we don't want, closes #4781.
Also closes #4780 by not publishing extended suggests package to final R repo on website.

[codecov]

Codecov Report

Merging #4768 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #4768   +/-   ##
=======================================
  Coverage   99.47%   99.47%           
=======================================
  Files          73       73           
  Lines       14556    14556           
=======================================
  Hits        14480    14480           
  Misses         76       76           

---

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 63632e6...f6affdd. Read the comment docs.

[mattdowle]

Have seen the failed appveyor-devel job is due to the -alike change in latest R-devel, which is fixed in master by #4770. So proceeding to merge.

[jangorecki]

New syntax for defining dependencies introduced in this PR gave us new DAG graph: https://gitlab.com/Rdatatable/data.table/-/pipelines/210901332/dag
That more precisely visualize dependencies between jobs.
It also let us to download a little less artifacts between different jobs. For deploy docker build jobs we didn't need website artifacts from integration stage but we do still need to wait for this stage to finish.

[jangorecki]

Another nice thing is that check summary table provides now more details:
https://rdatatable.gitlab.io/data.table/web/checks/check_results_data.table.html

[AdrianAntico]

@jangorecki

Hey guys, not sure if this recent update is causing the warning below or if my machine is acting weird. Nonetheless, I started receiving a warning today when running library(data.table). Code info is below:

> library(data.table)
data.table 1.13.2 using 32 threads (see ?getDTthreads).  Latest news: r-datatable.com
Warning message:
package ‘data.table’ was built under R version 4.0.3

> sessionInfo()
R version 4.0.2 (2020-06-22)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 18363)

Matrix products: default

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] data.table_1.13.2

loaded via a namespace (and not attached):
[1] compiler_4.0.2 tools_4.0.2   

[jangorecki]

@AdrianAntico This upgrade is not a cause of warning you are getting.
First of all you are using 1.13.2 which must have been installed from CRAN therefore nothing related to our CI pipeline which releases only odd version numbers.
AFAIU warning is caused by the fact that CRAN ships data.table built on r-release (4.0.3) and you use previous version 4.0.2. If you want to get rid of the warning, I think install.packages("data.table", type="source") should be sufficient. Alternatively you can also upgrade your R version to 4.0.3, that will also resolve the warning.

[AdrianAntico]

@jangorecki Thanks for getting back to me. I just reinstalled with type="source" and I'm no longer seeing the warning.