Tests 2127.58 and 2127.59 failed under Rdevel-strict-clang with gctorture

[mattdowle]

Follow up to #4628
See update I edited into the its top comment.
Unlikely to be latest R-devel per se because 1.13.0 passes all normal level checks with latest R-devel, both my me and by CRAN.
It doesn't seem like a gctorture problem on first glance. Maybe --enable-strict-barrier enables extra strict checks. Or it could be something in my local environment since this is a bespoke build and a bespoke library on my laptop.

• update CRAN_Release.cmd comment on the strict torture test to state it takes 74 hours to run.

That increase from 18 hours seems to me more likely due to changes in R; e.g. non-macro INTEGER, REAL could induce a gc inside them under torture. When I was watching it progress, it was slower on the long standing tests up to test 1,000, so it didn't feel like it was just because we have more tests now, although that's possible too as we do add new tests into old numbers.

[mattdowle]

I reran with Rdevel-strict-gcc under torture and all tests passed, taking 55 hours. Could be a difference between gcc and clang, or maybe it has become fixed in the meantime either in data.table-dev or in R-devel.
Will rerun with Rdevel-strict-clang.

[mattdowle]

Reran test 2127 in Rdevel-strict-clang with no problem under torture. Now rerunning all of tests.Rraw under torture which is expected to take 3 days: https://github.com/Rdatatable/data.table/blob/master/.dev/CRAN_Release.cmd#L365.

[mattdowle]

Rdevel-strict-clang torture rerun complete with no errors.

$ Rdevel-strict-clang

R Under development (unstable) (2020-07-14 r78854) -- "Unsuffered Consequences"
Copyright (C) 2020 The R Foundation for Statistical Computing
Platform: x86_64-pc-linux-gnu (64-bit)

> require(data.table)
Loading required package: data.table
data.table 1.13.1 IN DEVELOPMENT built 2020-10-13 05:55:42.470 UTC; mdowle using 1 threads (see ?getDTthreads).  Latest news: r-datatable.com
**********
This installation of data.table has not detected OpenMP support. It should still work but in single-threaded mode.
This is Linux. This warning should not normally occur on Windows or Linux where OpenMP is turned on by data.table's configure script by passing -fopenmp to the compiler. If you see this warning on Windows or Linux, please file a GitHub issue.
**********
>
> gctorture2(step=100)
[1] 0
> print(Sys.time()); started.at<-proc.time(); try(test.data.table()); print(Sys.time()); print(timetaken(started.at))
[1] "2020-10-13 00:16:28.670614 MDT"
getDTthreads(verbose=TRUE):
This installation of data.table has not been compiled with OpenMP support.
  omp_get_num_procs()            1
  R_DATATABLE_NUM_PROCS_PERCENT  unset (default 50)
  R_DATATABLE_NUM_THREADS        unset
  R_DATATABLE_THROTTLE           unset (default 1024)
  omp_get_thread_limit()         1
  omp_get_max_threads()          1
  OMP_THREAD_LIMIT               unset
  OMP_NUM_THREADS                unset
  RestoreAfterFork               true
  data.table is using 1 threads with throttle==1024. See ?setDTthreads.
test.data.table() running: /home/mdowle/build/R-devel-strict-clang/library/data.table/tests/tests.Rraw 

**** Full long double accuracy is not available. Tests using this will be skipped.

Running test id 1759      Test 1760 not run because this session either has no OpenMP or has been limited to one thread (e.g. under UBSAN and ASAN)
Running test id 2155.4         
Fri Oct 16 00:59:26 2020  endian==little, sizeof(long double)==0, longdouble.digits==, sizeof(pointer)==8, TZ==unset, Sys.timezone()=='America/Denver', Sys.getlocale()=='LC_CTYPE=en_US.UTF-8;LC_NUMERIC=C;LC_TIME=en_US.UTF-8;LC_COLLATE=en_US.UTF-8;LC_MONETARY=en_US.UTF-8;LC_MESSAGES=en_US.UTF-8;LC_PAPER=en_US.UTF-8;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=en_US.UTF-8;LC_IDENTIFICATION=C', l10n_info()=='MBCS=TRUE; UTF-8=TRUE; Latin-1=FALSE', getDTthreads()=='This installation of data.table has not been compiled with OpenMP support.; omp_get_num_procs()==1; R_DATATABLE_NUM_PROCS_PERCENT==unset (default 50); R_DATATABLE_NUM_THREADS==unset; R_DATATABLE_THROTTLE==unset (default 1024); omp_get_thread_limit()==1; omp_get_max_threads()==1; OMP_THREAD_LIMIT==unset; OMP_NUM_THREADS==unset; RestoreAfterFork==true; data.table is using 1 threads with throttle==1024. See ?setDTthreads.'
10 longest running tests took 131704s (50% of 261770s)
      ID      time nTest
 1: 1438 23752.501   738
 2: 1648 17780.170    91
 3: 1650 17153.054    91
 4: 1652 17082.302    91
 5: 1848 11594.922     2
 6: 1644 10436.518    91
 7: 1646 10356.955    91
 8: 1642 10356.281    91
 9: 1912  7445.288     2
10: 1639  5746.914   143
All 9983 tests in tests/tests.Rraw completed ok in 72:42:50 elapsed (72:42:10 cpu)
[1] "2020-10-16 00:59:26.88772 MDT"
[1] "72:42:55 elapsed (72:42:16 cpu)"