revdep DramaAnalysis failure after as.data.frame redirect

[tdhock]

revdep check machine says #5700 is causing new issues with DramaAnalysis checks

* checking examples ... ERROR
Running examples in 'DramaAnalysis-Ex.R' failed
The error most likely occurred in:

> base::assign(".ptime", proc.time(), pos = "CheckExEnv")
> ### Name: dictionaryStatistics
> ### Title: Dictionary Use
> ### Aliases: dictionaryStatistics dictionaryStatisticsSingle
> ###   as.matrix.QDDictionaryStatistics
> 
> ### ** Examples
> 
> # Check multiple dictionary entries
> data(rksp.0)
> dstat <- dictionaryStatistics(rksp.0, fieldnames=c("Krieg","Familie"))
Error in `<current-expression>` : node stack overflow
Calls: dictionaryStatistics ... as.data.table -> as.data.frame -> as.data.frame.data.frame
Execution halted
... 
* checking tests ...
  Running 'testthat.R'
 ERROR
Running the tests in 'tests/testthat.R' failed.
Complete output:
  > library(testthat)
  > library(DramaAnalysis)
  DramaAnalysis 3.0.2
  > 
  > test_check("DramaAnalysis")
  [ FAIL 2 | WARN 84 | SKIP 1 | PASS 255 ]
  
  == Skipped tests (1) ===========================================================
  * On CRAN (1): 'test-loading.R:6:3'
  
  == Failed tests ================================================================
  -- Error: (code run outside of `test_that()`) ----------------------------------
  <nodeStackOverflowError/stackOverflowError/error/condition>
  Error: node stack overflow
  -- Error: (code run outside of `test_that()`) ----------------------------------
  <nodeStackOverflowError/stackOverflowError/error/condition>
  Error: node stack overflow
  
  [ FAIL 2 | WARN 84 | SKIP 1 | PASS 255 ]
  Error: Test failures

@MichaelChirico @jangorecki can you please investigate?

[MichaelChirico]

I think it's something of a downstream bug -- they wind up with a duplicate-classed object here:

https://github.com/quadrama/DramaAnalysis/blob/bb7357b1afcf4d6b11ffd0cc6ac120ed5a9751d4/tests/testthat/test-dictionary-statistics.R#L50

Debugging as.data.table.data.frame() finds:

class(x)
# [1] "QDDictionaryStatistics" "QDHasCharacter"         "QDByDrama"             
# [4] "data.frame"             "data.table"             "data.frame"            
# [7] "QDByCharacter" 

That suggests a reprex:

x = copy(mtcars)
setattr(x, "class", c("data.frame", "data.table", "data.frame"))
as.data.table(x)
# Error: C stack usage  7974516 is too close to the limit

Filed quadrama/DramaAnalysis#173 downstream, and also confirmed that the duplicate class is also present on 1.16.4.

[MichaelChirico]

Actually, we don't need an example as quirky as "duplicate class", we just need the classes reversed:

x = mtcars
class(x) = c("data.frame", "data.table")
as.data.table(x) # on master
# Error: C stack usage  7969844 is too close to the limit

[MichaelChirico]

This no longer appears, e.g. as of Jan 22:

https://rcdata.nau.edu/genomic-ml/data.table-revdeps/analyze/2025-01-22/