Cryptic error message from subsetByOverlaps, maybe related to recent update to extractROWS?

[MalteThodberg]

My CAGEfightR package has recently produced a cryptic error message on the Bioconductor servers, that I cannot myself reproduce:

Running examples in ‘CAGEfightR-Ex.R’ failed
The error most likely occurred in:

> base::assign(".ptime", proc.time(), pos = "CheckExEnv")
> ### Name: trackCTSS
> ### Title: Create Genome Browser track of CTSSs.
> ### Aliases: trackCTSS trackCTSS,GRanges-method
> ###   trackCTSS,RangedSummarizedExperiment-method trackCTSS,GPos-method
> 
> ### ** Examples
> 
> library(Gviz)
Loading required package: grid
> data(exampleCTSSs)
> data(exampleUnidirectional)
> data(exampleBidirectional)
> 
> # Example uni- and bidirectional clusters
> TC <- rowRanges(subset(exampleUnidirectional, width>=100)[3,])
> BC <- rowRanges(exampleBidirectional[3,])
> 
> # Create pooled trakc
> subsetOfCTSSs <- subsetByOverlaps(exampleCTSSs, c(BC, TC, ignore.mcols=TRUE))
Error in h(simpleError(msg, call)) : 
  error in evaluating the argument 'x' in selecting a method for function 'extractROWS': no slot of name "NAMES" for this object of class "IPos"
Calls: subsetByOverlaps ... FUN -> extractROWS -> slot -> .handleSimpleError -> h
Execution halted

I haven't changed any code in CAGEfightR, so I assume it must be some upstream changes to the core Bioconductor packages. After looking around, there was recent change to the the extractROWS generic in this commit: 16f29b6, which is the same function throwing the error.

I am unable to determine what is going on with that commit, but it seems to remove a call to updateObject. The StitchedGPos I'm trying to subset should be up to date (GenomicRanges:::.get_GPos_version returns "current").

[LiNk-NY]

Hi Malte, @MalteThodberg

I attempted to reproduce the error on bioconductor/bioconductor_docker:devel but I wasn't able.
It seems that we may have old serialized instances of these objects on the BBS.
Hervé @hpages will comment with further insight.

Best,
Marcel

[MalteThodberg]

Now the error message mysterioulsy disappeared before I had the change to experiment much, I only managed to push a single update to circumvent the SummarizedExperiment: subsetOfCTSSs <- subsetByOverlaps(rowRanges(exampleCTSSs), c(BC, TC, ignore.mcols=TRUE))

Now I'm unsure whether that is actually what fixed the issue...