Comparing single base DNAString with single character causes stack error

[Shians]

Under Biostring 2.58.0 and R4.0.5 on both macOS and CentOS.

> Biostrings::DNAString("A") == "A"
Error: C stack usage  7969280 is too close to the limit

Preconverting either side of the comparison works fine

> as.character(Biostrings::DNAString("A")) == "A"
[1] TRUE
> Biostrings::DNAString("A") == Biostrings::DNAString("A")
[1] TRUE

[ahl27]

Sorry this took so long to follow up on...but thanks for identifying what ended up being a very interesting bug!

This isn't restricted to single character comparisons, it just appears so in Biostrings:

x <- XVector(10)
y <- sample(10)
x == y
## Error: C stack usage  7953792 is too close to the limit

This actually happens for any comparison of XVector objects...

XInteger(10) == XInteger(10)
## FALSE (could be TRUE idk)
XInteger(10) < XInteger(10)
## Error: C stack usage  7953792 is too close to the limit

It turns out that the internal comparison functions rely on the order function, but this unfortunately isn't defined for XVector objects. Insteadm it falls back to the default value of base::order. This obviously isn't a good idea for an S4 object, and it ends up crashing:

x <- as(sample(10), "XVector")
order(x)
## Error: C stack usage  7953792 is too close to the limit

selectMethod("order", "SharedVector")
## ...lots of output
## <environment: namespace:base>
## 
## Signatures:
##         ...           
## target  "SharedVector"
## defined "ANY"

The solution is we have to define an order method for SharedVector objects, which will live in the XVector package. I'm working on that now, will ping this thread when it's done.

[ahl27]

Fixed in Bioconductor/S4Vectors#127 and Bioconductor/XVector#6, still have some minor edits to do to them but should hopefully have this resolved in the near future. Sorry again for the slow response, and thank you for bringing this to our attention! Ended up being a much deeper bug than I expected 😅

[hpages]

Finally fixed in Biostrings 2.79.2. Thanks @ahl27 for #130