[R] write_parquet() has infinite recursion error when writing packageVersion() attributes

[tanho63]

Describe the bug, including details regarding any error messages, version, and platform.

Reprex:

x <- mtcars
attr(x, "arrow_version") <- packageVersion("arrow")
arrow::write_parquet(x, "x.parquet")
# Error: C stack usage  7974388 is too close to the limit
attr(x, "arrow_version") <- as.character(packageVersion("arrow"))
arrow::write_parquet(x, "x.parquet")
# works fine

Session Info:

R> sessioninfo::session_info()
─ Session info ─────────────────────────────────────────────────────────────────────────
 setting  value
 version  R version 4.3.2 (2023-10-31)
 os       Ubuntu 22.04.4 LTS
 system   x86_64, linux-gnu
 ui       RStudio
 language (EN)
 collate  en_CA.UTF-8
 ctype    en_CA.UTF-8
 tz       America/Toronto
 date     2024-08-19
 rstudio  2024.04.0+735 Chocolate Cosmos (desktop)
 pandoc   2.9.2.1 @ /usr/bin/pandoc

─ Packages ─────────────────────────────────────────────────────────────────────────────
 package     * version    date (UTC) lib source
 arrow         17.0.0     2024-08-17 [1] RSPM
 assertthat    0.2.1      2019-03-21 [1] RSPM
 bit           4.0.5      2022-11-15 [1] RSPM
 bit64         4.0.5      2020-08-30 [1] RSPM
 cachem        1.1.0      2024-05-16 [1] RSPM
 callr         3.7.3      2022-11-02 [1] RSPM
 cli           3.6.3      2024-06-21 [1] RSPM
 crayon        1.5.2      2022-09-29 [1] RSPM
 devtools    * 2.4.5      2022-10-11 [1] RSPM
 digest        0.6.33     2023-07-07 [1] RSPM
 ellipsis      0.3.2      2021-04-29 [1] RSPM
 fastmap       1.2.0      2024-05-15 [1] RSPM
 fs            1.6.4      2024-04-25 [1] RSPM
 glue          1.7.0      2024-01-09 [1] RSPM
 htmltools     0.5.7.9000 2024-03-23 [1] Github (rstudio/htmltools@30d13a1)
 htmlwidgets   1.6.3      2023-11-22 [1] RSPM
 httpuv        1.6.12     2023-10-23 [1] RSPM
 later         1.3.1      2023-05-02 [1] RSPM
 lifecycle     1.0.4      2023-11-07 [1] RSPM
 magrittr      2.0.3      2022-03-30 [1] RSPM
 memoise       2.0.1      2021-11-26 [1] RSPM
 mime          0.12       2021-09-28 [1] RSPM
 miniUI        0.1.1.1    2018-05-18 [1] RSPM
 pkgbuild      1.4.2      2023-06-26 [1] RSPM
 pkgload       1.3.3      2023-09-22 [1] RSPM
 prettyunits   1.2.0      2023-09-24 [1] RSPM
 processx      3.8.2      2023-06-30 [1] RSPM
 profvis       0.3.8      2023-05-02 [1] RSPM
 promises      1.2.1      2023-08-10 [1] RSPM
 ps            1.7.5      2023-04-18 [1] RSPM
 purrr         1.0.2      2023-08-10 [1] RSPM
 R6            2.5.1      2021-08-19 [1] RSPM
 Rcpp          1.0.13     2024-07-17 [1] RSPM
 remotes       2.4.2.1    2023-07-18 [1] RSPM
 rlang         1.1.4      2024-06-04 [1] RSPM
 rstudioapi    0.15.0     2023-07-07 [1] RSPM
 sessioninfo   1.2.2      2021-12-06 [1] RSPM
 shiny         1.8.0      2023-11-17 [1] RSPM
 stringi       1.8.2      2023-11-23 [1] RSPM
 stringr       1.5.1      2023-11-14 [1] RSPM
 tidyselect    1.2.1      2024-03-11 [1] RSPM
 urlchecker    1.0.1      2021-11-30 [1] RSPM
 usethis     * 2.2.2      2023-07-06 [1] RSPM
 vctrs         0.6.5      2023-12-01 [1] RSPM
 xtable        1.8-4      2019-04-21 [1] RSPM

 [1] /home/tan/R/x86_64-pc-linux-gnu-library/4.3
 [2] /usr/local/lib/R/site-library
 [3] /usr/lib/R/site-library
 [4] /usr/lib/R/library

A very strange error I suspect is related to #41969 ?

Component(s)

R

[nealrichardson]

Thanks for the report. This is bizarre behavior of packageVersion: it seems to be an infinitely recursive object.

> x <- packageVersion("arrow")
> typeof(x)
[1] "list"
> lapply(x, typeof)
[[1]]
[1] "list"

> lapply(x[[1]], typeof)
[[1]]
[1] "list"

> lapply(x[[1]][[1]], typeof)
[[1]]
[1] "list"

> x[[1]]
[1] ‘17.0.0.1’
> x[[1]][[1]]
[1] ‘17.0.0.1’
> x[[1]][[1]][[1]]
[1] ‘17.0.0.1’
> x[[1]][[1]][[1]][[1]]
[1] ‘17.0.0.1’

It's right there in the function definition: it basically does list(unclass(x)[[1]]), so it always returns a list

> `[[.numeric_version`
function (x, ..., exact = NA) 
{
    if (...length() < 2L) 
        structure(list(unclass(x)[[..., exact = exact]]), class = oldClass(x))
    else unclass(x)[[..1, exact = exact]][..2]
}
<bytecode: 0x1109cf1c8>
<environment: namespace:base>

I'll work around this, but this feels very wrong.