[Rd] Bug with `[<-.POSIXlt` on specific OSes

[Davis Vaughan]

I've got a bit more information about this one. It seems like it (only? not
sure) appears when `TZ = "UTC"`, which is why I didn't see it before on my
Mac, which defaults to `TZ = ""`. I think this is at least explainable by
the fact that those "optional" fields aren't technically needed when the
time zone is UTC.

I can reproduce this now on my personal Mac:

```

x <- as.POSIXlt(as.POSIXct("2013-01-31", tz = "America/Chicago"))

Sys.setenv(TZ = "")

x[1] <- NA

x

#> [1] NA


x <- as.POSIXlt(as.POSIXct("2013-01-31", tz = "America/Chicago"))

Sys.setenv(TZ = "America/New_York")

x[1] <- NA

x

#> [1] NA


x <- as.POSIXlt(as.POSIXct("2013-01-31", tz = "America/Chicago"))

Sys.setenv(TZ = "UTC")

x[1] <- NA
#> Error in x[[n]][i] <- value[[n]] : replacement has length zero

x

#> [1] "2013-01-31 CST"
```

Here are `sessionInfo()` and `Sys.getenv("TZ")` outputs for 3 GitHub
Actions platforms where the bug exists (note they all set `TZ = "UTC"`!):

Linux:

```

> sessionInfo()

R version 4.2.1 (2022-06-23)

Platform: x86_64-pc-linux-gnu (64-bit)

Running under: Ubuntu 18.04.6 LTS


Matrix products: default

BLAS:   /usr/lib/x86_64-linux-gnu/openblas/libblas.so.3

LAPACK: /usr/lib/x86_64-linux-gnu/libopenblasp-r0.2.20.so


locale:

 [1] LC_CTYPE=C.UTF-8       LC_NUMERIC=C           LC_TIME=C.UTF-8

 [4] LC_COLLATE=C.UTF-8     LC_MONETARY=C.UTF-8    LC_MESSAGES=C.UTF-8

 [7] LC_PAPER=C.UTF-8       LC_NAME=C              LC_ADDRESS=C

[10] LC_TELEPHONE=C         LC_MEASUREMENT=C.UTF-8 LC_IDENTIFICATION=C


attached base packages:

[1] stats     graphics  grDevices utils     datasets  methods   base


loaded via a namespace (and not attached):

[1] compiler_4.2.1


> Sys.getenv("TZ")

[1] "UTC"
```

Mac:

```

> sessionInfo()

R version 4.2.1 (2022-06-23)

Platform: x86_64-apple-darwin17.0 (64-bit)

Running under: macOS Big Sur ... 10.16


Matrix products: default

BLAS:
/Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRblas.0.dylib

LAPACK:
/Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRlapack.dylib


locale:

[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8


attached base packages:

[1] stats     graphics  grDevices utils     datasets  methods   base


loaded via a namespace (and not attached):

[1] compiler_4.2.1


> Sys.getenv("TZ")

[1] "UTC"
```

Windows:
This is the best I can get you, sorry (remote worker issues), but note that
it does also say `tz UTC` like the others.

```
version R version 4.2.1 (2022-06-23 ucrt)
os Windows Server x64 (build 20348)
system x86_64, mingw32
ui RTerm
language (EN)
collate English_United States.utf8
ctype English_United States.utf8
tz UTC
date 2022-10-11
```

And here is my Mac where the bug doesn't show up by default because `TZ =
""`:

```

> sessionInfo()

R version 4.2.1 (2022-06-23)

Platform: x86_64-apple-darwin17.0 (64-bit)

Running under: macOS Big Sur ... 10.16


Matrix products: default

BLAS:
/Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRblas.0.dylib

LAPACK:
/Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRlapack.dylib


locale:

[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8


attached base packages:

[1] stats     graphics  grDevices utils     datasets  methods   base


loaded via a namespace (and not attached):

[1] compiler_4.2.1


> Sys.getenv("TZ")

[1] ""


> Sys.timezone()

[1] "America/New_York"
```

-Davis


On Thu, Oct 6, 2022 at 9:33 AM Davis Vaughan <davis using rstudio.com> wrote:

> Hi all,
>
> I have found another POSIXlt bug while I've been fiddling around with it.
> This one only appears on specific OSes, because it has to do with the fact
> that the `gmtoff` field is optional, and isn't always used on all OSes. It
> also doesn't seem to be specific to r-devel, I think it has been there
> awhile.
>
> Here is the bug:
>
> ```
> x <- as.POSIXlt(as.POSIXct("2013-01-31", tz = "America/Chicago"))
>
> # Oh no!
> x[1] <- NA
> #> Error in x[[n]][i] <- value[[n]] : replacement has length zero
> ```
>
> If you look at the objects, you can see that `x` has a `gmtoff` field, but
> `NA` (when converted to POSIXlt, which is what `[<-.POSIXlt` does) does not:
>
> ```
> unclass(x)
> #> $sec
> #> [1] 0
> #>
> #> $min
> #> [1] 0
> #>
> #> $hour
> #> [1] 0
> #>
> #> $mday
> #> [1] 31
> #>
> #> $mon
> #> [1] 0
> #>
> #> $year
> #> [1] 113
> #>
> #> $wday
> #> [1] 4
> #>
> #> $yday
> #> [1] 30
> #>
> #> $isdst
> #> [1] 0
> #>
> #> $zone
> #> [1] "CST"
> #>
> #> $gmtoff
> #> [1] -21600
> #>
> #> attr(,"tzone")
> #> [1] "America/Chicago" "CST"             "CDT"
>
> unclass(as.POSIXlt(NA))
> #> $sec
> #> [1] NA
> #>
> #> $min
> #> [1] NA
> #>
> #> $hour
> #> [1] NA
> #>
> #> $mday
> #> [1] NA
> #>
> #> $mon
> #> [1] NA
> #>
> #> $year
> #> [1] NA
> #>
> #> $wday
> #> [1] NA
> #>
> #> $yday
> #> [1] NA
> #>
> #> $isdst
> #> [1] -1
> #>
> #> attr(,"tzone")
> #> [1] "UTC"
> ```
>
> The problem seems to be that `[<-.POSIXlt` assumes that if the field was
> there in `x` then it must also be there in `value`:
>
> https://github.com/wch/r-source/blob/e10a971dee6a0ab851279c183cc21954d66b3be4/src/library/base/R/datetime.R#L1303-L1304
>
> But this isn't the case for the `NA` value that was converted to POSIXlt.
>
> I can't reproduce this on my personal Mac, but it affects the Linux, Mac,
> and Windows machines we use for the lubridate CI checks through GitHub
> Actions.
>
> Thanks,
> Davis
>

	[[alternative HTML version deleted]]