`gt::md()` fails when applied to a blank string in new version v0.11.0 on Quarto

[ddsjoberg]

Prework

• Read and agree to the code of conduct and contributing guidelines.
• If there is already a relevant issue, whether open or closed, comment on the existing thread instead of posting a new issue.

Description

@remlapmot reported an issue to gtsummary after they updated to gt v0.11.0. By default, our headers in gtsummary are processed with gt::md(), and a blank string leads to an error in Quarto. There is no error when run interactively or using R Markdown. I tested in Quarto versions v1.4.549 and v1.5.54 (released a couple of days ago).

For now, I can update gtsummary to remove gt::md() when the string is blank.

Reproducible example

---
title: "tbl_cross reprex"
format:
  html:
    embed-resources: true
---

Example 

```{r}
packageVersion("gt")

mtcars[1:3, 1:2] |> 
  gt::gt() |> 
  gt::cols_label(mpg = gt::md(""))
```

Thank you!!

[remlapmot]

The error message (under Quarto version 1.5.54) in the terminal is as follows

==> quarto preview tbl-cross-reprex-md.qmd --to html --no-watch-inputs --no-browse



processing file: tbl-cross-reprex-md.qmd
  |...................................                 |  67% [unnamed-chunk-1]Error in `vapply()`:
! values must be length 1,
 but FUN(X[[1]]) result is length 0
Backtrace:
  1. global .main()
  2. execute(...)
  3. rmarkdown::render(...)
  4. knitr::knit(knit_input, knit_output, envir = envir, quiet = quiet)
  5. knitr:::process_file(text, output)
     ...
 30. gt:::dt_boxhead_build(data = data, context = context)
 31. base::lapply(...)
 32. gt (local) FUN(X[[i]], ...)
 33. gt:::process_text(label, context)
 34. base::vapply(...)
                                                                                                            


Quitting from lines 11-16 [unnamed-chunk-1] (tbl-cross-reprex-md.qmd)

Execution halted

There's no error with gt version 0.10.1.

[olivroy]

Thanks for the report. The culprit is possibly here #1578. in utils.R (line 600 ish)

[olivroy]

Found the issue charToRaw("") gives character(0). Base R (in)consistency strikes again! It is a shame when the length of something can change without you knowing. This time, at least the problem was created close to the error. But it is not always the case.

[olivroy]

@remlapmot @ddsjoberg I opened #1772 that should solve the issue. Would you mind doing some testing. I don't know if it would preferable to "cancel" md() processing for a single empty string.

md <- function(text) {

  # Apply the `from_markdown` class
if (length(x) == 1 && !nzchar(x, keepNA = FALSE)) {
  return(x)
}
  class(text) <- "from_markdown"
  text
}

[ddsjoberg]

Thanks for the quick fix @olivroy !!! amazing

I tested the example posted above, and the original issue found in gtsummary, and they both appear resovled with the update 🐐

[remlapmot]

Thanks for such a rapid fix.