Error: segfault from C stack overflow

[bhive01]

I have a data analysis pipeline who's last step is knitting an HTML report with all of the plots I've made. I've been running the latest tidyverse from github because I like across() and where(), but one dataset has routinely been giving me the error in the title and no other information.
I don't know how to debug this error or dig into the code to understand what function is causing the error. I've narrowed it down to the one file, but the minimal example I have is not minimal in any way and the data I have is proprietary so I've put it here for now... 😬 😬 😬
https://drive.google.com/file/d/1lhz7kOTZTKT7Dt0D5bcmbwSldiOLwe6F/view?usp=sharing

I've posted my session info here because it's huge:
https://gist.github.com/bhive01/86f54f459b287784274c16f0586f08bc
My tibble looks like this:

> moa
# A tibble: 3 x 16
  file                        check_map       SPECIES            COMNM     LCCD  TRLID   MATNM     data              TRT_order bar_color corr_df          pca_title                          PCAs             corr_mat corr_plot var_data        
  <chr>                       <list>          <chr>              <chr>     <chr> <chr>   <chr>     <list>            <list>    <list>    <list>           <chr>                              <list>           <list>   <list>    <list>          
1 Strawberry (NZS1802)_Maste… <tibble [5 × 5… Fragraria x anana… Strawber… NZPN  NZS1802 Albion    <tibble [1,483 ×… <chr [5]> <chr [5]> <tibble [60 × 4… NZPN   Strawberry - NZS1802   Alb… <tibble [120 × … <rcorr>  <gg>      <tibble [478 × …
2 Strawberry (NZS1802)_Maste… <tibble [5 × 5… Fragraria x anana… Strawber… NZPN  NZS1802 Camarosa  <tibble [131 × 1… <chr [5]> <chr [5]> <tibble [17 × 1… NZPN   Strawberry - NZS1802   Cam… <tibble [48 × 2… <rcorr>  <gg>      <tibble [189 × …
3 Strawberry (NZS1802)_Maste… <tibble [5 × 5… Fragraria x anana… Strawber… NZPN  NZS1802 San Andr… <tibble [1,233 ×… <chr [5]> <chr [5]> <tibble [50 × 5… NZPN   Strawberry - NZS1802   San… <tibble [130 × … <rcorr>  <gg>      <tibble [518 × …

This tibble is processed by the following functions moa %>% render_html_reports(.), which call these functions.

render_one <- function(df, COMNM, LCCD, TRLID) {
	rmarkdown::render(
		here('PerTrialReportDev.Rmd'),
		output_file = here("DEVoutputs", glue::glue("{COMNM}"), glue::glue("{LCCD}"), glue::glue("{TRLID}"), glue::glue("{LCCD}_{TRLID}_Experiment_Data_Summary.html")),
		params = list(TRLID = TRLID, input_data = df, trial_description = glue::glue("{LCCD}_{COMNM}-{TRLID}"))
	)
}

render_html_reports <- function(df) {
	df %>% 
		nest(trl_data = -c(SPECIES, COMNM, LCCD, TRLID))  %>%
		{pwalk(list(.$trl_data, .$COMNM, .$LCCD, .$TRLID), ~render_one(..1, ..2, ..3, ..4))}
			
	return(df)
}

The Rmd file is here:
https://gist.github.com/bhive01/6f68c3301dbfb75ba0063db2abbc7f64

The trial that I'm running has the most visualizations out of all, but I'm watching the RAM on my machine and it's not running out of RAM (R using 3 of 32 GB). I'm on a newish Mac Mini.

[bhive01]

An update.
Why is this here? At the time the seg fault only popped up once I updated to the latest vctrs and dplyr.
I've been trying to get at this all day and have had no success. I got rid of all of the development stuff except 2 non-CRAN packages ggtext and ggbiplot. My "new" non-development session_info() is here: https://gist.github.com/bhive01/d4cd5cf8f889bdc8a4e36554323fea83

I pulled my functions to pieces and even ran the RMD manually and it still crashes when making the outputs for the variable level plots. My hope is that it would happen consistently on a single plot and it would be obvious what the problem is, but that was not the case. I was able to get it to work once, but every time other than once it segfaulted. At different points.

I don't know how to dissect or debug compiled C code that I have no error message to guide me to what is wrong. It could be in knitr or pandoc or...?
Can anyone help guide me on how to break it down more?

[romainfrancois]

There isn't much we can offer with the available information.

Have you seen Jenny's rstudioconf talk ? perhaps it can help narrow it down some more.
https://rstudio.com/resources/rstudioconf-2020/object-of-type-closure-is-not-subsettable/

I understand this is frustrating but I honestly have no idea where to start.

[bhive01]

No problem Romain. I'm still digging and I am getting somewhere. It's just a very slow process. I'm used to a user-caused error where rlang spits back something that I can pick at and debug, but to just have a C segfault and no other information has made peeling the onion really hard.
I'm going to close this so it's not on your radar until I can get a smaller example going.