Fatal Error when running viewPLP() on Windows

[vertfreeber]

Describe the bug
Executing viewPLP() leads to a fatal Error in Rstudio after hanging for a while.

Additional context

Using RStudio on Windows 10. I noticed that in the new Version of PLP (starting from 6.0.0) the Method in viewShinyPlp.r contains the following line(l.133).

Sys.setenv("plpDatabaseSettings" = as.character(ParallelLogger::convertSettingsToJson(databaseSettings)))

In my case the String of the json of my databasesettings is 65 MB(!) big. The main problem here could be that Windows has a character limit on its system variables (See: https://www.rdocumentation.org/packages/base/versions/3.6.2/topics/Sys.setenv under details) Which I think is the reason my rstudio crashes.

Could this be the case? And is there a workaround?

Set up (please run in R "sessionInfo()" and copy the output here):`
R version 4.2.1 (2022-06-23 ucrt)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19043)

Matrix products: default

locale:
[1] LC_COLLATE=German_Germany.utf8 LC_CTYPE=German_Germany.utf8 LC_MONETARY=German_Germany.utf8 LC_NUMERIC=C
[5] LC_TIME=German_Germany.utf8

attached base packages:
[1] stats graphics grDevices datasets utils methods base

other attached packages:
[1] PatientLevelPrediction_6.0.4 FeatureExtraction_3.2.0 Andromeda_0.6.1 dplyr_1.0.10
[5] DatabaseConnector_5.1.0

loaded via a namespace (and not attached):
[1] zip_2.2.1 Rcpp_1.0.9 pillar_1.8.1 compiler_4.2.1 prettyunits_1.1.1 remotes_2.4.2
[7] tools_4.2.1 pkgbuild_1.3.1 bit_4.0.4 jsonlite_1.8.3 RSQLite_2.2.18 memoise_2.0.1
[13] lifecycle_1.0.3 tibble_3.1.8 pkgconfig_2.0.3 rlang_1.0.6 DBI_1.1.3 cli_3.4.1
[19] rstudioapi_0.14 curl_4.3.3 fastmap_1.1.0 rJava_1.0-6 withr_2.5.0 generics_0.1.3
[25] vctrs_0.5.0 hms_1.1.2 rprojroot_2.0.3 bit64_4.0.5 tidyselect_1.2.0 glue_1.6.2
[31] R6_2.5.1 processx_3.7.0 fansi_1.0.3 ParallelLogger_3.0.1 callr_3.7.2 SqlRender_1.10.0
[37] blob_1.2.3 magrittr_2.0.3 ps_1.7.1 ellipsis_0.3.2 assertthat_0.2.1 renv_0.15.5
[43] utf8_1.2.2 cachem_1.0.6 crayon_1.5.2

To Reproduce

plpResult <- loadPlpResult(file.path(getwd(),"singlePlp", "singledemo", "plpResult"))
viewPlp(plpResult)

PLP Log File
Problem is not logged.

[jreps]

That's super odd as this is the code:
server <- insertRunPlpToSqlite(
runPlp = runPlp,
externalValidatePlp = validatePlp,
diagnosePlp = diagnosePlp
)

connectionDetails <- DatabaseConnector::createConnectionDetails(
dbms = 'sqlite',
server = server
)

databaseSettings <- list(
connectionDetails = connectionDetails,
schema = 'main',
tablePrefix = '',
dbms = 'sqlite',
server = server,
user = NULL,
password = NULL,
port = NULL
)

where server should just be a string specifying where the sqlite location is. databaseSettings should be fairly small, definitely not 65mb. Is it possible to see what's making it such a large character?

[jreps]

Also, what version of ParallelLogger do you have? Perhaps that function being called is doing something to make the list bigger?

[jreps]

I just tried locally on my Mac and I see ParallelLogger::convertSettingsToJson is struggling I think due to the connection details object containing environments. I'm not sure why I've not noticed this issue in the past. Let me have a play around to see how to bypass this.

[jreps]

I've made a fix and just running the package check now. Once it passes, I'll pull in my local changes to the develop branch and then release it.

[vertfreeber]

Thank you for the quick response! I've already tried the changes, viewPLP runs perfectly on my system now :)

Thanks!