Environment checks in system progress preperation throws error in Positron

[milesalanmoore]

Hi stan-dev!

I have been trying out the new IDE from Posit, Positron, and have been really loving it. However, when I run a parallelized stan model (by specifying cores in the call to a brms model for example), rstan opens up a port in my browser with the following error message:

Cannot GET /var/folders/lm/4jkt0_jd37j4c_68xmrxp7f00000gn/T//Rtmpp3I1Pz/file5568b9aee29_StanProgress.txt

This behavior is OS agnostic and only applies to the case when a user runs chains in parallel as in:

library(brms)

# This runs without issue
brms::brm(
  mpg ~ hp * cyl, 
  data = mtcars)

# This opens the browser tab
m_mpg = brm(
  mpg ~ hp * cyl, 
  data = mtcars,
chains=4, cores=4)

For reference, see my issue on Positron's repo here.

In it, @juliasilge points out that this behavior is likely a result of rstan's environment checks with Sys.getenv("RSTUDIO"). See the code below, which would not be set to 1 for users working in Positron. Is it possible to generalize these checks a bit to include Positron, @juliasilge provides an example of what she thinks this might look like (linked above).

rstan/rstan/rstan/R/stanmodel-class.R

Line 517 in 9b8d8fe

!identical(Sys.getenv("RSTUDIO"), "1"),

I am happy to take a swing at implementing this if helpful, as well.

[danieltomasz]

Yes, it would be great to have this working on Positron :)