Error in if (latest_R_version == R_version)

[jessekps]

When using create_app with usual arguments and include_R=TRUE I get the error:

Error in if (latest_R_version == R_version) { :
argument is of length zero

I believe the error is in get_R.R on line 23:

latest_R_version <- readLines("https://cran.rstudio.com/bin/windows/base/", warn = F) %>%
    stringr::str_extract("[1-3]\\.[0-9]+\\.[0-9]+") %>% stats::na.omit() %>% unique()

The regular expression should start with [1-4], since R4.+ is now the current R version.

[AmyMikhail]

Coming to say I still have the same issue in January 2023 - but also just wondering why the regular expression could not be dynamic - why isn't the first part of the regular expression just [0-9]? That would last a good while longer...

[jessekps]

And if you don't feel like maintaining that code, you can simply rely on rversions. It seems to be actively maintained by some well known names in the R community.

> rversions::r_release()
version                date              nickname
4.2.2 2022-10-31 08:05:54 Innocent and Trusting

[vekinoir]

I am facing the same error with create_app with argument include_R = TRUE, R_version = "4.0.0"

Error in if (latest_R_version == R_version) { : 
  argument is of length zero

However when I change the 3 to 4 in the latest_R_version in trace(RInno::get_R,edit=T) and trace(RInno::code_section,edit=T)

I am still getting an error

Error in if (latest_R_version == R_version) { : 
  the condition has length > 1

How to fix this?

[matdillen]

latest_R_version is extracted from the html at https://cran.rstudio.com/bin/windows/base/ That page now seems to include a commented reference to 4.3.0 beyond the most up to date build (4.4.0). Hence you get a vector instead of a string and the listed condition throws an error.

You can fix the functions by adding the following code before running them (but after loading rinno):

body(code_section)[[4]][[3]][[2]][[2]][[2]] <- substitute(
  stringr::str_extract(readLines("https://cran.rstudio.com/bin/windows/base/",
                                 warn = F), "[1-4]\\.[0-9]+\\.[0-9]+")
)
body(code_section)[[4]][[3]][[3]][[2]] <- substitute(
  stringr::str_extract(readLines("https://cran.rstudio.com/bin/windows/base/old/", 
                                 warn = F), "[1-4]\\.[0-9]+\\.[0-9]+")
)
body(get_R)[[4]][[3]][[2]][[2]][[3]] <- substitute(
  stringr::str_extract("[1-4]\\.[0-9]+\\.[0-9]+")
)
body(get_R)[[6]][[2]] <- substitute(
  latest_R_version[1] == R_version
)
body(get_R)[[5]][[3]][[2]][[3]] <- substitute(
  stringr::str_extract("[1-4]\\.[0-9]+\\.[0-9]+")
)

Ugly and still at risk of breaking in the future if the cran pages change or R 5 comes out, but it works for now.

[FrancoMSuarez]

@matdillen Could you explain this solution better to me please?

[matdillen]

You can see it implemented for example here: https://github.com/AgentschapPlantentuinMeise/MIDSCalculator/blob/main/src/rinno_installer.R

The above script updates the get_R and code_section functions of rinno in the R session to hotfix bugs with R version numbers. A regex in both functions only goes up to R version 3 and the latest R version extracted from CRAN's website now always includes R-4.3.0 as well because of the (commented) line in the html below. Hence the script gets a vector instead of a string and crashes.

<!--<li>A <a href="rtest.html">pre-release</a> version for the forthcoming R-4.3.0 is available.</li>-->