file name conversion problem -- name too long?

[analisemacro]

Hi,

I have this basic problem using reticulate. I can´t run any kind of python code from notebook or rmarkdown document. For example,

---

title: "Untitled" output: pdf_document

library(reticulate)

x = 5

Wich gives this error message:

Error in dir.exists(x) : file name conversion problem -- name too long?

I think this a very basic problem, but I can´t solve.

Thanks for any help!

[JiaxiangBU]

I find the same error message.

Error in file.exists(in_file) : 转换文件名时出了问题--名字太长了？
Calls: <Anonymous> ... str_transform_all -> replacement -> processor -> file.exists

I think the problem is the function file.exist within rmarkdown package, I will try a substitute fs::file_exsit.

> packageVersion("rmarkdown")
[1] ‘1.13’

[JiaxiangBU]

I find a solution.

It looks like you have a problem because the path name has non-English characters in it. Stack Overflow

[EspenRosenquist]

I find a solution.

It looks like you have a problem because the path name has non-English characters in it. Stack Overflow

That is not exactly a solution, is it.

I know that it is HORRIBLE with non-english characters in directory names, but it is out of our control as end-users in an organisation. Besides, this is 2022 - when you should think that non-english characters are accepted citizens.

I come from a fairly normal society, but which happens to have æøå. People, including some of those organising file systems in large organisations, like using all letters available in our language. Both Windows and Linux have no problem creating and using directories containing spaces and latin letters with diacritics.

dir.exists fall flat on its face. Yes, the directory does exist, no questions asked. As this function is so deep in the toolchain, a LOT of higher level functions fail. E.g. rmarkdown::render

[kevinushey]

Assuming you're on Windows, R 4.2.x comes with native support for a UTF-8 locale, and so issues of this form should hopefully become much less common.