Finally found the real reason: package `project-rootfile` -> (add-to-list
'project-find-functions #'project-rootfile-try-detect t)
I don't know why project-rootfile involved in the Org mode Babel source
block execution path.
I checked out the source code of `project-rootfile`. Have not found any
clue.
Anyone have thoughts on this weird problem?


[stardiviner]           <Hack this world!>      GPG key ID: 47C32433
IRC(freeenode): stardiviner                     Twitter:  @numbchild
Key fingerprint = 9BAA 92BC CDDD B9EF 3B36  CB99 B8C4 B8E5 47C3 2433
Blog: http://stardiviner.github.io/


On Tue, Jan 14, 2025 at 6:49 PM <sam@crawf.uk> wrote:

> On 2025-01-14 05:29, stardiviner wrote:
>
> I tested using org-mode's `make repro` clean Emacs & org-mode environment
> instance. It works fine. I double checked my Emacs init file about R
> settings. Have not found suspicious config.
>
>
> So to confirm, you *cannot* reproduce the error with a clean config?
>
> Here is my reproduce test.
>
>
> I tried running this, and it seems to work as expected.
>
> The generated plot image "images/hist_rnorm.png" is NOT in
> "~/Org/Programming/R/images/hist_rnorm.png". It's in
> "~/Org/hist_rnorm.png".
>
>
> Some relevant settings that may be altered by your config are
> `ess-startup-directory` and `ess-startup-directory-function`. It might be,
> for example, that one of these is forcibly changing the R process' working
> directory to a project root ("~/Org", in this instance)
>
> By the way, setting `:dir` changes the *working* directory, which you
> might not want to be "images". There is a separate header argument,
> `:output-dir` that can be used for this purpose.
>
> Sam
>