Steps to reproduce in a freshly installed Windows 10 virtual machine hosted in Linux.
-
Download Emacs 29.1_1 or 29.2_1, see links:
http://ftpmirror.gnu.org/emacs/windows/emacs-29/emacs-29.1_1.zip
http://ftpmirror.gnu.org/emacs/windows/emacs-29/emacs-29.2_1.zip
-
Extract the zip and start Emacs with:
runemacs.exe -Q
-
Evaluate the following expressions:
(require 'package)
(add-to-list 'package-archives '("stable-melpa" . "https://stable.melpa.org/packages/"))
(list-packages)
-
Install ess and poly-R packages from the stable MELPA archive.
-
Evaluate the following expressions:
(let ((default-directory user-emacs-directory))
(normal-top-level-add-subdirs-to-load-path))
(setopt inferior-R-program-name "path to your R/bin/x64/Rterm.exe")
(require 'poly-R)
-
Open a relatively large Rmd document.
-
On an R code chunk, type C-return to start the inferior R buffer.
-
Finally, navigate the document via the Polymode menu with Next/Previous chunk entries.
Don't use key bindings, but the mouse and, in particular, use the same-type chunk entry in the menu.
May be after 10/15 chunk visits, like said above, Emacs crashes. Probably, the test works better if you do not disable the default Windows real-time virus & threat protection settings (active).
Specifically, for version 29.1_1, it just suddenly closes, so there are not many useful details.
For 29.2_1 it gets stuck. If you kill Rterm in Windows Task Manager, functionality seems to be restored, but only at keyboard level. Clicking with the mouse pointer is ineffective. I was unable to reproduce with ordinary R files in r-mode buffers, using the ESS menu eval entries, and similarly in Polymode using the key bindings, rather than the menu bar, I do not get errors.
As it seems, the bug is a mouse thing and is related to comint communication with the R process. The menu somehow queries R to navigate chunks and loses sync with the process, then it remains stuck waiting for an answer.
Steps to reproduce in a freshly installed Windows 10 virtual machine hosted in Linux.
Download Emacs
29.1_1or29.2_1, see links:http://ftpmirror.gnu.org/emacs/windows/emacs-29/emacs-29.1_1.zip
http://ftpmirror.gnu.org/emacs/windows/emacs-29/emacs-29.2_1.zip
Extract the zip and start Emacs with:
runemacs.exe -QEvaluate the following expressions:
Install
essandpoly-Rpackages from the stable MELPA archive.Evaluate the following expressions:
Open a relatively large Rmd document.
On an R code chunk, type
C-returnto start the inferior R buffer.Finally, navigate the document via the Polymode menu with Next/Previous chunk entries.
Don't use key bindings, but the mouse and, in particular, use the same-type chunk entry in the menu.
May be after 10/15 chunk visits, like said above, Emacs crashes. Probably, the test works better if you do not disable the default Windows real-time virus & threat protection settings (active).
Specifically, for version
29.1_1, it just suddenly closes, so there are not many useful details.For
29.2_1it gets stuck. If you kill Rterm in Windows Task Manager, functionality seems to be restored, but only at keyboard level. Clicking with the mouse pointer is ineffective. I was unable to reproduce with ordinary R files inr-modebuffers, using the ESS menu eval entries, and similarly in Polymode using the key bindings, rather than the menu bar, I do not get errors.As it seems, the bug is a mouse thing and is related to comint communication with the R process. The menu somehow queries R to navigate chunks and loses sync with the process, then it remains stuck waiting for an answer.