[Bug] No kernel persistance when switching between Quarto files

[mburtin]

Hi,

I’m not sure if there’s already an open issue on this (it looks releated to (#9595). When I tried using Quarto with Positron, I noticed that the “Kernel Action” doesn’t persist between documents.

Tests to perform:

• Open a Quarto document (Normally, it indicates that no kernel has been started but it's another issue)
• Click on ‘Reload kernel’
• Open a new Quarto document and note that the kernel action indicates that no kernel is active
• Also if you come back on the first document, it's indicated as none kernel started

Positron information:

Positron Version: 2026.04.0 (user setup) build 202
Code - OSS Version: 1.109.2
Commit: a5e20e179301ed1fbccd1691527c015000c95d10
Date: 2026-03-23T08:13:31.139Z
Electron: 39.3.0
Chromium: 142.0.7444.265
Node.js: 22.21.1
V8: 14.2.231.22-electron.0
OS: Windows_NT x64 10.0.22631

Misc note:
I often report bugs, but I’d like to thank you for all your work. I feel that once Positron is more stable and the main features are finalised, it will become my new favourite editor. I'm already starting to get the hang of it, although sometimes it's easier to use RStudio :)

[jmcphers]

Normally, it indicates that no kernel has been started

Are you using the Quarto inline output feature? We do plan to auto-start kernels in the future: #12616

I was not able to reproduce this in today's daily build -- maybe I've got a step wrong. What gesture/menu/button are you using in step 2? (is it "restart"?)

Positron Version: 2026.04.0 build 207
Code - OSS Version: 1.109.2
Commit: b76fc781b3a943d3674c502057db3d869506eb33
Date: 2026-03-24T07:42:32.864Z
Electron: 39.3.0
Chromium: 142.0.7444.265
Node.js: 22.21.1
V8: 14.2.231.22-electron.0
OS: Darwin arm64 25.3.0

[mburtin]

Here a demo:

Screen.Recording.2026-03-25.at.16.01.42.mov

[juliasilge]

Thank you for sharing that video! I am not able to reproduce that behavior when switching between .qmd files so far; when I switch back to a .qmd it shows the kernel running.

Do you also use non-Homebrew managed R? Do you also observe this category of problems with R installed in a more traditional way? I wonder if this problem is related to Homebrew.