Startup of console hangs the first time starting positron after upgrade

[rmflight]

System details:

Positron and OS details:

Positron Version: 2025.12.2 build 5
Code - OSS Version: 1.106.0
Commit: f5a10d0
Date: 2025-12-11T21:21:46.549Z
Electron: 37.7.0
Chromium: 138.0.7204.251
Node.js: 22.20.0
V8: 13.8.258.32-electron.0
OS: Linux x64 6.17.4-76061704-generic

Session details:

R 4.4.1 (have also seent it happen with R 4.5.2)

Describe the issue:

When I install an updated version of Positron, the very first time I start it, the console session hangs with "Preparing" (as shown in attached screenshot). This will continue for several minutes.

If I close Positron, and then open it again on the same project, the console comes right up. And will continue to come up reliably without issue, until I upgrade Positron again.

Steps to reproduce the issue:

1. Upgrade Positron on Deb based Linux (sudo dpkg -i ...)
2. Open a folder using Positron.

Expected or desired behavior:

Console should come up the same way, regardless if Positron has been updated or not.

I will note, this happens if I change the installed version of Positron, regardless of upgrade or downgrade. The logs below are from downgrading from 2025.12.2 to 2025.12.1, but the same issue persists, and the solution is again to close the application, and restart it.

I thought this was related to console can get stuck starting, but surface level it "feels" different. Maybe.

Were there any error messages in the UI, Output panel, or Developer Tools console?

Text files attached.

positron_developer_tools.txt

positron_output_preparing_hanging.log

[jmcphers]

@rmflight thank you for reporting, and especially for including the logs!

The main difference between the two logs is that, when the system is hanging in "preparing", it's installing bootstrapped extensions. This is something we do when you change Positron versions, since each version comes with its own bundle of extensions to install. I wonder if installation of bootstrapped extensions is interfering with startup in some way (are we e.g. restarting the host before startup can get off the ground? the logs don't suggest that this is the case, though).

[juliasilge]

I have observed this myself sometimes, but not every time I get a daily update.

[rmflight]

@jmcphers OK, the fact that in preparing, new versions of extensions are being installed makes sense. "preparing" comes up and is gone very quickly. Then "starting up" comes along, and it will sit there for 5 minutes if I let it. And it's doing something, because the laptop fans turn on to cool the cpu.

[juliasilge]

The next time someone internal runs into this (or you, @rmflight, if you are willing) let's check whether bootstrap installed extensions were getting updated.

We are working on #11144 this month to get more data on startup as well.

[rmflight]

@juliasilge I'm more than willing to upgrade to the next release (when it comes out), and try to capture that information. Is there something more than the logs I provided that's needed to get that information?

[juliasilge]

Unfortunately, whatever the nature of this problem is, it is resulting in basically no helpful startup logs being generated! 😩

What would be helpful to know for now is if this is only happening to folks when they are getting new versions of the bootstrap installed extensions; not every Positron update has new versions of those extensions. We are hopeful that #11144 will also help us get more data in a situation like this one.

[jennybc]

This just happened to me (and is something I experience regularly). The newly-launched-and-immediately-hung Positron is this:

Positron Version: 2026.02.0 build 22
Code - OSS Version: 1.106.0
Commit: 09eac9a
Date: 2026-01-08T07:20:15.913Z
Electron: 37.7.0
Chromium: 138.0.7204.251
Node.js: 22.20.0
V8: 13.8.258.32-electron.0
OS: Darwin arm64 25.2.0

And, given what I've been up to, it's pretty likely this was me upgrading from the just-previous daily. I'm actively using Positron alot. Offering these details in case we can establish if this daily increment involved:

folks when they are getting new versions of the bootstrap installed extensions

[jennybc]

I see this in the Extension Host output channel, but have no idea if it's relevant:

2026-01-08 15:22:17.548 [info] Extension host with pid 26945 started
2026-01-08 15:22:17.548 [info] Skipping acquiring lock for /Users/jenny/Library/Application Support/Positron/User/workspaceStorage/b193874f5b3c82d17baf771467e04920.
2026-01-08 15:22:17.554 [error] Error: ENOENT: no such file or directory, access '/Users/jenny/.positron/extensions/posit.shiny-1.3.3-universal'
    at async Object.access (node:internal/original-fs/promises:608:10)
    at async Object.realpath2 (file:///Applications/Positron.app/Contents/Resources/app/out/vs/workbench/api/node/extensionHostProcess.js:9532:5)

[jennybc]

Here's some interesting output from the Shared output channel.

positron-Shared-2026.02.0-build-22-09eac9a.log

There's a specific JSON blob about extensions that gets repeated over and over again, which makes the log super painful to look at.

I'm no expert in this area but it does seem like some of the bootstrap extensions are being upgraded (?).

[jennybc]

In my case, the console gets stuck at PREPARING and stays there indefinitely until I mercy kill it.

Screen.Recording.2026-01-08.at.3.54.35.PM.mov