-
-
Notifications
You must be signed in to change notification settings - Fork 116
New issue
Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? # to your account
After recent upgrade, GhostText disconnects from editor after 30 seconds #317
Comments
This is a limitation of the new Manifest v3 that Google has been forcing on us. I'll check if I can send a dummy message every 10 seconds to keep the connection alive. |
An update is out, but you need Firefox 128+ now since v24.7.1 Wait until July 9th for its release or download the beta now: |
I recently noticed the same problem described in this issue, that the connection between GhostText and my editor (Emacs, with atomic-chrome 20230304.112) has been dying after 30 seconds of inactivity. I am using the latest versions of Firefox and GhostText. Firefox: 128.0.3 I see that the I installed the older GhostText 24.6.7 from https://addons.mozilla.org/firefox/downloads/file/4299538/ghosttext-24.6.7.xpi, which seems to be the last release before you switched to Manifest v3, and it did not have the connection issue. |
Downgrading to v24.6.7 as @dhull suggested worked great for me! So awesome! Life changing to have such easy access to your favorite editor! I was not able to get the current version of GhostText working on the latest Firefox Beta, outlined the details on this comment. NOTE: Be sure to DISABLE UPDATES so that you don't have to re-install... |
Same issue on Firefox @dhull You can see an extension's logs/network connections in Firefox by going to It looks like the "ping" is sent from the background js worker to the extension js (through the "port"), not over the websocket from the extension to the server. FWIW I tried sending |
When v24.8.10 is available for download, please test it out and let me know. For me it worked correctly in both browsers |
Thanks! The new GhostText version v24.8.10 works for me in Firefox 129.0. The browser's connection to Emacs does not close after 30 seconds. |
@fregante works great! thanks! |
This is working on Firefox now as discussed above, but it has not been working in Chrome the last week. It loses the connection somewhat randomly between 1 and 5 minutes. Thank you for all your work on this! |
Do you mean it was working but stopped with the latest version? |
Yes, correct |
I noticed this a few months ago and thought that maybe I broke something. Instead, I just continue editing the text and then pasting it into unsynced tab through clipboard. Which kinda defeats the purpose. So I had less and less desire in using this plugin because of the issue. But today I finally tried searching if anyone else got the same problem, and it's so calming that I'm not the only one. Anyway, I updated Firefox (130.0) and the extension (24.8.10) to the latest version and let it sit for 5+ minutes. It didn't disconnect! Hurray! P.S. I use Neovim, btw (with this patch on top of subnut/nvim-ghost.nvim@4ce472f). |
v24.8.10 solved the problem for me (Firefox 130.0.1, Emacs atomic-chrome). |
Description and repro
Sometime in the last few days GhostText stopped working properly on both my Linux computers. I suspect it has to do with the recent GhostText update (Firefox says "Last Updated June 30, 2024" for GhostText on about:addons).
The content of the text area appears in my editor (Emacs using https://github.com/alpha22jp/atomic-chrome), and I can edit it, and if I close the editor buffer within 30 seconds everything works normally.
But almost exactly 30 seconds after I click the GhostText button in Firefox, the editor buffer closes (just as if I closed it manually), but GhostText in Firefox appears to still be active (green tickmark on GhostText icon) so I cannot easily enable the editing again without reloading the web page.
The 30s seems to be if the editor is idle. If I type in the editor more or less constantly, I can see the edits in the Firefox text area, and the connection lives for a little longer, around 55 seconds.
This could of course be an Emacs problem, but I tried snooping on the Websocket traffic between Firefox and Emacs using Wireshark, and I can see that Firefox/GhostText sends a "WebSocket Connection Close" packet exactly when Emacs closes the buffer.
I realize Emacs is not officially supported and I have not tested in Chrome, but it seems like a clear regression in GhostText (or maybe something triggered by a recent Firefox update)
Browser(s) tried
Firefox
Editor(s) tried
Emacs
The text was updated successfully, but these errors were encountered: