Firefox Community Forum

VPN is on. I can't access the internet. Error message is either: An error occurred during a connection to www.thewindowsclub.com. SSL received a record that exceeded the maximum permissible length.

Or: Connect to the internet You're offline. Check your connection.

I've googled the issue but it talks about ports and tunnels and going into settings. I've looked in the VPN Mozilla settings and there only seems the option to either include or exclude a website from the VPN. Nothing about different ports and testing. I've already written this message once but the VPN was on so it disconnected the internet and I lost this message! Trying a 2nd time now without the VPN

Any help? Or should I give up with it?

thanks Annie

Hello- Am having trouble with my settings( I think) Want to be sure Firefox is default browser, and use Adobe Reader as my PDF default. All I get is "download Firefox" ! Get all regular updates, so why can't I see "settings" ? Thank you

This problem has never happened before. This happened in the last few minutes. The whole screen suddenly went blank while I was trying to find out today's weather. I am suspicious of malicious interference with my computer. Any help you can give me will be reassuring.

A website I user, www.advisor.ca, displays a banner across the top which obscures some of the text in Firefox but displays correctly in Chrome. The attached images are for the page https://www.advisor.ca/tax/tax-news/court-dismisses-challenge-to-cras-administration-of-cap-gains-tax-proposal/. I expect this is one of those things that Chrome does 'wrong' but I would like to know if there is something I can do to make it display correctly in Firefox, of if FF could handle it in some manner.

Thanks.

yahoo mail keeps requiring me to get code from gmail to view my mail. problem may be that my cookies are deleted when i close firefox. windows 11 desktop. Thanks for Your help

Hi Moz Support,

I've had this happen several times before, but this time it really annoyed me.

I sent tabs from one Firefox instance to another on another device. I closed these tabs on the source device. These tabs were quite old this time, I'd not opened them in a while - this becomes important later.

On the destination device, let's call it "madbilly's destination Firefox", these tabs didn't arrive. After doing a couple of tests where I sent a tab from my mobile to this instance, I realised that this was a similar problem to what I'd had before, which is one where a Firefox instance/profile for some reasons "resets" itself and it seems to create a new instance in Firefox Sync. Tabs which have previously been sent to "madbilly's destination Firefox" don't arrive, because (I think) it's not registered in Firefox Sync as the same instance. These sent tabs, if closed on the source instance (which I did) are lost, unless one reopens the source instance and Ctrl-Shift+T reopens closed tabs. Unfortunately, in this case because the tabs were old and I hadn't opened them for a while there was no way to do this, because Firefox doesn't remember closed tabs which are not stored in memory (another bug, any fix?), except for looking in old session store backups which I'm going to do after I write this post, though I don't know if that will work.

Also, on Android, because it only remembers 10 closed tabs it's not really feasible to retrieve lost sent tabs this way either, and as there's no access to the session store backups the tabs are definitely lost.

Oddly, when this happens "madbilly's destination Firefox" usually appears twice in the send-tab-to menu, but in this case it didn't. To be able to send to this instance I had to sign out of Firefox sync on this instance and sign in again. Then the same instance appeared twice in the sync menu, the old one being now unreachable.

I'd like to know: 1. Why does this happen? 2. How to avoid it happening? 3. Is there a bug registered about this? 4. When will it be fixed and prevented from happening again.

I generally find that this happens when a Firefox instance isn't shut down properly, which on my old desktop I put down to it running out of resources, but on this shiny new laptop from work that doesn't make sense.

All help much appreciated thanks :)

Our team has already tried to fix the issue for a week now and none of the approaches seem to fix the issue. Does the Firefox dev department have any hints or clues for us?

Short description: Firefox network.lna.blocking blocks same-origin/loopback requests when the page is controlled by a Service Worker

Symptom: Login via OIDC fails in Firefox (the backend call after the IdP redirect is blocked). The error disappears if either network.lna.blocking=false is set or the Service Worker is disabled.

Environment: - Frontend (react) and backend run on the same machine (loopback/localhost)—no cross-network scenario. - Firefox with active network.lna.blocking (experimental Local Network Access feature, Chromium’s equivalent is Private Network Access).

Root Cause (Analysis): 1. apps/react/sw.js registers a global fetch listener via Workbox and calls workbox.core.clientsClaim(). This causes every open page to be immediately controlled by the Service Worker. 2. registerServiceWorker() is called unconditionally on every app start, regardless of authentication state—even on the login/OIDC page. 3. The app uses hash-routing (URL fragments like “#/oidc”); the login page and the rest of the app share the same document URL, so you can’t restrict the SW scope to “everything except login.” 4. Once a page is Service-Worker-controlled, every fetch—including requests not explicitly handled by Workbox routes—first goes through the SW’s fetch event (default behavior when any fetch listener is registered) before proceeding to the network if it isn’t intercepted. 5. Hypothesis: Firefox’s LNA implementation doesn’t correctly inherit the address-space classification (loopback/private/public) for SW-forwarded requests across the controlling client. A direct document fetch (loopback→loopback) is allowed, but an identical target fetch routed through the SW is blocked.

Reproduction (minimal): 1. Use Firefox with network.lna.blocking=true (the default in affected versions). 2. Load a page that registers a Service Worker with a fetch-event listener (e.g. via Workbox precacheAndRoute + clientsClaim()). 3. Ensure both the page and the target resource are served from localhost/loopback. 4. From the SW-controlled page, call fetch() against the loopback target resource → the request is blocked. 5. By comparison, the same fetch from a non-SW-controlled page succeeds. 6. Setting network.lna.blocking=false also allows the fetch to succeed under SW control.

Excluded Causes: - Different address spaces for frontend/backend (both are on the same loopback machine)—ruled out. - Misclassification due to redirect_uri or cross-origin navigation—the OIDC redirect is a top-level navigation (unaffected); only the subsequent backend call (token exchange/session validation in SsoLogin.tsx → ssoLoginSubmit) fails.

eg https://transportnsw.info/opal-view/#/account/cards is meant to show down arrow next to name - does not appear; go to Profile, Edit address - 'Save as delivery address' checkbox and Submit button do not work. Also had this on the StGeorge bank page https://www.ui.stgeorge.com.au/resolve/etb0/createCustComplaint - Next button does not work. other sites as well have buttons not working (about 20% of sites?) (Sorry both the above examples need accounts to test). Often (not always) if I open the page with Chrome it works. I think they were working ok after macOS update but after a Firefox update went to a very light screen - problem started then.

Also - using about:support the copy raw data button works; the copy text button does not.

Thanks