I'll load a page on one of a handful of specific sites. It'll seem to load fine, but the throbber will continue to spin for 20 seconds before external scripts actually run. During that delay, Firefox remains responsive, and the status "bar" will say "Transferring data from <some CDN>...". However, the Network panel claims that every request completes nearly immediately, and there's no network delay at all. The CDN mentioned is always the host of the last request made before the delay, /not/ the CDN that hosts the stubborn script. Wireshark confirmed that it's seemingly not a network problem (and there's no network activity save for some keep-alive during the delay), so the status overlay seems to be a red herring, though I don't know why it stays visible if there's no network activity. Taking a profile of the page load reveals that absolutely nothing happens during the delay. Inline scripts run, then there's a long pause with nothing but GC happening, and /then/ the first <script src> runs. So far I've only witnessed this on Twitter, YouTube, Dropbox, and GitHub — all places where it's particularly aggravating because the pages themselves are nonfunctional until some JS runs. The problem is consistent, but not /reliable/, which makes it a little hard to tell if anything I'm doing is helping. However: - It seems to be more frequent the longer I've had Firefox running. - I couldn't reproduce in a fresh profile, in either Nightly or stable. - I couldn't reproduce in Chromium. - No one else in my house seems to be having similar issues. - It still happens with all my extensions disabled. - I've gone through the modified settings in about:support and reset everything that seems even remotely relevant, to no avail. This is a very old profile, with a lot of history and open tabs (that I would prefer not to lose). But this has been going on for months, no one has any idea what could possibly be wrong, and it's driving me up the wall. :) Possibly also relevant: I'm on Arch Linux; I have HTTP/2 disabled because of problems with broken connections to Google servers; my cache is in /dev/shm/firefox-cache; I have IPv6 via HE, but none of the affected sites save for YouTube support v6.

eevee, Can you still reproduce using a current version?

It seems less common and less pronounced since electrolysis, but I do still see it in 55.0.3, yes. Only on Twitter, though; I don't think I've noticed it on GitHub or YouTube in some time.

There is also Bug 1373023 - [meta] Laggy text input on Twitter direct messages (we flush layout too often, and reflow is slow)

How many tabs? Does it reproduce with Firefox started in safe mode?

(In reply to Wayne Mery (:wsmwk) from comment #3) > There is also Bug 1373023 - [meta] Laggy text input on Twitter direct > messages (we flush layout too often, and reflow is slow) How is that relevant to this bug?

(In reply to Emilio Cobos Álvarez [:emilio] from comment #5) > (In reply to Wayne Mery (:wsmwk) from comment #3) > > There is also Bug 1373023 - [meta] Laggy text input on Twitter direct > > messages (we flush layout too often, and reflow is slow) > > How is that relevant to this bug? relevance would depend on what further details from the reporter, which so far we don't have

If the throbber spins, it means there is a whatever request still pending in the tab's load group (including possible iframes). A child log with LoadGroup:5,nsHttp:5,timestamp modules would be very helpful here.

for no response from the reporter on more data.