User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:55.0) Gecko/20100101 Firefox/55.0 Build ID: 20170504100357 Steps to reproduce: 1. Create a new profile and start Nightly using it. 2. Open the page of any YouTube video. Doesn't matter if it's paused or not. 3. Scroll down until comments no longer loading and you have reached bottom. Actual results: Memory rapidly goes up then returns to normal. A CPU thread (i7 4700MQ) reaches 100% (by the content process) and remains there until tab is closed. Reproducible most of the time. Captured profile: https://perfht.ml/2pMPr82 Expected results: After the page stops loading and the video is paused/finished, CPU usage should be low. Mozgression result: 12:58.08 INFO: Narrowed inbound regression window from [82c2d17e, 96605941] (3 revisions) to [d7e40bb8, 96605941] (2 revisions) (~1 steps left) 12:58.08 INFO: Oh noes, no (more) inbound revisions :( 12:58.08 INFO: Last good revision: d7e40bb852ea047a0e5f530bcdc04d29a1765001 12:58.08 INFO: First bad revision: 96605941c0021795376d9c6ec1e458de2fac329e 12:58.08 INFO: Pushlog: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=d7e40bb852ea047a0e5f530bcdc04d29a1765001&tochange=96605941c0021795376d9c6ec1e458de2fac329e
I apologize, the info in Comment 1 is _partly_ false. Blocking XHR of "www.youtube.com" using the same extension was what actually worked. Also, I was able to stop the issue by using an extension to block _only_ the comment section while the video was playing, therefore something is wrong with the comment section. If further info is needed, let me know.
The bug isn't present in Build ID 20170510183715. Mozregression (--find-fix) result: 8:15.34 INFO: First good revision: 8a43b69fbae255a2af98f11507862437d6303033 8:15.34 INFO: Last bad revision: ebbcdaa5b5802ecd39624dd2acbdda8547b8384d 8:15.34 INFO: Pushlog: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=ebbcdaa5b5802ecd39624dd2acbdda8547b8384d&tochange=8a43b69fbae255a2af98f11507862437d6303033 8:16.87 INFO: Looks like the following bug has the changes which introduced the fix: https://bugzilla.mozilla.org/show_bug.cgi?id=1354641
ffbug, Thanks for your help to fire this bug and keep figuring out the problem. Bug 1354641 is just backed out yesterday. Could you help test today's nightly to see if this bug is fixed by that backout. Thanks.
The issue seems to be absent in Build ID 20170511100157.
Thanks for confirming!