(In reply to :Gijs (he/him) from comment #4) > Luca, any idea about the fixed window in comment #0 and what is going on here? About the fixed window, I'm not yet sure exactly which part of Bug 16099200 patches may have affected the reproducibility of this issue ([1]), but the issue seem to be definitely related to the service workers and apparently I can still trigger it on Firefox 80 with a different STR: - In Firefox 79, where the issue can still be triggered, in about:debugging I see that there is an https://twitter.com/sw.js service worker listed, its status is stopped and it is listening for fetch events: - the first time the twitter bookmark is clicked, the service worker is still marked as stopped, and twitter isn't loaded in the tab - the second time the bookmark is clicked, the service worker in about:debugging goes from stopped to running, and twitter does load as expected in the tab - if I wait for the service worker listed in about:debugging to go to the stopped status again, I can also retrigger the issue as the first time without even restarting the browser - In Firefox 80, in about:debugging there isn't initially any service worker listed, then: - the first time the twitter bookmark is clicked, the service worker appears in the about:debugging list, it is marked as registering and then running, and the twitter url is being loaded - If I wait for the service worker listed in about:debugging to go to the stopped status again, I can trigger the issue as in Firefox 79 And so the issue doesn't seem to be really fixed, the issue with twitter not loading can still be triggered with a slightly different STR. On the other end something in 80 (not sure yet if it is actually Bug 16099200) may have changed something related to the status of the workers on startup or at least their status as visible in about:debugging (which wasn't expected, and I want to dig into that more, those patches were not supposed to affect that). I'm going to come back to this tomorrow and do some more digging, but I'm also needinfo-ing :asuth for his opinion about the possible underlying reasons for this issue, as well as about the difference that I spotted in about:debugging between 79 and 80 that I described above. [1]: a very big chunk of the changes introduced in those patches is actually locked behind prefs and disabled by default on all channels (and enabled only for the related test cases at the moment), I would have been less surprised if the fixed window was pointing to Bug 1642676 which landed in the same release and it was a dependency of Bug 16099200 patches.
Bug 1656310 Comment 5 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
(In reply to :Gijs (he/him) from comment #4) > Luca, any idea about the fixed window in comment #0 and what is going on here? About the fixed window, I'm not yet sure exactly which part of Bug 1609920 patches may have affected the reproducibility of this issue ([1]), but the issue seem to be definitely related to the service workers and apparently I can still trigger it on Firefox 80 with a different STR: - In Firefox 79, where the issue can still be triggered, in about:debugging I see that there is an https://twitter.com/sw.js service worker listed, its status is stopped and it is listening for fetch events: - the first time the twitter bookmark is clicked, the service worker is still marked as stopped, and twitter isn't loaded in the tab - the second time the bookmark is clicked, the service worker in about:debugging goes from stopped to running, and twitter does load as expected in the tab - if I wait for the service worker listed in about:debugging to go to the stopped status again, I can also retrigger the issue as the first time without even restarting the browser - In Firefox 80, in about:debugging there isn't initially any service worker listed, then: - the first time the twitter bookmark is clicked, the service worker appears in the about:debugging list, it is marked as registering and then running, and the twitter url is being loaded - If I wait for the service worker listed in about:debugging to go to the stopped status again, I can trigger the issue as in Firefox 79 And so the issue doesn't seem to be really fixed, the issue with twitter not loading can still be triggered with a slightly different STR. On the other end something in 80 (not sure yet if it is actually Bug 1609920) may have changed something related to the status of the workers on startup or at least their status as visible in about:debugging (which wasn't expected, and I want to dig into that more, those patches were not supposed to affect that). I'm going to come back to this tomorrow and do some more digging, but I'm also needinfo-ing :asuth for his opinion about the possible underlying reasons for this issue, as well as about the difference that I spotted in about:debugging between 79 and 80 that I described above. [1]: a very big chunk of the changes introduced in those patches is actually locked behind prefs and disabled by default on all channels (and enabled only for the related test cases at the moment), I would have been less surprised if the fixed window was pointing to Bug 1642676 which landed in the same release and it was a dependency of Bug 1609920 patches.