This question relates to my use of Firefox 88 on MacOS 10.14.6 on a dual-GPU retina MacBook Pro (15-inch, 2018). I'm using a dual external monitor set-up, and still runni… (čítať viac)
This question relates to my use of Firefox 88 on MacOS 10.14.6 on a dual-GPU retina MacBook Pro (15-inch, 2018). I'm using a dual external monitor set-up, and still running 10.14 due to 32-bit compatibility issues with some of my (scientific, bespoke) software.
I am pretty sure that there is a very odd, browser-level bug in alpha display on my system that leads to, for example, the overlay of YouTube's UI appearing peppered with row-wise stripes of other colours; or PDF page previews not displaying properly. There is nothing relevant displayed in either the (JS) console log; `dmesg` or MacOS's Console.app application. The appearance of the stripes / corruption is randomly different each time: it is different in private browsing mode to not, for example, and refreshing the page alters their exact appearance.
I have done the obvious things of disabling hardware acceleration -- no change -- trying in a new private window with no extensions / addons loaded -- no difference -- but I do not necessarily want to nuke everything and try again. This problem has existed for as long as I have used FF on this computer, but I am just reporting it now. Mostly it's fine, aside from one variety of captcha that I cannot solve!
I have no idea how to go about debugging this and alas I don't have the time to do it "properly" -- I am on a system with SIP disabled and can easily load up FF in lldb, for example, if someone could give me targeted advice. I am not a web developer and do not know the exact api call that causes this -- but it appears to be used in pdf.js' sidebar page overlay code, but not the actual page itself. Most PNGs, SVGs etc are fine.
Please see the attached series of screenshots – if it's not clear what I've done (the same youtube video in and out of private browsing mode with no extensions, a PDF viewer; and that page reloaded showing a different pattern), please don't hesitate to ask -- the corruption is pretty visible in all cases, I think. The fact that I can screenshot this and share it with you means, I think, that it is not an OS-level graphics driver bug and is probably a rendering bug.
Thanks for your help, and sorry if this is the wrong forum!