I too think they should drop everything else and work on this bug. But an example of how these kinds of bugs are tricky:
I have a similar bug at my job: Sometimes browsers delete our extension's database, or otherwise corrupt it. It's been an issue for years, but no one has been able to reproduce it. It's probably a 1 in 10 million bug.
I think it's a hardware bug. My "fix" was to backup a small, but key, part of the database to a separate storage mechanism browsers let us access. When the issue arises we can now try to detect the missing data and restore part of the db.
BUT! If this is actually a hardware bug, there is a chance that this additional database write will cause this big to occur even more often, as we now have to write to storage twice as often!
> Sometimes browsers delete our extension's database ... but no one has been able to reproduce it.
Are any of the bug reporters using Vivaldi? As for some reason that browser allows (and enables by default) clearing extension storage when clearing history/cookies via Delete Browsing Data.
This is something the Stylus addon dev noted when a user was reporting the addon wasn't remembering their settings and it was because the user had unwittingly wiped the extension storage due to the browser defaults.
(I actually use Vivaldi myself but virtually never use that feature so was unaware of the behavior until reading about the bug report)
"Delete" was an over simplification. The database seems to be there, but the data isn't. Dozens of engineers over the years have tried to find a way we could have accidentally cleared the data, but not the keys.
No user that has reported it has been willing to share their database with us (chrome stores it in a standard leveldb), as it contains private information.
But good to know that about Vivaldi! I didn't know that before, thanks!
I have a similar bug at my job: Sometimes browsers delete our extension's database, or otherwise corrupt it. It's been an issue for years, but no one has been able to reproduce it. It's probably a 1 in 10 million bug.
I think it's a hardware bug. My "fix" was to backup a small, but key, part of the database to a separate storage mechanism browsers let us access. When the issue arises we can now try to detect the missing data and restore part of the db.
BUT! If this is actually a hardware bug, there is a chance that this additional database write will cause this big to occur even more often, as we now have to write to storage twice as often!