Yeah, the blue lines under a random word in a perfectly correct sentence is such a waste of time ... did I mistype? Nope.

OMG yes. Pretty sure that bug has been around for something like a decade. Insane they haven't prioritized it, or I wonder if they hide behind the fact there doesn't seem to be any way to reliably reproduce it?

Someone just has to look really hard at the code and find the bug. Surely the relevant code can't be that long?

I remember back in the late 90s that, if you ignored the matter of hardware driver quality (and that is a big "if", no question) that open source software tended to be higher quality in general than a lot of commercial software. Not because of any moral characteristic per se, but just the "many eyes make bugs shallow" sort of thing. Since it was mostly only programmers using open source anyhow, if someone hit an annoyance, statistically speaking, there was a good chance that someone who could fix the problem had hit the same annoyance.

Then maybe in the 2010s commercial software at least caught up.

But it seems to be swinging back around to, if I want my software to effing work I want to be seeking out open source again. Statistically speaking, fewer of the users who may encounter problems can fix any problems they find, as the systems have gotten much larger, but it is still possible, and on the compensating side, no one on the emacs team is figuring out how to stuff AI where it doesn't belong [1] or how to monetize it via ads or any of the other exciting ways to arbitrage long-term software quality against short-term money.

It's an opinion, it is clearly highly path-dependent, and I won't deny this is just my impression... but it is something I've been noticing again lately. Especially as Windows seems to be heading down the catastrophe curve and this time I'm not sure they can stop it.

[1]: I'm not anti-AI at this point... but there are places where it belongs, and there are places it just doesn't, and stuffing it where it does not belong is not a win.