I started with Thunderbird.
I found a problem -- later identified as being server-side -- and tried as many others as I could find in the distro's repos: Evolution, Sylpheed, Claws, KMail, Balsa, GNUstep Mail.app, Geary, and more.
Evolution is better than it was and isn't quite so determinedly Outlook-like any more. (I do not like Outlook.)
Claws is pretty good, but it isn't multithreaded, so it hangs when collecting mail. This is very annoying.
Claws and Sylpheed desperately need to merge again. They are basically the same app, but with slightly different feature sets. AIUI the author of Sylpheed, Yamamoto Hiroyuki, refuses to accept patches/PRs. He really needs to get over himself and learn to act a bit more like Linus Torvalds did. This intransigence is crippling both programs.
It is the 21st century and I do not want a CLI/text-mode email app. They have their place, for instance if you need to do email over ssh. I do not. But I want something that readily scales to a large window, has a CUA UI, can show basic formatting, etc. So, no Mutt/Neomutt/Pine for me.
In the end, I went back to Thunderbird and I still use it today. It is, after considerable research and experimentation, the best FOSS email app there is.
It is cross-platform: I can and do use the same app on Linux, Windows and macOS.
It talks to everything. I have or have had it connecting to Gmail, Hotmail, Yahoo, AOL, Exchange Server, Groupwise, CIX, and more different accounts and servers than I can remember.
It does address books and calendaring as well.
It has integration with handy features like Google's various chat and note-taking services.
It uses standard storage formats that can be accessed from other apps.
It's big, it is a bit sluggish, and like Firefox Quantum, some add-ons no longer work. This is a foolish decision of Mozilla's. However, it still has a useful range of add-ons.
It handles secure email and encryption well.
Snags: it really needs a working sync function.
But, after a lot of time and effort, it remains best-of-breed for my needs.
I said "unfortunately" because Thunderbird has stagnated, has a bunch of long-standing bugs and annoyances. Like limiting the number of e-mail addresses one contact can have, the low quality of search, the lack of notification configuration and terrible architecture that causes freezes both foreground and background. It freezes entirely when opening "large" letters, moving or syncing a "large" amount of e-mail, when encrypting a few megabytes, when your (totally unrelated) smartcard is in use and so on and on, can't keep track of it all myself but those tickets have sat open for decades. (No, this is not a suggestion to use a stale bot, that's a cheap disrespectful feel-good way out)
From nice-to-have aspects, it sucks that I can't both S/MIME and GPG sign a single letter. Sometimes it's good to bridge the two systems and it should be technically doable.
None of the stuff you mention affects me personally, but it's the Pareto principle: we all only want or need 20% of the functionality, but a different functionality... so in the end, the pool of all users need 80% of the functionality. The other 20% or so is either there to support that 80% or it is something that is absolutely essential for a tiny number of people.
Which is what leads to OSes like Linux, which are huge and vastly complicated, but every obscure function that 80% of us will never use is crucial to someone somewhere... because Linux is a desktop and a laptop and a server and a router and a hypervisor, etc.
Tens of thousands of organisations agree on the usefulness of Linux, so it gets a lot of investment and R&D and development.
Only 1 company owns Thunderbird and it thinks that it's not important, so it gets very little: just a bit of bugfixing and maintenance. :-(
While actually, it is in its way a jewel of vital functionality that really should be celebrated and polished.
I'd love to work out what it _needs_ to get more airtime and investment.
Seriously, out the box, for me, it would be literally unusable. I know it sounds petty but the font choice, icons, layout and so on are below even my low standards.
Fortunately it supports plugins, and I found one that makes it completely fine. Perhaps that's the point.
OTOH, one of the reasons I don't use KDE is that for me, KDE has been fugly since v2 and still is. Going "flat" alleviated it, but only from fugly to merely ugly.
T'bird looks fine to me, and does what I need without a single addon -- but I can see where a few small addons could transform it from "very useful" to "world-beating". See my comment above. :-)
I needed more integrations for calendar and contacts with my self-hosted Nextcloud. I had to leverage plug-ins for Thunderbird and they were problematic but kind of worked most of the time with caldav and carddav data sources.
Earlier this year, I reset my Thunderbird profile and tried the updated non-plugin setups and it has been great. I know the overall roadmap and updates have not be dramatic but I can say the stability and core functions have really improved.
You should at least give it a shot.
Nothing else comes close.
Discussion on HN: https://news.ycombinator.com/item?id=25222166