Egregoros

@mWare @stalwartlabs they literally just released an update that requires a massive migration of data with a python script, requires backup+restore to do it right, and broke some shit like now forbidding login usernames to be something other than an email address (which is useful for making it harder for people to bruteforce your login)

and they still won't let you "undelete" mails unless you pay for their enterprise version.

it's just not worth it. I've already got my mails back by spinning up the backup from -24Hrs and using imapsync to get them back, and now I've got my mails all synced back into Aox again

I'll just switch back and move on with my life. I'll just have to add features to Aox myself. It shouldn't be so hard anymore really, and the code has been so stable all these years anyway... Stalwart just isn't ready for broader usage yet. Maybe when they finally hit 1.0 and stop rewriting everything.

@mWare @amenonsen aox has done one thing right for so many years: stores mail without losing it, and does it very efficiently while implementing all the important IMAP/Sieve extensions required to be useful.

It hurts my heart to think about trying to reinvent it.

> The source code is fully documented, to make life simpler as the team changes and expands. At the time of writing, we have over 300,000 test cases of different types, and are always extending the number and variety of tests.

How can you compete with that? Who else has that?

(also, where are those test cases? They're not in the code repo from what I can tell)

But the documentation in this codebase is so thorough that I have no doubt that if I can't figure something out I can get an LLM to answer my questions for me