The post yes, but not the comments at depth > 50.
No the max comment depth is generally lower now. However this doesnt affect comments created before upgrading.
Changing post.url
from varchar(512) to varchar(2000) really messed up database performance so lemmy.ml became unusable. Turns out that column statistics are removed when the type is changed, so we had to run analyze
as part of the migration. Seems like a bug in postgres.
https://github.com/LemmyNet/lemmy/issues/4983#issuecomment-2446945046
Dont think I did, it was really a team effort. And in the end the working solution was suggested by @phiresky@lemmy.world, thanks for that!
It will be rejected by the api (or by federation).
I find it very questionable that you publish this sort of hit piece against Lemmy without even bothering to ask for a comment from our side. This is not how journalism should work.
Effectively you are blowing the complaints of a single user completely out of proportion. It is true that we didnt respond ideally in the mentioned issue, but neither is it okay for a user to act so demanding towards open source developers who provide software for free. You also completely ignore that this is an exception, there are thousands of issues and pull requests in the Lemmy repos which are handled without any problems.
Besides you claim that we dont care about moderation, user safety and tooling which is simply not true. If you look at the 0.19.0 release notes there are numerous features in these areas, such as instance blocking, better reports handling and a new moderator view. However we also have to work on improvements to many other features, and our time is limited.
Finally you act like 4000€ per month is a lot of money, however thats only 2000€ for each of us. We could stop developing Lemmy right now and work for a startup or corporation for three or four times the amount of money. Then we also wouldnt have to deal with this kind of meaningless drama. Is that what you want to achieve with your website?
That instance list is built completely automatically by a crawler, no one approves instances before they are listed. In this case it was removed as soon as we became aware of it. Next time please make a pull request like that one, its much more effective than complaining.
I dont follow /c/worldnews so I dont see much of that. Also hexbear is federating now, so it might easily swing back the other way again.
I definitely didnt expect it, nor did I expect that there would suddenly be more than a dozen different apps. But its not a problem, the more choices users have the better. Those who like such clients can use them, thout it affecting anyone else. Plus monetization of apps could potentially help to fund development of Lemmy itself.
For instances with ads its pretty much the same, more choice for users. But I really doubt that model can have any success considering how many free instances are around which are run by volunteers. Defederation should be unnecessary assuming that ads are only shown to local users.
That particular improvement is actually mine. Lemmy was storing a lot of federation data which was completely unused so I removed it. However the 80% improvement is actually overstated, because not all data was migrated to the new table. So the db will grow a bit bigger over time, but still much smaller than before.
Phiresky made a lot of other sql optimizations which make Lemmy snappier and reduce CPU usage on the server. We don't have any benchmarks in that regard, but server load on lemmy.ml has gone down a lot since upgrading.
This would be neat to generate forum avatars, to show something similar to your real face.